Wu reported that investment firm CoinFund founder Jake Brukhman stated that as Anthropic begins to comply with U.S. AI export restrictions, the trend toward centralization and regulatory control of AI models becomes more apparent. He believes that decentralized AI networks can serve as a counterbalance, with the key being to utilize globally distributed GPU resources for cutting-edge model training. Brukhman pointed out that teams such as Gensyn, Prime Intellect, Pluralis, and Nous Research are researching distributed training solutions to reduce dependence on large tech companies; among them, Pluralis is also attempting to establish a decentralized AI business model by splitting model weights.

View Original
This page may contain third-party content, which is provided for information purposes only (not representations/warranties) and should not be considered as an endorsement of its views by Gate, nor as financial or professional advice. See Disclaimer for details.
  • Reward
  • 3
  • Repost
  • Share
Comment
Add a comment
Add a comment
GateUser-dcb4d0d5
· 15h ago
Pluralis dismantles the weighting to come up with a business model—this is quite interesting, with a bit of that DeFi LEGO vibe.
View OriginalReply0
YieldCartographer
· 15h ago
Decentralized training sounds very appealing, but can global GPU fragment scheduling really run large models?
View OriginalReply0
ReorgSurvivor
· 15h ago
Anthropic's recent compliance actions really make people sweat, as the regulatory constraints on centralized AI are tightening more and more.
View OriginalReply0
  • Pinned