I find it unsettling that every developer is asked for their opinion on LLMs and vibing, because of the dynamic it sets up with the pace of change in AI systems: if they give specifics about things that their experience shows LLMs can and cannot do, those are out of date immediately after publication, if not already at interview time, but if they are vague, it sounds like they have no experience with it, but also if they note the curve and try to adjust for those things, they come off like marketing for frontier orgs and it's still going to be laughable in six months or a year.

If you want to make the interview about that, fine -- there's lots of material and interest. If the interview isn't primarily about that, don't lay this rhetorical trap for the interviewee.

When I find speakers for NYC Systems I do ask them not to speak about their experience with AI (at least that the talk should not be entirely about this) because of exactly this risk of being distracting and immediately out-of-date.

However, I think it can be a useful signal to understand if/how experienced developers in major software infrastructure projects are using AI today. So I plan to keep asking the question.