Open Research Should Prioritize Artifacts and Speed

IanArawjo · x · 2026-07-16

[Repost] The original post argues that when establishing trust in research results, priority should be given to released artifacts—models or code—followed by reputation, rather than relying solely on paper reviews.

The author also believes that the pace of releasing results needs to accelerate: while papers remain important, faster communication methods hold more value. Using Percy Liang's open research around Marin as an example of a cutting-edge practice, the author notes there are many similar good cases.

The conclusion isn't that "traditional peer review has failed, so open research is doomed," but rather that we must adapt to new environments and redesign how open research is collaborated on and published.

Original post →

More from AGI Musings

AGI Musings channel →