Can Open Models Keep Up With the AI Frontier? | Nathan Lambert

▶ 9:45

The Chinese labs are very technically excellent [...] what they don't have is compute. [...] an average OpenAI researcher has 5000 H100 equivalents that they can access at any time

▶ 11:11

it's very hard for me to be like distillation is a major factor [in chinese model superiority] because, for example, in the fully open literature of people trying to do open post training recipes, all the open model builders are stuck using like a really outdated quen model. Because you can't figure out how to get the teacher right.