Hacker News11h ago187 points53 commentslive
Performance per dollar is getting faster and cheaper
Wafer achieved 2626 tok/s/node on AMD MI355X for GLM-5.2, over 2x cheaper than Blackwell, using MXFP4 quantization and speculative decode optimizations.
Read the full story atwafer.aiWhy this is in the Signal
LAXIMA AI Signal curates the highest-velocity stories across Hacker News, GitHub trending, and new Hugging Face / Replicate model releases — quality-filtered, deduplicated, and refreshed every four hours. This item surfaced from Hacker News with 187 points (by latchkey). We link straight to the original source above — see the full live feed.
More AI Signal briefs
- GHNative iPhone app for your Hermes agent
- HNAfricans Are Turning to Starlink
- HNGiant trees have no trouble pumping water to top branches
- HNEspionage Against the European Parliament
- HNJamesob's guide to running SOTA LLMs locally
- HNMemorizing session transcripts isn't useful
- HNValve open source the Steam Machine e-ink screen so you can make your own
- HNMarkets are competitive if and only if P = NP