Help brainstorm project proposals you would like to see get worked on! https://manifoldmarkets.notion.site/DRAFT-Manifold-30k-Community-Fund-201025bf8d294032b5e859e7b6236f0e
A proposal will resolve as 'YES' if it is officially submitted to the Manifold Community Fund (Manifund) and achieves at least the minimum required funding to commence the project.
Related questions
🏅 Top traders
# | Name | Total profit |
---|---|---|
1 | Ṁ482 | |
2 | Ṁ103 | |
3 | Ṁ83 | |
4 | Ṁ51 | |
5 | Ṁ4 |
100% Funded, assuming this counts as a podcast: https://manifund.org/projects/invest-in-the-conflux-manifold-media-empire
TikTok and other social media somewhat famously get people quickly hooked by learning what kind of content gets them engaged. AFAIK, a big problem for Manifold currently is new users churning. It seems worth it to try to directly optimize for user engagement. From what I understand, Manifold does already use AI for recommendations, but ‘only’ by showing similar markets via OpenAI embeddings, not some sort of RL model or something trying to balance showing new closer matches and more variety.
I’m thinking about submitting this myself. But while I have some ML experience, I’ve never touched recommender systems. So if anyone wants to steal the idea or team up, that’d be cool.
@1941159478 Oh interesting! I didn't know that they currently used OpenAI embeddings. There is certainly more tweaking that could be done but knowing that makes me a little less confident that it could be improved that much? (I also was thinking of pursuing the ML recommender system)