[GUEST ACCESS MODE: Data is scrambled or limited to provide examples. Make requests using your API key to unlock full data. Check https://lunarcrush.ai/auth for authentication information.]

![MacrocosmosAI Avatar](https://lunarcrush.com/gi/w:24/cr:twitter::1764410143217700864.png) Macrocosmos [@MacrocosmosAI](/creator/twitter/MacrocosmosAI) on x 4734 followers
Created: 2025-05-30 16:41:35 UTC

Subnet X proved decentralized LLM pretraining is viable. We are proud to release the technical primer for IOTA in advance of mainnet launch on June 2nd.

IOTA comprises a series of key innovations:
- Data- and Pipeline-parallel SWARM execution across heterogeneous and unreliable nodes.
- 128× activation compression for home-grade links
- CLASP: Contribution Loss Assessment via Sampling of Pathways
- Butterfly All-Reduce for O(1) sync bandwidth

Together, we believe they can combine to push the field of permissionless, performant decentralised training measurably forward.


XXXXXX engagements

![Engagements Line Chart](https://lunarcrush.com/gi/w:600/p:tweet::1928491993338167573/c:line.svg)

**Related Topics**
[token](/topic/token)
[mainnet](/topic/mainnet)
[llm](/topic/llm)
[decentralized](/topic/decentralized)

[Post Link](https://x.com/MacrocosmosAI/status/1928491993338167573)