Intel researcher: I'd never hire someone who thinks temp=0 guarantees LLM determinism
JFPuget · x · 2026-10-07
Intel principal researcher JF Puget says he would never hire someone who assumes temperature=0 guarantees full determinism in LLM inference. The quoted post he's responding to attributes LLM randomness to the statistical model and the temperature parameter.
Puget's jab points at a commonly missed fact: even at temperature 0, non-deterministic floating-point reductions on GPUs can make outputs vary across runs — an inference-stack engineering issue, not just a sampling-parameter one.
More from Infra
- AMD ships ROCm 10.1, targeting storage-to-GPU data movement as the new training bottleneck — AccBalanced · 2026-10-07
- Crusoe's Path: From Stranded-Gas Bitcoin Mining to Prefab Gigawatt-Scale GPU Datacenters — AccBalanced · 2026-10-07
- Intel to keep working with Elon Musk on Terafab push into cutting-edge chips — SumitGup · 2026-10-07
- Musk bets every 5GW of added US power equals roughly 1% GDP growth — XFreeze · 2026-10-07
- Burn 0.22 released: biggest Rust DL framework update yet, 6-15x faster rebuilds — JosephJacks_ · 2026-10-07
- Async training credit assignment called a key 'foomy' breakthrough for RSI — teortaxesTex · 2026-10-07