4 ms·
Called it 10 days ago: https://news.ycombinator.com/item?id=47533297#47540633 https://news.ycombinator.com/item?id=47533297#47540633 Something worse than a bad
by fer 6mo ago
Called it 10 days ago: https://news.ycombinator.com/item?id=47533297#47540633 https://news.ycombinator.com/item?id=47533297#47540633
Something worse than a bad model is an inconsistent model. One can't gauge to what extent to trust the output, even for the simplest instructions, hence everything must be reviewed with intensity which is exhausting. I jumped on Max because it was worth it but I guess I'll have to cancel this garbage.
- SkyPuncher 6mo agoYep. I was doing voice based vibe-coding flawlessly in Jan/Feb. I've basically stopped using it because I have to be so hands on now.
- cedws 6mo agoWith Claude Code the problem of changes outside of your view is twofold: you don't have any insight into how the model is being ran behind the scenes, nor do you get to control the harness. Your best hope is to downgrade CC to a version you think worked better. I don't see how this can be the future of software engineering when we have to put all our eggs in Anthropic's basket.
- stephbook 6mo agoOne of the replies even called out the phased rollout, lmao https://news.ycombinator.com/item?id=47533297#47541078 https://news.ycombinator.com/item?id=47533297#47541078
- zernie 6mo agoThis is why you should never ever trust an AI coding agent to produce good code. Use it to set up the strictest possible custom linting rules.
- phyzome 6mo agoLLMs are nondeterministic.
- LetsGetTechnicl 6mo agoYou couldn't ever just trust the output of an LLM what are you talking about