Positive users express excitement about recaps of self-distillation advances in AI algorithms, while negative users dismiss GPT-5.6 training claims as mere knowledge compression from larger models.
Based on 4 visible X reactions from 14 accounts; directional sample.
Ask a question below.
Published answers will appear here.
@willccbb This is the question nobody at OpenAI will answer publicly. If all you did was self-distill from a bigger model, the privileged information is the model itself. That's not a training breakthrough, it's knowledge compression. Different game entirely.
@willccbb @doomslide Love to read this article
@willccbb Nice
one of my goals this evening is to recap the recent news in self-distillation for @doomslide to celebrate his return. i hope it may ease his perennial skepticism about the lack of fundamental progress in algorithms
Positive users express excitement about recaps of self-distillation advances in AI algorithms, while negative users dismiss GPT-5.6 training claims as mere knowledge compression from larger models.
Based on 4 visible X reactions from 14 accounts; directional sample.
Ask a question below.
Published answers will appear here.