Tech's Dumbest Mistake
The tech industry is making a massive mistake. By believing AI can replace programmers, it's killing the very ecosystem that keeps innovation alive. We’re about to enter a world where:
- Junior programmers will be undertrained and over-reliant on AI.
- Companies that fired engineers will be scrambling to fix the mess AI-generated code leaves behind.
- The best programmers will be so rare (and so expensive) that only the wealthiest firms will afford them. But hey, if tech companies really want to dig their own grave, who are we to stop them? The rest of us will be watching from the sidelines, popcorn in hand, as they desperately try to hire back the programmers they so carelessly discarded.
Another cope article about AI goes viral on HN. This one's a more obvious cope, concluding with an optimistic scenario for seasoned-serious-nonAI programmers. And a strongly worded warning to the companies that would make the super-programmers super-rich.
There's a good intersection between the HN community and LLMs-are-dumb community. There are two types of people in the world - those that look for things LLMs can't do and those that look for things LLMs can do well.
Okay, but for real now. This article reads like it was written 2 years ago. It's clear they haven't tried out the newer LLMs enough to learn to drive them. This is also totally ignoring the rate of progress - think ChatGPT-3.5 to o3-mini. Not being aware of the scaling laws or ignoring them altogehter is surprising to me, considering how the predictions have been more or less accurate and there's no wall in sight yet.
I do think experienced programmer's skills will be in great demand for a while to come (2 years maybe?). The optimistic view of devs being invincible could be a remote possibility if LLMs have fundamental limitations on their capabilites and aren't going to be capable of complex, longer-term agentic tasks. Sama mentions
(AI) will not have the biggest new ideas, it will require lots of human supervision and direction, and it will be great at some things but surprisingly bad at others.
Maybe this will be one of the things it's surprisingly bad at. But I just don't see how people won't attack this problem until it's straigthened out ($500B ought to budge it, right?).
Embracing the change and moving forward by incorporating AI as a tool in the vast toolbox of humans is a more viable path forward.
Agency, willfulness, and determination will likely be extremely valuable. Correctly deciding what to do and figuring out how to navigate an ever-changing world will have huge value; resilience and adaptability will be helpful skills to cultivate.
