3 ms·
OpenAI's GPT-5 Pushed Back to Late 2025, but Promises PhD-Level Abilities
- sylware 2y agoAre they going to harness "tools" to their AIs?? Like a mathematical formal solver?
- notthatoldguy 2y agoThey may add any number of tools, but that will add exactly zero logic and reasoning to models — the main "weakness" of everything stupid marketing calls "intelligence" these days.
- sylware 2y agoThis is R&D, I guess they would have to design new ML methods. But I am not that convince you can do maths with only 2D layers of neurons.
- fuzzfactor 2y agoIf they only used live PhD-level human discourse, over topics that had been, or can be peer-reviewed in advance[0] before finding its way into a database, it might be more likely to seem more intelligent than training on average everyday data. Don't get me started on how much less advanced you should expect if random internet data is given any weight at all. [0] Rehearsal might be good, it could even include PhD's in Education since teaching the machine could almost be considered equivalent to professors preparing a curriculum for teaching regular students.