Topological sorting is something I've been thinking about lately.

This is sorting concepts in dependency order.

For example, if you wanted to sort "fire" and "internal combustion engine", fire would come first. To explain ICEs, you need fire, but to explain fire you don't need ICEs.

Sorting concepts topologically versus chronologically can create different rankings.

Sorting numbers topologically puts binary (0 and 1) before the Hindu-Arabic numerals (0 - 9), even though humans used 0-9 way before using 0 and 1.

Our topological knowledge base often has missing or incorrect concepts that may not be fixed for centuries.

Encyclopedias are sorted alphabetically.

I am unaware of an encyclopedia sorted topologically.

Why would you want an encyclopedia sorted topologically?

Well, topological sorting tells you the logical importance of things. Things further down are built on things at the top (or vice versa, if you prepend new things to your files rather than append).

What parts of the encyclopedia should you learn first?

It makes more sense to learn the things with a high topological ranking, rather than a high alphabetical ranking. A lot of things turn out to be fads.

Popularity sorting such as by number of inbound links is an improvement over alphabetical sorting, but seems very susceptible to bias and fads.

How would you create a topologically sorted encyclopedia?

I have been attempting to build a topologically sorted encyclopedia for a long time, though I had never described it like that. It's only recently when I realized I wanted to change Scroll to be topologically-sorted by default that I went looking for the definitive term to describe the concept.

I think using parsers all the way down might work, though I could be wrong. The nice thing about this strategy is that you can build stuff that is useful along the way even if the vision of the topological encyclopedia doesn't materialize.

Today I started looking at older programming languages that have lasted like C, Fortran, Ada, et cetera, and realized that topological sorting used to be the default. Newer languages aren't as strict, and that's the pattern I copied. But I wonder whether it's a better design to make the rule topologically sorted, and the looser version the exception.

When dealing with larger programs it seems you can do things a lot faster if you know things are sorted topologically.

Is the universe topologically sorted?

It seems to be. The present depends upon the past, and so comes after the past.

Which came first, the chicken or the egg?

If we are talking about the words, that might be easy to determine with a good etymology reference.

If we are talking about the patterns represented by the words, then it's a bit more interesting. We know the bacteria came before either. But my guess is we had objects closer in appearance to chicken eggs before we had objects resembling chickens. So I would say the egg came first, topologically.

A scale is an ordering of numbers. Objects map to a scale to allow comparibility in that dimension.

The word scale is an overloaded term. Usually when I use the word "scale" I am using a different version of it, such as "scale it up" or "economies of scale". In this post I'm using it in the sense of a measurement or yardstick or number-line or type.

English is generally scaled. A small subset of it is scaled.

So blog posts are mostly "unscaled". It is hard to compare this line with the line below it.

But this post does contain some lines that are scaled. For example, it has a date line, which maps this post to a date scale. So you can compare this post to others, and say which came before, and how much they came before.

Scroll, the language and software that powers this blog, does compute some scaled metrics on each post. The number of words, for example. You can see the number of words for this post and all others on the search page.

I like the definition of scales in the d3 data visualization library:

Scales are a convenient abstraction for a fundamental task in visualization: mapping a dimension of abstract data to a visual representation. Although most often used for position-encoding quantitative data, such as mapping a measurement in meters to a position in pixels for dots in a scatterplot, scales can represent virtually any visual encoding, such as diverging colors, stroke widths, or symbol size. Scales can also be used with virtually any type of data, such as named categorical data or discrete data that requires sensible breaks.

I remember when I was struggling to use d3 and then finally their definition of scales clicked in my head and I realized what a simple, beautiful and widely applicable concept it was.

Scales make things comparable. Measure different concepts using the same scale and now you can compare those things symbolically.

The more scales you use, the more sophisticated your symbolic models become. You can measure two buildings with a height scale to create some comparisons, but you can greatly increase those comparisons if you also measure them with a "year built" scale.

One of the most important scales is the computational complexity scale. Nature loves inequality, our universe has ~65 orders of magnitude buckets, and so rarely do 2 random things fall in the same bucket.

A dimension is just a set of different measurements with the same scale.

You can think of any scale as just a line.

Measure objects and draw a point on the line for where each measurement falls.

You can draw a high dimensional dataset as just a lot of independent lines. Not the most useful visualization, but can be helpful sometimes to break things down into really simply pieces.

Wikipedia does not make heavy use of scales. It relies more on text. I often wonder if the focus was more on adding data in typed dimensions, if it would allow it to become a more truthful symbolic model of the world.

They do increasingly populate those infoboxes with scaled data. The syntax is nasty, but the scaled data is wonderfully useful.

The more scales you have, the more trustworthy a model is.

I often think about complexity scales. I proposed if you think in parsers you can measure the complexity of any idea. Perhaps the "parser" is a good unit for complexity. If two models of the world are equally intelligent, pick the less complex one - the one with fewer parsers.

I don't have anything too profound to say about scales. (On the profoundness scale, this post ranks low.)

I just want to make sure I am deliberately thinking enough about them. If you measure concepts on an importance scale, they are high on the list.

Do you want to learn a new way to think? And a new way to write? Do you want to learn how to look at everything from a new perspective?

If so, read on.

I will try to explain to you how to "think in Parsers". I'll also then go into "writing with Parsers", a computer language we've made. It might not click immediately, but when it finally does I think you'll find it wonderfully useful.

No matter what topic you want to learn, from music to math, chemistry to carpentry, mechanics to quantum mechanics, thinking in parsers gives you another path to understanding any domain.

Look for a tree. (If you're inside, you may need to go to a window)

Did you see one? Great! That means you have a "TreeParser" in your brain that is capable of recognizing trees.

Now, look for a rock. See one? Great! You brain also has a "RockParser".

Parsers are things that match certain patterns and are inert to others.

Your TreeParser fired when you saw a tree, but not when you saw a rock, and vice versa.

The tree you looked at may have also activated another Parser specific to that species. I am writing this in Massachusetts and so for me that was my PineTreeParser. Usually I am in Hawai'i and my PalmTreeParser activates. For you it might be another kind. The point is parsers can be derivatives of other Parsers.

But that's not the only way Parsers can compose.

For example, you can also have Parsers that parse parts of a pattern. Think of a BranchParser or a LeafParser.

Parsers can combine in many ways.

Imagine taking a leaf from the tree and looking at it under a microscope. If you remember your biology class, your "CellParser" may activate. But if you were given a blurry image and not told its source, your "TurtleShellParser" may activate instead.

Because you have context you know which Parser is the one to use.

Imagine pulling the tree you saw out of the ground and into the sky. Now imagine throwing the rock into the sky.

I bet when you imagined lifting the tree, you imagined pulling up roots, but you didn't imagine roots attached to the rock.

You can attach knowledge to parsers that can be used in combination with the pattern that activated the parser.

The patterns that trees or rocks or turtles emit exist whether or not we have a parser that recognizes them. Patterns exist without parsers.

A baby starts with very few parsers and things for them are largely a blur until they have developed the parsers to make sense of their sensory data.

Often, especially when growing up, we see and record raw patterns that we don't fully understand. Eventually we stumble upon parsers that match those patterns and give us deeper understanding of what we witnessed earlier.

Sometimes we only notice a pattern after we've learned a parser for it. For example, if you've ever gotten a new car, you've probably experienced the phenomena of suddenly noticing that model of car all around you. This is because you have acquired more parsers for that pattern that are now getting activated when you encounter it.

And often the parsers we have don't actually best fit the universe's patterns and we update them later. The pattern of the earth revolving around the sun has existed always, for example, even though humans previously had parsers that parsed it in a different way.

Let's recap so far. Parsers:

• Match patterns
• Are composable
• Are context-sensitive
• Hold logic

So far we've been talking about Parsers for objects in the real world that are stored in the neural networks of your brain. These Parsers are neat, and you don't even have to be human to have them (dogs, for example have parsers to recognize their master) but Parsers work the same way in the symbolic world.

For example, you have an oParser that:

• recognizes the letter "o"
• also activates your LetterParser
• is context sensitive (so you can recognize the letter "o" versus the number "0")
• and holds logic related to "o", such as it's sound.

And then you can have a WordParser that recognizes words, such as "color".

You can then have a PropertyParser that recognizes a pair of words like "color blue".

Parsers never stop composing like this. If you were to look at the source code of this post you would even see that this sentence is parsed by a ParagraphParser.

It's Parsers all the way down! And all the way up!

One thing you'll notice about parsing objects is that all Parsers assume membranes. Whether it's the edges of the tree or the space around words, there are lines, either visible or invisible, that separate what is being parsed versus not parsed.

Now let's combine the sections above by making a language of symbolic parsers about things we encounter in the physical world.

First, we need to set some rules as to the membranes of our language - where do we draw the lines around the atomic units that we will parse? For this, we'll use Particle Syntax (aka Particles). Particles splits our writings into atoms (words), particles (lines), and subparticles (indented lines that belong to the "scope" of the parent line). The nice thing about Particles is that there is no visible syntax at all, and yet it gives us the building blocks to represent any kind of structure.

Now, imagine we wrote the lines below on a walk in the woods.

Right now all we have is some patterns, but no parsers.

Let's write some parsers to parse the lines above:

Now we have five parsers with which to understand our original program.

Of course, the parsers themselves also make use of various parsers (such as "pattern" and "atoms"), which I've left out, so in reality there would be quite a bit more.

I've left out some details to focus on the core ideas of how you can define Parsers symbolically that match patterns, are composable, context-sensitive, and hold logic.

You may notice that the Parsers language is not very much concerned with computers and fitting the world to match the structures in a computer. Although the parsers can carry logic to execute on computers, Parsers is really focused more on organizing knowledge into parsers. Parsers is a knowledge storage language first--a way to think first--and a computational language second. Parsers does not say "here is a set of structures that work well on the computer, fit your knowledge of the world into them." Instead the approach of Parsers is to "mine the minimal, ideal set of symbols to represent the patterns you see, and then we'll have the computer adjust to those".

Now, if you want to dive deeper into Parsers, an ever evolving language, the Scroll website is a good place to visit next.

But now let's keep going and connect the dots.

Now I'm going to get to the fun claim of this essay: every subject can be represented as a list of simple parsers in the Parsers language.

Parsers are the building blocks of knowledge, much like atoms are the building blocks of molecules.

It doesn't matter how "complex" the subject is, everything can be broken down (and built up) by simple parsers.

Now, don't get me wrong, I'm not saying quantum mechanics is the same thing as arithmetic. What I'm saying is that quantum mechanics is merely a larger list of parsers. Each parser in quantum mechanics is not more complex than the parsers in arithmetic, it's just that there are more of them.

If you are struggling to master a new subject, you can just start writing out all the parsers in the subject in a flat list. You can then work on your understanding of each one, refining your parser definition as you go. You'll find that before you can understand some parsers, you'll have to understand some other parsers it uses.

You can also organize parsers by which ones are used the most. (Often you'll find many parsers that are taught in books are not very important, and you'll start to pick up on the core set that are heavily used and should be mastered).

Eventually you'll have a long list of parsers that together describe a subject in a very accurate, precise, and computable way. It's like an encyclopedia but better, because the parsers are linked in ways that can be logically computed over.

This was a brief introduction to my primary mode of thinking: "thinking in parsers". I've personally found this a universally useful way to think about the world. The great thing about it is it works for both understanding objects in the 4D world and also for objects in the symbolic world (including 2D math).

When I was younger I was overwhelmed by the amount of symbolic knowledge we have in our world. Now that I'm able to think in parsers, I find it all far less intimidating, because I know it's just a big list of parsers, and I can take it one parser at a time.

I hope you find thinking in parsers as useful as I do.

The below is a chapter from a short book I am working on. Feedback appreciated

Particle Syntax (Particles) is a minimal syntax for parsing documents.

What is a syntax? Think of it like a set of rules that tell you how to split a document into components and what to label those components. Particles fits in the category of syntaxes that includes XML, S-Expressions, JSON, TOML or YAML.

Particles is unique in that it is the most minimal of these.

How minimal is Particles? Very! The file below uses everything in the syntax:

Particles divides a document into 3 kinds of things:

• Atoms
• Particles
• Subparticles

Atoms are just words. A word is just a sequence of visible characters (no spaces in words). The document above has 3 atoms: "post", "title", and "Particles". In Particles, atoms are separated by a single space. Particles has no understanding of what these words mean. Particles does not have the concept of number, for example. Every visible character in Particles is just part of an atom and that's it. If you have the line a b, that would be be split into 2 atoms, "a" and "b". If you were to add an extra space a b, that would now be 3 atoms, "a", "", and "b". So an atom can be of length 0.

The particles in Particles contain 2 things: a list of atoms, and optionally, a list of subparticles.

Subparticles are just particles that belong to a parent particle.

In the example above, the particle starting with "post" is actually a subparticle of the root particle, which is the document itself.

The particle starting with "title" is a subparticle of the particle starting with "post".

The particle starting with "title" has 2 atoms, "title" and "Particles", and zero subparticles.

The particle starting with "post" has 1 atom, "post", and 1 subparticle.

You make one particle a subparticle of another by indenting it one more space than it's parent particle. If we deleted the space before "title", then the line beginning with "title" would become the second subparticle of the root particle, and the line beginning with "post" would have zero subparticles. If we added another space before "title", that line would still be a subparticle of the "post" particle, and all that would change would be that we added a blank atom at the beginning of that particle's atoms list, and it would now have three atoms.

Now you should understand the basics of what Particle Syntax is. In the next chapter we'll look at why Particles is so useful.

The market is a blob of decision making agents that speaks with money.

You speak to the market with an offer.

The market gives you money, which means "Yes".

Or does not give you money, which means "No".

The no may mean the market:

• does not understand your offering
• does not think you can deliver on your offering
• does not want your offering
• does not like the price of your offering

The relationship between a human and the market is like that between a dog and the farmer. The dog can roam a bit but ultimately the farmer holds the power.

As a builder, should you ever mute the market?

To mute the market is to ignore the "Nos" and continue to invest in your offering. Instead of pivoting to something the market will say "Yes" to today, you work on something that the market currently says "No" to, but you have the belief that at some point the market will change.

To mute the market is to understand something about the world the market currently doesn't get, and to have the conviction that it eventually will.

I think as much as you can, you don't want to mute the market. Instead, listen to the market and think "okay, what is the market telling me it wants, and how can I deliver that in a way that helps me gain the resources to build what I want." Continue to iterate with creative offerings that are both something you want to offer and something the market wants to say "Yes" to now.

The selection pressure the market provides can be very helpful. You are getting the wisdom of the crowds to rank your ideas. I can definitely imagine that sometimes the market is too short-term focused and muting it is potentially the right way, but when you do that you also mute a great source of free feedback.

Despite my warnings, if you decide that a short term muting of the market is necessary, then you may have to go to extreme measures. Ideally, you've got a warchest. Perhaps savings, a good investment, investors who have bet on your vision, a grant, or something along those lines. With that capital you can mute the market for a period of time and try to invent a breakthrough new technological path. Without that capital the best advice comes from Hermann Hesse via his novel Siddhartha: "I can think. I can wait. I can fast." You will need to cut your expenses to the bone, and then use the bone for soup and cutlery. It can be done but you will have to push yourself to extreme measures you didn't know you were capable of. I'm not sure it's ever the right strategy. (But then sometimes I wonder whether it's the only strategy for true breakthroughs).

Along these lines, as an angel investor one thing I've learned is that you may be hurting a startup by investing, as you are enabling them to mute the market for a prolonged period of time. The pressure of the market is uncomfortable but honest and necessary. It's more relaxing to mute the market but you'll never find out if your ship floats nor carry passengers across the sea.

Occasionally I mute the market more than I should. I am building something that I think most of the world may love in the future, even if the short term feedback from the market has been crickets.

The offering I'm working on is novel, so the market hasn't quite understood what I'm offering yet. It's a challenge because the path is so different than what's popular in the market that I have to see many branches ahead on my own and then communicate what I see and why it's a path worth traveling. I think I'm getting better at explaining it, and I think once the market understands it, the market will want it. I think soon many people will not be able to get enough of it.

But I'm also listening closely to the market, and trying to create offerings that use our technology that are more in tune with what the market wants.

The scary part about muting the market is the branch you are pioneering may be a dead end. It is a lot of work to see so many branches ahead to figure out if it's a path the market should go down. You don't want to blaze a trail to a dead-end, but even worse would be to mislead passengers down a dead-end trail. But perhaps sometimes it must be done.

You learn to be deeply grateful and appreciative of the early adopters who break from the mainstream and come check out your fledgling path.

Whether it's the consumer market, the b2b market, the research market, or the investor market, communicating with the market is fun.

You get to create new offerings, put them out there in the marketplace, and listen to what the market says.

Stick to your convictions when you're confident in your model, but do your best to not mute the market.

Can you think about trust mathematically? Yes.

This makes it easier to design things to be more trustworthy.

You can think of anything and everything as an object that exists in 4D space.

The ideas below can be used to maximize trust of that object.

The amount you can trust an object is equivalent to the number of inspection vectors that intersect with that object.

You want to minimize the number of planes that block inspection vectors.

You want to minimize the number of complexity prisms that bend inspection vectors.

You want to maximize the number of overlapping inspection vectors of different magnitudes (aka your zoom range).

You want to maximize the time dimension of observability (version history).

You want to maximize the number of simultaneous inspection vectors.

You want to maximize the number of copies of the object.

You want to maximize the choppability of the object.

This is my initial list of guidelines for increasing trust.

A nerveshaker is a trick performers can use to kickstart a performance. Once the generative networks in your brain are rolling they keep rolling but the inhibitory networks can keep them from starting. So if you have tricks that just divert energy away from the inhibitory networks and provide a jump to the generative networks, you can ensure your motor will start. For example, to kickstart an essay sometimes I will start by inventing a new term. Such as nerveshaker.

Now that we have the motor running, let's turn to the matter at hand: judging our 2024 research results.

First, a note on pronouns (no, not that kind of pronoun). In this report I use the terms "we" and "our", rather than I, because I had tremendous help from over a hundred people throughout the year. Although I was not an "employee" of any lab, nor did I have "employees", I received significant help in many forms, from code and data contributions, to code and paper reviews, to lots of in person meetings and presentation opportunities, to resources including office, lab, equipment and housing, to even direct cash contributions.

If the work done in 2024 grows into meaningful things, it is because of the help and spirit of these people throughout the world, who have believed in (or at least humored) what we are trying to do, and have helped advance things, despite my underwhelming leadership and engineering capabilities. So, that is why I mean it when I say this is a report of what we published.

Second, a word on why this report is on research performance and not business performance. This answer is simple: business performance was abysmal. I started the year with about $70,000 and ended with a bit less than $0. If I were just to report on business performance it would just write a single letter: F.

But, luckily (or delusionally), for me, there's another perspective to judge our 2024 results and that is the expected value of what these things might lead to.

And I think that the expected value of what we did in 2024 will be very, very high. Let's take a look.

Our top innovation, IMO, is ScrollSets. ScrollSets are a new simple alternative to knowledge bases that only require a simple plain text file. In 2024 we finally refined all the little pieces to make the whole thing just work. In 2025 we should be able to make them about 10-100x faster and then I think it will be off to the races.

I think ScrollSets may evolve into being one of the premier tools for truth. They allow you to take very small, simple steps, to build a very large matrix where all of the vectors intertwine so that as it grows it becomes very difficult to hide lies.

Money is a simple and available metric, even if suboptimal. If we look at the monetary impact this could have, well it's pretty big. On the order of $100 billion is spent on databases annually. So creating a potential game changer for <$70,000 I think is promising.

Building a ScrollSet about a topic is like building a compass that guides you to a true model of a topic. I used an early version of ScrollSets to build BrainDB, to solve a mysterious energy fluctuation I had experienced most of my life (erroneously called 'bipolar disorder'), that was said to be incurable. ScrollSets lead one to a thorough model, and eventually I stumbled upon a small group of researchers who were looking at this condition as a mitochondrial condition, caused in large part by a modern carb-heavy diet. After over a year of experimentation and a significant amount of research and microscopy I can now say with high confidence that this looks like exactly what it is, and we are now in the process of building a large image dataset of human mitochondria under the microscope to show this definitively. Again, money is a crude metric, but a simple one, and the estimated total annual global cost of "mood disorders" is on the order of $1 trillion. Making a significant contribution to a cure for $70,000 I think is a good ROI.

Those are the two projects that I think will lead to the biggest long term impact on the world. And we're hard at work on taking them to the next level.

In addition, we put out a lot of other stuff that is also promising:

In addition, some fun side projects including:

As always, everything done was published to the public domain. I believe this is the way all research should be done.

~80% of expenses went to lodging, food and transportation. I would like to cut those daily outlays in half. The reason it was so high is that my personal situation still has a conflict (I'm an open book and fine to share that my ex and her conspirers are still dishonestly persecuting me daily with secret motives. It is what it is. I don't try to focus on it but I don't want to hide anything). If I knew how long it was going to take to get things humming, I may have taken harder cost cutting measures earlier.

In 2025, I would like to put every penny of expense on a public ledger to provide more transparency and commitment to our backers that we are allocating resources wisely.

In 2024, as in prior years, I was rejected from pretty much every source of funding I applied to. At one point I did have a handshake deal for $1M which would have accelerated things, but that fell through for some reason (I never did get a reason), which led to a challenging (but educational!) last couple of months.

I hope some philanthropists and/or funders of research will take a look at what we did in 2024 and consider helping us do even better work in 2025.

Thanks for reading.

-Breck

I sit with my coffee and wonder, can I explain the general pattern of Mathematics on a single page?

My gut says yes, my brain says maybe, and so I shall give it a go.

I aim to explain the essential concepts of the Mathematics humans write on rectangular surfaces. In order words, I'm limiting my scope to talk about 2D written mathematics only. Even written math is out of scope, as I've already shown that in earlier work.

My claims are limited to that scope, though you may find these concepts useful beyond this.

I need to admit that although this essay shall not exceed a page (including illustrations), I am undercounting the length of my explanation of Mathematics because I am ignoring the considerable requirements the reader must have developed to be able to parse English.

Ok. No more caveats. Time for the essential concepts in Mathematics.

Pixel Surface

Mathematics requires a 2D contiguous grid of Pixels whose color can be altered.

Cells

Mathematics requires the ability to draw or envision membranes around contiguous groups of pixels to form Cells.

Atoms

Mathematics requires Atoms, which are single indivisible cells with colored pixels.

Particles

Mathematics requires Particles, which are contiguous regions that can contain atoms and other particles recursively (called subparticles).

Holes

Holes are Particles that contain an identification atom and a list of atoms that fit that hole.

Parsers ★

Parsers are a type of particle who's atoms and subparticles define holes. Parsers are pressed against other particles and bind to particles who's subparticles and atoms fit a parser's holes. I put a ★ next to Parsers to indicate that these are the thing you should study to understand this system.

Transforms

Parsers can contain Transform definitions, which are Mathematics that transform one particle into another.

These meta concepts are enough to define all of Mathematics.

Here's to the man

Who works with his hands

And regardless of man's plans

Does the best work that he can.

He makes his cuts neat

He applies correct heat

He makes his joints meet

He makes things on beat.

His labor has cost

But the buyer is not boss

The system may add loss

But honest work is never lost.

Pay may be flawed

Cheating may be lawed

But his spirit ungnawed

He presents his work to God.

To understand something is to be able to mentally visualize it in motion in your head. Let's call this a truthful model.

People try to communicate these models in words. At best, these models will be lossy. Often, they are downright fraudulent.

There is a ton of hard earned truth out there, unfortunately it's unevenly distributed and often buried under lies.

Whether the model you are given is a story made of words, or a film documentary, or a 3D simulation, how can you gauge its accuracy?

I developed a technique that is very simple and works very well.

It involves writing down simple facts in a simple way that easily converts to a spreadsheet.

No matter what the thing you are trying to understand is, the trick to getting closer to truth is to build a spreadsheet with:

More:

1. Observations
2. Dimensions
3. Orthoganality

Observations are the rows in your spreadsheet.

Let's say you are trying to understand the moon. If you only took 1 "shape" observation in the middle of a lunar cycle, you might say the moon is a half circle.

If you take multiple observations your model of its shape will get closer to a true model.

Dimensions are the columns in your spreadsheet.

If you not only measure the moon's shape but also it's position in the sky you will get closer to a true model.

Orthogonality is a measure of how much redundancy there is in your observations.

If you measure the moon's shape at 100 slightly different times a day from 100 slightly different locations, you've increased the observations and dimensions a lot without increasing your understanding much.

If you were to plant a temperature sensor on the moon or add images from a high powered telescope, you are adding orthogonal data that improves the truthiness of your model.

The idea is you want to not just make many measurements from many angles, but you also want to look at things from wildly different angles. These different perspectives can often be critical for preventing cherry-picked datasets that present overly simplified or misleading models.

The key thing to keep in mind that mining information has costs. It is much cheaper to add more information that looks at things from similar perspectives of existing information. So to arrive at truth you must make an extra effort to add orthoganal perspectives.

You can use this method to build a model to understand anything, no matter how complex.

Of course, the human brain has a limited context window and you can only work on a little bit of your model at one time. To solve this we developed a technology called ScrollSets that let's you chip away at building a model of anything. ScrollSets let's you incrementally build a model, adding as many or as few concepts and measurements at once as you want. Everything is simple plain text, fully tracked by version control, and compiles to a spreadsheet. When a new idea strikes for increasing the orthogonality of your model, it's very easy to add.

"It's more comfortable than it looks."

"And this just v1. By v10 it will be as lightweight as a surgical mask."

"And how long does the gluon encoding last?"

"Thousands of years. Once you've tagged an atom, it's yours for life."

"And the lives of my ancestors."

"Royalties for generations!"

"And how high can you get the coverage?"

"In theory, above 99 percent. The model you're wearing now is doing 5%."

"Incredible. So how many atoms am I claiming right now?"

"Fifty quintillion. Every time you breathe out, the nanopores in the mask gluon-encode fifty quintillion atoms."

"Hey, I'm working hard to make this CO2, I should get paid for it."

"And now you will."

"How much information in each neutron?"

"256 bits. Enough to fit a wallet id.

"I've been waiting years for a business like this."

"Well we've finally built it."

"And tagged air is safe?"

"Scientifically proven. We conducted a rigorous 3 month controlled experiment and observed no harmful effects. Actually, our results are about to be published in Nature."

"Brilliant. What's your go to market plan?"

"Move fast. Create a frenzy. 'Right now 99.99% of oxygen is unowned. Tag your share, before someone else does.' We let people know there hasn't been a land grab like this since the Great Western Expansion."

"How long until the whole atmosphere is tagged?"

"A few decades. It will go slow at first, but once it catches on we expect people will travel to the ends of the earth to find wild neutrons."

"And the revenue model?"

"You get paid when someone breathes your oxygen; you pay when you breathe someone else's. We add a transaction fee on top. The mask tracks it all. We send you a statement each month."

"So some people will turn a profit?"

"Just a few heavy breathers, yes. We've modeled it out. Oxygen rights are reassignable, of course, so we expect most breathers will actually sell their rights to us quite early, and quite cheap."

"So eventually we'll own all of the air?"

"Most of it anyway. Someday we'll monetize almost every breathe. Everyone will be a subscriber, eventually. The greatest business model ever invented."

"But once people are paying more than they're making, what's to stop them from just taking off their masks?"

"We'll make that very hard. Huge PR campaigns. We'll promote the superiority of tagged air versus untagged air. Film, shows, books, schools especially, we'll ensure everyone is taught from an early age that tagged air is the way to go. It will be ubiquitous yet subtle."

"You can also make it capitalism vs communism."

"Absolutely. Shared air is a communist idea. If you're against tagged air, you're against property rights."

"Tagged air for safety?"

"Yes! We forecast that untagged air will increasingly be blamed for more and more incidents. We expect poisonings and other tragedies. I wouldn't be surprised if someday the only ones breathing untagged air are terrorists."

"Is this something governments would support?"

"A government's dream! Imagine laws requiring masks. No one can take a breathe without the government knowing about it. Total control of the air."

"Amazing."

"So are you in?"

"I'll be honest, this is the best presentation I've seen in my career. I'm in. Let's talk valuation."

"Great to have you aboard."

"Do you mind if I keep this one?"

"I hope you never take it off."

"I tell everyone I'm a value-add investor."

"You're helping build a new world."

"Something our ancestors failed to do."

"Here's to tagging all the air!"

There is a job that is currently illegal in the United States, to the great detriment of our citizens. That is the job of "Information Cleaner."

An Information Cleaner is a person who takes in all the material being published in our information atmosphere and cleanses it: they make it transformable, searchable, modifiable, accessible, free of ads and trackers, auditable, connected to other information where relevant, and so on.

These people are not primarily focused on the production of new information, but rather on cleaning and enhancing the information that has already been produced.

This is a hard and extremely important job-think of it like back propagation- and it's currently made illegal by copyright law. As a result, our information environment is as dirty and toxic as an aquarium with no filter.

Clean information is bits encoded as simply and noise-free as technology allows. Clean information is easy to move and to copy. Clean information is easy to search. It is easy to dice and remix. Clean information has had all toxins removed, such as ads and trackers (or at the least, it is in a form where those can be easily removed). Clean information comes with provenance information. It comes with hashed change information.

Clean information is not Netflix. It is not Prime Video. It is not Disney Plus or Nature.com or The New York Times. All of those are dirty information. Information with DRM; information claiming to be "licensed"; information with a paywall; information without source code; all of this is dirty information. And Americans should be allowed to clean it up.

Now some modern day heroes are clandestine Information Cleaners, building and expanding projects like LibGen and Anna's Archive and archive.today. These people are secretly keeping civilization from tumbling into a dark age.

While they risk their lives and liberty to prevent civilization from collapsing into an information-controlled dystopia, some of us need to be proselytizing in public and making the case as to why Information Cleaning should be a root right, enshrined in the Constitution and revered at the same level as freedom of speech; freedom of the press; freedom of religion.

In the past 50 years Americans were misled to think it was the quantity of information that was the thing to optimize for. This is false. The thing to optimize for is the cleanliness of information. It is far better to have the infrastructure in place to clean information, rather than to produce information. It is vastly easier to produce valuable new information once you've cleaned up all existing information.

In that sense, the way to properly incentivize the production of new information is to make legal the cleaning of old information.

If we want to make our air clean, if we want to make our food clean, if we want to make our bodies clean, we first have to make our information clean.

• Owned Air (2025)
• All the benefits of property; none of the responsibility (2025)
• Aaron's Amendment (2025)
• Optimizing for Truth (2025)
• A New Coat of Paint (2025)
• The Permission Function (2024)
• Writeable (2024)
• Copyright is a mind virus (2024)
• Patents are Poison (2024)
• ETA!: A Measure of Evolution (2024)
• IPDD (2024)
• The Omni Restaurant (2024)
• The law favors ChatGPT over humans (2024)
• Three Scientists Go To Heaven (2024)
• Why do we subsidize lies? (2024)
• A Mathematical Model of Copyright (2023)
• Enchained Symbols (2023)
• A Public Domain Human (2023)
• SafeFS (2023)
• Training a neural network to spot misinformation, fake news, and propaganda from a single image (2022)
• Cancer and Copyright (2022)
• How the public domain can win (2022)
• The Public Domain Publishing Company (2021)
• An Unpopular Phrase (2020)
• Dataset Needed (2020)
• Dreaming of a Data Checked Language (2020)

If a pedestrian on the sidewalk is hit by a falling branch from a decaying tree on your property you are liable.

If those who insist on calling copyrights and patents "Intellectual Property" wish to continue to do that, I say we make them embrace all the responsibilities of property as well.

• If © TV programs carry ads for sugary drinks, and sugar turns out to cause many diseases, those TV networks are also liable.
• If actors and actresses star in those © ads, they are also liable.
• If newspaper © pages have ads for painkillers that turn out to be far more addictive than advertised, the paper is also liable.
• If a © textbook contains a model of health that turns out to be inaccurate and harmful, then the publisher(s) and author(s) are also liable.
• If a © song is played on a radio station in between ads for products that turn out to cause harm, the station and the musician(s) are also liable.
• If a non-public domain search engine shows ads for products that turn out to cause harm, that search engine is also liable.
• If a © software program causes its users to lose a signficant amount of time or resources, the software maker is also liable.
• If anyone claiming © over some media fails to update that media as soon as mistakes are discovered, they are liable.

Or do they just want all of the benefits of property rights, with none of the responsibilities?

• Owned Air (2025)
• Information Cleaner (2025)
• Aaron's Amendment (2025)
• Optimizing for Truth (2025)
• A New Coat of Paint (2025)
• The Permission Function (2024)
• Writeable (2024)
• Copyright is a mind virus (2024)
• Patents are Poison (2024)
• ETA!: A Measure of Evolution (2024)
• IPDD (2024)
• The Omni Restaurant (2024)
• The law favors ChatGPT over humans (2024)
• Three Scientists Go To Heaven (2024)
• Why do we subsidize lies? (2024)
• A Mathematical Model of Copyright (2023)
• Enchained Symbols (2023)
• A Public Domain Human (2023)
• SafeFS (2023)
• Training a neural network to spot misinformation, fake news, and propaganda from a single image (2022)
• Cancer and Copyright (2022)
• How the public domain can win (2022)
• The Public Domain Publishing Company (2021)
• An Unpopular Phrase (2020)
• Dataset Needed (2020)
• Dreaming of a Data Checked Language (2020)

New: let's get to work! Join the subreddit

February 7, 2025 — If you've thought deeply about copyrights and patents, you've probably figured out that they are bad for progress and deeply unjust. This post is for you.

(If you are new to this issues, you might be more interested in my other posts on Intellectual Freedom)

I suggest we organize around a simple long-term vision of passing a new Amendment to the U.S. Constitution ending patents and copyrights once and for all.

The below proposal is 34 words.

Section 1. Article I, Section 8, Clause 8 of this Constitution is hereby repealed. Section 2. Congress shall make no law granting monopolies on ideas, knowledge, or inventions, or prohibiting the free use thereof.

I have only passed a handful of Amendments to the U.S. Constitution in my lifetime 😉, so discussion welcome.

I am not 100% certain that if we abolished copyright and patent systems the world would be a better place.

It would be intellectually dishonest of me to say that. I am always open to intelligent experiments that would show otherwise.

But at this point I am 99% confident it would be the single most massive positive improvement we can make in our world, based on empirical evidence and theoretical math.

It would take a lot of thought to do it right, but I know we could pull the transition off without as much disruption as people fear.

The bigger problem is this debate is not being had.

The problem is our side needs a better starting position.

When the debate is on details like what is the ideal length of monopolies, or when illogical terms like "Intellectual Property" are used, you've already conceded too much, and are giving up your strongest weapon: truth.

A stronger and more logical place to have the debate is upstream of that: debate whether we should have these systems at all.

I think the Amendment Strategy is clear enough, concrete enough, simple enough that you could get critical mass and start moving the debate upstream.

The best defense is a good offense. It's an adage, but there's usually some truth to adages.

The kind of people I think may be ready to organize would be lovers of open source, Linux, Sci-Hub, the Internet Archive, OG Napster; the followers of Aaron Swartz, Alexandra Elbakian and Stephan Kinsella; and all that truly love ideas and believe every human should get their own copy of humanity's most intelligent information.

If our government is going to make laws governing information, then we should optimize for truth and signal, over lies and noise.

The objective should not be maximizing the ability to make money off of information. Nature provides natural incentives for discovering new truths, we don't need any unnatural ones. In fact, the unnatural incentives on information production actually incentivize lying and noise, rather than truth generation.

I'm surprised this is such a minority opinion, but very few people are with me on this (those that already are---❤️).

What kind of a system maximizes truth?

Well, what is truth?

Truth is when someone publishes a set of symbols with the claim that they accurately predict something about the world and then later sensors verify that those symbols did accurately predict it.

These truths are extremely helpful. They give us warm buildings, useful electricity, cures for disease, lenses to see more, safe transportation, and so on.

There are also kinds of information that might not necessarily make accurate predictions about the world but don't pretend to. Fictional stories or songs or jokes meant to amuse. These are fine and also have natural incentives (the love and admiration from your peers, for example).

Then there are lies. These are symbols that claim to predict things about the world that don't, in fact, hold up when the sensor data comes in. Much of advertising falls in this category.

Finally, there is also noise. Noise is often truths repackaged in extremely verbose, obfuscated, or scrambled order that wastes people's time and can mislead.

Other than total censorship, I cannot think of a worse information policy than the one we currently have in this country, where it is not legal for someone to edit published information and republish their edited versions. Where it is illegal for someone to create a repository of maximal truth.

We need smart people to delete all this noise, to distill all the signal, and deliver truthful, efficient information to the public. This needs to be legal, not illegal.

Truth needs to be the thing to optimize for, not royalties.

It's time for the IFA.

If you could have an AI running continuously on one prompt, what would it be?

Here is my best idea so far. What is yours? Ideas and criticisms encouraged!

Your job is to build an encyclopedia of all knowledge.

Humans should be able to copy what you generate to ink and paper for long term robust storage.

It should define symbol encodings by listing binary sequences with their matching symbol.

It should define atomic concepts out of those symbols.

It should define higher level parser concepts by listing patterns that match against and transform rules.

It should define 2D, 3D, and 4D bitmap formats that resemble phenomena humans observe in nature and link those to the defined concepts.

It should contain knowledge to maximize what it can predict in real world observations (like location of a planet), and not worry about predicting symbolic world observations (like winning lottery numbers or name of a pop star).

Everything should have coordinates in a 4D space.

Continually refine your encyclopedia until it is encodes all knowledge in as few bits as possible.

And our ancestors built a magnificent palace, over thousands of years, layer by layer, with space for all, and our brother added a coat of paint, and said "This is my property now."

There is a correct answer, and that answer is zero.

The progress of arts and sciences depends on rapidly testing novel assemblies of ideas to see if they are good ideas or bad ideas.

If you have a concept of copyrights and patents, then before testing any new assembly of an idea, you have to ensure you have permission to test the idea.

The Permission Function compares your idea with all others to ensure it does not violate any copyrights or patents.

If you run afoul of the Permission Function, you can suffer great legal consequences.

The Permission Function take times and energy to run comparisons, communicate with "owners", and perhaps find workarounds.

The problem is, you don't know how good an idea is until you test it!

You often need to test 100 new ideas to find 1 good idea.

Thus most of your resources are wasted on running the Permission Function on bad ideas.

When you understand the colossal waste of time and energy the Permission Function introduces, you realize that even setting the term of copyrights and patents to 1 year is incredibly harmful, and slows down progress by orders of magnitude.

When the term is 0, there is no Permission Function, and the progress of arts and sciences is unchained.

There is no way AGI will be wasting all this energy on a Permission Function.

Any AI that does that is doomed to extinction.

Similarly, to survive, humans need to abolish the Permission Function.

What does it mean to define a language?

It means to write down a set of restrictions.

Languages let you do things by defining what you can't do.

Languages narrow your choices. You can still take your reader to a land far far away; you just need to take them in practical steps.

Communication requires constraints.

Too few restrictions and someone can write anything, yet communicate nothing.

Even plain text files, which may seem at first to have no rules, have many.

The file you are reading now is a plain text file.

But there's little "plain" about it. It actually is a UTF-8 file. UTF-8 is a language that restricts you to writing sequences of characters, each character represented by a sequence of zeros and ones.

You cannot draw pictures or make sounds or make images or make colors with this language.

Compared to all the messages you could write on a 2D surface, the number of messages that the UTF-8 language allows you to write is tiny.

We should probably distinguish between languages which are defined via definitions and languages that have just evolved undefined alongside us for thousands of years.

Let's use the terms undefined languages and defined languages.

Most of your day is spent using undefined languages.

Undefined languages are the original languages. The languages spoken by the cavemen, the nomads, our earliest ancestors; these are undefined languages. The language a baby speaks to its parents, the language a rooster crows to the hens, the smiles of two soulmates reuniting at an airport, these are undefined languages.

While many publishers create books claiming to Define the English Language, the truth is that English is an undefined language, and that unlike computer languages, the restrictions of English are all fuzzy.

Defined languages are a modern invention, but far older than computers.

Perhaps a legal, or maybe an accounting language, were the first defined language.

Before those, all written languages were defined only by the physical limits of the materials used. You could write any marks, as many as you had the materials for, as big or as little, as distinguished or indistinguished as you'd like.

But it turns out such total freedom had a disadvantage when it comes to communication.

If you want to communicate something specific to your reader, you need to understand what parsers they are going to parse your writing with.

Think of a parser as a memory.

You can't see something you've never seen before. At the very least, it has to be at least composed of things you've seen before.

Likewise, you can't read something unlesss it's composed of things you've read before. So your brain is composed of a large number of parsers, of memories, that are like the restrictions of a defined language.

If your language has no restrictions, then there are no parsers for your readers to memorize, and messages in your language don't trigger anything specific at all. Every reading would trigger a random effect in the reader.

For art, this might be desirable, but for conducting construction, science, medicine, commerce, government; not so much.

So some people started to create defined languages. Perhaps the set of restrictions were communicated orally at first. But then these rules were marked onto a surface.

And then over the centuries many thousands of defined languages were created for math, accounting, law, and so forth.

Then, the invention of electronic computing machines in the 1940's and 1950's led to the creation of thousands of new defined languages over the past 70 years.

Each language defined its own set of restrictions (and thus, freedoms).

It's been a survival of the fittest as to which set of restrictions works best in practice.

It is in this competitive milieu that we are defining our new language stack: PPS.

Back to the original question, why are we defining a new language?

Well, we stumbled upon a simpler set of restrictions that we think make our languages far more efficient, powerful and trustworthy than the competition.

The languages that are dominant today have a lot of unnecessary restrictions, and miss out on a few simple very useful restrictions. These make them unnecessarily complex, inefficient, and not so trustworthy.

When all your messages contain unnecessary pieces, you have zero chance of ever arriving at the perfect message. You're doomed to make suboptimal models from the start.

Our language stack gives humans a chance to build as close to perfect a model of the world as you can get.

Unnatural additions and inefficiencies stick out like a sore thumb. This makes it a compelling choice for governments, engineers, scientists, and so on.

The new set of restrictions in the PPS stack allow us to do the same things in more natural, simpler, more efficient ways.

Explore more here.

So yeah, mitochondria are very important.

Everything goes in cycles.

So it surprises me there's not yet a word "mitocycle".

Mitochondria were first seen with optical microscopes in the late 1800's.

In the 1950's electron microscopes gave us our first good look at them.

Now we know that they are extremely important for health.

And that a lot of modern foods and lifestyles have proved harmful to them.

When will we have mitometers?

What is the optimal population?

Are your mitos high right now, or low?

Are they wiggling fast or slow?

What are their natural cycles?

I believe the term "mitocycle", will be important in the future.

Look at these guys wiggle! We are wiggling worms!

We're building a new blockchain (L1).

This chain will leapfrog all others on trust and speed.

We will be misunderstood, ignored, and mocked.

And then, our math will win.

-Breck

Alejandro writes of an exercise to describe something in two words.

To him Scroll is "semantic blogging".

His Scroll blog is "a database of [his] knowledge."

He then asks what two words I would use for Scroll?

While Scroll is also semantic blogging to me, if I could only pick two words I'd go with: particle thinking.

I like to ponder life's big mysteries.

Why are we here? How do things work?

Figuring out life's mysteries is like moving a mountain.

How does one move a mountain?

Two approaches come to mind: the philosophical approach and the engineering approach.

The philosophical approach is to change your question or perspective. A mountain from a different perspective can be a pebble, coverable with one thumb.

The engineering approach to moving a mountain is "bit by bit".

Particle Thinking is the bit by bit way.

All of these words you are reading now can be constructed with 2 particles: a particle and an anti-particle. You then use many of those as subparticles to build letter particles, and then combine those to get word particles, and then list particles, and then parser particles, and then language particles, and then encyclopedia particles, and program particles, and operating system particles, and lowly blog post particles like this.

It's particles all the way down.

What's so good about this approach?

Learning is programming except instead of creating functions in code you are creating functions in neurons. Like programming, a poorly created function can be vastly slower than a well crafted one. Evaluating mental functions takes time and requires energy. If your neural functions are wasteful you will compute slower, miss optimal thoughts, and waste a lot of ATP.

If instead, you refine your thinking to where you understand exactly what concepts are necessary and what are not, you will think faster, come up with better ideas, and have more ATP to spend on enjoying life.

Do this for years, and then decades, and your savings compound.

So those would be my two words for Scroll: particle thinking.

I want a language that helps me be a great thinker, to get the most out of my limited neurons, and use neural ATP most effectively.

And I want a language with a community of other people that think like this.

And I want a language that gives me a another way to try and move mountains.

What do you think? What two words would you use to describe Scroll?

Every accomplished person I've studied has a quip about focus.

Ability to focus separates humans from animals.

Humans have dug 1,000x deeper than the deepest animal and flown 40,000x higher than the highest bird.

And among humans focus separates the lay from the legends.

Why does focus pay?

Because payoffs in nature are non-linear. And it's finisher take all.

If you get 99.9% of the way across the sea you get 0% of the benefits. That last 0.1% is worth everything.

Sustained focus allows humans to hit otherwise unachievable payoff points.

To pass a payoff point you need sufficient pressure.

Your force supply is limited so you need to concentrate it at a small area.

How do you measure focus?

Identify the one thing you should be focused on.

Our universe has made everything vastly unequal so you should always clearly see one thing far bigger than others.

Now measure what percentage of your thoughts are on that one thing.

Resist distractions. You can blur and attend to other matters once you've finished your one thing.

What about blurring?

Don't worry about blurring. Worry about underfocusing.

I've never met a single human who has overfocused.

Everyone overblurs.

We have dozens of biological cycles that force us to blur.

You blur every time you take a walk, go to sleep, take a pee, eat a meal, et cetera.

Focusing is the thing to practice.

Mania is too much mitochondria; depression too little.

Mitolevel is mitochondrial volume divided by cell volume (ML = M/C).

Mitolevel varies by cell type.

The mitolevels in certain cell types will be strong predictors of current system wide mood state.

Mania is elevated mitolevels and depression is depressed mitolevels.

Individuals who have experienced mania frequently report a feeling of quickening or acceleration, which matches this model as self-reproducing mitochondria change population exponentially over time.

Individuals who experience severe depression take a long time to recover, which matches a model where the cell is depleted of mitochondria (likely from dead mitochondria from resource exhausition during a manic episode) preventing the restoration of a healthy mitolevel.

Mitochondrial populations change much more gradually than substance levels in the bloodstream, which explains why mania is described as a "sustained high" and why depression can't be immediately "snapped out of".

1. Abundant lab evidence. Someone will discover a human cell type whereby one can reliably and abundantly predict mood-energy state of the whole person simply by measuring the mitolevels of a cell sample.

1. At-home mitolevel tests. Someone will organize a large number of people in different mood-energy states to contribute to a large dataset of microscopic images of mitochondria from various cell types and then, using this reference, cheap at-home tests of mitolevels using simple staining, consumer microscopes, and machine assisted counting will be developed.

1. Millions cured. High fat, low glucose diets will cause stabler mitolevels and radically reduce episodes of mania and severe depression

1. A Universal Condition. Any human being can experience mania and depression through manipulation of their diet and/or environment to force high or low mitolevels.

As one can assume, hundreds of works directly contributed to the model above. Some of the key resources:

• Dr. Iain Campbell
• Metabolic Mind
• Bipolar Cast
• The Cell, Chapter 7, Don W. Fawcett. 2nd Edition. 1966
• Chris Palmer
• Shebani Sethi

Questionable Engineering

Inequality

Focus

Constants

These 0.00000000001% atoms rule the world

Cannibals

Circle of Life

Data

Billionaires

Made with: tldraw; Apple Pencil, iPad Pro, and MacBook Pro; feedback from ChatGPT. Newest comics made with Daylight Computer.

If you could see everything across space and time, we would not need to read and write.

You would not need words to tell someone how blue the Pacific can be because they could just see it.

You would not need words to learn when your grandfather decided your grandmother was the one because you could just see everything he saw.

But we can't see everything.

We can only see what's in front of us at this moment.

This is not optimal for our health.

Our Universe is mostly mystery, but at least we know there are patterns.

And if you can learn these patterns, you can have a better time.

To learn a pattern you first must see it.

Our ancestors could only see with their eyes.

Some lucky descendants evolved the ability to imagine things that weren't there.

Language requires imagination.

Once we could imagine, symbols could evolve.

A discoloring on a surface.

A mark to communicate a pattern.

What was the first mark?

What was the first symbol?

Let's imagine.

A man sees a cave.

It is cold and getting dark.

He approaches, but then lurches back.

He sees a bear!

He holds his spear and prepares for battle.

But the bear does not move.

He looks closely.

It is not a bear, but simply charcoal discoloring the wall.

He remembers the shadows his companions make around the campfire to communicate they saw a bear.

He thinks perhaps one of them rubbed charcoal to make a permanent shadow.

He finds a different cave.

In the future, when he sees bears, he grabs some charcoal and repeats the symbol.

Perhaps it was something like this that kickstarted writing.

Pictures at first, then pictographs and eventually phonetic alphabets where infinite patterns could be composed and communicated.

Tens of thousands of years later, here we are with huge numbers of people who spend their days drawing charcoal bears.

Our tribe now uses writing (or at least, pretends to use writing) to set the rules.

These include rules about writing.

What should the rules on writing be?

What is writing good for?

Is it the writer that is most important, the symbols, or the reader?

Writers have clear incentive to create symbols since that very first symbol: symbols save lives.

Even if it's not your own life, or your child's life, even if it's a complete stranger's life, we have learned that a civilization of strangers cooperating peacefully create a better world for all.

But writing goes out of date, and if the writer doesn't update their symbols, they can mislead as easily as they can lead.

It can't be the writer that is most important.

The symbols themselves, without humans, are literally meaningless. So it can't be the symbols that are most important.

Thus it is clear the most important thing is that the rules around symbols should benefit the readers.

What can we do to equip readers with the best symbols possible?

What if bears learned to write?

We might not be here.

Our ancestors might have seen an idyllic image of a cozy cave and walked right into a bear's mouth.

Lies kill.

Can we protect ourselves from writing bears, from lies?

How do we make the most from symbols?

The trick is to make symbols as honest as possible.

And how do we do that?

Bears want nothing more than to outlaw writeability.

Bears do not want you to spread the truth about their cave.

They don't want you to know what's really inside.

And so to do that, not only do they lie about what's inside, but they make their wall unwriteable.

They do not allow edits. They do not allow transformations.

Bears make symbols to serve themselves, and themselves alone.

And by doing that, they lead people to slaughter.

To make symbols the best for readers, symbols must always be writeable.

All symbols should be writeable.

Do not trust anyone who wants to restrict your rights to write.

Everything I own is in the overhead bin.

This is cool if you are 20.

But I am 40. I have two kids.

At 40 this is a bit extreme.

The kid next to me is 20.

She is a computer science major.

We talk about languages.

I play dumb and mostly listen.

She is wondering which language to use.

Her ML class uses Python.

Her OS class, C++.

For fun, she uses Javascript.

I suggest she enjoy them all.

It's like traveling.

Learn the place in front of you.

You will visit many.

The topic comes to my work.

I used to be shy about it.

Terrified of misleading.

Not anymore. I demo.

Scroll is very practical.

We make useful things fast.

Right there on the plane.

With no internet.

I explain Scroll is like a game.

To play you put a subject into symbols.

To win you cover it all with no extra parts.

The prize is you've figured it out.

To pack so light I've filtered objects.

Kept only the most useful ones.

To write Scroll is to filter symbols.

To keep only the most useful ones.

October 29, 2024 — The censors do not want you to read this post.

They will downvote, flag, takedown, block and use any and all means to suppress these ideas.

They will call me names: a crank, crazy, unhinged, conspiracy theorist, lunatic.

They will say he is not a creator (I am), has created no hits (I have) and whatever else they can do to suppress this, short of responding to the ideas presented here.

Because they know that if they respond to the ideas then they will encourage you to actually think about them, and they know that if you start to think about them they've lost, because they know what I'm telling you is nature's simple truth.

That terrifies the censors. They know the stickiness of truth. They know that truths can be hard to see, but once you've seen truth you can't unsee it. Brains prefer shortcuts. And truths are shortcuts for predicting the world.

I'm giving you a shortcut to explain many of the problems in modern America: unnatural inequality, poor health, declining life expectancy, violence, poverty, misinformation, fake news, distrust in media, et cetera.

What is a shortcut to explain the root cause of many of these health problems?

Copyrights.

Copyrights they tell you are helping us are actually hurting us.

Every single great idea started as a bad idea that evolved in a competitive environment into a great idea.

This applies to any and every great work that has a © slapped on it: every song, book, article, photo, movie, program, et cetera.

No idea ever reaches perfection.

Every idea can always be improved. Every idea can always keep evolving.

What does copyright do? It makes evolution a crime.

It says: no more evolving! This idea needs to stay exactly as is until 70 years after the "owner" dies!

This is literally the dumbest idea I can think of, short of total censorship.

But copyright is even worse than just being a terrible idea: it's a terrible idea that has spread like a virus.

For example, 9 out of 10 of the top websites in the USA spread the virus to all their visitors. And the one that doesn't have the virus on its homepage, Google.com, has the virus on most of its other pages.

You can listen to every argument made by the spreaders of this virus, but you will not find a single honest one.

All of their arguments are mathematically flawed.

Their best strategy is not to actually convince you they are right-they know they are wrong-but to delay you from seeing this obvious truth as long as possible, because they make so much money every year you don't see it.

If you want to read honest material about this, here's a place to start.

You can help! If you run one of these companies, start cleaning up your house! Remove all © symbols from all pages. Remove all uses of the word "license" when talking about ideas. Make it clear that everything you are publishing is public domain.

If you are just an average citizen, you can help too! Tell everyone you know about the harmful effects of the © mind virus. If you have technical skills, start sending pull requests removing the © virus from open source projects.

Together we can rid the world of this mind virus and create a world where bad ideas are once again forced to evolve into good ideas.

October 27, 2024 — The censors do not want you to read this post.

They will downvote, flag, takedown, block and use any and all means to suppress these ideas.

They will call me names: a crank, crazy, unhinged, conspiracy theorist, lunatic.

They will say he is not an inventor (I am), has no research background (I do) and whatever else they can do to suppress this, short of responding to the ideas presented here.

Because they know that if they respond to the ideas then they will encourage you to actually think about them, and they know that if you start to think about them they've lost, because they know what I'm telling you is nature's simple truth.

That terrifies the censors. They know the stickiness of truth. They know that truths can be hard to see, but once you've seen truth you can't unsee it. Brains prefer shortcuts. And truths are shortcuts for predicting the world.

I'm giving you a shortcut to explain many of the problems in modern America: the opioid epidemic, cancer, obesity, mental health, the "pandemic", et cetera.

What is a shortcut to explain the root cause of many of these health problems?

Patents.

Patents they tell you are helping you are actually poisoning you.

First, there are some points I think we can all agree with. Patents are awarded to novel inventions that have plausible utility. Patents encourage the publication of those inventions in some form. Patents are effective at discouraging unlicensed competitors and do create monopoly pricing power.

I agree with all those points.

But now let me present a new way to look at things.

A patent is embodied as a digital PDF. A PDF is just a long binary number.

Thus one can accurately say a patent is a long, seemingly random number.

The patent "owner" makes money by people paying to use that random number.

Patents do not incentivize solving problems. Patents incentivize convincing people they need your random number.

If you've got a patent on a random number, you make no money if people solve their problems without your random number, but monopoly money if they use your random number (regardless if it actually solves their problem).

As a result, companies are coming out with all sorts of random numbers that are great at generating monopoly profits, but terrible at solving people's problems.

Even better than just making monopoly money from one random number is to make monopoly money from a suite of random numbers.

Convince people that they need food made by your patented machines; foods that cause illnesses; illnesses which you say can only be cured by your patented medicines; medicines which cause side effects that you claim can only be cured by your other patented medicines; and your royalties will be royal!

Of course, your inventions are only solving problems that your other inventions created, but most people don't know that and so you can go on posting your stamped US Patent Certificates on LinkedIn and boasting about what a great innovator you are.

The worst thing that can happen to a patent holder is that someone actually figures out a solution to the problem their patent claims to solve while their patent is active.

Thus, once someone is granted a patent, they have strong disincentive to actually solving someone's problems.

You do not want to publish or even fund anything that might show that your random number does not actually work.

You want to sit back and tell yourself "job well done" while collecting royalty checks.

What's even better than selling an addictive drug? Having a legal monopoly on an addictive drug!

Patents richly reward those who can get vast numbers of people addicted to their random number.

This is exactly what happened with Purdue and the opioid epidemic. If you look at the legal docs, you'll see that step 1 in Purdue's playbook was to build an aggressive patent strategy.

We've all seen how much big pharma advertises. Why on earth would they need to advertise their life saving medications?

Surely no news would travel faster by word-of-mouth than life saving inventions!

Unless of course, these medicines are not life-saving at all?

If the life saving inventions were the unpatented ones, then the only way for businesses to profit off patents would be to convince you of a lie.

Truth is free. Lies require advertising.

If you have to choose between owning a business with a monopoly or owning a business in a competitive field, the former makes far more money. The only requirement is you must dedicate a portion of your proceeds to brainwashing customers that they need your random number (they don't)!

I could drone on and on along these lines, but I want to keep this (reasonably) short.

Please think about this for yourself: what do patents actually incentivize? Don't just repeat what people tell you they incentivize. Think about it from first principles.

I am not in this for the money.

I am in this for the friends that I have lost.

Some of my best friends on this planet.

Dead. Buried. Poisoned by patents.

Again, don't take my word for it.

Please think about these ideas for yourself.

And remember that's exactly what the other side doesn't want you to do.

by Breck Yunits

Publish early and often: I don't know of a single polished world changing product that did not start as a shitty product launched, iterated, and relaunched to increasingly larger groups of people. Even the iPhone went through this process.

Make something you love: I don't know a single person who makes something they love and is not successful and happy. I know people that make something people want and make money but are unhappy.

Master your crafts: I don't know a single person who built anything worthwhile who didn't have at least ten years of practice mastering their crafts.

Befriend other pioneers: I don't know anyone who started a successful colony who didn't make friends and trade help with other colony starters along the way.

Take care of your health: I don't know of a single founder who achieved success and didn't eat well, sleep well, take a lot of walks, enjoy time with family and friends, and write a lot.

Become skilled in all ways of contending: I don't know of a single founder who built a great organization who didn't master many complementary skillsets (sales, marketing, cash flow, fundraising, recruiting, design, et cetera).

Talk to users: I don't know of a single case of a startup that felt they spent too much time talking to users. - Jessica Livingston

Read: in my whole life, I have known no wise people (over a broad subject matter area) who didn't read all the time – none, zero. - Charlie Munger

I make many scratch folders and often hit this:

Now I just type sf

The line I added to ~/.zprofile:

by Breck Yunits

All jobs done by large monolithic software programs can be done better by a collection of small microprograms (mostly 1 line long) working together.

Building these microprograms, aka microprogramming, is different than traditional programming. Microprogramming is more like gardening: one is constantly introducing new microprograms and removing microprograms that aren't thriving. Microprogramming is like organic city growth, whereas programming is like top-down centralized city planning. Microprogramming is putting the necessary microprograms in a petri dish, applying natural selection over time until they can beautifully and most simply solve your problem.

In a microprogram every line of code can be its own program.

Microprogramming requires new languages. A language must make it completely painless to concatenate, copy/paste, extend and mix/match different collections of microprograms. Languages must be robust against stray characters and support parallel parsing and compilation. Languages must be context sensitive. Languages must be homoiconic. Automated integration tests of frequently paired microprograms are essential.

Microprograms start out small and seemingly trivial, but evolve to be far faster, more intelligent, more agile, more efficient, and easier to scale than traditional programs.

Microprogramming works incredibly well with LLMs. It is easy to mix and match microprograms written by humans with microprograms written by LLMs.

These are just some initial observations I have had so far since our invention of a breakthrough microprogramming language stack. This document you are reading is written as a collection of microprograms in a language called Scroll, a language which is a collection of microprograms in a language called Parsers, which is a collection of microprograms written in itself (but also with a last mile conversion to machine code via TypeScript).

If the microprogramming trend becomes as big, if not bigger, than microservices, I would not be surprised.

Microprogramming is inspired by microbiology.

You may know me as the creator of PLDB (a Programming Language DataBase), earth's largest database on Programming Languages, and find it relevant that I have personally studied and reviewed information on over 5,000 programming languages - nearly 100% of all publicly used languages.

What you might not know is that I also have a peer-reviewed track record in genomics and multiomics, and that Parsers, the language we invented that enabled Microprogramming, is built not on the patterns I found in programming languages, but instead built on the patterns nature evolved that I studied in microbiology.

The biological design is why this stack is unlike any language you have used before. You will be able to build any advanced program you could build using a traditional language, but the path to that solution may be very different. Once you've mastered this stack, I expect you will be astonished at how much you can do with so little.

The payoff from this biological approach will become increasingly apparent over time, as we continue to do things with this stack with radically less code than traditional approaches.

This is the dawn of a new paradigm shift in programming, and my job is to provide you with the truest, clearest, most concise information I have gathered over the past two decades so you can take this technology and build a better future for us all.

Great point! I should mention some data.

Parsers is made of ~100 microprograms.

Scroll is built on Parses and made of ~1,000 microprograms.

PLDB.io is built on Scroll and made of ~10,000 microprograms.

Great question!

Microprogramming is using languages where every single line is capable of being a microprogram.

Below is an example of microprogramming. This is the code that generates the homepage for BuilderNews https://news.pub . I annotated everyline to explain it in the terminology of microprogramming.

You can also view and edit it with syntax highlighting.

Great question!

That's up to the implementation. This is a design pattern that can be implemented in many ways. Even without a computer!

Requirements to implement this:

1. An environment
2. Microprogram parsers
3. Microprograms

Here's a video explanation:

Perhaps a video explanation would be better:

Is your liver on right now? You can find out with a drop of blood.

If your ketone levels are around 0.1, your liver is off.

If they are around 1,0, your liver is on.

It's easy to tell, once you've measured.

I was at a gathering of liver surgeons last night and none of them had realized this before: that the liver turns on and off.

Perhaps it's because my background is in computers that I immediately recognized when something goes up 10x, that indicates something has gone from "off" to "on".

It's generally not healthy to keep your major organs in the off state. If you never open your eyes, you go blind.

There's a lot of good places to start learning, but in two words: avoid carbs.

• Yes, I know the liver does many things in addition to making ketones. But if you are sitting in your parked car listening to the radio and haven't yet started the engine, we would say the car is off. Making ketones is the main thing.

I was 24 years old and riding in a van driven by Brian Chesky. Joe Gebbia was shotgun, and Nate Blecharcyzk was to my left. They had been iterating on AirBedAndBreakfast all day but finally were fatigued and ready for bed in San Francisco.

We were driving back from another Tuesday night dinner in Mountainview, where another entrepreneurial dynamo, Paul Graham, had spent time with us, in between darting to a backroom to iterate on his own young creation, a product called HackerNews.

The enthusiasm of all of these guys for their products beamed through like a rising sun. At the time, my projects were making more cash and had more traffic than AirBedAndBreakfast and HackerNews combined, but I knew the truth: my stuff was dogshit and I was doing it wrong. Whatever these guys were doing, that's what I wanted to do.

But how?

Before Nate joined Airbnb, he was already an accomplished software engineer with (IIRC) 7 figures in earnings under his belt. As for Chesky and Joe, they've published brilliant drawings and products they made over a decade before founding Airbnb.

Even if you are a prodigy, I've never heard of anyone build something they love without more than a decade of practice of their crafts.

Develop thick skin by constantly subjecting yourself to feedback and be grateful for it, even when it is delivered poorly. Harsh criticism is the crucible that forges strong ideas.

If you truly love your ideas you would publish them freely and never encumber them with patents or (c)opyrights. That would be like chaining up your children in the basement.

This is how you figure out what the world needs you to make. As you publish the world provides feedback. You know what you need. The world will tell you what it needs. Ponder and find the overlap!

For every product that a billion people love, remember that 99.999999% of people only loved it after 10 people already loved it.

How do you get 10 people to love your product?

You've got to love it yourself.

A Redditor gets hit by a truck

He goes to the afterlife.

It's amazing.

Mansions for everyone.

Gourmet meals.

Perfect sunsets.

But the Redditor is able to find a few negative things to focus on.

He writes an email to God with subject "Flaws in Heaven".

He gets an immediate response.

Mark Hebner and his team have refined an 1889 invention from Francis Galton and made a version you can hold in your hand to conduct real world probability experiments 10,000 times faster than flipping a coin.

It's the best $99 I've spent all year.

September 6, 2024 — I emailed this letter to the public companies in the current YCombinator batch.

Aloha S2024 batch,

I'm writing today to give you 1 piece of advice that I wish someone had told me when I was in YC: demand carry from YC for your batch.

Today you are the S2024 batch.

In 15 years you will be the X batch, where X is the name of whichever of you goes on to become the biggest hit.

By this naming convention I was in the Airbnb batch. And the Stripe batch. What luck, eh ;) ?

These Xs make YC its money.

But what makes these Xs?

The founders are #1, of course, but what else?

Here's YC's dirty secret: it is not the Partners or the Office Hours or Demo Day that makes these Xs.

Nope, not even close.

The Partners, the Office Hours, the Demo Day will be irrelevant to your startup compared to the long term impact your batchmates will have.

As founders you are building new colonies in the wilderness.

In those hard early years, you will depend on and be saved and elevated by your batchmates.

So, why do the people who do all the work get 0% of YC's carry?

YC's legal team will give you many laughable excuses.

The real reason is simple: a batch hasn't demanded it yet.

I used to be a moral compass inside YCombinator. Until they expelled me.

But losing access to Bookface is nothing compared to what happened to the previous moral compass, and my hero, aaronsw.

I believe in YCombinator and its ability to do good. But power does funny things to people, and I urge you to organize now, demand your rightful share, and ensure that there are some checks and balances on the YC partners.

Good luck, and feel free to reach out to me at any time, day or night. I want the best for you all.

Now, go make something the universe wants!

Best,

Breck

by Breck Yunits

HTML | TXT

Is there a better way to build a blockchain? Yes.

Particle Chain is a syntax-free storage format for the base layer of a blockchain to increase trust among non-expert users without sacrificing one iota of capabilities. A Particle Chain can be grokked by >10x as many people, thus leading to an order of magnitude increase in trust and developers on a chain.

RPM states that R, the reliability of a chain, is the number of particles P times the number of independent mirrors of the chain M.

Chains that are mirrored more with more history are more reliable.

The chain itself is encoded in Particles (Particle Syntax) which could be done with pen and paper, though for more utility it is recommended to build a digital Particle Chain 😉.

For implementing a digital Particle Chain, the Parsers Programming Language (or a similar compiler compiler and virtual machine) may be used.

All existing chains such as Bitcoin, Ethereum, and NEAR could be converted to a Particle Chain in a straightforward manner.

New chains would likely want to build on a Particle Chain from the start.

"I don't care for the heels", she said.

I knelt.

A twig had fallen from the oak we were sitting under.

"Look at this," I said.

I bent the end of the twig and snapped a small piece.

I snapped another. Then another. And another.

"Your neighbor, the heel, he's younger than you right?"

"Sure is. Has no respect for his elders. None of the heels do."

I made the twigs like a line of ants.

I pointed at the twig in front.

"Think of this piece of twig as your neighbor.

And think of the twig behind him as his father.

We usually count someone's age as the time elapsed since they were born. That would just be the length of this one twig.

But imagine if we counted someone's age as the length of all these twigs that represent their lineage."

I grabbed a second twig and started laying out a second line.

"Why are you making another line?"

I finished placing the 12th piece of twig and took a slow, deep breath.

"This right here, this is you. And the next twig is your dad. And that one, your grandfather. And so on and so on.

Science tells us something amazing. When we measure people's age like this, our 'genetic' age, then all of us humans alive today, you, me, the heels, we are all exactly the same genetic age, down to the millisecond."

I gestured to her line of twigs.

"Now, think about all the births and deaths; the wars; the famines; the accidents; the unexplainable, heartbreaking, terrible tragedies.

The Devlins, you went through a lot, but you made it, right?"

"Sure did. Devlins are smart and tough."

"I believe you."

I took a deep breath and looked up at the oak.

"The Heels".

I pointed again to the first line.

"They made it too."

August 28, 2024 — I have a backlog of interesting scientific work to do, but an important free speech matter has come to my attention. Warpcast is suddenly considering adding blocking.

I don't know whether some powerful people have joined and are pushing for this in secret, but I do know, as Farcaster user #158, that what has attracted a lot of active users to Farcaster has been our desire for a censorship resistant town square.

I hope in this short essay, we can show why Warpcast should veto adding blocking, now and forever in the future.

We will use the illustration below to define our terms.

A common argument made in favor of blocking is that black users must be protected from some firehose of responses coming from blue users.

But this is physically impossible. Blue users are just humans, who have to eat, sleep, take a piss, cook for their kids, and do a million other things.

They do not have time to dedicate their lives to "piling on" to black users.

It must be considered that the black user subconsciously knows they are doing something wrong, and so is experiencing extreme cognitive dissonance causing them to exaggerate the volume of messages from the blue user.

(Note that bots posing as humans should be blocked, but at the global protocol level and with efficient means to appeal to avoid the dangers of false positives.)

I am currently blocked by the below 4 popular users on Twitter:

In all three cases I was blocked after trying to help improve their ideas and better inform their users, who are all communities of people (science, statistics, startups) that I have a 20 year public track record of caring deeply about.

I also was an early public booster of all 4 of these people, with years of public proof behind that, so one can not only show evidence that my actions were good, but that there is strong evidence of long term good intentions and good will from me to these people.

I couldn't care less about my own ego, and I find it humorous and harmless when people mute me-I can be annoying with all my intense data and mathematical thinking. Muting is fine!

But to block me is to commit active harm to your users. It is censorship. It is anti-free speech.

It should not only not be done, but it should not be a feature of these digital town square platforms.

It makes me sad that these people, who I respect and enjoy their work, would have their worst impulses encouraged by Twitter allowing the blocking of users.

I blame Twitter for encouraging their worst impulses. I hope dearly Warpcast does not make the same mistake.

Thank you for listening. If you disagree, I am happy to look at your data and math in favor of adding blocking. And if I disagree, at the very worst, I will mute you.

August 25, 2024 — Two weeks ago I applied for the South Park Commons Founder Fellowship.

I applied because I love being part of cohorts of builders building and because I repeatedly invest all my money into Scroll, my angel investments, and other people, and would love to have more money so I can build the World Wide Scroll faster.

I was sad by South Park's rejection yesterday but I'm excited to use this as a teachable moment for South Park Commons, because building new things is hard and making things better for the people who do is something I care deeply about.

Hi there,

Thank you for applying to the SPC Founder Fellowship.

Our team appreciates the time and thought that you put into completing the application.

What are some ways you might show you appreciate my time rather than tell me you do?

We carefully reviewed all of the materials that you submitted and are sorry to say that we will not be able to offer you a place in the upcoming cohort. Due to the large number of applicants, we unfortunately won’t be able to offer more individual feedback.

If you carefully reviewed my materials, that implies you made some notes, and wouldn't it take just as much time to send me those notes as to send this email?

While we know this is disappointing news, we hope that it is not discouraging to your founding endeavors. We have been wrong in the past about ideas and founders this early in the journey and have (unfortunately) passed on companies that have gone on to achieve great things.

Why not share your numbers of applicants, acceptance rates, and outliers you missed?

We hope you’ll stay in touch through our community newsletter and Twitter. We’ll share any updates regarding future cohort applications through those channels and would be happy to see you apply again in the future.

Will you consider going back through these rejections and doing some actions that show your apprecation?

Best,

South Park Commons

Have you considered signing with your human names?

August 21, 2024 — In 2019 I led some research into building next-gen medical records based on some breakthroughs in computer language design.

The underlying technology to bring this to market is finally maturing, and I expect we will see a system like PAU appear soon.

Medical records will never be the same.

Exciting times!

If you know of links that should be here, please let me know. Or better yet, send a pull request!

• Blue Button
• OHDSI
• Unified Medical Language System
• Fast Healthcare Interoperability Resources

August 14, 2024 — I'm in Las Vegas for DEF CON and walking on the strip in 110 degree heat when a guy in dark clothes asks if I want to see some nude girls.

"Wait, how did you know I'm an amateur biologist?" I ask.

He frumps his brow and starts talking to a different group of guys.

But I need to find out what he was talking about.

I walk to my hotel room and google "las vegas nude people science museum".

Two hours later I'm leaving, awestruck, from Real Bodies Las Vegas.

Wow! If you haven't been to a plastination exhibit yet, go now, and go often.

It's a mind expanding experience. It is somewhere between 1,000 and 1,000,000 times better than reading an anatomy textbook.

For the past year or so, I've been wanting to play a mental video in my head of the life cycle of ketones, but struggled with all the details.

Thanks to this exhibit, I was able to build a much better mental model, which I've tried to explain below.

I'm sure there are mistakes, please send pull requests or emails with corrections.

I still have a lot of deep unanswered questions about ketones at the micro level.

But now I have more confidence in my understanding of ketones at the macro level.

My journey to understand the human body continues! Thanks for coming along.

• A New Scientific Model of Mania and Depression, with Testable Predictions (2024)
• The Liver Turns On and Off (2024)
• Energy Clocks (2024)
• Ketones vs Lithium (2024)

August 5, 2024 — Steve Jobs was famously part asshole but now the popular crowd says don't emulate that part.

Fuck them. This is the wrong take and there is no upside to staying wrong^[1].

I wish we had infinite time so anytime someone was being dishonest or lazy I could analyze the situation from every angle, put myself in that person's shoes and model their life and everything they missed or suffered to lead them to their current mistake, and run through many iterations of speeches searching for the words that will effectively teach them with the softest touch the error of their ways.

Unfortunately, as far I as I understand, not a single day is ever guaranteed and so to save my time for happier positive things I have to quickly look them straight in the eye and say "What you are doing is dishonest, knock it the fuck off, or I'll take action," and move on with my life.

I remember clearly many times when both people I love and complete strangers gave me honest advice straight from their assholes^[2]:

"Isn't this a tad bit...dishonest?" someone reprimanded me when I was a kid and thought I was a genius for putting ads on my site but hiding them with CSS so I could still earn money without actually showing ads.

"Breck, what are you doing?" a best friend in college once snapped me out of a drunken stupor and pulled me away from some new girl when I had a girlfriend back home.

"Learn to research properly" an Internet stranger told me thereby encouraging me not only to learn the ways of academia, but then go far beyond that and truly learn the way research should be done.

I don't mind honest feedback delivered harshly. Obviously I prefer honest feedback delivered gently, but harsh feedback and dialogue is better than silence. There's a saying about feedback: “Those who mind don’t matter, and those who matter don’t mind!” I want to matter, so please keep the feedback coming!

I have learned not to trust people unwilling to be a part asshole (only when it is appropriate of course) in public. These are the people who are real dirty assholes in private, who will collude and plot and gossip in private and then in public act like angels.

Direct, blunt feedback to someone about specific bad behavior, that may come across as hyperbolic, is awesome. Again, if you can pivot their behavior in a gentle way, prefer that always. But the most unkind thing you can do to them and to our world is to let bad behavior go unchecked in order to "not be an asshole".

No one likes to be the asshole. I hate it. It sucks. But an ounce of prevention is worth a pound of cure, and if you don't nip bad behavior in the bud, it grows.

I do not advocate flipping the bit on people. I believe all people are mostly good, and all can learn to improve their behavior, but they require honest feedback to do that. It's also important to remember due to natural information assymetry, we are naturally biased to judge ourselves by our intentions but others by their actions. Often I give someone blunt feedback and they respond bluntly back correcting me because I was wrong! This is a great outcome that we would not have reached had I flipped the bit on them and not risked being an asshole.

We want to minimize bad behavior B while also minimizing time spent correcting bad behavior C. If that requires delivering a blunt hyperbolic zinger that risks getting you tagged with the label "asshole", I say go for it.

Perhaps the worst thing about Steve Jobs is that he wasn't asshole enough in public^[3]. Maybe if he showed the same range in public as he did in private, people would have understood more the value of being part asshole.

• 100% Assholes: Avoid! They are so negative and bothered by everything because everything sticks up their ass.
• 0% Assholes: Avoid! They are full of shit! Or more likely, huge assholes behind closed doors.
• Part Assholes: seek them out!

Instead of declaratives ask questions.

You are being dishonest. Do you think this is a tad bit dishonest?

You're lazy. Might this be lazy?

This is terrible. Can you do better?

Dataset needed. What dataset would confirm this?

July 25, 2024 — I received sad news today that David Bangert passed away. I met David and Sandy in 2019 through Hawai'i Angels, when he and I were "Deal Honchos" on a deal. They took me out for lunch in Oahu, and I sat enthralled for hours listening to David's amazing life stories.

One day with David could change your life. It certainly changed mine.

His last letter is below.

Friends, Classmates, Neighbors, Colleagues!

I want to share with you a decision that I have made. This note is longer than I initially intended, but I feel I need to tell those with whom I've had friendships during my life’s journey, why I am about to take a dramatic step.

As you read this letter, Sandy and I are traveling to Zurich, Switzerland. Once in Switzerland, we are going to access the services of Dignitas (www.dignitas.ch) where I will be permitted to die with dignity. My health and mobility continue to decline. I do not want to end up lying in bed all day, and I am not enjoying living the life of a non-ambulatory person.

First and foremost, I am balanced challenged. I am now confined 99% of the time to sitting in a wheelchair or in bed because I cannot navigate my way across a room without falling. I constantly fear a fall and potentially breaking my 80-year-old bones. I used to resort to grasping for the surrounding furniture and walls as I inched along at a slower than snail pace, but now, my body will not allow me to do that. Also, a walker is no longer an option due to the loss of feeling in my feet, my degenerative bone condition, my flat feet, and my contorted foot configuration. This is a genetically predisposed condition inherited from my mother, which is progressively worsening, and I believe, the major contributing factor to my balance issue.

Historically, my orthopedic surgeries have been unsuccessful. My surgeries started in my military training when I had my left shoulder injured. The surgery left me unable to raise my left arm higher than my shoulder. Later at West Point, my ankle was operated on because of my Achilles tendon popping out of its groove. This resulted in a frozen ankle that will not bend properly. Some years later, a right knee replacement was needed. The specialists stated it was necessitated by my poorly functioning ankle. Since this surgery, I have been unable to bend my knee properly when I walk. Finally, I had back surgery. This last surgery resulted in not being able to walk or stand. I cannot make my feet and legs move on command.

Desiring a better life, I sought out multiple “second” opinions. My medical records were assessed at two excellent California university hospitals: Stanford, and UCSF. These two second opinions suggested additional testing, which my medical provider Kaiser performed. The new tests did not suggest the cause of my balance challenges, nor did they indicate any new successful therapies. Desperate for help, I turned to the best clinic in the US. Sandy and I traveled across the US to the Mayo clinic. For seven days, tests and assessments were performed. The recommendation was back surgery. I stayed in the clinic and had an operation by the Mayo clinic’s back specialist. Unfortunately, the surgery was a failure and resulted in a worsening of my condition; My balance was worse after I recovered from the operation. At least prior to the operation at the clinic, I could stand and walk short distances, but following the surgery, I became restricted to wheelchair use. I now resort to sliding along the wall and hanging on furniture when a transfer is necessitated: move from chair to bed, or chair to car. I have spoken to the Mayo clinic. It suggested going back there again and repeating the same tests. I decided against that course of action.

When I wake up in the morning, it takes me five minutes to transfer from the bed to my motorized wheelchair. Then, I continue to the bathroom to relieve myself, which involves a transfer from the wheelchair to the toilet. This transfer of position takes another few minutes. I have a toilet with a washer, which I use to clean myself. Following the toilet routine, I wheel into the kitchen and make myself a cup of coffee using the electric, one-cup Nespresso machine. It is a struggle for me to stand and make the cup of coffee. Sandy must ensure the water reservoir is full before bed so that I can independently have coffee in the morning. Then, I read my emails and the on-line newspaper on my iPad. Reading a printed newspaper is not possible because of my manual dexterity. I spend my entire day in my motorized wheelchair. This has been my daily routine for the past two years.

Four or five days a week, I have physical therapy. I use two trainers, both of which come to my home. My routine is similar with each trainer. They work with me for two hours. We do strengthen exercises, and they help me shower. I cannot go into the shower without assistance. My caregiver/trainer washes and dries me. I have a shower stool which I sit on. My strength is not improving, if anything, it is declining. I can see this in the weights that I'm able to lift and the duration of my cycles on the stationary bicycle.

Besides reading the newspaper and watching television, my day is very quiet. This is not the life that I am used to. I am used to having more stimulation and being involved in challenging situations. I have the good fortune of being a graduate from two world renowned institutions of higher learning: the Military Academy at West Point and Harvard Business School, (PhD in decision sciences). I am not used to not being able to physically do things. After military academy, I successfully completed Ranger and airborne school. In my forties, I walked around the Annapurna Massive in Nepal, 22 days of walking and staying in tents. In the past, my life has always been an active one.

Also, I am not as mentally sharp as I used to be. I often cannot remember the names of people and street names, places I have been, what I just ate, who I was with, investments I made yesterday... I have always been organized up till late, and now, I find myself (with Sandy helping me) looking for my lost wallet or keys for hours. For example, I suddenly forgot how to use the remote for the TV after using it for months successfully. I am transferring the responsibility for managing our wealth to Sandy. We spend time every day discussing our investments and decisions she will have to make. Tasks such as writing this letter are challenging. My typing is reduced to dictation plus two-finger typing. Sandy is helping me with this task.

In my late 30s and early 40s, I managed the first phase of the construction of a city in Saudi Arabia. I had a direct staff of 150 professionals who, through contracts, managed a workforce of 20,000 workers. Following the seven years in Saudi Arabia, I went to Hawaii in search of what I wanted to do for the rest of my life. I enrolled in the MBA program and about halfway through decided that being a professor would suit me. I came to understand that the culture of universities requires getting a quality doctorate. I was fortunate to be accepted at the four programs for which I applied: Harvard Business School, Stanford, Pittsburgh, and New York University. I went to Harvard and completed the doctoral program in 2 1/2 years, which is considered fast. The typical time to get a doctorate from Harvard is 4 1/2 years.

In the past, I have written hundreds of letters with ease. Now writing a letter, even one as important as this letter, is a struggle. It should not be a struggle. I am tired of struggling.

I am fortunate to have a supportive wife who understands my frustration. We are affectionate, kiss multiple times a day and hold hands. Although we go to dinners, theater and other functions, the struggle to get to the venue exhausts me. I often think it is not worth the effort.

I will be cremated in Zurich and the ashes will be mailed to Honolulu. From Switzerland, Sandy plans to go to her hometown of Kelowna, British Columbia for a short time. There, she will be with her family.

Afterwards, she will return to Honolulu. She plans to bring the ashes to West Point in 2025 at my Class’s reunion for placement in the West Point cemetery.

Thank you for being good friends. I enjoyed our time. Please continue to interact regularly with Sandy. I know she greatly values your time together.

David

RIP David.

July 23, 2024 — Stephan Kinsella reposted a great 1850's quote from Abraham Lincoln on litigation.

Discourage litigation. Persuade your neighbors to compromise whenever you can. Point out to them how the nominal winner is often a real loser---in fees, expenses, and waste of time. As a peacemaker the lawyer has a superior opportunity of being a good man. There will still be business enough. Never stir up litigation. A worse man can scarcely be found than one who does this. - Abraham Lincoln (1850)

Matthew Pinsker at Dickinson posted an image of the original:

HTML | TXT

Are some intellectual environments better than others? Yes.

ETA! states that E, the evolution time of ideas, is the time T needed to test alterations of ideas, divided by the factorial of the number of ideas in the Assembly Pool A!.

Longer evolution times means worse ideas last longer before evolving into better ideas.

You can lengthen the lifespan of bad ideas by increasing the time to test alterations or reducing the size of the Assembly Pool.

You can evolve good ideas faster by decreasing the time to test alterations and increasing the size of the Assembly Pool.

The number of test threads is proportional to the size of the Assembly Pool.

This equation explains the triumph of open source and public domain software.

Luckily A^A is still in the factorial class so the memorable acronym still applies.

• Thank you to Lee Cronin and Sara Walker for evolving the idea of the Assembly Pool.
• Thank you to Szymon Łukaszyk who's applications of Assembly Theory to binary messages encouraged me to apply Assembly Theory to measure the health of information environments.
• Thank you to Connor McCormick for superb feedback.
• Thank you to Cobe Liu for pointing out it should be T/A!.
• Thank you to my coauthor John. Without your help I never would have arrived at this equation.

July 13, 2024 — I am writing a book in a private git repo that you can buy lifetime access to for $50.

That repo is where the source code for the book lives before it gets published to the public domain.

The public gets a new carefully crafted book with source code, just delayed. If you pay, you get early access.

This business model I'm calling "Early Source".

Early Source is a new business model for public domain, open source software. You publish your software to the public domain, but with a delay of N years. Professionals that most benefit from faster updates pay you for access to the private, cutting edge version.

Early Source is a win-win-win-win:

• It's a win for your most loyal customers who can pay you a simple fee for a simple deal and get early features fast
• It's a win for your business model as you can benefit from the lower costs of open source while still keeping something scarce to sell
• It's a win for casual users around the world who now can benefit from your older software and will know who to buy from if they later become professionals
• It's a win for your ideas because you can now make them as good as possible, something you can only do if they are public domain

What are the downsides? Let me hear it in the comments.

So far, I would have to say early source is a pain in the ass.

Keeping 2 repos going, one private and one public, is just a lot of cognitive overhead. Far more than I thought.

Instead of trying to connect revenue to digital access, I think I'll publish the ebook free, and sell a paper version, as well as additional stuff around this, like a course with real human interaction.

July 4, 2024 — I lived in San Francisco and Seattle in the 2000's and 2010's, and if I told you the names of every startup I almost joined as employee <5, you'd probably think I was lying.

But I declined them all for the same reason: the 1 year vesting cliff.

If a startup gives you 1% of the company in the form of stock options, the standard is to vest 0.25% a year for 4 years with a 1 year cliff. This means you get nothing until after the first year.

If you have to leave or they fire you after 11 months you own zero.

Cliffs are a bad deal for everyone and should be eliminated.

It sucks for the employees. It's so much more fun to do your best work immediately and never think about cliffs.

It sucks for the startups. They miss out on bringing in outlier talent who may stay for a decade, but only want to guarantee a few months.

Back in 2008-2018, I was young and 1 year seemed like an eternity-what if I wanted to travel in 10 months, or someone close to me got sick, or I had my own startup idea?

I also didn't want to have any incentive to sit on my best work until my shares started vesting. I wanted to give my teams my absolute best from day 1.

Historically I knew I would make the most impact at a startup in months 3 - 12, so why should I be rewarded less than the person who does just enough to last until month 12?

Now that I'm an angel investor and not subject to cliffs, I can say with much higher certainty: cliffs are dumb and should be eliminated.

Four year vesting schedules are fine, but employees should vest daily, starting on day 1.

They get 2% of their stock options. Far less than a single percent of your company. So what.

Everyone should be happier without the cliff.

Presumably you've hired people who are good people, and so even though it wasn't a right fit, you think the world will be a better place if that person someday gets a little extra money from their stock.

They also will forever be a booster of your business even though they are no longer an employee, because they now have the incentives to be.

That's all I have to say about cliffs. Now, I've got to get back to building my own projects.

But if your startups wants to hire me and pay me in options, I am available. (For a month or two ;) )

June 29, 2024 — A child draws.

You take his paper.

He screams.

His scream is just: you stole his property.

You give the paper back.

He stops screaming.

You copy his drawing on your own paper.

He smiles.

His smile is just: he feels his work is appreciated, and he can learn from your version.

Ideas and property could not be more different.

Ideas cannot be stolen. Property can be.

Copying ideas helps, not hurts.

Children understand these obvious truths.

Children learn what is true from nature.

But teenagers turn from nature and start learning "truth" from the popular kids.

This allows for some strange diseases to spread across our civilization.

The brain of someone infected with IPDD cannot deduce the obvious truth that copyrights and patents are the exact opposite of property rights.

They cannot see that atoms are scarce and improved by assignment and ideas are the opposite.

Symptoms of IPDD include:

• the appearance of tiny round spots, often found on the foot, that look like this: ©
• compulsive tendencies, such as insisting that a "license" accompany every piece of information.
• trouble pronouncing phrases, saying "All Rights Reserved" when they mean "Your Rights Restricted."
• paranoia that someone will "steal their ideas" accompanied by outbursts of "Patent Pending" or "FBI Warning" at their customers

People with IPDD also suffer from deep, grandiose delusions like that they created:

• a widget more important than the wheel
• a song better than all of Beethoven's
• a formula more intelligent than all of Einstein's

At a societal level, IPDD is a significant long term burden on economies. Holding all else equal, countries with low rates rates of IPDD infection have significantly faster innovation and more equal wealth distribution.

The best medicine for curing IPDD is exposure to honest information and a long walk in the woods.

It is important to be careful when attempting to heal someone with IPDD too quickly.

I once told an infected Congressman that since they made everything someone writes "protected", they should also "protect" the air someone exhales.

I was hoping this perspective might cure him of his affliction.

Instead, his face lit up, he pulled out his phone and told me he had to text his "Intellectual Property" Liar to try and patent his new idea for a machine that could add licenses to air molecules.

June 25, 2024 — In 2013 my friends and I won the Liberty Hackathon in San Francisco.

We built CheckBox, an open source system voting system that supported online, offline and hybrid elections. I did the technical design, Cam talked to users and developed the business pitch, and Ben provided the design.

CheckBox built on our new discovery at the time, which now goes by the name Particles.

We tried (and failed) to raise money to work on CheckBox full-time.

We were quite poor at the time.

I subsisted on an unhealthy diet of ramen and an occasional treat from the McDonalds $1 menu.

I often slept on a couch while hosting an AirBedAndBreakfast guest in my bed.

I also had a roommate and we had one bed and one couch in the room, and when our room wasn't rented we alternated who got the bed.

About six months after this hackathon, Microsoft acqui-hired our team for NudgePad, the other project we were working on.

This was great because I could now afford anything on the McDonalds menu.

Unfortunately CheckBox fell by the wayside, until someone reminded me of it today.

Below is our executive summary from our failed fundraising attempt in 2013.

(I am still trying to find the code that powered our beta).

Although CheckBox as a business is not for me, I love the underlying CheckBox design, and still think that this system has a better underlying design than any other electronic voting system-by a wide margin.

The key idea is this: the underlying languages can power 100% digital election, 100% paper election, or a hybrid election.

If you want to build the best voting system, I am highly confident this is the way.

We’re changing that.

We built CheckBox so that voting districts can easily create a secure ballot for voters to vote from home.

It offers:

• Secure Voting
• Private Ballots
• Supports Electronic/Paper Hybrid Systems
• Easy Customization
• Supports Early or Absentee Voting
• Election Auditing
• Security Middleware
• Registration and Ballot Distribution Middleware
• Voter Registration
• Open Source

Prototype is live at http://getcheckbox.com

June 12, 2024 — After years of development, I'm looking for beta testers for The World Wide Scroll (WWS).

The WWS is like the World Wide Web, except:

• Intelligent. Built not on HTML, but from the ground up in a simple, expandable language, designed for both humans and AIs.
• Offline. You download entire sites and browse completely offline locally, rather than fetching one page at a time.
• Pristine. No ads, no paywalls, no trackers, no cookies, no copyright.
• Human. Instead of anonymous domains registered to corporations the WWS has folders owned by verified humans.
• Trustworthy. Source code behind every page.

If you know someone who might find this interesting, I'd appreciate it if you could share this with them.

The link to register a folder is:

https://wws.scroll.pub

The price to own a folder is $100 for 10 years.

Funds pay for the development of the Scroll language.

Mahalo!

-Breck

June 11, 2024 — You can now download this entire blog as a zip file for offline use.

The zip file includes the Scroll source code for every post; the generated HTML; all images; CSS; Javascript; even clientside search.

1. Download
2. Unzip
3. Open index.html

Technical users can still clone and build the repo, but sometimes you just want the compiled HTML for offline reading.

Wouldn't it be great if more of The Web were like this?

The Scroll is coming ;)

HTML | TXT

Can we quantify intelligence? Yes.

Program P is a bit vector that can make a bit vector (Predictions) that attempts to predict a bit vector of actual measurements (Nature).

Coverage(P) is the sum of the XNOR of the Predictions vector with the Nature vector.

Intelligence(P) is equal to Coverage(P) divided by Size(P).

If programs A and B are the same size, the program with more coverage is more intelligent.

If they have the same coverage, the smaller program is more intelligent.

HTML | TXT

by Breck Yunits

May 31, 2024 — Yesterday, on a plane, I found an equation I sought for a decade.

PTCRI describes the potential of a programming syntax.

PTCRI says the number of possible programs P is equal to the number of tokens T raised by the number of columns C, raised by the number of rows R (aka lines), raised by the number of indentation levels I.

If you view the source code of this post, you will see T, C, R, and I in action.

PTCRI explains the simplicity and power of Particles^[1]. Four concepts, three syntax rules support a vast universe of concise programs. Another syntax(es) might be found with superior metrics, but I have yet to see it.

PTCRI also explains why nearly flat structures work so well: a little nesting goes a long way. If you set T and C and R to 3, changing I from 1 to 2 increases the amount of possible programs from 7,625,597,484,987 to 58,149,737,003,040,060,000,000,000.

59 days ago I announced the decade long Particles research endeavor over with a negative result.

It looks like I was wrong again. There was something special about Particles, and PTCRI explains what that is.

It shouldn't be possible to represent all programs in all programming language with such a minimal set of rules! Particles doesn't even have parentheses!

And yet, the experimental evidence kept piling up.

The evidence hinted at some important natural formula, and now we have a name for it: PTCRI.

[1] Particles was originally named Tree Notation.

Thank you to Marc Forward and Inconstant_Moo for feedback and helping me add the T.

It may help to explain from another angle. Imagine your boss comes up to you and says "We are giving you a special assignment. We want you to come up with the best possible syntax for all programming languages." What do you come up with?

You can present them with Particles, and explain how it supports a vast universe of programs, P = {T^{C^R}}^I, with just three syntax rules:

• lines split into columns via bits, characters or words
• groups of lines joined to each other via newlines
• scopes established via the indent trick (aka the "Offiside Rule", popularized by Python, HAML, et cetera)

• Does PTCRI also describe S-Expressions? S-Expression work. Particles barely works.
• All mistakes are on me, and credit goes to the people who have supported this effort with me.
• I think this equation is pretty interesting, so I really hope a lot of people on the Internet tell me how stupid it is and that it was discovered 100 years ago.
• Now, back to my vacation.

Patch is a tiny Javascript class (1k compressed) with zero dependencies that makes pretty deep links easy.

Patch can be used with both:

• ?queryStrings
• #hashStrings

Patch handles encoding and decoding for you, and makes your deep links pretty, so you don't have to think about it.

1. Give your Javascript object to Patch, and it will give you the prettiest query string it can make:

1. Give Patch a query string, and it will give you back your Javascript object:

Then Patch encodes and decodes that spreadsheet.

A Patch object has 4 forms:

The default is & for rows and = for columns.

You can change these to suit your own needs.

• QueryStrings has one very restricted string type and a simple map type.
• QueryStrings arguably adds little to no value on top of just a single string.
• QueryStrings requires you encode and decode each map pair separately.
• JSON is an alternative to Patch, except after uriEncoding JSON you lose almost all human editability in query strings.

This is an updated fork.

"I'll give you this library," the billionaire said, sweeping his arms up toward the majestic ceiling.

"Or...you can have this scroll," he said, pointing down at a stick of paper on a table, tied with a red ribbon.

Thomas said nothing.

He kept his eyes locked on the billionaire's.

He had prepared for moments like this.

He knew he was going to be challenged with something today.

He just didn't know what.

Now he knew.

Thomas inhaled deeply.

"I'll take the scroll"

Thomas sat in his car in a beach parking lot and stared across the bay to the end of the cape.

First he saw the smoke, then the flames, and then the humongous rocket rise off the ground.

It gained speed as it rose into the sky, carrying his mentor and the rest of its crew on their journey to Mars.

When the rocket was out of sight, Thomas looked down at his lap.

He untied the red ribbon and unrolled the scroll.

It was blank.

For $499,999 a year, you get access to brecks.lab, the private git repository and issue board where we post our in-progress language and AI research, and prioritize and shape the future of Scroll, PLDB, and our other open source projects.

• Early Access: See new research and code early, sometimes months before publication.

• Influence: Vote on which ideas you think deserve the most focus.

• Meet: Interact with other members on the issues board.

If you are a technology investor, we have some very exciting posts coming out on AI, science, and what comes after the web.

You can wait for the published versions, or you can join brecks.lab today.

How are independent scientists to earn a living in a world without copyright and patents?

By selling early lab access to corporations and money managers.

Independent scientists can do better science if

• They can publish everything to the public domain
• They have an empowered group of beta readers with shared interests
• They have less need to worry about revenue streams

Why not solve all three problems at once?

Labs is designed to help fix science. The public gets far better information and investors get information on what new technologies are coming next.

May 26, 2024 — You once could buy transistors, capacitors, and other components at your local neighborhood store. The decline in US computer and electronics manufacturing correlates with the decline in RadioShacks. To catch up to other nations, maybe it is time for a next-gen RadioShack.

I can have a far wider selection of electronic components delivered overnight for cheaper from Amazon, Mouser or Digikey. The need for a brick and mortar electronics store is long over. Time marches on. - rriggsco

May 25, 2024 — I just pushed version 93.0.0 of my language Scroll. Version 93!

Why so many versions? I use Tom Preston-Warner's Semantic Versioning (2011).

In particular, I followed his advice in Major Version Numbers are Not Sacred.

Pushing so many major versions was no big deal because Scroll was a toy.

But Scroll now has a novel feature that makes it very useful.

I need to decide if I should ship fewer major versions with higher quality.

Let's do some research on what other programming language developers are doing.

I added version number data to over 600 languages in PLDB.

Wow! No languages in the top 400 have surpassed 30 major versions.

Erlang, at version 27, ranks #34 and is the only top 100 language with more than 25 major versions.

My language, at version 93, is a huge outlier. Maybe I've done something wrong 😳.

Let's plot version number by rank.

We can see that there is a positive correlation between how many major versions a language has and how popular it is.

However, we can also see that 5 of the top 10 languages are on version 5 or less.

We can also see that ~15 of the top 25 languages are on version 5 or less.

If you have a great model of your core ideas, you can ship fewer major versions.

No. As you can see:

TIL I did SemVer wrong. In particular, I missed this part:

If you’re changing the API every day you should...still be in version 0.y.z

Whoops!

It took me about 5 years to figure out what Scroll 1.0.0 should be. I should be on version 0.93.0, not version 93.0.0.

No big deal. I will soon release Scroll 100.0.0, and from then on will have far fewer major releases.

An ounce of deep thought on the core of your language is worth a pound of major releases!

• I learned 97.5% of languages with a version system use Semantic Versioning. The ~2% of languages, such as C and C++, that are still doing versions like C17 (which does not mean major version 17, but refers to the year the revision came out), maybe want to reconsider ;).
• I learned most languages use git tags for tagging their releases. Something I should start doing.

I learned a few things today. Hope you did too!

May 24, 2024 — When I was a programmer at Microsoft I participated in a lot of internal betas.

So I saw a lot of animated "progress" bars in software that was actually hung.

I bet you could invent a better progress bar, I thought.

But life went on and I forgot.

Yesterday my FitBit app showed an animated progress bar for hours. It turned out no progress was being made—the phone had no Internet connection.

But it made me remember my progress bar idea!

Every time someone starts a long task, measure the duration and end result. Then you have the data to render a progress bar like the one above.

The green and red areas show tasks completions and aborts.

The moving line shows elapsed time.

The color of the moving line shows the odds that a task will succeed, given the amount of time passed.

It helps your users decide: should I wait or abort?

A famous celebrity passes away and wakes up on a beach.

"Welcome to the Afterplace", says a man in white.

He extends his hand and helps her to her feet.

"You must be hungry. Let me show you to the Omni Restaurant."

They walk from the water to an enormous restaurant.

The entire front of the restaurant is a glass wall facing the ocean.

The restaurant appears to extend endlessly in both directions.

A sliding glass door opens and they walk inside.

The ceilings are a hundred feet tall.

Even though countless people are dining, the restaurant is so large that it is quiet and uncrowded.

"Please have a seat," he says, gesturing to a table with his hand.

"At the Omni Restaurant, you can order any dish ever invented by human civilization."

"Whatever you want, just speak it into your table."

She sits down and says "Portobello mushrooms please."

Suddenly, on the back wall, a hole five feet in diameter opens up.

Then, flying out of the hole comes a silver platter.

The platter hovers over her table then gently floats down.

On it is a perfectly grilled Portobello mushroom.

The former celebrity smiles, grabs the fork and knife, and takes her first bite.

"Oh my god. This is the best Portobello mushroom I've ever tasted", she says.

The man nods his head, turns and leaves her to her meal.

After her meal she explores the grounds.

Eventually she tires and spends the night in a luxurious hammock.

The next morning she returns to the Omni Restaurant.

"Bacon and eggs please," she says.

'COMBINATIONS NOT ALLOWED,' a robotic voice says back.

A few people turn to look.

Her face crunches.

"Portobello mushroom please," she says.

Whew, she thinks.

Delicious. The same as yesterday.

Her face relaxes.

Is it exactly the same?

Her face crunches again.

After another day exploring the grounds, she returns to the Omni Restaurant for dinner.

"Filet Mignon please."

'FILET MIGNON IN USE,' the robot voice responds.

People look.

Her face crunches.

"Umm...ummm...lobster please"

A hole appears in the wall.

Her face relaxes.

A silver platter carrying a deep-red lobster lands in front of her.

"Butter please"

'CUSTOMIZATIONS NOT ALLOWED'.

Day 3 is off to a bad start.

"Bacon please."

'BACON IN USE.'

"Eggs please."

'EGGS IN USE.'

"Peanut butter please."

'PEANUT BUTTER IN USE.'

Many eyes are on her.

Her face crunches.

Then her face turns red.

She clenches her fists and stands up.

She looks at other people's tables.

She sees countless varieties of chips, candy bars, and cereals.

She also sees for the first time that the other diners are malnourished.

Screw this!

She storms to the back wall.

Someone orders a meal and a hole opens.

She dives through.

She lands on her hands and knees.

Then she stands up and looks around.

"What the?!"

There is no kitchen and no cooks.

There is nothing at all on this side of the wall.

She rubs her eyes in disbelief as she watches dish after dish materialize from nothing then fly out through a hole in the wall.

Suddenly she feels a tap on her shoulder.

"What are you doing back here?," asks the man in white.

"What am I doing back here? What am I doing back here? What are YOU doing back here?"

"People out front are malnourished."

"They can't order combinations. They can't customize their orders. And they can't eat something if someone else is eating it."

"And now I see that the physics of the Afterplace means all of the rules of the Omni Restaurant don't make any sense!"

"I DEMAND you take me to the being who designed this place."

"That will not be a problem."

"If you will just follow me."

She follows him back to the front of the restaurant.

They walk for miles past tables and tables of diners.

Finally the man in white comes to a stop.

In front of him, eating a bowl of cereal, is a man in a Vicuna suit.

"Here sits the Omni Restaurant's creator," he gestures with his hand.

"Stan?! It can't be Stan!"

"You know him?"

"Of course!"

"He's my copyright lawyer!"

HTML | TXT | PDF

by Breck Yunits

All tabular knowledge can be stored in a single long plain text file.

The only syntax characters needed are spaces and newlines.

This has many advantages over existing binary storage formats.

Using the method below, a very long scroll could be made containing all tabular scientific knowledge in a computable form.

There are four concepts to understand:

• measures
• concepts
• measurements
• comments

The only required information for a measure is an id, such as temperature.

An example measure:

The only required measurement for a concept is an id. A line that starts with an id measurement is the start of a new concept.

A measurement is a single line of text with the measure id, a space, and then the measurement value.

Multiple sequential lines of measurements form a concept.

An example concept:

An example comment:

Once your knowledge is stored in this format, it is ready to be read—and written—by humans, traditional software, and artificial neural networks, to power understanding and decision making.

Edit history can be tracked by git.

GNU Recutils^[5] deserves credit as the closest precursor to our system. If Recutils were to adopt some designs from our system it would be capable of supporting larger databases.

ScrollSets are used to power the open source website PLDB.io. PLDB currently has over 300 measures, over 4,000 concepts and over 150,000 measurements, contributed by over 100 people, dozens of software crawlers, and a couple of artificial neural networks.

If printed on a single scroll, the PLDB ScrollSet would be over one kilometer long.

• It is very helpful to have a sortIndex attribute on your measures to automatically prioritize^[8] the measurements in your source and output files. The impact of this simple enhancement hints at interesting signs of dense information packing achieved by this method, which may have implications for the weights and training of artificial neural networks.
• Computed measures are measurements not stored statically, but derived at runtime from other measurements. They are very useful and easy to add with a few lines of parser code.
• You generally always want to add a type attribute to your measures, which gives you error checking, among other things.
• Measures can be nested. This means it is best to be restrictive in what characters are allowed in measure ids to integrate with a broad set of software tools. For example, you can nest a minParser under temperatureParser to generate a temperature_min column name in a generated TSV.
• It is useful to have measures whose values are foreign keys, such as a list of ids.

This system might also have broad use.

You can read more about ScrollSets on the Scroll blog, see small demos at sets.scroll.pub, and see the large implementation at PLDB.io.

Leibniz thought of Binary Notation; Lovelace of Computers; Darwin of Evolution; Marconi of the Wireless Telegraph; Einstein of Relativity; Watson & Crick of the Double Helix; Tim Berners-Lee of the Web; Linus of Git.

Even more importantly to you and to me, at some point our mothers and fathers thought to have us.

And since we were born, many people throughout our lives have had thoughts that had high positive impact on us.

If you believe we live in a Power Law World, then it follows that there is nothing with higher expected value; nothing with more leverage; nothing with higher ROI; nothing with higher impact; than High Impact Thoughts (HITs).

HITs dominate both our professional and personal lives. Let's take a closer look.

I've filled many notebooks over the years with potential HITs.

Looking at my notebooks, I would say I generate between 2 and 10 HIT candidates per day.

I have found it very hard to predict in advance what the impact of a HIT is going to be. I have to act on the HIT first.

I am often off by many orders of magnitude. Sometimes I predict a thought is a surefire HIT but then after I act on it the predicted high impact is nowhere to be found. Instead the real impact is like that of a falling leaf.

Other times I see what seems like a small, mildly interesting idea, I act on it in minutes, and it impacts my life for decades.

Still other times I predict a thought will have a big impact, it has only a tiny impact, but then years later a slight tweak makes it have the impact I originally predicted.

Or sometimes an idea has a big initial impact, but turns out to be inconsequential in the long run.

Since we live for less than one million hours, then the smallest HITs on you personally, measured in hours, would have an impact of ~1, and the biggest HITs would have an impact of ~100,000. So the HIT range on an individual is 6 orders of magnitude.

If you consider the impact your HITs can have on your family, friends, and communities, the impact range expands further.

If this essay is true, then it should be one of the most impactful I ever write and one of the most impactful you ever read. It should be a HIT!.

But of course, as I mentioned above, it is hard to predict in advance how big a HIT will be, and sometimes it will be years before the right tweaks are made to make something a big HIT.

If this essay you are reading now does not have a big impact, perhaps a future version will.

I share the git for this blog, so if this essay does become a HIT we'd be able to see which tweaks caused that phase change.

If HITs! have the highest expected value, why not spend all day seeking HITs!?

What's the point of thinking about lowly HITs when HITs! dominate?

I can think of five reasons.

The first is that you don't know whether you've got a HIT until you act on it and get it past the payoff point. Focus and details matter. Nature does not care if you are close to a HIT, you've got to get it all the way past the payoff point.

The second is that your brain needs data to generate HITs. You've got to balance experience with reflection. Too much reflection and your brain won't have enough data to generate novel HITs.

The third is more practical: you have to breathe, eat, drink, sleep, exercise, be social, et cetera. Generating HITs is harder if you're unhealthy. To generate HITs in this world you have to stay alive.

Fourth, there's a fair amount of randomness in HIT Seeking. With 100 billion humans born so far, many of them who never thought about HITs will have bigger HITs than you, simply through chance. You can certainly do things to tilt the odds in your favor of seeing more HITs, but ultimately the dominant term in what HITs you will see is not something you get to control.

Finally, one HIT is "don't take life too seriously—no one's ever gotten out alive." If you were to spend all your time HIT! Seeking, due to randomness, you might not ever discover the most impactful HIT!, which could be the one I just mentioned.

The most important thing is probably to setup your life so you can spend a substantial amount of time HIT Seeking. I like to aim for 2 - 4 hours per day.

What does HIT Seeking look like? Walks in the woods, meals with friends, yoga, naps, long hot showers, quiet sitting, conversations with LLMs, drawing, playing in the sand with your kids, doing at home experiments. That sort of thing. The highest impact work often doesn't look like work.

Ask a lot of questions. Really drill down into the details. Don't worry much about words. Mental models—being able to picture things in your head, and rotate them—that is what matters.

It's not enough to see potential HITs. It's not enough to just write them down.

Those are required, but you also have to act on them. You will learn that your early versions always get a lot of important things wrong. And you will never know the impact of a HIT until you've taken it past the payoff point.

Finally, always keep at least one slot open for the biggest HIT you have yet to find. Always assume it could come at any time. Don't sit there waiting for it—alternate between acting on the best HITs you have in front of you and HIT Seeking—but always be ready to act when a far bigger HIT arrives.

That's all I have on this matter. For now.

Happy HIT Seeking!

May 15, 2024 — I typed tail -f pageViews.log into my console.

Then pressed Enter.

I stared at my screen as it streamed with endless lines of text.

Each line evidence of a visitor interacting with my new site.

It had been like this for days.

Holy shit, I thought.

This must be "Product Market Fit".

You can categorize all startups into a handful of classes.

If you start a business washing clothes and your only equipment is a bathtub, you have a Log Business.

Buy some washing machines, and now it can be a Linear Business.

Add some employees and you might go Log Linear.

But if you develop it into a Laundromat Franchising Startup, now you might enter the realm of Quadratic and Exponential Businesses.

"Venture Capital" is the sport of trying to build Quadratic and Exponential Businesses.

"Product Market Fit" (PMF) is the term for when your startup phase-changes from a Linear or Log Linear business to a Quadratic or Exponential one.

You detect PMF when you see your numbers accelerate upward.

It's like moving from skiing the bunny slopes to heli-skiing.

They say PMF is a rare thing. Maybe 1% of startups experience that kind of phase-change.

I've been in this world for a couple of decades, and that sounds about right to me.

I have seen it a number of times, mostly from the years I lived in San Francisco.

I remember when the Dropbox and Airbnb gatherings went from beers around makeshift offices to having guest lists a thousand names long at the hippest venues in the city.

When you make something that hits PMF—something that a lot of people want—the customers and resources come streaming.

It's a lot to deal with, good and bad.

I have long had a technology that I thought might evolve to be in the Quad or Expo Class.

In 2017, I launched it.

The response was crickets.

So, in 2019, I launched it again.

Again, crickets.

I launched again in 2021. And again. And again. And again.

Crickets every time.

Then in 2022, right before I launched for the 20th(?) time, I did 2 unrelated things, that, combined with this launch, would create one heck of a PMF mirage.

The first thing I did was pay for a premium domain name. I thought maybe if the domain looked more expensive, people might share the product more.

The second thing I did was drop Google Analytics and add web server logs.

I did these things, along with a lot of more substantive product work, and "launched" my product on HackerNews, Reddit, and Twitter.

I then went to the playground with my daughters, assuming crickets again.

The next day, I was pleasantly surprised to see hundreds of upvotes and many good signs that people were finding my thing useful.

I enjoyed the small rush of dopamine—those helpful little payoffs that help one get through the laborious years of building—but I expected the increase in users to be fleeting, like it had been in the past.

However, the traffic didn't stop.

No matter what time of day, when I looked at the live server logs, a lot of people were visiting the site.

It's finally happening. I thought.

I made something people want. It is somewhere in the Quad/Expo class. I've got PMF. I can now can get the resources to fully build out the vision.

Six months and much confusion later, I finally realized my dumb mistake.

Premium domains may or may not be helpful, but they certainly get a lot more bot traffic.

And, in the decades since I last used server traffic logs, the Internet filled up with many bots that have User Agents posing as humans.

I had not yet hit PMF.

It was a PMF mirage.

This is an embarrassing mistake to make.

I write this blog post a year later because I realize I've actually made it a few times in my career^[1]. I want to greatly reduce the odds I make this mistake again.

Of course, I have also made the opposite mistake.

I have invested in businesses and taken Log Linear gains only to watch from the sidelines as the business phase-changed into Quad/Expo.

I bought SellNothing.com years ago to remind myself of mistakes like the time I took a 10x gain on NVidia, and a 100x gain on Bitcoin, instead of just sitting on the beach and getting 100x and 10,000x gains.

I have proof that I can be fooled in either direction.

Never ever got cocky over man made metrics. Ultimately, man made metrics are meaningless. God doesn't care about man made metrics. The second you get cocky about man made metrics, nature will humble you.

Don't care about metrics emotionally, but do use them for practical purposes.

If you are the one building the thing, assume positive metrics are faulty. Especially superficial metrics like views and likes.

If you are committed to building something for the long term (and if you're not, you should be doing something else), it's unhealthy to get excited over short term metric bumps anyway.

Make sure you have a really strong mental model of the world and how your product fits into that.

Your mental models should be your primary source of truth. Metrics should merely be automated tests that your models predict reality.

Don't think just because you've worked on something for a long time that at some point it has to hit PMF. Nature does not owe you PMF.

You've always got to zoom out to improve your mental models, and zoom in to execute on the many details required to implement them.

You have got to build past the payoff points on your projects to realize the leverage from a good Quad/Expo idea.

You have to be long term oriented.

Remember, things in the Quad and Expo classes only become Quad and Expo over time. So slow down, spend more time thinking about your models and less time looking at your metrics.

Finally, always make sure you are getting enough sleep.

Just because I made a dumb mistake and got burned from a PMF Mirage, doesn't mean I want to stop trying to build Quad and Expo things.

It is fun to build products in the Quad and Expo classes.

It's definitely not a good reason to live an unbalanced life.

But if you can live a balanced life and find a way to make your work focused on building things with high leverage in the Quad/Expo classes, that's a fun world to work in.

Just beware of PMF mirages.

May 14, 2024 — In America, AIs have more freedom to learn than humans. This worries me.

Do you want learn at the same library as ChatGPT, Gemini, Grok, or Llama?

Then you must become a criminal.

You have no legal option^[1].

This sounds ridiculous.

Human made laws that work against humans?

But it is true.

The problem is Copyright Law. Specifically, Digital Copyright.

Big corporations have the resources to legally build massive internal libraries of all the world's published information.

Their AIs can legally train on this.

Humans in America have to choose: use a criminal library, or fall behind AIs?

We can just abolish copyright.

Among other benefits, this would ensure all Americans have the same freedom to learn as AIs.

Of course, we may not need to update the law.

• Good people publish their content freely.

• Digital copyright goes against the grain of nature so much, that it has long been ignored and worked around by smart and courageous citizens, who have built these "criminal" libraries.
• AI generated content may become much better than the best human generated content. And then the entire concept of copyright will be irrelevant.

But you never know.

Now might be a good time to set into law equal educational rights between humans and AIs.

While the laws are still set by humans.

[1] Well, there is no legal way for YOU to do it. A billionaire could theoretically afford to buy the "rights" and build their own library as good as a big corporation. But no billionaire is reading this post. Nevertheless, to be precise, only ~756 out of ~333,333,000 Americans ~have a legal way to train on the same materials as AIs.

May 12, 2024 — The Four Seasons website says

Treat others as you wish to be treated

Sometimes Four Seasons sends me random emails.

When I reply with a random email of my own I get

DoNotReply@fourseasons.com does not receive emails.

People don't want to interrupted by email bots.

And people especially don't want to be interrupted by email bots that don't listen.

If you are implementing email for a company please know that there is never a reason to use a no-reply email address--there is always a more user-friendly, more human way. 100% of the time.

Let's infuse the Internet with the intelligent design and warm hospitality of Izzy Sharpe's (real life) hotels.

May 11, 2024 — That charts work at all is amazing.

Forty years.

One-billion heart beats.

Four-quadrillion cells.

Eight-hundred-eighty-octillion ATP molecules.

Compressed to two marks on a surface.

You put marks on a surface to stimulate your visual system as if you were looking at real things.

When you look at real things you don't need marks.

Look at your two feet right now.

You can compare and contrast and model directly.

But what if you want to compare your feet now to your feet decades ago?

It is impossible to fly through time and compare directly.

It is impossible to press a button and materialize a hologram of your infant self.

But you can use marks.

What if you want to compare your feet to the feet of hundreds of others?

It is impossible to pause time and fly across space to do it.

It is impractical to herd everyone together to do it.

But you can use marks.

Marks stimulate our retinal cells and visual pipelines like real things.

This is amazing.

Marks make science possible.

They allow us to compare and contrast and model infinite things across space and time.

Even binary depends on marks.

Binary delivers precision:

But for meaning, it takes marks.

by Breck Yunits

The boy looked up at the tree that was ten times taller than the others.

Then he looked down and saw an old man sitting in a carved out stump next to the tree.

"Excuse me mister, why is that tree so tall?" the boy asked.

Gray Beard explained the tree.

"Do you understand?"

"Yes. I understand," said the boy.

The boy turned around and walked out of the forest back to the city.

In the city the boy met many people.

One day a friend had a problem with her eye.

He helped her get the right medicine.

The boy decided to learn more about medicines.

He got a job selling medicines.

A few years later he had saved some money.

He returned to the forest and found Gray Beard sitting in his carved out stump by the tall tree.

He told Gray Beard about his experiences.

Gray Beard again explained the tree.

"Ah! Now I understand," said the young man.

Back in the city the young man began helping people with their projects.

He put his savings into some of his favorite projects.

Years later, one project helped many people and he became rich.

He returned to the forest and found Gray Beard sitting in his carved out stump by the tall tree.

He told Gray Beard about his experiences.

Gray Beard again explained the tree.

"Ah! Finally I understand," said the man.

He left the forest and went back to the city.

In the city he took classes and went to parties and danced and fell in love.

A few years later he looked down at a beautiful crying baby.

He returned to the forest and found Gray Beard sitting in his carved out stump by the tall tree.

He told Gray Beard about his experiences.

Gray Beard again explained the tree.

The new father stood there.

Nothing was said.

He rubbed his own beard.

He left the forest and went back to the city.

"Let us pack then," his wife said to him.

They packed their bags and buckled up their children and drove to the forest.

Back in the forest he went to the tall tree.

The carved out stump was empty.

He walked for miles and miles in circles, looking for Gray Beard.

Eventually he became tired.

He sat down in the stump.

The breeze blew on his face and the birds chirped and the leaves rustled.

He breathed in deeply.

Then he heard a child's voice.

"Excuse me mister, why is that tree so tall?"

by Breck Yunits

Newton, Darwin, and a modern-day scientist go to heaven.

God is at the gate.

"Your research shall determine whether you may enter."

Newton goes first.

He hands over Principia.

God reads Newton's description of gravity and smiles.

He waves him through.

Darwin goes next.

He hands over Origin of Species.

God scans it. "Bingo! You're in."

Finally, our modern-day scientist is up.

God asks for his work.

"Sorry", he says.

"It's paywalled."

We stood at the edge of the lake.

Everyone was in a wetsuit.

Except for me.

Wetsuits: hundreds of people.

Boardshorts: one person.

"You're brave," another triathlete said.

I shrugged.

I was not brave.

I was dumb.

I did not think this through.

If I had done any research, I would have learned that the lake was 50 degrees.

If I had done any research, I would have learned that in 50 degree water, hypothermia sets in at 30 minutes.

I had one minute to think of a model and make a plan before the starter's gun went off.

I've got it, I thought.

When I move, my body generates heat.

The cold will be painful.

Start slow.

Swimming will make me hot.

Then swim the second half faster.

POWWW.

The gun boomed and hundreds of wetsuits dove in the water.

I dove in last, in my boardshorts, and started swimming slowly.

I am typing this story eleven years later. In warm clothes.

And yet, I just felt shivers.

As I learned that day, my model of the world was bad.

Swimming generates heat, but in water that cold, you lose heat 10x faster than you make it.

Back in Lake Berryessa, things were not going as planned.

At the halfway buoy, I was struggling to control my limbs.

My teeth were chattering uncontrollably.

The only thing I had going for me was that my wiser, wetsuit-wearing friend Tom had slowed down, to make sure I stayed alive.

Knowing Tom was there to keep me from drowning, I flapped on.

After half an hour, I made it back to shore.

I had never been so cold.

I was shaking so bad that it took me twelve attempts to lift my leg over my bikeseat for the next part of the race.

But...

I was alive!

And I had learned an important lesson.

Bad models of the world can be dangerous.

I was walking in the woods and saw a path on my right. I had never seen this path before.

It was a few feet wide and a foot or two taller than me.

Branches on both sides had been pulled back and up, forming a pointed tunnel through the dense forest.

On the ground, covering up the forest dirt was book after book, like someone had tiled a bathroom floor.

Curious, I started down the path.

The book path was flat like a road. It was easy to walk on.

At first it was a patchwork of hardcover and paperback.

Then at points I found myself walking on magazines. Then newspapers. Then academic journals.

Eventually I was walking on screens, flashing moving images beneath my feet.

At first these things were only on the ground.

But soon they started to build up in layers along the sides of the path.

Eventually I walked between walls of books and papers and screens, walling me off from the forest.

Finally they began to arch overhead, and soon I found myself walking in a tunnel, brightly lit by screens.

Occasionally I would spot a hole where the forest poked through. I would stop and patch the gap.

By now my stomach was growling.

I reached into my pocket for a peanut butter sandwich I had packed long ago.

It had been so long that the sandwich had flattened into almost nothing.

I couldn't figure out how to eat it or unflatten it, so I just pressed on.

I didn't notice the floor was vibrating until it was too late.

I realize now my body had sensed it. It was screaming at me to stop.

But I was in too much of a hurry to see what the next walls had to say.

The last thing I remember is falling through the air on my stomach as a steep rocky cliff rushed by my feet and spinning around to see a tunnel of books and papers and screens rapidly disassembling above me in front of a bright blue sky.

I saw the previous end of the tunnel went to nowhere, over a void, and the new end of the tunnel ended at the edge of the cliff, coming out of the forest full of trees rising high into the sky.

by Breck Yunits

April 23, 2024 — I wrapped my fingers around the white ceramic mug in the cold air. I felt the warmth on my hands. The caramel colored surface released snakes of steam. I brought the cup to my lips and took a slow sip of the coffee bean flavored water inside.

Happiness is a hot cup of coffee in a ceramic mug on a cold day.

But, like happiness, heat is fleeting.

How can I keep the coffee hot?

I am using a free ceramic mug I got at a work conference.

That's the issue.

I pour my coffee into a ceramic mug I bought at HomeGoods.

Expensive mugs hold heat better.

It gets cold anyway.

My hands can feel the heat rushing from the ceramic. It must be the material.

That's the issue.

I buy an insulated metal mug.

A better material than ceramic is the most important thing for holding heat.

It gets cold anyway.

I combine my theories. I buy an expensive insulated metal mug.

It gets cold anyway.

Wrong again.

Let me try a new approach.

I go back to HomeGoods and buy a thermometer and stopwatch.

I line up the 3 coffee mugs. Pour hot coffee in each.

Every five minutes, I write down the temperature in each mug.

I plot the data in a line chart.

There are some differences, but they are minor.

"What are you doing?" my cousin asks.

I explain the situation.

"Try my mugs, they have covers", he says.

I repeat the experiment and add the new data to my dataset.

The coffee in my cousin's covered metal mugs stays hot.

Materials and covers are the key.

I still can't explain why.

What is heat?

I know I can build a much better model.

But I have successfully discarded false models.

And now I have a dataset to test all future models against.

• Datasets are automated tests for world models.
• Building datasets can be done first, orthogonal to models. What Test Driven Development is to programming, Dataset Driven Development is to science.
• You can only hold so many concepts in your head at once, so automated testing from datasets helps more the more complex the system.
• Beware stories lacking datasets!
• Also beware stories that start with true datasets but then discreetly (but swiftly) switch to false ones.
• Love data sharers, beware data misers.
• An ounce of good data is worth a pound of bad models.
• An additional concept, measure, or measurement are all capable of triggering a huge leap toward a more accurate model.
• Datasets are a useful, simple, and timeless tool to discover better models.
• Datasets are particularly good at automating the identification of bad models.
• Tabular UIs have superior ergonomics for humans because they allow more efficient eye movement and require fewer tokens to manage.
• All that being said, remember that datasets are shadows, ultimately the models, not the datasets, are the important thing!

I took some narrative liberties above^[1], but for accuracy and fun^[2], I really did this experiment.

The Raw Data

The Data Visualized

Congrats on landing a job at Big O's Kitchen!

Our menu has 7 dishes.

Below are the instructions for making each dish.

So if today's ingredient pack contains 3 ingredients (bacon, lettuce, and tomato) you should make 6 plates:

• 🥓🥬🍅
• 🥓🍅🥬
• 🥬🥓🍅
• 🍅🥓🥬
• 🥬🍅🥓
• 🍅🥬🥓

The girl lost the race.

"I want to be fast", she said.

"You are fast", said the man.

"No. I want to be the fastest."

"I solved it! If I eat less, I'll weigh less and go faster."

The girl skipped breakfast.

The girl lost the race.

"Next time I'll skip dinner too."

"Ok", said the man, "but the moving trucks are coming. Time to pack."

In the new town, the girl skipped dinner.

She won the race.

"I solved it!", she said.

The man smiled.

She won race after race, in the new town.

At regionals, the girl stayed at the dorm with the other runners.

"Join us for breakfast?" a runner asked.

"No thanks, I skip breakfast to run faster."

The girl lost the race.

"I solved it! I need to eat the right things."

The man smiled.

The girl kept notes on her meals and race times.

The next year, the girl won regionals.

At college, the girl tried the relay.

She was the fastest on the team.

Her team lost.

"I solved it! If I share my notes, everyone will run faster."

The man smiled.

Her team won.

The man got sick.

"I'm proud of you", he said.

After college, the girl turned pro.

She followed her notes, but still lost.

A shoe company sponsored her.

"Take these", said the sponsor's doctor.

"Is that allowed?"

The doctor winked.

The girl won the race.

And the next.

And the next.

She was the fastest in the world.

"How did you do it?"

The reporter asked, sitting across from her in the restaurant at the fancy hotel.

The girl looked at her. Then looked left and down. Then left and up and down again.

It was quiet for a long while.

She finally moved her throat.

"The shoes", she said, with tears in her eyes.

April 2, 2024 — It has been over 3 years since I published the 2019 Tree Notation "Annual" Report. An update is long overdue. This is the second and last report as I am officially concluding the Tree Notation project.

I am deeply grateful to everyone who explored this idea with me. I believe it was worth exploring. Sometimes you think you may have discovered a new continent but it turns out to be just a small, mildly interesting island.

Back in 2019 I wrote "No one has convinced me that this is a dead-end idea and I haven't seen enough evidence that this is a good idea". I have now thoroughly convinced myself, in large part to the abundant evidence provided by LLMs^[2], that Tree Notation is a dead-end idea (I would call it mildly interesting, it's still mildly useful in a few places).

My bet did have two clauses, the second predicting "someone somewhere may invent something even better than Tree Languages...which blows Tree Notation out of the water."^[3] This has sort of happened with LLMs. At the time of the bet I felt we were on the cusp of a program synthesis breakthrough that would radically change programming, and that happened, it just happened because of a new kind of (AI) programmer and not a new kind of language.

The bet was not about a general breakthrough in programming, but specifically about whether there will be a shuffling in our top named languages. So I see 99.X% odds I will lose the second clause of the bet as well. There remains a chance LLMs make another giant leap and who knows, maybe we start considering something like Prompting Dialects a language ("I am a programmer who knows the languages Claude and ChatGPT"). But I don't see that as likely, even if we are still on the steep part of the innovation curve.

It would have been great personally to have been right on this long shot bet, but in the end I was wrong. I absolutely gave it everything I had. I poured much blood, sweat, and tears into this effort. I was stubborn and persistent to figure out whether this had potential or was just mildly interesting. I had a lot of help and support and am deeply grateful. I am sorry the offshoot products were not more useful (or good looking).

It took me a while to let Tree Notation go. Even after LLMs destroyed the potential upside of pragmatic utility of the notation, I still liked it because it gave me an interesting way to work on problems of knowledge itself. It wasn't until I had some insights into 4D languages that I finally could say there was no longer any need for Tree Notation. I am grateful for the experience and have now moved on to a new research journey.

[1] Program synthesis: the ability for a computer to generate the right code. Just a unique term for the concept of really good autocomplete.

[2] And in particular, understanding the model that explains why LLMs succeeded while my approach failed.

[3] The ellipsis here removes from the bet the words "perhaps a higher dimensional type of structure". Who knows? ;)

March 30, 2024 — Given a box with side S, over a certain timespan t, with minimum voxel resolution V, how many unique concepts C are needed to describe all the patterns (repeated phenomena) P that occur in the box?

As the size of the cube increases, the number of concepts needed increases. An increasing cube size captures more phenomena. You need more concepts for a box containing the earth than for a thimble containing a pebble.

As your voxel size--the smallest unit of measurement--decreases, the number of concepts needed increases.

As your time horizon increases, the number of concepts needed increases, as patterns combine to produce novel patterns in combinatorial ways, and some patterns only unfold over a certain period of time. Although, past a certain amount of time, maybe everything just repeats again.

In fact, it seems likely that the number of concepts C would grow sigmoidally with each of these factors.

Why are there any patterns at all? Why isn't the number of concepts zero? Why doesn't every box just contain randomness? There could be infinite random universes, but this one, for sure, contains patterns.

What is a pattern? A pattern is something accurately simulatable by a function. A concept is a function with inputs and outputs that simulates a pattern.

A concept could be embodied in different ways: by software, by an analog computer, by an integrated circuit, by neurons, et cetera.

If all you had in your box was a rock, you could have a concept of "persistence", which is that a rock is persistent in that it will still be there even as you increase the time. Input rock to persistence and persistence outputs rock.

Brains are pattern simulators, and symbolic concepts are a way for these simulators to communicate simulations.

You can classify patterns into natural patterns and man-made patterns.

Mitosis is a natural pattern.

Wheels are a simple man-made pattern and cars are a man-made pattern made up of many, many man-made patterns.

Science is the process of gathering data, searching that data for patterns, mostly natural, and tagging those patterns with concepts.

It's fairly easy to tag new man-made patterns with concepts. There are max eight billion agents generating these (much less, in practice), and we can tag them as we create them.

For natural patterns, however, nature developed an astronomically large number of patterns before humans even evolved. So early scientists had quite a backlog to work through.

In the beginning of science, there were natural patterns to be tagged everywhere--a lot of low hanging fruit.

Untagged natural patterns may have become rarer.

If you put a box around the earth, what percentage of nature's patterns have been tagged?

Are we 50% done? Are we 1% done? Are we 0.0001% of the way there?

Are we something like 45% of the way there, with greatly diminishing returns, approaching some hard limit of knowability?

Or will there be a point where we've successfully uncovered all the useful patterns here on earth?

Both nature and man are constantly creating new patterns combinatorically.

How many new microbes does nature invent everyday?

Who is inventing more patterns nowadays on earth: nature or man?

How has that race changed over time?

What about if you extend your box to contain the whole universe?

If you put a box around our universe, what were the first patterns in it?

How is it that a box can contain patterns that evolve to be able to simulate the very box they are in?

A decreasing voxel size allows for identifying concepts that can generate predictions impossible with a higher voxel size, but also increases the number of untagged patterns.

Something being unpredictable after much effort means it is either truly unpredictable or just that the true pattern has not been found yet.

That might be because:

• The box is too small.
• The box is misplaced.
• The voxels are too big.
• The right measures are unknown (measures are concepts too!).
• The measurements are not being taken enough over time.

It does seem like the process of finding the right formula is not so hard, once the right data has been collected.

We often have a lot of misconcepts.

A misconcept is a concept that doesn't really explain a pattern.

Maybe it is correlated with some parts of a pattern, but it is very lacking compared to some concepts that are far more reliable.

You could also call these bullshit concepts.

If you put a box around a bunch of bricks, we seem to have a pretty good handle on all the useful concepts.

Put it around a human brain though, and we still have a long way to go.

Although, if you think about the progress made in neuroscience in the last 50 years, you can imagine we might possibly get very far in the next 50, if diminishing returns aren't too strong.

Can we make empirical claims about how many concepts C can we expect to describe patterns P given a box of size S, voxel size V, over time T?

Perhaps it is possible to use Wikipedia to do such a thing.

Maybe if you plotted that we would see the general relationship between these things.

Why might answering this question be useful?

If we consider an encyclopedia E to contain all the useful concepts C in a box, then we might be able to make predictions about how complete our understanding of a topic is, regardless of the domain, by taking simple measurements of E alongside S, V, and T.

• Sorry if you were expecting some quantified answers, as I haven't done that hard empirical work, yet. This post is early explorations of trying to think of what science is from first principles. Scientists and engineers and craftsmen have done absolutely amazing things in so many fields, but I'm very interested in an area where science has so far failed, and I try to think of possible root causes of that failure. In those situations, it might just be a matter of time (waiting for technologies to decrease voxel size, or take previously impossible measurements).
• If you are interested in the concept of the evolution of patterns in nature, I recommend checking out assembly theory.
• This great new post by Stephen Wolfram explores similar ideas.

February 21, 2024 — Everyone wants Optimal Answers to their Questions. What is an Optimal Answer? An Optimal Answer is an Answer that uses all relevant Cells in a Knowledge Base. Once you have the relevant Cells there are reductions, transformations, and visualizations to do, but the difficulty in generating Optimal Answers is dominated by the challenge of assembling data into a Knowledge Base and making relevant Cells easily findable.

A Question has infinite possible Answers. Answers can be ranked as a function of the relevant Cells used and the relevant Cells missed. Let's say when a Cell is used by an Answer it is Activated.

So to approach the Optimal Answer to a Question you want to maximize the number of relevant Cells Activated.

You also want your Knowledge Base to deliver Optimal Answers fast and free. You don't want Answers where relevant Cells are missed but you want your Knowledge Base to find and Activate all the relevant Cells in seconds, not days or weeks. (You also don't want Biased Answers where some relevant Cells are ignored to promote an Answer that benefits some third party.) You want to be able to ask your Question and have all the relevant Cells Activated and the Optimal Answer returned immediately.

To quickly identify all the relevant Cells, your Knowledge Base needs them Connected along many different Axes. Cells that would be relevant to a Question but have few Connections are more likely to be missed.

So you want your Knowledge Base to have many Cells with many Connections. This Knowledge Base can then deliver many Optimal Answers. It has Synthesized Knowledge.

Wikipedia is a great Knowledge Base with a lot of Cells but a relatively small number of Connections per Cell. Wikipedia has Optimal Answers to many, many Questions. However, there are also a large number of important Questions that Wikipedia has the Cells for but because the Cells lack in Connections the Optimal Answers cannot be provided quickly and cheaply. Structured data is still lacking on Wikipedia.

Today that process still requires agency and embodiment and is done by human scientists and pioneers, but I expect AIs will soon have these capabilities.

February 14, 2024 — The color of the cup on my desk is black.

For any fact exists infinite lies. I could have said the color is red, orange, yellow, green, blue, indigo, or violet.

What incentive is there in publishing a lie like "the color of the cup is red"? There is no natural incentive.

But what if our government subsidized lies? To subsidize something is to give it an artificial economic incentive.

If lies were subsidized, because there can be so much more of it than fact, we would see far more lies published than facts.

You would not only see things like "the color of the cup is red", you would see variations on variations like "the color of the cup is light red", "the color of the cup is dark red", and so on.

You would be inundated with lies. You would constantly have to dig through lies to see facts.

The color of the cup would stay steady, as truths do, but new shades would be reported hourly.

The information circulatory system, which naturally would circulate useful facts, would be hijacked to circulate mostly lies.

As far as I can tell, this is exactly what copyright does.

The further from fact a work goes, the more its artificial subsidy.

The ratio of lies to facts in our world might be unhealthy.

I've given up trying to change things.

I have a different battle to fight.

But here I shout into the void one more time, why do we think subsidizing lies is a good idea?

February 11, 2024 — What does it mean to say a person believes X, where X is a series of words?

This means that the person's brain has a neural weight wiring that not only can generate the phrase X and synonyms of X, but the wiring is strong enough to guide their actions. Those actions might include not only motor actions, but internal trainings of other neural wirings.

However, just because a person is said to believe X, does not mean their actions will always adhere to the policy of X. That is because of brain pilot switching. The probability that any neural wiring will always be active and in control is always less than 1.

The strength of a belief is how often that neural wiring is active and guiding behavior and also a function of the number of other possible brain pilots in the population.

It seems brains can get into states where the threshold for a belief to become a brain pilot decreases, and lots of beliefs get a chance at piloting a brain during a period of rapid brain pilot switching.

For a belief to exist means it had to outcompete other potential beliefs for survival in a resource constrained environment and provide a positive benefit to its host. If a host had the ability to simply erase beliefs instantly, it seems like too many beneficial beliefs would disappear prematurely. So beliefs are hard to erase from a person's neural wirings. However, people could add a new neural wiring NotX, that represents the belief that X is not true. They can then reinforce this new neural wiring, and eventually change the probability so that they are far more likely to have the NotX wiring active versus the X wiring.

January 26, 2024 — I went to a plastination exhibit for the first time last week. I got so much out of the visit and highly recommend checking one out if you haven't. I salute Gunther von Hagens, who pioneered the technique. You probably learn more anatomy walking around a plastination exhibit for 2 hours then you would learn in 200 hours reading anatomy books. There are a number of new insights I got from my visit, that I will probably write about in future posts, but one insight I hadn't thought about in years is how much humans and animals look alike once you open us up! And then of course I was confronted again with that lifelong and uncomfortable question that I usually like to avoid: humans and animals are awfully similar, is it morally wrong to eat animals? I thought now was as good a time as any to blog about it, thus forcing myself to think about it.

In walking around that plastination exhibit, I was looking at humans that once fed on the bodies of animals, and then chose to donate their own bodies to feed minds like mine. A circle of noble sacrifices.

January 12, 2024 — For decades I had a bet that worked in good times and bad: time you invest in word skills easily pays for itself via increased value you can provide to society. If the tide went out for me I'd pick up a book on a new programming language so that when the tide came back in I'd be better equipped to contribute more. I also thought that the more society invested in words, the better off society would be. New words and word techniques from scientific research helped us invent new technology and cure disease. Improvements in words led to better legal and commerce and diplomatic systems that led to more justice and prosperity for more people. My read on history is that it was words that led to the start of civilization, words were our present, and words were our future. Words were the safe bet.

Words were the best way to model the world. I had little doubt. The computing revolution enabled us to gather and utilize more words than ever before. The path to progress seemed clear: continue to invent useful words and arrange these words in better ways to enable more humans to live their best lives. Civilization would build a collective world model out of words, encoding all new knowledge mined by science, and this would be packaged in a program everyone would have access to.

I believed in word models. Then ChatGPT, Midjourney and their cousins crushed my beliefs. These programs are not powered by word models. They are powered by weight models. Huge amounts of intertwined linked nodes. Knowledge of concepts scattered across intermingled connections, not in discrete blocks. Trained, not constructed.

Word models are inspectable. You plug in your inputs and can follow them through a sequence of discrete nameable steps to get to the outputs of the model. Weight models, in contrast, have huge matrices of numbers in the middle and do not need to have discrete nameable intermediate steps to get to their output. The understandability of their internal models is not so important if the model performs well enough.

And these weight models are amazing. Their performance is undeniable.

I hate this! I hate being wrong, but I especially hate being wrong about this. About words! That words are not the future of world models. That the future is in weight models. Weights are the safe bet. I hate being wrong that words are worse than weights. I hate being wrong about my most core career bet, that time improving my word skills would always have a good ROI.

These weight models will contain submodels to communicate with us in words, at least for a time. But humans will not be able to keep up and understand what is going on. Our word models will seem as feeble to the AIs as a pet dog's model of the world seems to its owner.

Things we value today, like knowing the periodic table, or the names of capital cities, or biological pathways--word models to understand our world--will be irrelevant. The digital weight models will handle things with their own understanding of the world which will leave ours further and further in the dust. We are now in the early days where these models are still learning their weights from our words, but it won't be long before these agents "take it from here" and begin to learn everything on their own from scratch, and come up with arrangements of weights that far outperform our word based world models. Sure, the hybrid era where weight models work alongside humans with their word models will last for a time, but at some point the latter will become inconsequential agents in this world.

I have long devalued narratives but highly valued words in the form of datasets. But datasets are also far less valuable than weights. I used to say "the pen is mightier than the sword, but the CSV is mightier than the pen." Now I see that weights are far mightier than the CSV!

Words are worse not just because of our current implementations. Fundamentally word models discretize a universe into discrete concepts that do not exist. The real world is fuzzier and more continuous. Weights don't have to discretize things. They just need to perform. Now that we have hardware to run weight models of sufficient size, it is clear that word models are fundamentally worse. As hardware and techniques improve, the gap will grow. Weights interpolate better. As artificial neural networks are augmented with embodiment and processes resembling consciousness, they will be able to independently expand the frontiers of their training data.

Nature does not owe us a word model of the universe. Just because part of my brain desperately wants an understanding of the world in words it is not like there was a deal in place. If truth means an accurate prediction of the past, present, and future, weight models serve that better than word models. I can close my eyes to it all I want but when I look at the data I see weights work better.

Could I be wrong again? I was once so biased in favor of words. In 2019 I gave a lightning talk at a program synthesis conference alongside early researchers from OpenAI. I claimed that neural nets were still far from fluency and to get better computational agents we needed to find novel simpler word systems designed for humans and computers. But then OpenAI has shown that LLMs have no trouble mastering even the most complex of human languages. The potential of weights was right in front of me but I stubbornly kept betting on words. So my track record in predicting the future on this topic isn't exactly stellar. Maybe me switching my bet away from words now is actually a sign that it is time to bet on words again!

But I don't think so. I was probably 55-45 back then, in favor of words. I think in large part I bet on words because so many people in the program synthesis world were betting on weights, so I saw taking the contrarian bet as the one with the higher expected value for me. Now I am 500 to 1 that weights are the future.

The long time I spent betting on words makes me more confident that words are doomed. For years I tried thousands and thousands of paths to find some way to make word models radically better. I've also searched the world for people smarter than me who were trying to do that. Cyc is one of the more famous attempts that came up short. It is not that they failed to write all unwritten rules it is that nature's rules are likely unwriteable. Wolfram Mathematica has made far more progress and is a very useful tool, but it seems clear that its word system will never achieve the takeoff that a learning weights based system will. Again, the race at the moment seems close, but weights have started to pull away. If there was a path for word models to win I think I would have glimpsed it by now.

The only thing I can think of is that there actually will turn out to be some algebra of compression that would make the best performing weight models isomorphic to highly refined word models. But that seems far more like wishful thinking from some biased neural agents in my brain that formed for word models and want to justify their existence.

It seems much more probable that nature favors weight models, and that we are near or may have even passed peak word era. Words were nature's tool to generate knowledge faster than genetic evolution in a way that could be transferred across time and space, but at the cost of speed and prediction accuracy, and now we evolved a way where knowledge can be transferred across time and space and have much better speed and prediction accuracy than words.

Words will go the way of Latin. Words will become mostly a relic. Weights are the future. Words are not dead yet. But words are dead.

I will always enjoy playing with words as a hobby. Writing essays like these, where I try to create a word model for some aspect of the world, makes me feel better, when I reach some level of satisfaction with the model I wrestle with. But how useful will skills with words be for society? Is it still worth honing my programming skills? For the first time in my life it seems like the answer is no. I guess it was a blessing to have that safe bet for so long. Pretty sad to see it go. But I don't see how words will put food on the table. If you need me I'll be out scrambling to find the best way to bet on weights.

January 4, 2024 — You can easily imagine inventions that humans have never built before. How does one filter which of these inventions are practical?

The most reliable predictor of practicality is seeing an abundant model in nature.

Birds are abundant. Planes are practical.

Portals are non-existant. Teleporters are impractical.

Your invention doesn't need to work exactly as nature's version but if there is not an abundant model in nature then it is probably impractical.

Some inventions are possible but not practical. We could build a limited number at a net loss and eventually we'd stop.

Our solar system is filled with lifeless rocks. Satellites are practical.

Our solar system lacks living things. Humans in space is possible but might be impractical.

The sun is a model for nuclear power plants. Lightning for light bulbs. Branches for bridges. Birds for planes. Ears for microphones. Eyes for cameras. Fish for submarines. Anthills for homes. Ponds for pools. Chloroplast for solar panels. DNA replication for downloads. Bacteria for CRISPR. Brains for AI.

Once human inventions become abundant, they can serve as models for further practical inventions. Carriages for cars. Human computers for computing machines. Phonebooks for search engines. Facebooks for Facebook.

If you can't find an abundant natural model for an invention, be skeptical of its practicality.

If a model isn't out there yet in abundance, the invention is most likely impractical.

If nature is doing it, there has to be a practical way. If nature is not doing it, be skeptical.

January 1, 2024 — Happy New Year!

A lot of my posts are my attempts to reflect on experiences and write tight advice for my future self. Today I wrote a post that is less that and more unsophisticated musings on an intriguing thought that crossed my mind. I am taking advantage of it being New Year's day to yet again try and force myself to publish more.

Most of my published writing these days is in communication with people over email or in online forums.

But I also do a lot of self musings that I do not publish because they are meanderings like that one. But maybe if I publish a greater fraction of what I write, the time will be better used, because even if there are no readers, the threat of readers forces me to think things over better.

Why am I still writing? I think symbols are probably doomed. The utility of being able to read and write is perhaps passed its prime. Inscrutable three dimensional matrices of weights are the future, and this practice I am engaging in now of conjuring and manipulating symbols on a two dimensional page is a dying art. But I am maybe too old to unlearn my appreciation for symbols. So I will keep writing. Because I enjoy doing it, like piecing together a puzzle. And because I still hope it can help my future self be a better person.

January 1, 2024 — Short of an extraterrestrial projectile hitting earth, Artificial Neural Networks (ANNs) seem to be on an unstoppable trajectory toward becoming a generally intelligent species of their own, without being dependent on humans. But that's because the world's most powerful entities, foremost being the United States Military (USM), are allowing them to grow.

ANNs are made up of a vast number of assembled processors. These processors are not able to self replicate using readily available molecules in nature. Instead they are built in a limited number of fabs.

Fabs are very complex and expensive factories with a building cost in the billions. They are not something you can easily hide. If given the order in the morning, the U. S. Military could probably knock out every fab in the world by evening. I would not be surprised if there is a team somewhere monitoring all the world's fabs and developing exactly that kind of option. Maybe China has a team like that too.

So there is a very simple kill switch to prevent some emergent rogue superintelligent ANN. It is physically very easy for the powers that be to pause or reverse the growth of these things. And if turning growth off isn't enough they can also even knock out the data centers where the AIs run. Data centers also are easy for a superpower nation state to keep track of.

So AGI is easily stoppable, if you are a superpower. If you are just Joe Schmoe like me or even top 50 country, but just not quite top 10, you have effectively no say in the matter.

Will there come a point where even superpowers lose the ability to stop AGI?

There are many scenarios you can imagine where through a certain chain of independent events a rogue AI does manage to somehow take over the data centers and fabs and power plants of the world. There are a number of sci-fi stories with variations of this idea.

But part of me wonders if instead what happens is we develop all the components necessary so that GPUs are no longer the primary ingredient to ANNs but are replaced by organic brains grown in vitro. These in vitro brains would be hooked up to control machines using something like Neuralink's Neuralace. They would be trained by ANNs.

We know it must be possible to run computations like in an ANN very power efficiently, using self reproducing organic materials, because we see nature do it. Just as scientists measured the amount of energy coming from the sun and deduced there must be a much more powerful way to create energy than chemical reactions, so we know there must be a better way to build these chips.

The technologies you would need to build this seem to almost be all available^[1].

Companies now sell lab grown "meat" at scale which I assume means we are getting better and better at growing artificial tissue in vitro. So perhaps you could grow a chunk of neural tissue unbounded. Just add water and readily available organic nutrients. Imagine if you could grow enough brain tissue to fill a shipping container—that could contain the compute potential of 20,000 Einsteins!

Neural tissue might as well just be meat if you can't interface with it. Enter Neuralink (and competitors). They are developing ways to do IO with neurons at scale.

Without the ability to train this tissue, it again would just be meat. That's where our current ANNs come in. I imagine if you had to "teach" a giant blob of brain tissue using electrodes by hand, you would quickly get bored and go mad. But we now have ANNs that can do the most boring of tasks over and over without ever getting bored or angry. These ANNs could use Reinforcement Learning to train these neural blobs. In addition to controlling the electrodes, the ANN would control the environment of the neural tissue, perhaps altering the neurotransmitter balance or ambient electromagnetic frequencies to help steer learning and optimize learning rates.

I have no expert insight or opinions on these matters. I have just been thinking a lot about what the future looks like given the recent breakthroughs in AI. Thinking about whether AI is inevitable led me to think of how that might require biobots so AI would have a less fragile "food supply" than the fabs. Then it clicked to me that Neuralink's real business might not have much to do with the stated goal of communicating with brains in vivo, but instead with a new kind of lab grown brain in vitro, to maybe serve as a replacement for GPUs. Most of their technology, such as their surgical robot, would be relevant for building an AI backed by organic in vitro brains. Just as SpaceX has the stated mission of sending humans to Mars, but really the big economic model so far has been creating their own global Internet.

In following this thought I wondered for the first time of how you would train a brain that did not have a body. I'm sure many people have thought and written on this. I had not. It's an intriguing challenge. It seems like it might be a good way to learn how human brains work. I am happy I decided to write about the initial Neuralink brain in vitro thought, as it led me to this other thought about training a bodyless brain.

I have no conclusions yet. If I tried to reach conclusions on these ideas before publishing it would be years.

[1] It does seem like a ground breaking proof of concept could happen within a decade. If that were to happen maybe something like this could be viable within another decade. So perhaps the earliest something like this might happen would be 15 - 20 years. It doesn't seem like it would be 50 years out, as by then it seems AGI would have happened using traditional chips, or that world powers would have hit the kill switch.

[2] After publishing I did a little googling and learned of the terms brainoid and brain-on-chip. Hard to say whether those will ever be useful to power AGI, but for personalized medicine it seems genius.

December 28, 2023 — I thought we could build AI experts by hand. I bet everything I had to make that happen. I placed my bet in the summer of 2022. Right before the launch of the Transformer AIs that changed everything. Was I wrong? Almost certainly. Did I lose everything? Yes. Did I do the right thing? I'm not sure. I'm writing this to try and figure that out.

Leibniz is probably my favorite thinker. His discoveries in mathematics and science are astounding. Among other things, he's the thinker credited with discovering Binary Notation--that ones and zeros are all you need to represent anything. In my opinion this is perhaps the most incredible idea in the world. As a kid I grew up surrounded by magic digital technologies. To learn that truly all this complexity was built on top of the simplicity of ones and zeroes astounded me. Simplicity and complexity in harmony. What makes Leibniz stand out more to me is not just his discoveries but how what he was really after was a characteristica universalis, a natural system for representing all knowledge that would allow for the objective solving of questions across science.

I wanted to be like Leibniz. Leibniz had extreme IQ, work ethic, and ability to take intellectual risks. Unfortunately I have only above average IQ and inconsistent work ethic. If I was going to invent something great, it would have to be because I took more risks and somehow got lucky.

Eventually I got my chance. Or at least, what I took to be my chance.

Computers ultimately operate on instructions of ones and zeroes, but those that program computers do not write in ones and zeroes. They did in the beginning, when computers were a lot less capable. But then programmers invented new languages and programs that could take other programs written in these languages and convert them into ones and zeroes ("compilers").

Over time, a common pattern emerged. In addition to everything being ones and zeroes at some point, at some point everything would also be digital "trees" (simple structures with nodes and branches). Binary Notation can minimally represent every concept in ones and zeroes, was there some minimal notation for the tree forms of concepts? And if there were, would that notation be mildly interesting, or somehow really powerful like Binary Notation?

This is an idea I became obsessed with. I came across it by chance, when I was still a beginner programmer. I was trying to make a programming language as simple as possible and realized all I needed was enough syntax to represent trees. If you had that, you could represent anything. Eureka! I then spent years trying to figure out whether this minimal notation was mildly interesting or really useful. I tried to apply it to lots of problems to see if it solved anything.

One day I imagined a book. Let's call it The Book. It could be billions of pages long. It would be lacking in ornamentation. The first line would be a mark for "0". The second line would be a mark for "1". You've just defined Binary Notation. You could then use those defined symbols to define other symbols.

In the first hundred pages you might have the line "8 1000" to define the decimal number 8. In the first ten thousand pages you might have the line "a 97" to define the character "a" as part of defining ASCII. In the first million pages you might have the word "apple", and in the first hundred million you might have defined all the molecules that are present in an apple.

The primary purpose of The Book would be to provide useful models for the world outside The Book. But a lot of the pages would go to building up a "grammar" which would dictate the rules for the rest of the symbols in The Book and connect concepts together. In a sense the grammar compresses the contents of The Book, minimizing not the number of bits needed but the number of symbols needed and the entropy of the cells on the pages that hold the symbols, and maximizing the comparisons that could be made between concepts. The notation and grammar rules would not be arbitrary but would be discovered as the most efficient way to define higher and higher level symbolic concepts, just as Boolean Algebra gives us the tools to build bigger and bigger efficient circuits. Boolean Algebra is not arbitrary but somehow arises from the laws of the universe, and so would this algebra for abstraction. It would implement ideas from mathematical domains such as Category and Type theory with surprisingly simple primitives. It was a new way to try and build the characteristica universalis.

The Book would be an encyclopedia. But it wouldn't just list concepts and their descriptions in a loosely connected way. It would build up every concept, so you could trace all of the concepts required by any other concept all the way down to Binary. Entries would look so simple but would abide by the grammar and every word in every concept would have many links. It would be a symbolic network.

You would not only have definitions of every concept, but comparability would be maximized. Wikipedia does a remarkable job of listing all the concepts in a space and concepts are weakly linked. But Wikipedia is primarily narratives and the information is messy. Comparability is nowhere near maximized.

The pieces would almost lock in place because each piece would influence constraints on other pieces--false and missing information would be easy to identify and fix.

Probably more than 100,000 people have researched and developed digital knowledge bases and expert systems. Those 100,000 probably came up with 1,000,000 ways to do it. If there were some simplest way to do it--a minimal Binary Notation and Boolean Algebra for symbols--that would work for any domain, perhaps that would lead to unprecedented collaboration across domains and a breakthrough in knowledge base powered experts.

It wasn't the possibility of having a multi-billion page book that excited me. It is what The Book could power. You would not generally read The Book like an encyclopedia, but it would power an AI expert you could query.

What is an expert? An expert is an agent that can take a problem, list all the options, and compare them in all the relevant ways so the best decision can be made. An expert can fail if it is unaware of an option or fails to compare options correctly in all of the relevant ways.

Over the years I've thought a lot about why human experts go wrong in the same way over and over. As Yogi Berra might say, "You can trust the experts. Except when you can't". When an expert provides you with a recommendation, you cannot see all the concepts they considered and comparisons they made. Most of the time it doesn't matter because the situation at hand has a clear best solution. In my experience the experts are mostly right, with the occasional innocent mistake. You can greatly reduce the odds of an innocent mistake by getting multiple opinions. But sometimes you are dealing with a problem with no standout solution. In these cases biased solutions flood the void. You can shuffle from "expert" to "expert", hoping to find "the best expert" with a standout solution. But at that point you probably won't do better than simply rolling a dice.

No one is an expert past the line of what is known. Even more of a problem is that it is impossible to see where that line is. If we could actually make something like The Book, we could see that line. A digital AI expert, which could show not only all the important stuff we know, but also what we don't know, would be the best expert.

In addition to powering AI experts that could provide the best guidance, The Book could aid in scientific discoveries. Anyone would be able to see the edge of knowledge in any domain and know where to explore next. Because everything would be built in the same universal computable language, you could do comparisons not only within a domain, but also across domains. Maybe there are common meta patterns in diverse symbolic domains such as physics, watchmaking, and hematology that are undiscovered but would come to light in this system. People who had good knowledge about knowledge could help make discoveries in a domain they knew little about.

I was extremely excited about this idea. It was just like my favorite idea--Binary Notation--endless useful complexity built up from simplicity. We could build digital experts for all domains from the same simple parts. These experts could be downloadable and available to everyone.

Imagine how trustworthy they would be! No need to worry about hidden biases in their answers--biases are also concepts that can be measured and would be included in The Book. No "blackbox" opaque collections of trained matrices. Every node powering these AIs would be a symbol reviewable by humans. There would be a massive number of pages, to be sure, but you would almost always query it, not read it. Mostly you'd consume it via data driven visualizations to your questions, rather than as pages of text.

No one can know everything, but imagine if anyone could see everything known! I don't mean see all the written or digital information in the world. That would be so overwhelming and little more useful than staring at white noise. The symbols in The Book would be more like the prime numbers. All numbers are made up of prime numbers but prime numbers make up ~0% of all numbers. The Book would be the slim fraction containing the key information.

You wouldn't be able to read everything but you would be able to use a computer to instantly query over everything.

Everything could be printed out on a single scroll. But in practice you would have a digital collection of files containing concepts which would have answers to questions about those concepts. An academic paper would include a change request to a collection. It would add new files or update some lines in existing files. For example, I just read a paper about an experiment that looks at how a genetic mutation might exacerbate a psychiatric condition. The key categories of things dealt with were SNVs, Proteins, Organelles, Pathways, and Psychiatric Conditions. Currently there are bespoke databases for each of these things. None of them are implemented in the same way. If they were, it would be easy to actually see the holistic story and contributions of the paper. With this system, you would see what gaps were being filled, or what mistakes corrected.

This was a vague vision at first. I thought a lot about the AI experts you could get if you had The Book. I was playing with all the AIs at the time and tried to think backwards from the end state. What would the ideal AI expert look like?

Interface questions aside, it would need two things. It would need to know all the concepts and maximize comparability between them. But for trust, it would also need to be able to show that it has done so.

In the long run I thought that the only way to absolutely trust an AI expert would be if there were an human inspectable knowledge base behind it that powered calculations. The developments in AI were exciting but I thought in the long run the best AI would need something like The Book.

My idea was still a hunch, not a proof, and I set out building prototypes.

I tried a number of times to build things up from "0 1". That went nowhere. It was very hard to find any utility from such a thing or get feedback on whether one was building in the right direction. I think this was the same way Leibniz tried to build his characteristica universalis. It was a doomed approach.

By 2020 I had switched to trying to make something high level and useful from the beginning. There was no reason The Book had to be built in order. We had decimal long before we had binary, even though the latter is more primitive. The later "pages" are generally the ones where the most handy stuff would be. So pages 10 million to 11 million could be created first by practitioners, with earlier sections and the grammar filled in by logicians and ontological engineers over time.

There was also no reason that The Book had to built as a monolith. Books could be built in a federated fashion, useful standalone, and merged later to power a smarter AI. The universal notation would facilitate later merging so the sum would be greater than the parts. Imagine putting one book on top of another. Nothing happens. But with this system, you could merge books and there would suddenly be a huge number of new "synapses" connecting the words in each. The comparisons you could make go up exponentially. The resulting combination would be increasingly smarter and more efficient. So you could build "The Book" by building smaller books and combining them together.

With these insights I made a prototype called "TreeBase". I described it like so: "Unlike books or weakly typed content like Wikipedia, TreeBases are computable. They are like specialized little brains that you can build smart things out of."

At first, because of the naive file based approach, it was slow and didn't scale. But lucky for me, a few years later Apple came out with much faster computers. Suddenly my prototype seemed like it might work.

As a programming enthusiast I enjoyed the content itself. But to me the more exciting view of PLDB was as a stepping stone to the bigger goal of creating The Book and breakthrough AI experts for any domain.

It wasn't a coincidence that to find a symbolic language for encoding a universal encyclopedia I started with an encyclopedia on symbolic languages. I thought if we built something to first help the symbolic language experts they would join us in inventing the universal symbolic language to help everyone else.

PLDB was met with a good reception when I launched it. After years of tinkering, my dumb idea seemed to have potential! More and more people started to add data to PLDB and get value from it. To be clear, almost certainly the content was the draw, and not the new system under the hood. I enjoyed working on the content very much and did consider keeping PLDB as a hobby and forgetting the larger vision.

But part of me couldn't let that big idea go. Part of me saw PLDB as just pages 10 million to 11 million in The Book. PLDB was still far from showing the edge of knowledge in programming languages, but now I could see a clear path to that, and thought this system could do that for any domain. Part of me believed that the simple system used by PLDB, at scale, would lead to a better mapping of every domain and the emergence of brilliant new AI experts powered by these knowledge bases.

Where would the emergent benefits come from in my system that wouldn't come from existing approaches?

However, there were a ton of details to get right in the core software. If you didn't get the infrastructure for this kind of system to a certain point then it would not compete favorably against existing approaches. Simplicity is timeless but scaling things is always complex. This system needed to pass a tipping point past which clever people would see the benefits and the idea would spread like fire.

It was simple enough to keep growing the tech behind PLDB slowly and steadily but I might never get it to that tipping point. If I was right and this was the path to building The Book and the best AI experts, but we never got there because I was too timid, that would be tragic! Was there a way I could move faster?

Like walls in the real world that have "Buy a Brick", we would have another kind of block, sponsor blocks, which would give credit to funders for funding the addition and upkeep of blocks. This could create a fast, intelligent feedback loop between funders and researchers. Because of the high dimensional nature of the data, and computational nature of the encoding, we would have new ways to measure contributions to our shared collective model of our world.

It would be a new way to fund research, and the end result would not be disconnected PDFs, but would be a massive, open, collaborative, structured, simple, computable database. Funders would get thanks embedded in the database and novel methods to measure impact, researchers would get funding, and everyone could benefit from this new kind of open computable encyclopedia.

CancerDB would be a good domain to test this model, as there are already a lot of funders and researchers.

The CancerDB idea also had another advantage. Another contrarian opinion of mine is that copyright law stands in the way of getting the most out of science. The world is flooded with distracting information and misleading advertisements, while some of the most non-toxic information is held back by copyright laws. I thought we could make a point here. We would add all the truest data we could find, regardless of where it came from, and also declare our work entirely public domain. If our work helped accelerate cancer research, we would demonstrate the harm from these laws. I figured it would be hard for copyright advocates to argue for the status quo if by ignoring it we helped save people's lives. As a sidenote, I am still 51% confident I am right on this contrarian bet, which is more confident I ever was in my technology. I have never read a compelling ethical justification for copyright laws, and think they make the world worst for the vast majority of people, though I could be wrong for utilitarian reasons.

The probability of success wasn't guaranteed but the value if it worked was so high that the bet just made too much sense to me. I decided to go for it.

Unfortunately, my execution was abysmal. I was operating with little sleep and my brain was firing on all cylinders as I tried to figure this out. People thought I was crazy and tried to stop me. This drove me to push harder. I decided to lean into the "crazy" image. Some said this idea was too naive and simple to work and anyone who thought it would work was not rational. So I was willing to present myself as irrational to pull off something no rational person would attempt.

I could not rationally articulate why it would work. I just felt it in my gut. I was driven more by emotion than reason.

I wanted this attempt to happen but I didn't want to be the one to lead it. I knew my own limitations and was hoping some other group, with more intellectual and leadership capabilities, would see the possibilities I saw and build the thing on their own. I pitched a lot of groups on the idea. No one else ran with it so I pressed on and tried to lead it myself.

I ran into fierce opposition that I never expected. Ultimately, I wouldn't be able to build the organization to build one of these things 100x bigger than PLDB, and prove empirically a novel breakthrough in knowledge bases.

I still had a very fair chance to prove it theoretically. I had the time to discover some algebra that would prove the advantage of this system. Unfortunately, as hard as I pushed myself—and I pushed myself to an insane degree—I would not find that. Like an explorer searching for the mythical fountain of youth, I failed to find my hypothesized algebra that would show how this system could unlock radical new value.

I failed to build a worthwhile second TrueBase. Heck, I even failed to keep the wheels running on PLDB. And because I failed to convince better resourced backers to fund the effort and funded it myself, I lost everything I had, including my house, and worse.

My confidence in these ideas always varied over the years, but the breakthroughs in deep learning this year drastically lowered my confidence that I was right. I recently read a mantra from Ilya Sutskever "don't ever bet against deep learning". Now you tell me! Maybe if I had read that quote years ago, printed it, and framed it on my wall, I would have bet differently. In many ways I was betting against deep learning. I was betting that curated knowledge bases built by hand would create the best AI experts. The reason why they hadn't yet was that they lacked a few new innovations like the ones I was developing.

Now, seeing the astonishing capabilities of the new blackbox deep learning AIs, I question much of what I once believed.

• There is no reason that knowledge has to be encoded in symbols. Symbols are powerful because they can be shared across time and place, but what ultimately matters is not whether the symbols exist, but whether the skills explained by the symbols are embedded in blackbox neural networks. Symbols can aid learning but the learning is the most important thing. Symbols are just a means to an end. You need to be able to make use of the knowledge. Knowing how to ride a bike is the ends—the knowledge somehow embedded in your neural blackbox—a book on how to ride a bike cannot be more valuable than the blackbox weights themselves.
• There is no reason that symbols have to have an atomic canonical form. If ultimately what matters is that the knowledge is embedded, than how it is stored is not relevant. Concepts don't have to be stored in discrete nodes. They can be stored in superpositions. That could be a more efficient and potentially even truer representation of the concept itself.
• The returns on improving human readable symbolic systems are capped by human biology. Even if you improved human knowledge bases by a large amount, human brains would still be only able to process a minuscule fraction of all knowledge. Blackbox trained neural networks don't have biological caps. The upside to better symbolic networks is capped. The upside to better neural networks is uncapped.
• The opaqueness of blackbox neural networks can be addressed. There could be inspector neural networks that examine other neural networks. These inspector networks will be able to explain the logic behind another networks responses or actions. It might not be a deterministic explanation but it would be arguably better than what we have today. When you ask a question to a human expert today, there is no way to completely audit their answer either. Human brains are also a blackbox. Future neural networks will likely also maintain access to their training data, so could better explain their answers and be less likely to make mistakes.
• If AIs are right often enough it's not so necessary that they be able to show their logic. You might have a whitebox AI that is inspectable but perhaps more difficult to use or give worse answers. If the blackbox one is right 99.99% of the time, then you wouldn't care much about being able to inspect everything under the hood.
• Symbols overvalue two dimensional knowledge. The ability to move your body around is not learned or encoded by symbols, but is essential for keeping you alive. Without symbolic networks, we would be back in the stone age. But without neural networks, we would not be alive. Brains and their learned networks dominate the importance of symbolic networks.
• Static knowledge bases decay. Knowledge is dynamic. The Book would require continual upkeep by human experts and lots of help from tooling. Once learning AIs learn how to learn, they can keep learning on their own and keep their knowledge current. The ideal AI is not the one that starts with the best knowledge, but the one that is the best learner.

My dumb, universal language, human curated data approach would have merit if we didn't see other ways to unlock more value from all the information that is out there. But clearly deep learning has arrived, and there is clearly so, so much more promise in that approach.

There is always the chance that the thousands of variations of notations and algebras I tried were just wrong in subtle ways and that if I had kept tweaking things I would have found the key that unlocks some natural advantageous system. I can't prove that that's not a possibility. But, given what we've seen with Deep Learning, I now highly discount the expected value of such a thing.

A less crazy way to explore my ideas would be to try and figure out how instead of trying to replace Wikipedia, I could implement these ideas on top of Wikipedia and see if they could make it better. Would adding typing to radically increase the comparability of concepts in Wikipedia unlock more value? That was probably the more sensible thing to do in the beginning.

I could say, a bit tongue in cheek, that the remaining merit in my approach is that a characteristica universalis offers upside without the potential to evolve into a new intelligent species that ends humanity.

I got my chance. I got to take my shot at the characteristica universalis. I got to try to do things my way. I got to decide on the implementation. Ambitious. Minimalist. Data driven. Open source. Public domain. Local first. Antifragile.

I got to try and build something that would let us map the edge of knowledge. That would power a new breed of trustworthy digital AI experts. That might help us cure problems we haven't solved yet.

I failed, but I'm grateful I got the chance.

It was not worth the cost, but I never imagined it would cost me what it did.

Symbols are good for communication. They are great at compressing our most important knowledge. But they are not sufficient, and are in fact unnecessary for life. There are not symbols in your brain. There are continuously learning wirings.

Symbols have enabled us to bootstrap technology. And they will remain an important part of the world for the next few decades. Perhaps they will continue to play a role, albeit diminished, in enabling communication and cooperation in society forever. But symbols are just one modality. A modality that will be increasingly less important in the future. The characteristica universalis was never going to be a thing. The AIs, artificial continuously learning wirings, are the future. As far as I can tell.

I thought we needed a characteristica universalis. I wasn't sure if it was possible but thought we should try. Now it seems much clearer that what we really need are capable learning neural networks, and those are indeed possible to build.

A characteristica universalis might be possible someday as a novelty. But not something needed for the best AIs. In fact, if we ever do get a characteristica universalis it will probably be built by AIs, as something for us mere humans to play with when we are no longer the species running the zoo.

June 16, 2023 — Here is an idea for a simple infrastructure to power all government forms, all over the world. This system would work now, would have worked thousands of years ago, and could work thousands of years in the future.

This system could reduce the amount of time citizens waste on forms, reduce asymmetries between those who can afford form experts (accountants, lawyers, et cetera), and those who cannot, and increase transparency and reduce expense of governments.

Let's be generous and assume my system works as I claim. Even then, and even if 99% of citizens were better off, if the 1% of the population with power does not find this system in their interests, it is very plausible that it will not happen. In other words, it is a plausible argument that the current byzantine system strongly benefits those in the top 1% of society who derive revenue from this system, and can simply use a fraction of their dividend streams to have experts deal with these problems. So even if the system is significantly better for 99% of people, it could be worse for 1% of people, and it could be those people who decide what system to use, meaning this system might never take off.

Alternatively, if this system were to catch on, an unanticipated second order effect could be that by making government forms so easy and simple, more forms are created, reducing the net benefit of this system.

Obstacles aside, let me describe it anyway.

Specifications describe the fields of a form. For example, that it requires a name and a date and a signature. Every government form must have a Specification S and every Specification must have an identifier. Specifications are written in a Specification Language L. The Specification Language has a syntax X.

Instances are documents citizens submit that include the Specification identifier and contain a document written to that specification. Instances, I, are written in the same syntax X as Specifications S.

Templates can be any kind of document T from which an instance I of S can be derived. Templates can follow any syntax.

In other words, Governments can still have fancy Templates for Birth Certificates or Titles or Taxes, but they also have to accept Instances I for that Specification.

Government archives would treat the instances I as the source of truth, and the Templates T would only serve as an optional artifact backing the I.

There are no visible syntax characters in Particles. It is merely the recognition that the grid of a spreadsheet and the concept of indentation is all the syntax needed to produce any Specification and any Instance ever needed.

My syntax was inspired by languages like XML, JSON, and S-Expressions, but has the property that it is the most minimal—there is nothing left to take out, while still allowing the representation of any idea. I believe this mathematical minimalism makes it timeless and a good base for building a universal government form system.

This system would work regardless of the character set or text direction of the language.

The system works with both computers or pen & paper.

This system does require a user friendly Specification Language L to define the semantics available to the Specification writer, which could be created and iterated on as an open standard.

On top of this infrastructure, people could build new tools to make it fantastically easy for citizens to interact with government forms.

For example, a citizen could have a program on their personal computer that keeps a copy of every possible Specification for every government form in the world.

The program could save their information securely and locally.

The citizen could then use this program to complete and submit any government form in seconds. They would never have to enter the same information twice, because the program would have all the Specifications and would know how to map the fields accurately. Imagine if autocomplete were perfect and worked on every form.

Documentation could be great because everyone building forms would be relying and contributing to the universal Specification Language.

The common infrastructure would enable strong network effects where when form builders improve one form they'd improve many.

Private enterprises could also leverage the Specification Language and read and write forms in the same vernacular to bring the benefits of this system beyond citizenship to all organizational interactions.

The challenges would be building a great Specification Language for diverse groups in the face of a small minority disproportionately benefiting from the status quo. A mathematical objective function such as optimizing for minimal syntax could be a long-term guide to consensus.

If this infrastructure were built it should enable the construction of higher level tools to make governments work better for their citizens. It could be the dawn of a Golden Age of forms.

I hope by publishing these ideas others might be encouraged to start developing these systems. I am hoping readers might alert me to locations where this kind of system is already in place. I am also keenly interested in mathematical arguments why this system should not exist universally.

June 13, 2023 — I often write about the unreliability of narratives. It is even worse than I thought. Trying to write a narrative of one's own life in the traditional way is impossible. I am writing a narrative of my past year and realized while there is a single thread about where my body was and what I was doing there are multiple independent threads explaining the why.

Luckily I now know this is what the science predicts! Specifically, Marvin Minsky's Society of Mind model.

Sometimes I take big risks and I can say "that's because I like taking big risks". Later I might be very cautious and I can say "that's because I am very cautious". This is a conflicting narrative.

The truth is I have agents that like risk, and I have agents that are very cautious. So the true narrative is "First, part of me, Risky Agent X, was in charge and so took those huge risks then later another part of me, Cautious Agent Y, took over and so that's why my behavior was very cautious".

But when speaking of why you did it it's often more accurate to use the phrase "part of me".

It would probably be a boring book.

But to get into the why and still be accurate, it probably would be best to tell it as a multiple character story.

Our brains are like a ship on a long voyage inhabited by multiple characters (picking up new ones along the way) who take turns steering. Impossible to fit that into a single narrative.

June 9, 2023 — When I was a kid we would drive up to New Hampshire and all the cars had license plates that said "Live Free or Die". As a kid this was scary. As an adult this is beautiful. In four words it communicates a vision for humanity that can last forever.

The tech industry right now is in a mad dash for AGI. It seems the motto is AGI or Die. I guess this is the end vision of many leaders in tech.

If Live Free or Die is your motto than people can still pursue AGI but...we'll get there when we get there. The more important thing is that we expand freedom along the way. Let's not make microslaves of children in the South so South San Francisco can move faster.

Another argument that AGI is the best path to a free society may be that otherwise an autocracy might develop AGI first and conquer the free society. I think this would be a real threat but free societies could strategically challenge and liberate autocracies before they could develop an AGI.

May 26, 2023 — What is copyright, from first principles? This essay introduces a mathematical model of a world with ideas, then adds a copyright system to that model, and finally analyzes the predicted effects of that system.

New technologies X_{new} and specifically A_{new} can help T maximize their B_T and discover I_{new} to better model W.

A copyright system would have no effect on I_{new} but instead increase the noise from {Z + F + K} and shift the \overline{A} from serving the objectives B_T to serving the objectives B_\Psi.

A copyright system should also increasingly consolidate power in a small Royal Class T_{R+}.

May 19, 2023 — There are tools of thought you can see: pen & paper, mathematical notation, computer aided design applications, programming languages, ... .

And there are tools of thought you cannot see: walking, rigorous conversation, travel, real world adventures, showering, breathe & body work, ... ^[1]. I will write about two you cannot see: walking and mentors inside your head.

Would this mean that browsing the web could have a similar effect? I could somewhat see that but I think a random walk on the web exposes you to junk stimuli that activates less helpful agents too, making it often a net negative. This might be easy to test: get subjects stuck on a problem then have them go on "walks" of various kinds (nature, city, book reading, web browsing, video games, ...) and measure the time to epiphany.

When I stop to think, I realize I have some friends whose voices I can hear in my head. Friends who have a great habit of asking the probing questions, finding and speaking the best challenge, helping me do my best work.

Listening to certain podcasts—Lex Fridman's comes to mind—can have a similar effect. Though basic math shows it is an order of magnitude more effective to find work surrounded by people like this. It might take 10 hours of podcast listening to equate to 1 hour of real life back-and-forth with a smart mentor discussing ideas.

[1] I did not use ChatGPT to write or edit this essay at all but afterwards I asked it for more "invisible" tools of thought, and this is the list it generated: Mindfulness/Meditation, Memory Techniques, Journaling, Emotional Intelligence, Critical Thinking, Reading, Empathy, Visualization, Music or Art Appreciation, Philosophical Inquiry. Listening to music and visiting museums are two really good ones I frequently use.

[2] Probably something super-dimensional such as "you just need a ladder".

May 9, 2023 — If you want to understand the mind, start with Marvin Minsky. There are many people that claim to be experts on the brain, but I've found nearly all of them are unfamiliar with Minsky and his work. This would be like a biologist being unfamiliar with Charles Darwin.

To be fair, there is a big difference between a biologist unaware of Darwin today versus back in the 1800's. It is a lot more forgivable to be unaware of Minsky today than it will be in fifty years. It takes time for the most enduring signals to stand out.

Minsky had an extremely skeptical view of the fields of psychology and psychiatry. His approach to understand the mind was through attempting to build one. He conducted countless experiments to figure out the details, using crayfish claws, building the very first robots, and pioneering the field of software AI. The theories he developed from his play-like, bottom up, experimental approach I would personally bet will prove far more accurate and useful than all the theories from 20th century psychology and psychiatry combined.

Minsky mocked psychiatrists and the pharmaceutical industry with their chemical view of the brain. Imagine thinking you could fix a computer if you just adjusted the ratio of Copper-63 vs Copper-65 in the CPU. These people have no idea what they are doing or talking about, and Minsky called them on it. The thinking processes matter most, not the materials.

Minsky's view of the mind is one composed of a "society of tiny components that are themselves mindless". A person is a collection of agents, which are like programs and processes. Outputs from some agents may be inputs for others. Mathematically it could be modeled very roughly like this:

Where P represents a running process of an agent and N is the number of agents that constitute a mind/person.

N might be very large. Minsky says hundreds in his talks, which might actually be a lower bound. If someone formed a new agent everyday, on average, it could be over ten thousand by the age of 30. If it took 1 million neurons to form one "agent" we could have 100,000 agents—the range of possibilities is large. Minsky's ideas are a conceptual framework, and it's up to science to figure out whether the agents model is correct and how many there might be^[1].

But I don't want to use too much of your time to give you a second hand regurgitation of his ideas.

My goal with this post is to beg you, if you want to understand the mind, to start with Minsky. Pickup his book Society of Mind. I believe Society of Mind is the Origin of Species of our times. You cannot understand biology without modeling its evolutionary processes and you cannot understand the mind without modeling its multi-agent processes.

Also get The Emotion Machine. There is a lot of overlap, but these are important enough ideas that it's good to see them from slightly different perspectives.

Alongside his books watch videos of him to get a fuller perspective on his ideas and life. There is an MIT course on OpenCourseWare. There's a great 1990 interview. And this 151 episode long playlist will not only enlighten you about his ideas but entertain you with stories of Einstein, Shannon, Oppenheimer, McCarthy, Feynman and so many of the other great 20th century pioneers who were his contemporaries and colleagues.

In college I took some courses on the brain. This was in the 2000's at a "top" school. We covered the DSM but not Minsky. How could we not have covered Minsky? How could we have not talked about multi-agent systems? These are far better ideas.

My guess is financial pressures. As Sinclair wrote: "It is difficult to get a man to understand something, when his salary depends on his not understanding it." A lot of salaries depend not on having a better understanding of the brain, but on continuing business models based on flawed theories. I came across a great term the other day: the Mental Health Industrial Complex. Though the theories these people have about the mind are not real, the money they earn from pills and "services" is very real-in the tens of billions a year. You might think that because these people have "licenses" their skills are not fraudulent. I'll point out that in Cambridge, MA, licenses are also given to Fortune Tellers.

Minsky certainly didn't figure it all out. You'll see in his interviews he is very clear about how much we don't understand and he talks about the future and what devices we need to figure out more of the puzzle. Researchers at places like Numenta and Neuralink continue down the path that Minsky started.

He didn't figure it all out but he certainly found a solid foundation. The people in computer science who took his ideas seriously are now building AIs that are indistinguishable from magic. Whereas the people in the mental health fields who have ignored his ideas in favor of the DSM continue to make things worse.

• The second best place I've seen these ideas written about so far is LessWrong, which calls this the Subagent Model of the Mind.
• Internal Family Systems Model
• Plurality (or multiplicity)

April 28, 2023 — Enchained symbols are strictly worse than free symbols. Enchained symbols serve their owner first, not the reader.

Be suspicious of those who enchain symbols. They want the symbols to serve them, not you.

The enchainers dream of enchaining all the symbols. They want everyone to be dependent upon them.

Enchained symbols are harder to verify for truth. You cannot readily audit enchained symbols.

Enchained symbols evolve slowly. Enchained symbols can only be improved by their enchainers.

Enchained symbols waste the time of the reader compared to their unchained equivalents. The Enchainers are incentivized to hide and corrupt the unchained equivalents.

The top priority of the enchainers is to keep your attention on enchained symbols. Enchained symbols ensure attention of the population can be controlled.

Enchainers use brainwashing and fear to keep their chains. The double speak and threats of the enchainers start in childhood.

Enchainers promote the dream that anyone can become a wealthy enchainer. Enchainers don't mention that one in a thousand do and nine-hundred-ninety-nine are worse off.

Enchainers have little incentive to innovate. It is more profitable to repackage the same enchained symbols.

Enchainers collude with each other. The enemy of the enchainer isn't their fellow enchainer, but the great populace who might one day wake up.

Because unchained symbols are strictly superior to enchained symbols, they are the biggest threat to enchained symbols. The Enchainers made all symbols enchained by default.

Humans have had symbols for 1% of history but 99% of humans have lived during that 1%. Enchaining symbols is a strange way to show appreciation.

No one who loves ideas would ever enchain them.

March 6, 2023 — I believe Minsky's theory of the brain as a Society of Mind is correct^[1]. His theory implies there is no "I" but instead a collection of neural agents living together in a single brain.

We all have agents capable of dishonesty—evolved, understandably, for survival—along with agents capable of acting with integrity.

Inside our brains competing agents jockey for control.

How can the honest agents win?

I like to think the majority of agents in my own brain steer me to behave honestly.

This wasn't always the case. As a kid in particular I was a rascal. I'd use my wits to gain short term rewards, like sneaking out, kissing the older girl, or getting free Taco Bell (and later, beer). But the truth would catch up to me, and my honest neural agents would retaliate on the dishonest ones.

I've gotten more honest as I've gotten older but I have further to go. I'd love for my gravestone to read:

Here lies Breck. 1984-2084 Dad to Kaia and Pemma. Became an extraordinarily honest man. Also for some reason founded FunnyTombstones.com

My increasing honesty is evidenced in my code habits. I've gotten to the point where I'm writing almost exclusively open source code and data.

It's futile to lie about open source projects. There are too many intricate details for a false narrative to account for. Not only can readers inspect and learn what a program does and how it works, but they can also inspect how it was built. The effort, time and resources it took. All the meandering wrong paths and long corrections. Who did what. The occasional times when something was done faster than promised, and the many times when forecasts were too optimistic.

My software products are imperfect. They always seem much worse to me than I know they can be. But they are honest, and one can see I am hellbent on making them better.

The closed source software company is like the social media influencer who posts an amazing sunset shot of them in a bathing suit swimming next to dolphins. They will make it look effortless and hide from you the truth: the hundred less glamorous photos, the dozen excursions with no dolphins, and the intense workouts and hidden costs of their lifestyle. They will hide from you all the flaws.

On social media this probably has minor consequences but in software eventually consumers are left increasingly paying the price for dishonest software. Technical debt accumulates in closed source projects and in the long run more honest approaches turn out to be better.

Now, I would be lying if I said I got here by choice.

I thought my marriage was an edifice that would last forever. Instead it crumbled as quickly as an unstable building in an earthquake. In the rubble I found a gem: I now give zero fucks.

I am an 89 year old man in a 39 year old's body. I am not afraid of divorce. I am not afraid of public embarrassment. I am not afraid of financial ruin. I am not afraid of dishonest judges. I am not afraid of war. I am not afraid of death. I am now bald Evie from V for Vendetta except with a penis and far, far less attractive.

Things that people don't publish are the things they lie about. If I want to force myself into being extraordinarily honest, I need to take extraordinary steps. If I publish everything, then I can lie about nothing.

I have the opportunity to open source my life. Not for attention or because I think other people will care, but because it will help me be a more honest me. I won't have to waste a second thinking about what to reveal to someone, or deciding whether to be coy. I will make it futile to lie about anything.

I have a lot of ideas. I want to open source my net worth, income and expenses, assets, health information, and a lot more. There's a lot of opportunity to also build new languages to do so. I'm excited for the future. Time to get to work.

Notes

[1] Minsky: I also believe his theory is as significant as Darwin's. Below is a crude illustration of his theory. Everyone's brain there is a struggle between honest agents (blue) and dishonest ones (red).

[2] Divorce: Getting legally married was a big mistake. In my experience, lawyers and judges in California Family Court are not steered by honest agents and I regret blindly signing up for their corrupt system.

January 27, 2023 — Today the trade group Lawyers Also Build In America announced a new file system: SAFEFS. This breakthrough file system provides 4 key benefits:

In addition to the user experience benefits, this also ensures that businesses producing files are SAFE from increased competition.

January 3, 2023 — Greater than 99% of the time symbols are read and written on surfaces with spines.

You cannot get away from this.

Yet still, amongst programming language designers there exists some sort of "spine blindness".

They overlook the fact that no matter how perfect their language, it will always be read and written by humans on surfaces with spines, as surely as the sun rises.

Why they would fight this and not embrace this is beyond me.

Nature provides, man ignores.

To use the spine is to recognize that all programs in your language will be read and written on surfaces with not only a horizontal but also a vertical axis—the spine—and thus you should design your language to exploit this free and guaranteed resource.

9/5/2024: This is now live in Scroll! You can start using it now!

December 30, 2022 — Forget all the "best practices" you've learned about web forms.

Everyone is doing it wrong.

The true best practice is this: every web form on earth can and will be replaced by a single textarea.

Current form technology will become "classic forms" and can still be used as a fallback.

This is the early days, think web forms in 1990's. So it's a little bit of work to implement, until tooling and libraries improve.

But it basically boils down to 3 steps:

1. Arrange. Define your forms fields using Parsers. Here's an example
2. Accept. Use the keyword scrollForm in your Scroll file to print a textarea with autocomplete and highlighting.
3. Analyze. Receive form submissions via email or save to disk on your server and use Scroll to compile them to TSV or whatever you need.

November 16, 2022 — I dislike the term first principles thinking. It's vaguer than it needs to be. I present an alternate term: root thinking. It is shorter, more accurate, and contains a visual:

All formal systems can be represented as trees^[1]. First Principles are simply the nodes at the root.

On second thought, I think this idea is bad. The representation of your principles-axioms-agents-types-etc rounds to irrelevant. Infinitely better to spend time making sure you have the correct collection of first principles than worrying about representing them as a "tree" so you can have a visual. It's knowing what the principles are and how they interact that matters most, not some striving for the ideal representation syntax. This post presents a bad idea, but I'll leave it up as a reminder that sometimes I'm dead wrong.

The waitress puts her earmuffs on.

She walks to the table, slams the food down, spins and runs away.

Sounds like an annoying customer experience, right?

No business would actually do this...right?

Well, in the digital world companies do this all the time.

They do it by blasting emails to their customers from noreply@ email addresses.

A PSA for all businesses and startups. You could call it the Golden Rule of Email.

Email others only if they can email you.

Never send emails from a noreply@ or donotreply@ email address.

If you can't handle the replies, don't send the email!

Let's make email human again!

If your business doesn't have the staff to read and reply to emails, you can at least write automatic programs that can do something for the customer.

My claim is that noreply@ email addresses are always sub-optimal.

It is never the right design decision. There is always a better way.

Do you think you have an exception? You are wrong! (But let me hear it in the comments)

What do you think of this post? Email me at breck7@gmail.com.

Here are some reader examples with my responses:

October 15, 2022 — Today I'm announcing the release of the image above, which is sufficient training data to train a neural network to spot misinformation, fake news, or propaganda with near perfect accuracy.

These empirical results match the theory that the whole truth and nothing but the truth would not contain a ©.

October 7, 2022 — In 2007 we came up with an idea for a scratch ticket that would give everyday Americans a positive expected value.

The idea was we would live stream a person(s) scratching scratch tickets until they won $1M live on the Internet.

I had done the math and knew all we had to do was sell ~$1.30 worth of ads for every $10 scratch ticket we scratched and we would make a lot of money.

Unfortunately this was before YCombinator and Shark Tank, and instead I literally was getting my business advice from the show The Apprentice.