Loading activity…
@simran_s_arora
building ai systems, cs phd @stanford @hazyresearch, incoming assistant professor @caltech, svp of research @togethercompute
AI
Together says its compiler serves trillions of tokens daily and adapts to traffic shifts within hours of recompilation.
AI
Unlike models with fixed tokenizers, H-Net learns to group raw bytes within the model, Cartesia says.
XAI
Its creators also report 10–21× faster unconditional generation on three datasets, with comparable or better generative perplexity.
XAI
A post says SanSi runs Ouro-1.4B for up to eight loops and can answer after any loop.
AI
A user says the company did not exist three years earlier and credits its reported headcount to M&A and AI.
AI
AMD says its code walkthrough takes developers from a baseline GPU kernel through optimizations including asynchronous memory transfers and Tensor Data Movers, showing how each technique affects performance.
AI
MIT researchers examine dynamic compression to let RNNs revisit and reorganize memory states during sequences.
