Hi Sean, I was actually thinking of the exact same idea you just suggested i.e. reward high volume reviewers with priority reviews on their PRs. This is a great idea for reducing the 'drive-by' contributor where they just submit PRs without investing in whittling down the review queue. It would also encourage already active contributors to review more frequently. There is of course a risk that this may be 'gamed' as well via agents or other means so in addition to AI slop PRs, we may end up getting AI slop reviews. But the biggest benefit would be incentivizing trusted contributors who care about their reputation to spend more energy in reviews vs working on their own PRs. This benefit cannot be understated as the biggest constraint is review capacity not code generation.
Overall, strongly +1 in favour of this particular measure you proposed. Thanks, Sameer Mesiah. On Sat, 20 Jun 2026 at 01:15, Ghaeli, Sean <[email protected]> wrote: > Some stats I'm seeing plotted out: > > https://gist.github.com/seanghaeli/8fb41609f15c2ab403949cefbe43ff40 > > Looking at the current queue of open PRs, there's a long-tailed > distribution for PR age (first plot). I expected a strong correlation > between PR age and its size (second plot), but the connection is quite weak > (except for very large PRs spanning 1k+ lines). I'm interpreting PR size as > a proxy for PR complexity. > > Seeing this, I think that there's room for a stronger forcing function to > our review process to make that PR age distribution much steeper without > diluting review quality. This goes back to my suggestion of explicitly > weighting PR authors' level of contribution to other PR reviews, which > would act as a natural forcing function. > > I'm not sure how to quantify it, but I suspect contributor retention would > improve with quicker feedback loops on their work. > > > --------------------------------------------------------------------- > To unsubscribe, e-mail: [email protected] > For additional commands, e-mail: [email protected] >
