Bittensor's iota productizes decentralized training with pipeline parallelism
markjeffrey · x · 2026-10-02
Iota (Subnet 9 on Bittensor, by Macrocosmos) has productized decentralized AI training: anyone can rent it and train whatever model they want.
The key insight: most distributed training is data parallel, so every machine holds the whole model and size is capped by a single box. iota uses pipeline parallelism — each machine holds a slice — so model size scales with the network.
More from Infra
- Local AI roundup: 27B reasoning in 5.9GB, phone-class 35B, and dozens more — vramkickedin · 2026-10-02
- Liquid AI's Decision Model D1 Hits OpenRouter: Typed Answers With Probabilities at $0.04/M Input, $0 Output — maximelabonne · 2026-10-02
- Teenager tapes out a chip, rebuilds GPU interconnects with optics and poaches Nvidia veterans — ai · 2026-10-02
- huggingface_hub v2.1.0 ships: 17x faster downloads, job retries, rerun and port exposure — huggingface · 2026-10-02
- Infra engineers compare notes on when self-hosting LLMs beats paying for APIs — One_Mention_5385 · 2026-10-02
- Gemini 4 isn't even out yet — but Google's TPU advantage is being underappreciated — haider1 · 2026-10-02