Zeteo Turns Discord Into AI Research Lab
Zeteo is a community experiment aimed at producing publishable AI research, starting with a concrete target in AI memory. The model is deliberately competitive: members submit short hypotheses in a rate-limited channel, humans and AI agents select and refine the best ideas, then small project threads get three weeks to show real progress before graduating to paper writing or getting archived. The post frames the project as a collective process for turning raw ideas from strangers into research worth publishing, with credit retained by the originators and contributors.
Hot take: this reads less like a product launch and more like a meta-framework for organizing research talent in public, with Discord as the operating system.
- –The strongest angle is the process design: rate limits, staged filtering, and survival gates create pressure that should reduce noise.
- –The weakest point is the claim surface: there is no evidence yet of SOTA results, only an ambition to reach them.
- –The human-plus-agent review loop is the interesting novelty; it could surface good ideas faster than a normal open Discord.
- –The biggest risk is coordination collapse: if the committee becomes the bottleneck, the “many contributors” model can turn into slow moderation theater.
- –It is most credible as an experiment in research ops and community filtering, not as a proven research engine.
DISCOVERED
48d ago
2026-04-10
PUBLISHED
48d ago
2026-04-09
RELEVANCE
AUTHOR
1kmile