Jeffrey Wang
Running frontier LLMs is slow, but that's the tradeoff we make to get more intelligent output. But if you think about the text (tokens) that an LLM produces (or that a human produces, for that matter), you might have an intuition that a lot of it does not actually require high intelligence....
Jeffrey Wang
The world of GPU programming for AI has come a long way since I worked on writing a CUDA-based matrix library back in 2009. Both NVIDIA hardware and the CUDA ecosystem have evolved dramatically and are now the basis for the majority of AI compute in the world today (hence the $4T+ market cap)....
Jeffrey Wang
As LLMs become larger, more capable, and more ubiquitous, the field of mechanistic interpretability -- that is, understanding the inner workings of these models -- becomes increasingly interesting and important. Similar to how software engineers benefit from having good mental models of file...
Maximilian Speicher
Tales of Design & User Experience (ToDUX) #13 ⁂ Dear readers & friends, It’s been a while, and I’m sorry for that. I’ve been suffering from a mild case of writer’s block for the better part of the past year. But I’ve managed to turn things around and could finally finish my...