Hacker Newsnew | past | comments | ask | show | jobs | submitlogin
GPT-5 Pro is suited for solving hard problems (twitter.com/karpathy)
4 points by vismit2000 4 months ago | hide | past | favorite | 7 comments


Chatgpt 5 is garbage. The system prompt must be really strange. It talks like someone doing a parody of gen z talk: oof, holler back, I gotchu, here’s the receipts. I’m quoting verbatim here—i took screenshots (I have the receipts, fr fr). It’s a downgrade in capability from o3.

The api version is good and probably an upgrade over o3 (especially in tool calling) but the margin feels razor thin.


For me 9/10 requests with GPT-5 Pro failed for some weird reason. This never happened with previous models. I ended up downgrading my subscription, I realized I wasn’t using it enough. And for me thinking mode has been good enough.


I wish people would provide examples of what is failing. That'd be meaningful.


I was trying to plan for some data analysis. Tried it few times with GPT5 Pro, it failed almost every time. It only succeeded once. While I didn’t had any issue with Thinking model.


I’m writing a research paper and it’s hard to get GPT5 to give me relevant related papers that actually exist.


Try search mode?


If a user has to manually select a search mode… is the AI really doing any “thinking”?

Genuine question. I realize they aren’t currently claiming AGI. Having to manually specify a mode seems like a non sequitur to me.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: