relevant re deepseek https://gwern.net/complement

model commoditization is good news for perplexity

There’s a lot of misconception that China “just cloned” the outputs of openai. This is far from true and reflects incomplete understanding of how these models are trained in the first place. DeepSeek R1 has figured out RL finetuning. They wrote a whole paper on this topic called…

Aravind Srinivas

model commoditization is good news for perplexity

https://x.com/aravsrinivas/status/1884130395437908205?s=46

design @splits | /latenightcrew /event-pass /feud

indeed, it's good news for the entire ecosystem except for those making frontier models and hardware

as a consumer it feels like a huge boon