[tptacek]

A lot of people, including people in attendance, do too. I think the reason is that people read Paul Graham posts about startup management as if they were written on stone tablets. Sometimes I think he intends for them to be taken that way (and in some cases, I think I get why), but this was not one of those cases, and the discourse ran away with it.

That said: there's a real phenomenon Graham and Chesky were grappling with, and if you've done startups for awhile --- startups, in particular, because they give you the vantage point of seeing a company's management processes develop from zero --- you've almost certainly seen it yourself. Not enough has been written about it! The point Graham was trying to make isn't banal (or wrong).

It's just not fully formed, and is being taken that way.

[extr]

I've worked in startups a long time myself. Personally I've never seen this problem, feels like it's advice literally specifically for Chesky or maybe a founders at a couple dozen other unicorns. In my experience it's a way more common problem in the other direction. Founder can't let go of the details and is mucking about in everyone's work. Or have literally just become bored with the company and can't even be bothered to hire the professional executives that are supposedly such a problem. Perhaps a hot take, but IME most professional executives/managers/MBA types are actually pretty solid people to work with and do a good job.

[danielmarkbruce]

There is an implicit assumption that the founder will make good decisions. If someone is "mucking about" in everyone's work, presumably it's phrased that way because they are making a bunch of stupid decisions? Not much (there are counter examples) will fix a ceo making dumb decision after dumb decision.

[tptacek]

Both things are true: there are founders who lose Github merge privileges and start meddling because they haven't defined a long-term job for themselves, and there are most definitely pasteurized processed business units that get hired and run the exact same performative playbook at startup after startup. It's not an either-or thing.

[extr]

I'm not saying that both don't exist, I'm just saying PG's advice is so narrowly applicable as to be effectively "bad advice" for most people reading it.

[tptacek]

As this article points out: Graham's article has probably been misconstrued, and is best read as "this is something you should consider and pay attention to", and not a directive for everyone to go "be in founder mode". But to be clear: I also don't think Graham's post is especially good, though I think the issue he's engaging with is important and widely slept on.