Home/AIAccelerating Block Low-Rank Foundation Model Inference on MemoryConstrained GPUsmatt_d·Hacker News·AI·July 16, 2026Read full article →Related ArticlesOpenAI’s o1 correctly diagnosed 67% of ER patients vs. 50-55% by triage doctorsdonsupreme · Hacker News · 2mo agoAccelerating Gemma 4: faster inference with multi-token prediction draftersamrrs · Hacker News · 2mo agoA couple million lines of Haskell: Production engineering at Mercuryunignorant · Hacker News · 2mo agoUsing “underdrawings” for accurate text and numberssamcollins · Hacker News · 2mo agoProgramBench: Can language models rebuild programs from scratch?jonbaer · Hacker News · 2mo ago