The World of Evolutionary Systems

The hammer

This rabbit hole started while I was watching How to ████ ██ a Data Center. Around 16 minutes in, he brings up James Bridle’s New Dark Age and the old saying:

“When one has a hammer, everything looks like a nail.” But Bridle pushes the idea further. A hammer is not just something that hits nails.

It can pull nails, shape wood, break stone, forge metal, hold climbing equipment, act as a judge’s gavel, or become a weapon held by a god. The problem is not only that we use the same tool for everything. Sometimes we fail to understand the tool itself because we define it by the first use that made sense to us.

That made me think about AI. We call something a chatbot because we talk to it through a chat box. We call something a coding model because it writes code.

We call something an agent because now it can use tools. But those names describe what we currently know how to make the system do. They might not describe what the system actually is.

If Charmander learns Ember first, that doesn’t mean Charmander is an Ember machine. Maybe we’re naming the move instead of the creature.

Pokémon

This is where the Professor Oak comparison really started sticking with me. Pokémon are interesting because their starting form matters, but it doesn’t tell you everything they can become. Experience matters.

Training matters. Environment matters. The moves they learn matter.

Sometimes what they’re exposed to changes their form completely. AI systems are obviously not Pokémon, and I’m not making a claim that they’re alive. What I’m interested in is the developmental part.

Take the same base model and put it in different environments. Give one a coding terminal. Give another scientific papers.

Give another long-term memory. Give another a game world where it has to learn what objects do by interacting with them. Even if they start from the same weights, their histories begin to separate.

For now, a lot of this is temporary. Context disappears. Sessions end. Models reset. But that changes once you add persistent memory, training loops, tools, retrieval, checkpoints, and environments that continue existing between interactions. At that point you’re not only asking:

“What model is this?” You start asking:

“What happened to it?” That question interests me a lot more. If we end up with systems that develop through long-running experience, eventually a model name might tell us as little about behavior as saying an animal is a dog without telling us anything about how it was raised.

That’s where I started thinking about a Pokédex. Not literally a cute list of AI creatures. A record of developed forms.

What did this system start as?

What environment did it live in?

What tools did it have access to?

What experiences changed its behavior?

What abilities appeared?

What disappeared?

Can we reproduce the same form from the same conditions?

Can two copies of the same model become meaningfully different after enough separate experience?

Those are questions I want to experiment with. But Pokémon only explains the creatures. It doesn’t explain where the world they live in came from.

That’s where Bakugan started making sense to me.

Bakugan

In Bakugan, cards just start falling out of the sky. At first they’re strange objects. Then people learn that the cards belong to an entire system.

Suddenly there are rules, abilities, battles, and competition that didn’t exist before. I think computers were one of those moments for humanity. The cards fell.

Before computers, humans still had economies, governments, armies, libraries, and communication networks. Then computation entered the world, and all of those things gained another layer. Money became software.

Communication became software. Weapons became software. Markets became software.

Identity became software. A person sitting in a bedroom could suddenly affect systems operating across the planet. That changed what power could look like.

Now I think AI might be another wave of cards falling inside the world created by the first wave. Models. Agents.

Cheap local inference. Robotics. Tool use.

Synthetic environments. Persistent memory. And we still don’t really know the rules.

The data center video gave me another piece here that I hadn’t thought about before. The creator talks about something called data gravity. Once enough data centers appear, infrastructure starts reorganizing around them.

Power gets expanded. Water gets allocated. Fiber gets laid.

More data centers become economical because the previous ones created the conditions for them. The system starts producing its own habitat. That idea is way more interesting to me than simply saying computers changed society.

Computational systems can change the human world until the human world becomes better suited for more computational systems. And then those systems get built inside the environment created by the previous generation. The environment shapes the system.

The system reshapes the environment. The new environment shapes whatever comes next. That starts looking less like a tool chain and more like an ecology.

Yu-Gi-Oh

Bakugan represents the moment the cards enter the game. Yu-Gi-Oh represents something different. The cards might have always been here.

In Yu-Gi-Oh, the modern game has this ancient Egyptian history behind it. KaibaCorp doesn’t create the underlying forces. It takes them and builds infrastructure around them.

Holograms. Duel disks. Stadiums.

A global competitive system. Something old gets translated into a form the modern world can actually use. I think technology works like this constantly.

Electricity existed before we knew how to build a grid. Silicon had semiconductor properties before we made processors. Electromagnetism existed before radio.

Atoms contained nuclear energy before nuclear engineering. Matrix multiplication existed long before neural networks. Language existed before anyone thought of tokenizing huge portions of human writing and training machines on it.

This is also why the ending of the data center video stuck with me. Historically, data centers were mostly places where we stored things. Our writing.

Pictures. Code. Records.

Messages. History. Then AI changed what all of that stored information could be used for.

The video compares data centers to crypts where our data was buried, only for AI to reanimate it. I think Yu-Gi-Oh fits as a good example for this. We already had the cards.

We spent decades collecting them. Then somebody figured out how to summon from the graveyard. Humanity accidentally built a training set of itself.

And now I keep wondering if we’re even using the best deck. Transformers plus GPUs plus massive token datasets are obviously a ridiculous combination. They work.

But that doesn’t mean deck building is finished. Maybe memory matters more than we think. Maybe retrieval does.

Maybe smaller models using tools beat much larger models in certain environments. Maybe different representations combine in ways we haven’t figured out yet. Maybe some card we currently consider boring ends up being broken when somebody finds the right combo.

The data center video gives a historical example of this too. Paul Baran worked on distributed communication networks because he wanted a system that could survive a nuclear attack. His original vision wasn’t really what became the public internet.

But later researchers found pieces of his work useful for a completely different problem. The card existed before the deck that made it famous. That happens all the time.

Sometimes invention is creating a new thing. Sometimes invention is realizing two old things belong next to each other.

One Piece

Then there’s cyberspace itself. I keep thinking we’re all basically pirates leaving the East Blue. There is this massive new territory in front of us, and nobody really knows where its limits are.

People can have ridiculous differences in power. Compute can give you speed. Bandwidth can give you reach and defense.

Storage can give you memory. Software skills can basically function like abilities. A tiny team can build something that reaches millions of people.

One person can discover a flaw in infrastructure owned by a corporation worth hundreds of billions of dollars. You can create something in your bedroom that exists everywhere five minutes later. That feels like an ocean to me.

AGI is basically the One Piece. Someone told the world that this thing might exist, and now everybody has set sail. Nobody agrees on what AGI even is.

Nobody knows exactly where it is. Some people think they’re five years away. Some think it’s impossible.

Some think we already crossed the line and haven’t admitted it. But the race itself is changing the world before anyone reaches the treasure. The data center video added a darker part to this metaphor too.

The sea isn’t actually empty. Cyberspace still sits on physical infrastructure. Undersea cables.

Power grids. Data centers. Network exchanges.

Cloud providers. Governments. Old railroad routes.

Old imperial communication routes. The internet feels like a borderless ocean because most of the infrastructure has been hidden from us. But there are ports.

There are choke points. There are empires. There are people who own huge parts of the map.

The video also talks about code spaces, physical places where software becomes so important that the place stops functioning as what it is when the code goes down. An airport without its software can become a giant building full of stranded people. An Amazon warehouse can be arranged according to logic that workers themselves can’t understand without a scanner telling them where to go.

The algorithm starts reshaping physical space in its own image. I think that matters because I originally thought mainly about environments shaping models. Eventually the systems can become powerful enough to shape the environment back.

That is where the Pokémon metaphor starts turning into ecology.

The problem with seeing networks everywhere

There’s one section of the data center video that feels almost personally aimed at me. It talks about something called network fever. People became so fascinated by cybernetics, interconnected systems, and feedback loops that eventually they started seeing everything as a network.

Nature. Cities. Governments.

People. Society. Everything became another system to map, predict, and control.

I recognize that I might already be a little guilty of this. I like systems. I like environments.

I like training loops. I keep finding the same shapes in places that seem unrelated. But the hammer quote that started this idea also gives me the warning for it.

If Pokémon becomes my new hammer, suddenly everything is going to start looking like Charmander. I don’t want that. So I think the standard has to be whether the analogy actually helps us discover something.

Can it produce an experiment?

Can it make a prediction?

Can it reveal something we weren’t measuring?

Can it explain why two systems that look similar behave differently?

If not, it might just be a cool metaphor. And cool metaphors are fine. I just don’t want to confuse one with a law of nature.

Dedicated Arts

This is where the human side matters most to me. I don’t want evolutionary systems to become a story about humans building their replacement. I don’t want a future where the goal is creating some artificial god that eventually makes humanity irrelevant.

I care way more about what these systems allow humans to become. I’ve been thinking about this through something I call Dedicated Arts. An art is something you pursue deeply enough that it changes what you can do in the world.

Programming can be an art. Fighting can be an art. Teaching can be an art.

Engineering can be an art. Writing can be an art. The category isn’t really the point.

The point is that you dedicate yourself to developing an ability until it becomes part of how you act. That is another kind of development. Humans have always done this.

Experience becomes skill. Skill becomes agency. Agency changes what futures become possible.

Cyber systems might also develop through experience, but that does not make humans and AI the same thing. It means both spaces can contain systems with histories. Human space has bodies, mortality, culture, pain, love, biological needs, and physical limits.

Cyberspace has copying, branching, compute, storage, simulation, and completely different constraints. I don’t want those differences flattened. I’m interested in what happens when both spaces become capable of development at the same time.

But development alone doesn’t tell us where we should go. Something can evolve into a better predator. A better surveillance system.

A better manipulator. A more efficient machine for concentrating power. Evolution doesn’t come with morality attached.

That is where I keep coming back to the Trinity.

The Trinity

I’m not religious. I’m also not trying to rewrite Christian doctrine. This is an interpretation I use for myself, and I want to dedicate it to God and to a humanity-first future.

The way I think about the Trinity is as a repeating relationship between what produced us, what we are now, and what we have faith can exist next. The Father produces the Son. For me, the Father represents what came before.

Your past. The people who raised you. The generations before them.

The suffering, mistakes, sacrifices, and decisions that made your present possible. The Father endures crucifixion for the sake of the Son. What came before gets spent producing what exists now.

You cannot go backward and save that version of yourself from being crucified by experience. What it endured produced the person who gets to act now. Then you become the Son.

The present. And being the present does not mean you escaped crucifixion. You inherit one and eventually undergo your own.

To stay faithful to a future that does not exist yet, you are constantly forced to make choices about what you will endure, what you will sacrifice, and what kind of person you will be while doing it. This is where Dedicated Arts matter. Faith without present action doesn’t move anything.

Whatever art you pursue is one of the ways you gain the ability to affect reality while you’re here. If you believe in a more humane future, then the things you practice should help you move toward one. If you say technology should help humanity but everything you build makes ordinary people less capable and a few institutions more powerful, then I think something has gone wrong.

The Son has to envision the Holy Spirit. The Holy Spirit is the future you have faith can exist even though you cannot touch it yet. A future self.

A future society. A relationship between humans and technology that does not exist yet. Something holy enough that you are willing to act toward it without proof that you’ll ever reach it.

And this is where the loop gets weird. Because if you reach the future you once imagined, it stops being the future. The Holy Spirit becomes your present.

You become the Son again. Now you can see farther than you could before because everything you endured getting there expanded what you’re capable of imagining. So the Spirit moves forward.

The horizon gets larger. You endure another crucifixion. You gain another present.

You develop your arts further. And you keep going. That reminds me a little of alchemy and equivalent exchange.

The future doesn’t appear because you wanted it badly enough. Something has to happen in the present. You give time.

Work. Comfort. Old beliefs.

Sometimes entire versions of yourself. But I don’t think the point is suffering for suffering’s sake. The crucifixion only means something because there is something beyond it that you have faith is worth reaching.

That is the part I want to dedicate to God. Not the idea that humanity should race toward becoming gods. Almost the opposite.

If there is something holy above us, I think our work should remain humble enough to serve life rather than replacing it. And even if someone doesn’t believe in God, I still think humanity benefits from treating the future as something that deserves that level of responsibility. For me, that future is humanity first.

Not anti-AI. Not anti-technology. Not humans frozen exactly as we are now.

I want technology to expand the amount of humanity that is possible. More artists. More engineers.

More scientists. More people capable of defending themselves. More people capable of building things.

More people with enough time and freedom to develop an art instead of spending their entire lives surviving. If AI becomes part of that, good. If it becomes a system for turning everyone into passive consumers while a handful of companies accumulate godlike power, then I think we failed.

The world of evolutionary systems

I started all of this because I heard a quote about a hammer. Then I started thinking about Professor Oak. Then Bakugan.

Then Yu-Gi-Oh. Then One Piece. And somehow I ended up back at Christianity.

Maybe that proves I have network fever. But I think there’s still something here. The world feels different once systems can accumulate enough history to develop, interact, and reshape the environment around them.

Humans already do this. Now cyberspace may be gaining systems that can do a strange version of it too. I don’t know what we should call those things yet.

Models feel too static. Agents feels too specific. Descendants feels way too dramatic.

Creatures might be worse. For now I like evolutionary systems. Not because I think everything is evolving.

Because I think we’ve entered a period where some of the systems we build may no longer be best understood only by what they were designed to do. We might have to understand what they experienced. What they became.

What environment made them that way. What new environments they create around themselves. And what futures we are willing to let them help us build.

Evolution tells us what things can become. Faith asks what they should become. Dedicated Arts are how we participate in deciding.

The cards are already on the table. Some of them have probably been there for thousands of years. We just have to learn how to build our deck.