Michael Smith writes about the distinction between tools that make us smarter vs dumber. I thought the comparison between abacuses and calculators was memorable:
Learning how to use an abacus trains your brain to internalize it. Arithmetic becomes faster and more reliable over time, and the mechanisms behind why different strategies work become obvious and intuitive. Eventually you don’t even need the physical abacus anymore. Whereas with a calculator … those mental skills sort of fade away over time. And you will always need a calculator for math: it never becomes part of you the way an abacus does.
This topic has many dimensions, which means it lends itself a little too easily to simplification.
Obviously, enlightening tools are better than stultifying ones. And obviously, certain educational benefits only come through unpleasant hard work. So lets have demanding tools that educate us.
However, also obviously, there’s value in tools that easy to use. So let’s make tools pleasant and effortless, and save education for classrooms.
Also, somewhat obviously, new tools are often not simply easier. They make one kind of hardness go away but introduce a new kind of hardness, which is educational in a new way. Right now, for instance, there exist people who are expert at writing code, but who are so bad at prompting LLMs to generate good code that they still claim it cannot be done!
In other words, there are a lot of obviously true points at play but they all point in different directions. Analogies are great in this situation, because they are just a specific, memorable peg for a particular set of tradeoffs.
So let’s follow the analogy. The idea is, an abacus is better than the calculator because it helps you internalize arithmetic. I buy that idea. That’s why I have a slide rule by my desk, in the hope it will help me internalize logarithmic relationships. (It’s not working.)
But…what are you really internalizing? Memorably, Feyman tells a story about initially losing in a mental calculation competition vs an abacus salesmen, but then ultimately winning as the problems because more complex, specifically because the abacus encouraged a mental skill whih was too rote and procedural, and did not promote insight:
A few weeks later, the man came into the cocktail lounge of the hotel I was staying at. He recognized me and came over. “Tell me,” he said, “how were you able to do that cube-root problem so fast?”
I started to explain that it was an approximate method, and had to do with the percentage of error. “Suppose you had given me 28. Now the cube root of 27 is 3 …”
He picks up his abacus: zzzzzzzzzzzzzzz— “Oh yes,” he says.
I realized something: he doesn’t know numbers. With the abacus, you don’t have to memorize a lot of arithmetic combinations; all you have to do is to learn to push the little beads up and down. You don’t have to memorize 9+7=16; you just know that when you add 9, you push a ten’s bead up and pull a one’s bead down. So we’re slower at basic arithmetic, but we know numbers.
Right now, many worry that LLMs will make us get worse at writing code. I think they probably will. But they may also be inviting us to get better at something deeper.