If this was true, YC would see a trend of companies failing not because they failed to find a product/market fit, but because they had a fit and failed to execute when a key engineering employee left.
And yet YC is forever telling people to focus on "building something people want", above all else. Jessica Livingston just gave an interview listing the things that caused startups to fail; it was a short list, and included "founder breakup" and "failing to build something people want", but not "key engineer leaves".
This squares with ~15 years of experience, mostly in startups, a significant chunk of it in the valley. Recruiting is important, team building is important. But the value of any one "key" developer is lower than your comment makes it out to be. Loss of a key engineer is, for most companies, even in highly technical spaces, easily survivable.
Orthogonally, I'd also suggest you think of a Venn diagram. Draw a circle for "highly effective and appropriately specialized engineer". Now figure out where the circle is for "wants to found a company", and the circle for "intrinsically capable of founding a company", and the circle for "has life circumstances compatible with founding a company".
I get to talk to a lot of developers --- I hire them, at what I believe is a reasonably fast clip, and I work at a consultancy to software development shops --- and I think this notion that everyone wants to found a startup is the product of a lot of HN echo. Most developers do not in fact want to start companies. Starting a company is stressful and, believe it or not, even if you can wangle your way into being a founder many times in a row, it isn't the most reliable path to retiring wealthy.
I'm not sure that failure is necessarily the form in which such a problem would manifest itself - I've been around enough businesses to know that they can muddle through personnel changes. But it tends to impede short term progress and create stress for those remaining after a key employee leaves - the sorts of problems the alleged no-poach agreements allegedly sought to mitigate.
Now there may be something else that motivated Altman's essay, but it is almost certainly related to YC and the idea of improving employee equity appears to have been seen as a potential solution to some trend in their data. I think that YC came to believe turnover was impacting rates of return, and YC has had to think about ways to mitigate it. They're at the point now where they would have that data and some companies in their portfolio are mature enough where brain drain has a significant impact. For example, they could be running regressions around on former employees of portfolio companies applying to YC against the return rates of investments in the former employers.
There's the pursuit of better returns in there somewhere, and I reserve the right to pull another theory out of my ass at a later date.
And yet YC is forever telling people to focus on "building something people want", above all else. Jessica Livingston just gave an interview listing the things that caused startups to fail; it was a short list, and included "founder breakup" and "failing to build something people want", but not "key engineer leaves".
This squares with ~15 years of experience, mostly in startups, a significant chunk of it in the valley. Recruiting is important, team building is important. But the value of any one "key" developer is lower than your comment makes it out to be. Loss of a key engineer is, for most companies, even in highly technical spaces, easily survivable.
Orthogonally, I'd also suggest you think of a Venn diagram. Draw a circle for "highly effective and appropriately specialized engineer". Now figure out where the circle is for "wants to found a company", and the circle for "intrinsically capable of founding a company", and the circle for "has life circumstances compatible with founding a company".
I get to talk to a lot of developers --- I hire them, at what I believe is a reasonably fast clip, and I work at a consultancy to software development shops --- and I think this notion that everyone wants to found a startup is the product of a lot of HN echo. Most developers do not in fact want to start companies. Starting a company is stressful and, believe it or not, even if you can wangle your way into being a founder many times in a row, it isn't the most reliable path to retiring wealthy.