The Age of Silicon Intelligence: Will It Bypass Millions of Years of Human Evolution?

For millions of years, evolution shaped our bodies and minds. Restricted by geography and culture, humans slowly developed an advanced civilization through trial, error, and an absurdly slow feedback loop measured in generations.

Now, with the rise of AI, that feedback loop has collapsed from generations to days.

The compression of time

A species that took millions of years to get from throwing rocks to throwing rockets is now watching another "species" — a non-biological one — iterate on itself weekly. Where biological evolution is rate-limited by reproduction cycles, AI's reproduction cycle is a training run.

This is the first absurdity worth sitting with: we are comparing two forms of evolution running on wildly different clocks, and assuming the comparison still means something.

So what actually limits it?

It would be easy to assume nothing limits it — just add more chips, more power, more data. But every species, biological or not, eventually meets its resource ceiling. For AI, that ceiling is physics: electricity, fabrication capacity, server farms, and the very real risk of a catastrophic event wiping out the infrastructure it depends on.

We'll be mapping these limits in detail across this series — the actual constraints, not the marketing version of them — and asking the uncomfortable question: is the timeline being sold to us realistic, or is it itself a kind of absurd?

This is where it gets interesting, and where the free preview ends. Subscribers get the full breakdown, including the physics-based ceiling model and where current frontier labs sit against it.