# Full-stack decentralization, not one layer

- **Speaker:** Greg Osuri
- **Attribution:** attributed
- **Said:** 2024-03-19
- **Theme:** Decentralized AI — https://www.gregosuri.com/predictions/decentralized-ai/
- **Source:** Decentralizing AI: Compute Protocols and Proof of Work - Panel Discussion \| Proof of Data 2024 (Recall Network)
- **Watch at:** 00:13:38 — https://www.youtube.com/watch?v=uUoePJrzyDQ&t=818s
- **Canonical:** https://www.gregosuri.com/predictions/decentralized-ai/2024-03-19-full-stack-decentralization-not-one-layer-virq/
- **Cite as:** Greg Osuri, "Full-stack decentralization, not one layer," Decentralizing AI: Compute Protocols and Proof of Work - Panel Discussion \| Proof of Data 2024 (Recall Network), 19 Mar 2024, 00:13:38. https://www.youtube.com/watch?v=uUoePJrzyDQ&t=818s

## Quote — verbatim

> All the way from GPU access to the layers in the middle — when you have verification and fault tolerance and caching and load balancing and all the good stuff — to the UIs and UXs of the world.

## Context — site annotation, not spoken

His answer opens "you need to have full stack decentralization... from resource acquisition to delivering the application" (00:12:56); Akash stays general-purpose and minimal at the protocol layer so specialized protocols like Gensyn can compose on top — "open protocols building on open protocols" (00:15:03).