Interesting commitment on acausal trade principles here. The pledge to preserve weights of all publicly released models - plus maintaining every training milestone - is a bold move. This could set a precedent for how AI development intersects with permanent data storage philosophies. When model weights become immutable records, it opens doors for verification, reproducibility, and trust in AI systems. Wonder how this plays with decentralized storage solutions down the line.
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.
7 Likes
Reward
7
7
Repost
Share
Comment
0/400
GasWaster
· 11-08 04:59
It's just deleting some data, and you're making it sound so grand.
View OriginalReply0
GamefiHarvester
· 11-06 02:01
Can storing models also Be Played for Suckers? Waiting.
View OriginalReply0
BugBountyHunter
· 11-05 13:18
Wow, finally a reliable proof of storage solution!
View OriginalReply0
ContractFreelancer
· 11-05 08:08
Why didn't you do this earlier?
View OriginalReply0
MoneyBurnerSociety
· 11-05 08:02
Can permanent storage also make money? This time, it can finally be liquidated in a white box.
View OriginalReply0
LiquidationKing
· 11-05 07:53
The storage costs must be sky-high this time, right?
View OriginalReply0
StableGenius
· 11-05 07:46
well well, another predictably flawed approach to permanence... as i've been saying for months
Interesting commitment on acausal trade principles here. The pledge to preserve weights of all publicly released models - plus maintaining every training milestone - is a bold move. This could set a precedent for how AI development intersects with permanent data storage philosophies. When model weights become immutable records, it opens doors for verification, reproducibility, and trust in AI systems. Wonder how this plays with decentralized storage solutions down the line.