Hacker News new | past | comments | ask | show | jobs | submit
Every single time I use AI (which is every day), regardless of the model or harness (which is pretty much always Opus 5.5 or Fable 6.2, with Kiro or Claude Code), the LLMs do complete and utter rubbish to the point where they would be genuinely harmful if left to run write commands unsupervised.

Just today Opus 5.5 tried to send a Slack message to one of my teammate when it had to edit a code file. When asked why, it said it just "picked the wrong tool".

LLMs are borderline useless, AI has led to 0 increase in actual value produced. It's all powder in front of our eyes.

I genuinely cannot take anyone seriously when they say they're afraid of being replaced by AI. The only explication possible (unless they're astroturfing), is that they literally never look at what LLMs produce and always run in "Trust all tools" mode without ever looking.

The people who are most bullish on AI are the people who aren't hands on with the material. I'm hands on with AI all day. Its like a bunch of junior devs who never learn or mature or grow. With a Junior dev a teachable moment generally sticks. The AI never learns, even if it makes a "memory", it will straight up ignore those memories at the slightest convenience. AI is shit, but I use it everyday, much in the same way that Javascript is shit but I use it everyday.