| Light on evidence, still true enough to be worth discussing. I think that sleep deprivation and the high failure rates are connected but not that the causal arrow isn't so simply drawn. People can be well-rested and still have terrible ideas. It might go the other way. It could be that, because these ideas are so terrible, people throw insane amounts of effort behind them to avoid being blamed when they fail. When you have a death march (and most VC darlings are whole-company death marches) you're going to see a split between those who disengage and drop out, and the "A for Effort" strategists who believe that 110-hour work weeks will be noticed above and have them "rescued" (promotion in a big company, EIR gig in VC) when their current project fails. Generally, founders are going to take the "A for Effort" strategy so that, if they fail, they can launch their careers in VC. Employees generally don't have those exit options, but they can't exactly leave before their bosses do in most of these companies. What we have is an economy of asymmetric risk. Just as hedge fund managers share profits but not losses, and therefore have a higher-risk tolerance than the principals, the decision makers in the VC-funded world benefit disproportionately from obscene risk. Being able to say that one was in Facebook early is career-making for a VC, while 0x and 5x are essentially identical in the personal career calculus. Founders share some of the risk (they suffer if the thing fails) but at least take part in the upside. Employees get mainly the downside: mediocre equity and no control, nonexistent career support, high risk of job loss and painful hours. This leads to a risk fetishism (young people! fail fast! sleep is for the weak!) that doesn't make for a good place to work, nor does it produce broad-based success (VC is an underperforming asset class). It produces terrible ideas, reckless business expansion, and the extreme long hours of the startup ecosystem. TL;DR: the sleep deprivation is a symptom of a deeper problem (risk fetishism arising out of the principal-agent problem in VC). |