[pembrook]

This kind of advice is similar to the "get rich day trading stocks" narrative. It only sounds realistic if you are ignorant of the statistical probabilities involved (it seems most people are).

A vast majority of VC funds produce weak or negative returns. And this is after diversifying their fund investment across 10+ start-ups and assuming that 90% are going to be losers compared to putting that money in public equities.

You can't work for 10+ companies at once like a VC can. And even if you could, the odds are still against you. The idea that you'll be able to pick ONE winner at an early stage is, quite frankly hilarious and naive.

[jacquesm]

Another major reason that you won't be able to pick the winner is because you only get to pick once, and then you are booked for a long time. If you think a better one comes along you now have a sunk cost, and you'll be starting all over again, with a very good chance that your 'better' one will end up being worse. So the odds are very much against you if you are evaluating start-ups serially.

The better way to do it is to evaluate a whole pile of them at once, and then to pick the best one that you can find. And you're going to have to do a lot of work to evaluate those options, about as much as though your future depends on it, because it does. If you're not prepared to put in that kind of work then it really is just a lottery, and you're most likely better off to just take a job that pays you roughly what you are worth on the market, in the longer term that + a good savings regime will be a much surer path to some serious cash than buying lottery tickets at an opportunity cost of 300-500K each.

[echevil]

I think the right way to think about this problem is treating picking the winner at an early stage as a probability. Even though the value of the probability is small, you'd still want to maximize it when you're picking a startup to join, and you can do much better than picking randomly. You can continue to evaluate the probability after joining a startup.