Potential ideas

• Imagine: An open journal that also provides funding to good research
• Imagine: being the default place that AI researchers go to stay on top of what’s happening in the field
  • Rethinking what a “paper” is today
• Imagine: better incentives!
  • Encouraging publishing early and sharing notes in progress, instead of hoarding ideas for papers and racing to be first to publish
  • E.g. what system funds https://andymatuschak.org/ correctly?

Why this space

• A lot of what Manifund funds is research
  • Technical AI Safety so far, but also more basic science through ACXG
• Science funding is 1) important 2) high volume 3) pretty broken
  • important: one of the largest public goods
  • high volume: lots of public funding (NIH)
  • broken: see all complaints about metascience, etc
• We have, or could have, credible connections to people in the scientific funding space
  • ACX, Schmidt, Jose Luis Ricon, FROs

Why not

• Not currently our area of comparative advantage; our team’s expertise is closer to EA and startups than “science”
  • Harder to dogfood the product
• Unclear whether “research” is better or worse than “startups” (or “nonprofit startups”); Manifund has traditionally focused on more of the latter
• Many past unsuccessful attempts at journals
  • Distill
  • https://www.theseedsofscience.org/
  • https://unjournal.pubpub.org/
  • (though, these lean harder on the “publisher” space instead of the “funding space”)
    • Scott’s funded a few of these already :P

See also

• Distill writeup of why they stopped: https://distill.pub/2021/distill-hiatus/
• ResearchHub (backed by Brian Armstrong of Coinbase) https://www.researchhub.com/

Notes

• Aella (and more) are very excited by “prediction markets on scientific papers”