by @johnandersen777.bsky.social
All conversations about decentralized compute — DePIN, distributed GPU networks, decentralized cloud, trustless execution, peer-to-peer infrastructure, and projects like Akash, Filecoin, Render, and more.
Codes By Febin How do you verify a computation without running it again? I’m exploring this with rust-stark-zkvm, an open-source Rust project for STARK-based verifiable computation. Code: github.com/CodesbyFebin/rust-stark-zkvm What would you want to prove? #Rust #OpenSource #zkvm #codesbyfebin
Ilia Gusev A distributed training job asks for 8 GPU pods. 7 schedule, the 8th sits Pending, and all 7 hang on all-reduce while the GPUs bill. We wrote up why the default scheduler does this on Sept 9. One thing needs an update. 🧵 podostack.com/p/kueue-gan...