Among the things I find interesting - the guy was the co-founder of what is now Google Maps (after Google's acquisition). However, now, he has chosen to join Facebook instead of setting out on his own new venture.
Perhaps he could have a bigger impact if he founded a new company.
If you have a brand new or revolutionary idea, that may be true. But if you have an idea that fits into what somebody else has done, then joining that company and using it as a lever to magnify the reach of your work might be more effective. Another way to look at it is that in a startup you usually have to worry about a lot of things that are not core to the idea, such as marketing and productization. If you have other people to do that for you (as I'm sure Facebook does) then you are more free to focus on implementing your ideas to the best of your ability.
I fail to see how he could possibly hope to achieve that.
I think Lars could get funding from just about any investor in the valley, if he wanted to start a new venture. But Norvig said it best (paraphrasing): "If you want to accomplish interesting things, go where the data is." Right now, FaceBook has one of the most interesting data sets on the planet, possibly rivaling Google's. If Lars wanted to replicate that, he needs to build a product that's useful enough to convince everybody to give him their data. That's non-trivial. It's easier just to join a company that already owns the data and then figure out interesting things to do with it.
I suppose you could have something like the PC revolution, where the industry was still based around manipulating data, but the data lives on your personal property. But that ship has sailed: personal computers have been available for 35 years, so pretty much all applications that rely on one person's data being manipulated in isolation by themselves have already been discovered. The interesting work now is when you collect patterns of data among many users, and that requires that many users be willing to give you their data.
You say it like our personal information is the only data in town.
Of course, any app needs data, but it can be anything.
Personally, the major reason I'm still interested in startups is because I haven't yet succeeded at one. It's the challenge that's the thrill. There's a whole lot of bullshit that goes along with that, though, and if I already had a major success under my belt, I'd be more than happy to go work for the hot flavor-of-the-decade company where I can spend my time innovating and let someone else deal with the bullshit.
I'm sure Lars is well-off as he is, but my understanding is that an order-of-magnitude more money is a real possibility.
If you do a startup, you spend a LOT of time thinking about managing and hiring people -- at least after the jump.
Startup are great places to potentially make a lot of money. After doing a few myself, including some pretty decent deals -- I don't think I'll do another unless I simply have some Google-like breakthough.