PyTorch to ship new stateless RNG API after giffmana's year-long design push
giffmana · x · 2026-11-17
PyTorch core dev ezyang revealed an experimental (soon public) stateless RNG API that fixes the current interface's problems. giffmana shared the backstory: his way of contributing is to "complain until something happens" — he prototyped his ideal RNG design, used it in his research codebase, and wrote a detailed doc explaining why the current API couldn't support it.
The PyTorch team returned with a largely new design close to what he wanted, and after iterating on rough edges, the redesign is now landing.
Related event: PyTorch to ship experimental stateless RNG API(3 posts)→
More from Infra
- Altman explains 'double counting' warning: chipmakers all assume OpenAI will buy their supply — toptickcrypto · 2026-11-17
- RightNow launches inference API, claims 14x NVIDIA performance via GPU kernels — ycombinator · 2026-11-17
- Redditor Defends Strata: Free Tool Lets Average Hardware Run Large Models Amid Community Backlash — BoogerIsACult · 2026-11-17
- Baseten open-sources Rust/CUDA DeepSeek engine; agent on Vera Rubin finds optimizations speeding Blackwell by 10% — baseten · 2026-11-17
- Wici Infer offers free large-model serving on consumer hardware via VRam+Ram+SSD — AlexDorofeev · 2026-11-17
- Benchmark shows new e-means clustering 100x slower than Lance's hierarchical k-means — leland_mcinnes · 2026-11-17