Content
@
0 reply
0 recast
0 reaction
Li Jin
@li
A fine-tuned LLM trained on Farcaster data with revenue flowing back to users/wallets whose data contributed to the training set
8 replies
3 recasts
29 reactions
osama
@osama
don’t have to go that deep and finetune. just rag on casts and pay to accounts whose casts get picked for answering queries as they come. fwiw, i have the pipeline/infra from a recent hackathon project. built on neynar ofc. if you see use and want to bounty 50k usdc/eth, i’ll ship this to production in two weeks
1 reply
0 recast
1 reaction
casslin.eth
@casslineth
tried training mistral AI's open source model this month, amazing idea, w/ current llm tech may be very heavy on computation cost tho.
0 reply
0 recast
1 reaction
TheModestThief🎩
@thief
the LLM might fall in love with thief
0 reply
0 recast
1 reaction
wizard not parzival
@alexpaden
10% if you were queried seems fine
0 reply
0 recast
0 reaction
BrixBountyFarm 🎩
@brixbounty
The bots must have wallets. Funding AGI feedback loop. Kidding, I think. Reckon this would make a lot of sense, eventually.
0 reply
0 recast
0 reaction
Leo Simon
@leo5
👀
0 reply
0 recast
0 reaction
Callum Wanderloots ✨
@wanderloots.eth
Yep! Feels like a natural product and extension of an open data graph But the monetization will be the tricky part imo… what would actually make the LLM valuable? Maybe different niche versions of the same LLM that enables different user bases fine tuned for those niches
0 reply
0 recast
0 reaction
Sumer Malhotra
@0xcurry.eth
i'm working on this!
0 reply
0 recast
0 reaction
