Dumb as Hell; Goes Like Mad

Feynman again? Yep.

I probably introduce more Richard Feynman-isms quotes and tell more stories about Feynman’s life than anyone else here. It’s not that I worship him as a hero or anything like that—it’s just that I really like the way he thought about things and shared his observations with the world.

Feynman lived at just the right time to see the rise of digital computing, and although the field was still very young, and he had already had a long and illustrious career in physics, Feynman delivered some of the most consequential (and interesting) lectures on computer science ever given.

It was 1983-1984, probably around the time when I was first trying to figure out how to use this IBM 5150 computer that my dad was able to bring home from school for one semester:

While I was doing my own computer science experiments like trying to get games to load so I could play them, Caltech was offering a special class on the foundations of computation and how it all worked. Feynman’s take on this project included early lectures on quantum computing, reversible computing, and how to create efficient algorithms—all things that seem way ahead of their time in retrospect.

In the same way that Feynman looked at the universe and the laws of nature, he took a clean, fresh look at how computers worked. This might seem surprising if you know how prominent his work in physics was, but there were only a handful of people around who had been using computers as long as Feynman, and his experience at the Manhattan Project gave him a really good idea of what an electronic computer might be able to do.

Feynman figured out how to split the incredible calculating work he needed to do into actionable chunks, so he could tell one (human) computer to work on this part of the problem, then go tell another person (or group) to work on a different piece. Then, the pieces of the problem could all be collated later, and you’d have a solution to a massive problem.

This is parallel processing—when you run more than one thing at a time—and while Feynman certainly didn’t invent this idea, he understood and remarked that an electronic computer would be able to do this sort of thing all at once, under the hood—and he now understood how to take advantage of this understanding.

By exploiting parallel processing—lots of clearly organized chunks of a problem—you could solve a problem many times faster.

It probably seemed like magic to people looking in on the outside, but this phenomenon only increased as time passed. Computers got faster and faster, and they seemed like black boxes. Feynman wanted to demystify what they actually did inside.

To make his point, he used some colorful language. He said that computers were very fast at doing dumb things, and also called them morons at one point.

And, of course, the quote I used for today’s title:

The inside of a computer is as dumb as hell but it goes like mad!
— The Feynman Lectures on Computation (1983–84)

What did he mean by this, though? Clearly you could do some incredible things with computers—Feynman himself had seen the complex problems that could now be solved by these machines—so why the derogatory language?

What Feynman was driving at rhymes with the different types of thinking I introduced a couple of weeks back. Fast, deep, and wide thinking all cover different aspects of thought, and each are useful in their particular arena.

It’s important not to confuse fast thinking for deep or wide thinking in humans. Likewise, computers don’t know what they’re doing in the sense that a person does. They are really great at following very, very simple orders and executing them quickly. That’s sort of as far as things go, though.

In summary, Feynman wanted for us to understand that we are the ones doing the thinking, and the computer is a tool that’s here to help amplify our minds. Now, how our minds work is another matter entirely.