Ultimately, the author (the engineer writing the PR) should always fully own the impact and consequences of their code. This has always been true, but it’s become even more critical with AI-assisted coding and more junior engineers who primarily test only for functional correctness.
Just like AI now helps nearly everyone write code, it makes sense for AI to handle an initial pass at reviewing code too. However, there’s significant value in human reviewers providing an outside perspective—both to build shared understanding of the codebase and to catch higher-level issues.
TLDR; AI should do the first pass on reviews, but another engineer should also review the code for context and collaboration. Ultimately, though, the author still fully owns every line that gets merged.
Glad to hear the product resonates! Hope your team likes it, and please share any feedback as you try it out. If helpful, I can also start a direct slack channel between our teams, just reach out to contact@mrge.io with your slack emails!
thanks for the feedback and glad to hear that parts of our platform resonate. let me know if we can help onboard the team in the future if that makes it easier, it should be quick to switch as we also have our own cli. right now, our billing will be per author. our free trial is 2 weeks--but if you start it and don't trigger any/do any reviews we're happy to restart it later for you. just contact us at contact@mrge.io!
thanks for bringing this up, and totally understand the concern. we are committed to security, and we never write/access your code without your action--the only reason that setting is necessary is so that you can merge/1-click commit suggestions from the AI directly from the code suggestions it's posted.
We would be happy to try except when it has write/merge permissions .
One click and auto merge are nice to have. Having the bot (and your company) able to deploy any code changes to production (by accident, via hack, etc) is a no go.
Suggest making them optional features and just having code comments/repo read version.
Not sure if it’s possible - but if the permissions could exclude specific branches that would be ok as well.
But needs to be no way a malicious actor could write/merge to main.
sorry to hear that it didn't catch all the issues! if you downvote/upvote or reply directly to the bot comment @mrge-io <feedback>, we can improve it for your team.
We take all these into consideration when improving our AI, and your direct reply will fine tune comments for your repository-only.
That's good to know, but —assuming my sample of size 1 isn't a bad outlier, I should really try a few more— there's another problem: I don't think we'd be willing to sink time into tuning a currently-free subscription service that can be yanked at any time. And I'm in a position to say it is highly unlikely that we'd pay for the service.
(We already have problems with our human review being too superficial; we've recently come to a consensus that we're letting too much technical debt slip in, in the sense of unnoticed design problems.)
Now the funny part is that I'm talking about a FOSS project with nVidia involvement ;D
But also: this being a FOSS project, people have opened AI-generated PRs. Poor AI-generated PRs. This is indirectly hurting the prospects of your product (by reputation). Might I suggest adding an AI generated PR detector, if possible? (It's not in our guidelines yet but I expect we'll be prohibiting AI generated contributions soon.)
totally get where you're coming from--many big open source repos have also been using it for a while and have seen some FP but have generally felt that the quality overall was worth it. would love to continue having you try it out, but also understand that maintaining a FOSS project is a ton of work!
if you have specific feedback on the pr--feel free to email at contact@mrge.io and i'll take a look personally and see if we can adjust anything for your repo.
nice idea on the fully AI-generated PRs! something in our roadmap is to better highlight PRs or chunks that were likely auto-gened. stay tuned !
Just like AI now helps nearly everyone write code, it makes sense for AI to handle an initial pass at reviewing code too. However, there’s significant value in human reviewers providing an outside perspective—both to build shared understanding of the codebase and to catch higher-level issues.
TLDR; AI should do the first pass on reviews, but another engineer should also review the code for context and collaboration. Ultimately, though, the author still fully owns every line that gets merged.