The truth? Ugh. The pitch is that, even if you fail, you'll learn a ton and be slotted into a much more senior role in your next job. The reality? Yes, you do learn a lot. However, you get the same kind of job you would've gotten before the startup. Not more. That 2.5 years you spent building a complex product from scratch? Fucking wasted. In your next job, you're still going to be assigned to legacy maintenance, and random bugs and tickets like any other junior, not given real work and a meaningful role.
I consider this pretty dangerous, because there's nothing worse than gaining 2-3 levels in ability, but having to linger at a low level because your startup's failure (which may or may not have been your fault; in my case, it wasn't) invalidates the experience completely. You might fare better, psychologically, if you don't go through the startup and gain levels of experience that you won't be permitted to use in the next job. It's easier to function if you're at-level than if you gain 5 years of experience in 15 months but no one believes you.
It's pretty humiliating and you might break. (This was my Google experience. Granted, I had a rotten manager, but the psychological load of coming off a failed startup didn't help the situation.) I think the whole "nothing bad happens when your startup fails" pitch that you hear from VCs is disingenuous. VCs fix some peoples' careers after startup failure, but that's not going to happen for an early employee (E #2, not really a founder) of a company that didn't even make it to the A round.
This isn't just an issue that I faced. I know plenty of hiring managers who worry about ex-startup people, not because they take the failure to mean anything about the person (for some good news, that stigma is mostly gone) but because they know that people who did startups are (a) a rank level or few beyond where they'd be slotted based on seniority, and (b) have chips on their shoulders. It's a valid concern. When it's an acqui-hire, the founders are usually tired of startup stress and ready to take a (still well-paid and high-status) subordinate role, but after a startup just fails you get someone who feels a need to prove himself and won't be happy being re-slotted to a second-year junior programmer.
So... to answer the question, it's not an easy road. After a short spell at Google, I consulted with a horrible startup (that I've taken off my CV, and that inspired many of my cautions about startups, see: http://michaelochurch.wordpress.com/2012/07/08/dont-waste-yo...) that wanted to hire me as VP/Eng but only if I disparaged, in writing, the work of 10 engineers, many of whose work I knew nothing about. (I turned that offer down.) Then I worked in finance for a while, did a bit of consulting (including lottery design and technology selection) and have done some R&D (new product development and data science) in Baltimore. My life has gotten better, and with the blogging and the fact that I'm starting to know a few things about in-demand fields (machine learning, Haskell, Clojure) I'm now finding reasonable demand for what I can do. It's been a lot of work to get back on track, but I have, and I'm probably stronger for having had to fight this hard.
I think that, in the long run, the experience is good for you. The year after (or few) can be hell. The pitch to the effect of "even if it fails, you'll get a great job afterward" is not really true. It's true for founders who line up with marquee investors and whom the investors really like, but in the average case and for people of average means and connections, the outcome seems to be neutral to negative.
My case is just one data point (and an atypical one) but I think that most veterans of failed startups (early employees hit the hardest, then founders, then late employees) will admit that their careers have been set back, not accelerated, by them. It's certainly not fatal, but it causes a loss of time and, as you get older, you realize just how finite that is.