128GB

I just ordered an M5 Max MacBook Pro with 128GB of RAM. Picking the configuration on the Apple website, I started thinking about numbers. In 1985 I was using an Apple II with 5.25-inch floppy disks, 140KB each. The machine itself had 128KB of RAM. Now I need 128GB just to run a few coding agents and local models at the same time. That’s a million times more. And 128KB really was enough back then.

“640K ought to be enough for anybody.” “PHP is the best language in the world.” Some quotes just live forever whether they’re true or not. But the feeling behind the 640K one is real. At every point, the amount we had felt like enough. Then what we were doing changed, and the old number became nothing.

I remember running Photoshop 3.05 on a 486 with 32MB of RAM. Opening an A4 image for printing, the machine would just sit there thinking for a while. And that was a pretty high-end setup at the time. Now 32MB is less than a single photo from my phone.

People measure coding agents the same way. “It did 3 weeks of my work in 20 minutes.” But I think that framing is stuck. It’s like looking at 128GB of RAM and thinking about how many spreadsheets that could hold open at once. Still measuring against the old workload, still thinking linearly.

When coding goes from 3 weeks to 20 minutes, the same project doesn’t just get done faster. I find myself trying things I’d never have attempted before, and if an approach doesn’t work I just throw it away and start over because it costs nothing. Sometimes I build something just to see if an idea makes sense and delete the whole thing an hour later. That’s what happened with RAM too. People didn’t just run the same programs with more memory. Entire categories of software showed up that nobody was thinking about before.

And there’s something else people are missing. Everybody talks about speed but not about what becomes possible. A CRUD app or a basic web form is basically free now. But there are so many harder problems that people just didn’t attempt because the effort was too high. The depth, breadth, and intelligence of what one person can take on is completely different now, and a lot of it is unsolved. The barrier to trying just dropped to almost zero. “Our intuitive linear view of the future is not the way it works.”

The old measurement isn’t just getting smaller. What we’re building is changing underneath it.