X · @teortaxesTex
· X / Twitter
another DeepGEMM update what is funny is that DeepSeek's efficiency (and margin) is a moving target. GLM can optimize DSA, why can't they keep pushing…
another DeepGEMM updatewhat is funny is that DeepSeek's efficiency (and margin) is a moving target. GLM can optimize DSA, why can't they keep pushing their V4 stack post-deployment? We knew nothing of DSpark before the paper.