AI Predictions - What's coming?
Emergence is the concept that something new and unexpected arises from simpler parts coming together - and the new thing has properties that none of the individual parts have on their own.
I wrote about some examples in Emergence And Opportunities.
As an emergent system develops, we enter into a new era we are entering into a new state. Almost like a shift from water to gas.
It is structurally different to the one before it.
As we can see with AI, we are moving into a new state.
The structural way the world works is going to be very different in the following ways:
The cost of acting on information collapses
Before the Industrial Revolution the cost of mechanical work was the binding constraint. Steam power reduced that cost. Everything reorganised around cheap mechanical energy - factories, cities, railways, global trade, new class structures.
The binding constraint now is cognitive work - the cost of processing information, making decisions, generating outputs, coordinating complex activities. AI is collapsing that cost in the same way steam collapsed the cost of mechanical work.
When a binding constraint collapses everything built around its scarcity gets reorganised. Every institution, profession, and power structure that exists because cognitive work was expensive is under pressure.
Individual capability expands dramatically
In the pre-AI age a single person’s productive capacity was limited by the hours they could work and the cognitive load they could sustain.
In the AI age one person can do what previously required teams. A solo founder can build what previously required an engineering department. A single researcher can survey literature that previously required a research group. A single advisor can serve clients at a quality previously requiring a whole firm.
This isn’t marginal improvement. It’s a qualitative shift in what individuals can accomplish without institutional support.
State Changes
We can see parallels in the massive change we are seeing with how water changes state. With heat it becomes gas, without heat it becomes ice.
The state change doesn’t happen at once, but it happens in stages. Water becomes gas with bubbles. Ice starts forming in specific parts then spreads until it all becomes ice.
Transitions aren't smooth and linear, they're lumpy, they nucleate in specific places first, and the early pattern tells you something about the mature state.
When water heats, it rises to 99 degrees, then when it reaches the thrashold, it breaks out, first in small parts, small concentrated bubbles form, then more, and it becomes more volatile.
The early pattern of a transition encodes real information about the mature state - because the forces shaping how something nucleates are the same forces that will shape what it becomes. This gives us real predictive power.
Nucleation is where and how the new structure first starts to form
Nucleation
The nucleation points matter because
It shapes what grows from it. The orientation - the values, assumptions, technical choices, power structures - of the nucleation point gets baked into everything that grows from it. The internet nucleated in academic culture and retained an open, distributed, non-commercial character in its fundamental architecture for decades because of that.
Early nucleation creates advantage that compounds. The first place the new structure forms attracts more resources, more talent, more development. The nucleus grows faster than anywhere else. Being the nucleation point isn’t just being first - it’s being the seed that everything else organises around.
Absorption or State Change
Not all state changes are enough. Sometimes it gets absorbed.
The printing press was something that was invented in China, but it didn’t produce the effect it did in Europe because it was absorbed by the existing structure.
The same happened for many other technologies. China invented printing. China invented gunpowder. China invented the compass. All technologies that drove state changes in Europe.
The irony the historical pattern suggests: the system that absorbs AI most smoothly into its existing structure may end up less transformed and ultimately less advantaged by it
What we can learn from state changes?
The key thing we can learn from state changes is that the nature of the technology in the small scale, and how it breaks up, gives us an insight and predictive abilities in how the new state is likely to be.
It sounds obvious when you think about it, but it has big implications.
AI for example can be described as dendritic - because it needs data centers, energy, and power, it will likely mean that the future system is likely to see a consolidation of power along those lines.
Open-source AI technology is actually quite different. It is low cost and can be run locally.
So the question is about the battle between how those two interact.
That AI's physical infrastructure requirements will produce concentrated power while open source produces a flatter distribution, and the battle between them determines the structure of the AI era.
Nucleus & Grains Alignment
When a new dendritic structure is forming, the nodes of where they are started are incredibly important, which is why there is such a race. The early parts matter
But they can also be more brittle.
And where something starts, doesn’t necessarily mean it has the innate advantages that persist.
Transitioning to a New State
The next 5 years feel more chaotic than the last 5 -not because things are getting worse but because we're in maximum turbulence during the transition. Multiple nucleation events simultaneously, institutions failing to keep pace, old frameworks failing before new ones consolidate.
How to apply these ideas personally?
Most people have not priced in emergence and state change happening - understand these is a massive advantage. The absorption assumption is the default.
You are probably making decisions based on the old state's logic. The state is changing. The gap between what most people assume and what is actually happening is where the significant opportunities are right now - and that gap closes as the state change becomes obvious to everyone.
Where are you seeing bubbles or new states forming?
Ignore the chaos and noise of the state change - The people who understand that what looks like chaos is actually formation - a new structure emerging rather than the old one simply breaking - will make better decisions than those who don't.
Which directions do you see things happening? You can see what’s happening in pockets and apply them universally.
The temptation is to go where the existing power and money is concentrated. But that's the old crystal. The opportunity can be identified at the nucleation sites.