You seem to share a common statistical misconception about success rates. If 10% of startups succeed, that doesn't mean that if you start a startup, your chances of succeeding are 10%. They are either much higher or much lower.

This is clearer if you consider a statement like "10% of men are over 6 feet tall." There's no one who has actually has a 10% chance of being over 6 feet tall. 10% of people have a 100% chance, and the remaining 90% have a 0% chance.

For the sort of person who has sufficient drive to get rich at all, starting a startup is a much more reliable way to do it than the overall success rate implies.

The frequentist way you're doing probability in that comment is not valid since the success or failure of one particular startup is not a repeatable experiment. The only valid interpretations are (1) the frequentist probability across a group of startups and (2) the degree of belief, or bayesian probability of a single startup's success. In this framework it is entirely reasonable to assign a 10% probability of success to one particular startup. I agree that if you knew beforehand that you're in the small group of people whose success rate exceeds your standard for reliability, then for that person a startup would be a reliable way to get rich. The problem is, of course: how do you know which group you're in?

I didn't mean to imply that for the right sort of person the odds are 100%. But they're pretty good. Probably over 30% and maybe as high as 50%. And since failing is usually pretty quick you could easily try 3 startups in 5 years.

As it turns out there is an easy way to know which group you're in: ask us. Like all venture investors, it's our job to answer that question, and we work very hard to try to do it well.

I would have to agree. For all the belly-aching about Wall Street in NYC's tech community, you're guaranteed a significant income for a number of years.

For this equality to hold:

    E(Wall Street hacker's earnings) == E(startup founder's earnings) 

you need an extremely high payoff, since

    E(startup founder's earnings) == P(startup success) * payoff of startup.

and we know the probability of success at a startup is low.

Everyone on HN needs to be honest with themselves: this kind of high payoff almost never happens, in aggregate. And I haven't even included a risk premium, which would make the required payoff higher!

If wealth capture is a main concern, you're better off building a nest egg at a hedge fund or large corporation for 10 years. If you want to start a startup, do it for the culture and technological freedom. From a statistical view, however, a startup is not the optimal way to become rich.

EDIT: mynegation reminded me of risk-adjusted returns. Thanks!

Even that equality is too weak. Ideally you want the equality of _risk adjusted returns_ (See 'Sharpe Ratio' - http://en.wikipedia.org/wiki/Sharpe_ratio)

The risk adjusted returns of a venture-capital funded startup for a a founder with assets < 100k and non-entrepreneurial opportunities of > 100k is about negative 100-200k. In other words, you'd pay money to not be a venture-capital funded startup founder in that scenario.

I believe pg meant there's a much larger percentage of rich people who were once part of a startup vs rich people that worked like everyone else. I'd say that's a fact right there.

PG wouldn't mean that. He knows that

    P(becoming rich | working at a startup) != P(worked at a startup | is rich)

I don't want to parse statements here, but it's pretty clear PG suggests that starting up or working at a startup is the best way to become rich. I'd wager that, in practice, people become rich for a variety of reasons, with none being more effective than any other.

-----

EDIT: Above, I'm trying to point out how the fraction of rich entrepreneurs vs. rich employees provides us with no information by itself.

However, we can develop a distribution of probabilities for becoming rich based on the career you chose, using the careers of those who are already rich. I'd suppose this is PG's real point, and while I can't verify it right now I'd bet entrepreneurs become rich at a higher frequency than non-entrepreneurs.

Can't say we are saying anything much different but let's focus on what seems important. So there's a variety of ways to get rich and we want to know which one is easier: work for someone else vs start your own company. My point is that 99.9% of the world population work for someone else while only a really small percentage started a company. And only a really small percentage of the world population is rich. Even if only of 1% of the rich people started their own company, there's still truth to what the article said. So, I was saying it's a fact that your chances to get rich are much higher having your own company than working for someone else. Which I also believe is the point in the article.

You're right, we're agreeing violently. Above, you said that more rich people started a company than worked for others; here you're applying Bayes. That first assertion is not verifiable, but you're definitely right that starting a company increases your likelihood of becoming rich.

Also, apologies for my rude tone. :(

That depends entirely on your definition of rich. While P(earnings > $1M/year) is probably going to increase, P(earnings > $100k/year) is probably going to decrease. It's also unclear whether your expected earnings are going to increase, and note that most people's loss function is not even linear (e.g. you're not going to be 10x as happy earning $1M than earning $100k -- and you're going to be much less happy when you're broke).

It's ok, I could have been much clearer (but then I would also repeat what's in the article). I have the feeling people commenting here did not read the whole article or are just nitpicking, it's one of the best articles about the history of work and its current status I have ever read, everyone should read this.