AI: Cheap Intelligence

AI's deep meaning isn't any one app — it makes a kind of intelligence cheap and general, and it speeds up every other force.

The Future of Technology · Lesson 10 · 8 min read

For every force so far, a human still had to supply the thinking — design the chip, plan the power grid, interpret the genome. That thinking was the scarce, expensive bottleneck gating how fast everything else could move. What happens to all the other forces when the thinking itself — analyzing, designing, predicting — becomes cheap and available on demand? Hold that question; it is why so many people treat AI as pivotal.

AI makes a kind of intelligence cheap and abundant

Strip the hype. AI’s deep meaning is that tasks which used to require human cognitive work — finding patterns, predicting, drafting, solving structured problems — can increasingly be done by machines, cheaply and at scale.

This is the same shape as the other forces. Just as cheap compute and cheap energy crossed thresholds and unlocked new uses, cheap cognitive work does too: once a useful capability becomes nearly free, you stop rationing it and start applying it everywhere.

It is general — and it amplifies the other forces

AI applies across essentially every domain (the mark of a deep force), but its special role is as an amplifier. It helps design better chips (compute), forecast demand and discover materials (energy), and read and design genomes (biology). It is the force that acts on the other forces, which is why it can speed the whole system up at once rather than adding just one more capability.

Watch AI amplify the other three:
• Compute — AI helps design and lay out better chips, which makes compute cheaper, which makes AI cheaper: a loop.
• Energy — AI forecasts demand, balances grids, and helps search for better battery materials.
• Biology — AI predicts what a protein or DNA sequence will do and proposes designs, compressing years of trial-and-error.
In each case AI did not replace the force — it sped it up.

Powerful, imperfect, and ours to steer

Now the honest part, which is the whole track’s ethos in miniature. Today’s AI works by predicting plausible patterns, not by knowing the truth — so it can be confidently wrong or inherit bias from its training data. It is a superpower with a to-do list: neither magic nor doom.

Used deliberately — with verification, judgment, and care — cheap, general intelligence is arguably the most leveraged of the forces, because it makes humans and the other forces more capable rather than removing the need to think well. You get the leverage by steering it, not by trusting it blindly.

Suppose a small team wants to develop a new battery material. Before cheap intelligence, a few human experts could test only a handful of candidates a week — their thinking was the bottleneck. With AI: the model proposes thousands of plausible candidate materials, predicts which are worth trying, and the team spends its limited lab time only on the best few — then verifies the winners physically, because the model predicts plausibility, not truth. The force that got cheaper was the thinking, and it accelerated an energy problem. Same pattern works for chips and genomes — which is why AI is described as the force that speeds up the other forces.