Horses and Engines

I have friends (both pro- and anti-) interested in how AI is advancing; this paraphrased anecdote by Andy Jones is an interesting way to think of the times we’re in.

For centuries, a lot of human advancement came from horses—or rather, their motive force/labor.

In 1700, along came engines. They get better at a pretty steady clip of about 20% a year for about 200 years.

Until roughly 120 years ago, the horses, if they were capable of noticing, were unbothered. Horse populations and their roles in our transport and work grew as well.

And then? A small decline. And a bigger decline. And finally, between 1930 and 1950, 90% of the horses in the U.S. disappeared.

Progress in engines was steady; equivalence to horses was sudden.

OK, back to Greg speaking for Greg: We’re not horses—we can move up the value chain; they couldn’t. And, no, this isn’t some deterministic claim that humanity will suffer their fate. This is nothing more (or less) than an example of steady, incremental progress being dismissed until it’s too efficient, too effective, or just too damn cheap to ignore.

Then things happen quickly.

Are we at that point with AI? Depends who you ask. Unemployment among early-stage white-collar workers is up, and the lower unemployment rate that college grads have enjoyed vs. their non-degreed counterparts is the lowest it’s been in decades. If you’re say, 28 or younger? Yeah, it feels like the stable’s emptying out a bit.

The structure of my own field mirrors consulting, law, and a lot of other professional knowledge practices: An anthill of juniors bill mountains of hours while smaller numbers of senior folks deliver strategic value—or at least what passes for it. Senior folks historically needed the juniors to handle rote and near-rote tasks—things the seniors don’t want to do and, in any case, few clients are willing to pay their four-digit-an-hour retail rates. Plus, there are optics to consider: You want Mr. Four-Digits-Per-Hour in a tailored suit leading the engagement, even if he can knock out a press release.

Right now, the seniors are starting to experience a world where the army of juniors becomes a company, a platoon or even just a squad. Equity holders sing along: Feels good, man.

Depending on the circumstance, you can 10x a senior knowledge worker’s value creation with the AI tools that exist today. Some of that is speed; a lot of that is not paying AI for healthcare, sick days, office drama, or demands for career advancement. It’s compelling, even as the people who feel compelled are just one or two 20% improvements away from their own career crisis.

So many think they’re out of the meteor’s blast radius. Dinosaurs in the La Brea tar pits probably thought the same thing.

My point (Thank God, some of you are probably thinking) isn’t AI doomerism—I’m long on AI’s capabilities now and into the future.

But when progress is steady, tipping-point change is inevitable. And while resilience in the face of that change may not be 100% in your control, I’m guessing the number isn’t zero, either.

Become resilient. Learn new tools. Don’t pretend the meteor will land somewhere else.

It’s better than the glue factory.

Greg W. Brooks. First published on Facebook, Dec. 9, 2025.