Hacker News Reader

Bro, it sure if you noticed. ChatGTP isnt that great at coding end to end. It can regurgitate common examples well, but if your working on large technical code bases it does more harm that good. It need constant oversight, why don’t I write the code myself. We are at an infrastructure limit, not sure we are going to see order of magnitude improvements any more.

View on HN

I no longer write code. I’ve been a swe for over a decade. AI writes all my code following my instructions. My code output is now expected to be 5x what it was before because we are now augmented by AI. All my coworkers use AI. We don’t use ChatGPT we use anthropic. If I didn’t use AI I would be fired for being too slow.

What I work on is large and extremely technical.

And no we are not at an infrastructure limit. This statement is insane. We are literally only a couple years into LLMs becoming popular. Everything we see now is just the beginning. You can only make a good judgement call of whether we are at our limit in 10 years.

Because the transition hit so quickly a lot of devs and companies haven’t fully embraced AI yet. Culture is still lagging capability. What you’re saying about ChatGPT was true a year ago. And now one year later, everything you’re saying isn’t remotely true anymore. The pace is frightening. So I don’t blame you for not knowing. Yes AI needs to be managed but it’s at a point where the management no longer hinders you and it instead augments your capabilities.

Reply on news.ycombinator.com