The answer, of course, is money.
But Dixon prefers to camp out on algorithms and behavioural science rather than providing the simple answer, and if you read between the lines (and you're familiar with the startup scene, VC, and suppressed developer wages by collusion + H-1B visas + "talent shortage") it's not hard to figure out why.
If you stay in finance for some time, you have probably made a lot of money by the time you get out. The plan is to work yourself to death as an investment banker and hope you made enough to retire (or become a VC, like him) at 35.
Work years for a startup, and you might not have made anything beyond living expenses. The upside is that the working culture will probably suck less and winning the startup lottery will make even more money. Also, scamming people and being a sociopath is optional (but see Uber).
It's also a case where I think starting young has significant advantages. Life gets pretty good once you have bootstrapped revenues and a business somewhat on autopilot. Especially for the young, as they're relatively free and can enjoy location and time freedom more fully.
That way the financial system in the form of VCs can extract maximum benefit for minimum expenditure.