Get the latest tech news
GPT-5: Overdue, overhyped and underwhelming. And that's not the worst of it
A new release botched … and a breaking research new paper that spells trouble
(Some argued that this is because GPT-5 is still using an older models for generating images, but given that the new thing was supposed to be tantamount to AGI and “fully multimodal” that hardly seems like a compelling excuse.) Reading the abstract (Chain of Thought reasoning is “a brittle mirage that vanishes when it is pushed beyond training distributions”) practically gave me deja vu. GPT-5 may be a moderate quantitative improvement (and it may be cheaper) but it still fails in all the same qualitative ways as its predecessors, on chess, on reasoning, in vision; even sometimes on counting] and basic math.. Hallucinations linger.
Or read this on Hacker News
