Content
@
0 reply
0 recast
0 reaction
Stepan Gershuni
@stepa
web3 can be useful to prevent AI misuse risks in a few ways: 1. Distributed incentives (more people can contribute to ML R&D, including data, code, labeling, rlhf) 2. Cryptography (hard guarantees of authenticity, privacy, provenance) 3. Governance and decision-making
0 reply
0 recast
1 reaction
