Given OpenAI’s declining performance after his being sidelined and then departing, interested to see what they do. Should be a clear demonstration of who was really driving innovation there.
Probably will be an unpopular opinion here but I think declining performance is more likely related to unclear business models backed by immature technology driven by large hype trains they themselves created.
Unpopular because it does not follow the OAI hate train but I think this is a pretty solid take. There is real value in LLM but I believe the hype overshadowed the real cases.
They're probably just scaling back resources to the existing models to focus on the next generation. I feel like I have seen OpenAI models lose capability over time and I bet it's a cost optimization on their part.
100% OpenAi performance is decreasing. I basically use Claud sonnet exclusively and canceled my OpenAi subscription for personal use. my company still uses them because you cant currently fine-tune a Claud model, yet.
