Discussion (3 Comments)Read Original on HackerNews

goingbananas•about 2 hours ago

LLM diminished returns are here, disguised as intentional slow down

emmettbt•about 1 hour ago

I very much agree. There is need for more innovative llm training methods than just give it more data.

sakompella•about 1 hour ago

theyve been talking about their upcoming astra models for a while, this is neoliberal self-regulation talk

OpenAI Is Slowing Down Its AI Training