I hear a lot of non-technical people saying things like, "if you talk to people in the know, it's going to revolutionize science in 5 years." But my personal experience has been that it's pretty difficult to really leverage effectively. For all the talk of it making the programming career disappear, the best I've managed to do with it is some shell scripts and a few bug fixes.
But I'd like to hear the State of HN on AI in 2025. Are you killing it with AI? Are you pretending to kill it with AI for that AI-powered paycheck? Are you still figuring out how to grok it?
The coder models continue to get better month over month. Give it 1-3 years and a lot of the lower-end dev jobs will be better done by an LLM.
Turns out humans aren't really that good at programming
My opinion is that it's great for senior guys, horrible for younger ones. If you know what can go wrong, you can use it the right way. As a junior, you usually don't and in the long term you will loose a lot of time and develop much less skills with AI.
I don't see it replacing people in my field anytime soon, if ever. I do see it replacing a lot of people in say web programming, in the same way that most people don't need to do their website now because they can just use a service like Wix. For anything non-trivial, there is 0 chance that AI will work within the next 5 years.
In the end, it's just a tool. You still need people to know what to do, the how is usually easier.
fnord77•3mo ago