Content
@
https://warpcast.com/~/channel/aichannel
0 reply
0 recast
0 reaction
shoni.eth
@alexpaden
AI research call summary #8: 04/25/2025 === We opened with a mini-app crash-course. Easiest path: drop Farcaster’s LLM one-shot prompt into Cursor and let the agent scaffold your repo. Pull only the Neynar endpoints you need—keeps context lean, tokens cheap. Cursor vs. VS Code is back on the table. Copilot’s new Agent mode basically clones Cursor’s killer features. If you hate Microsoft bloat, VS Codium + Copilot Agent is looking viable; otherwise Cursor still wins on polish (some alts: windsurf & cline). Connor dropped the big update on Negation Game. Scroll DAO integration is live—116 sign-ups, 25 active voices. “Cred” (500 on join) is play-money that only grows when you share how you’d change your mind. Positivity is free; down-votes require reasons.
2 replies
1 recast
9 reactions
shoni.eth
@alexpaden
Long game: move from human delegates to AI “digital delegates” that spend your cred. Reputation will use Connection-Oriented Cluster Matching (civil-resistant upgrade over quadratic voting). On-chain comes last—iterate off-chain until the game-theory holds. Why this matters: if we can tie token value to well-reasoned knowledge graphs, we unlock a funding flywheel for biotech and basic research—paying scientists for insight, not just products. Side threads: new Farcaster mini-app that gamifies jailbreaks proves every current LLM is still socially hackable. Vitalik’s “Steer the Ship” essay is now mandatory reading for AI governance. Next steps: Allen’s finishing his mini-app bootcamp; everyone else, go post a rationale at scroll.negationgame.com and feel the UX. Orbiter hackathon is live if you want free IPFS hosting + bounties.
1 reply
0 recast
3 reactions
shoni.eth
@alexpaden
this is for 04/22/2025-- my mistake.
0 reply
0 recast
0 reaction
