"AI problems at Google" is just an example. I find hard technical problems interesting. A problem is hard if I don't know in advance whether or not I will be able to solve it. AI is hard. Making a good OS is hard. Startups avoid hard problems, because they have enough things to worry about already.

I'm not saying the job at Google with Peter Norvig (ah) is perfect. I was just pointing out another reason why I'm hesitant to make a startup.

That's funny. I work at a place that's supposedly full of hard technical (and creative) problems. But they're always someone else's problems.

And for every hard problem at a "real" job, there's 10 annoying ones. I bet the startup ratio is probably similar, but they're YOUR problems (hard and annoying), not someone elses.

I agree, I love working on hard problems. Sort of. What I've found is that I tend to make decisions for a variety of reasons, and am most decisive when I have a either broad convergence of motivation, or a singularly powerful one.

If you're asking yourself wether to pursue a job at google or starting a startup, you'll want to try enumerating as many reasons as you can for each one. You won't likely get a definitive answer, but the excercise will get you thinking about what other things are important to you. If you're lucky, one will turn up to be the obvious choice.

In my case, I have slid technically to the point that preparing for one is the same as preparing for the other.