A eulogy for coding in the age of GPTs. I have LOTS of thoughts, but this piece by James Somers for The New Yorker was a wonderful read. I usually try to avoid overly quoting from an external piece, but this section captures the crux too well:

Perhaps what pushed Lee Sedol to retire from the game of Go was the sense that the game had been forever cheapened. When I got into programming, it was because computers felt like a form of magic. The machine gave you powers but required you to study its arcane secrets—to learn a spell language. This took a particular cast of mind. I felt selected. I devoted myself to tedium, to careful thinking, and to the accumulation of obscure knowledge. Then, one day, it became possible to achieve many of the same ends without the thinking and without the knowledge. Looked at in a certain light, this can make quite a lot of one’s working life seem like a waste of time.

But whenever I think about Sedol I think about chess. After machines conquered that game, some thirty years ago, the fear was that there would be no reason to play it anymore. Yet chess has never been more popular—A.I. has enlivened the game. A friend of mine picked it up recently. At all hours, he has access to an A.I. coach that can feed him chess problems just at the edge of his ability and can tell him, after he’s lost a game, exactly where he went wrong. Meanwhile, at the highest levels, grandmasters study moves the computer proposes as if reading tablets from the gods. Learning chess has never been easier; studying its deepest secrets has never been more exciting.

Computing is not yet overcome. GPT-4 is impressive, but a layperson can’t wield it the way a programmer can. I still feel secure in my profession. In fact, I feel somewhat more secure than before. As software gets easier to make, it’ll proliferate; programmers will be tasked with its design, its configuration, and its maintenance. And though I’ve always found the fiddly parts of programming the most calming, and the most essential, I’m not especially good at them. I’ve failed many classic coding interview tests of the kind you find at Big Tech companies. The thing I’m relatively good at is knowing what’s worth building, what users like, how to communicate both technically and humanely. A friend of mine has called this A.I. moment “the revenge of the so-so programmer.” As coding per se begins to matter less, maybe softer skills will shine.

Somers goes on to talk about how what he thought he’d teach his children about programming has changed. I think about that too. I have kids, and coming from a programming/computer science background, I’ve always assumed I could give them a leg up in life by teaching some of these skillsets early.

But the proliferation of LLMs and AI have changed my thinking. More than how to hack together a bash script to do in one click what used to take five, I think learning how to break down problems into chunks that a computer can tackle—systems thinking—is a more relevant skill to acquire early. Paired with actual application (either through directly writing code or interacting with an LLM), I’ll likely net the same effect I was hoping for.

I have a deep love of programming. GPT-4 (or whatever variant you’re using) has only deepened that love. Having an always available, deeply knowledgeable, and easy-to-use thing that I can just throw questions or problems at feels a bit like magic. Not because it teaches me something I couldn’t ultimately learn on my own, but rather the sheer speed at which it increases my learning has become its own high that I now look forward to chasing whenever I sit down in front of the keyboard.

Thursday, 16 November 2023

Read more about this site, or follow via email, RSS, JSON.