DeepSeek’s latest R1 model has become very popular (as it is good and cheap) and the take that it shows that less compute is required to train a model and therefore less compute will be needed for future AI models lead to Nvidia having the largest company-specific daily drop in market cap ever.
A good article about DeepSeek on Stratechery goes through several aspects and makes it easier to understand what’s going on, even for those of us that are not AI researchers.
