AI reviews are getting very good at surfacing bugs that are not the immediate thing you are fixing, but are close to it, and this is surfacing the Boy Scouts on your company.
The bugs I’m talking about are small bugs or rough edges that are easy to understand if you are already touching that area.
I’ve seen two types of responses to this type of review:
- The ones who acknowledge it and fix the bug,
- The ones who say, “this was there before, it is not part of my fix” or “we’ll get to it later.”
I don’t think every extra AI comment should expand the scope of the MR, but I’m talking about the easy stuff. The cost of doing those fixes in the past was high enough that we could tolerate people saying “we’ll do it later,” even though we knew what later actually meant: NEVER. Now there are no excuses.
This is happening more often. Human reviewers miss this kind of thing all the time, not because they are bad reviewers, but because it’s hard for them to hold the full context window in their head. This is where AI is giving us superpowers.
These fixes compound over time, and with them we make the product better. That’s one more reason I’m bullish on AI code review.
Be a scout. Leave things better than you found them. There are literally no excuses now.
PS: To paraphrase Volkswagen, on the road of product there are Boy Scouts and there are tourists. Boy Scouts wanted.
I’m always on the lookout for great product engineers (full-stack). If you want to work on a remote first, non-bullshit, no-politics, flat, let’s-get-things-done kind of team, drop me a quick message.