IMHO "100% freedom" is utopia. You don't have 100% freedom even on a start-up. In a corp like Google you have managers, and in a start-up you have customers...

Having said that, if you work for a big high-tech corp (like I did, although I never worked for Google per se), you will find that it is hugely difficult and frustrating to push your own idea through the system.

That would be Y Combinator.

One huge difference with YC is the funding amount. YC essentially pays a low salary for a few months to build a prototype. That narrows the talent pool down to people with very minimal financial requirements (couldn't be more than ~5% of the total Great Hackers in the US).

Google could afford to take all the other ~95% of Great Hackers by paying a high salary until the project is either killed or a success. The big problem is that Google doesn't fire people for failing, they just move them around like most big companies do. They suffer from the "good effort" syndrome, where people are rewarded for "trying" even if when they fail to produce "something users want".

5%? Really? There's a huge class of young entrepreneurs, who have limited financial requirements (notice, I'm not using the word desires), and slightly older hackers who have greater requirements (car/house/etc) but probably have worked long enough to supplement the YC income with their own stock options, savings, etc.

If you really want to build something new and interesting, you should be willing to make financial sacrifices to make it happen. If you're comfortable, you probably won't be as hungry.

But our goal is to hook people up with investors, and investors are eager to invest in anything promising. Most great hackers can make something promising enough to interest investors in 3 months, esp. with our advice about how to do it.

Do you have resource and IP sharing among startups?

Well obviously it happens, seeing how Infogami/Reddit deal occurred. They probably have a lot more than between the average set of startups, at any rate.

Riches if you succeed and no job if you fail is like a startup.

It's still no substitute for people who want to do real startups. But a good salary during the effort is a great way to attract all the great hackers that need reliable income.

Getting a big check or a pink slip at the end is ideal.

You might have a better chance of succeeding within a company though, particularly a company with massive computational resources and accumulated IP. Scalability would be much easier to deal with. There's also the issue with name recognition. Having your service on something like Google Labs would give it lots of publicity. Finally, your service could be added as a feature to a popular service such as gmail say.

Makes me think of the issues PG raised in "The Power of The Marginal". I think when you're an outsider it seems like those things are huge advantages, but when you have them they feel more like baggage.

Interesting idea. Rather then taking on all of the risk (and reward) by doing a startup that risk would be spread across all of the employees in the organization. Sounds like the organization would exist to provide "startup insurance".

This is different from pursuing a startup because you are allowed to make use of massive resources and whatever IP the company has accumulated.

So basically something like the classic Bell Labs, Xerox PARC, etc. environment? But you have to make money, not innovation? If that's what you mean, no thanks...

This would not be a research lab. The goal is to create something that makes money not publications necessarily. You can try to do something really novel just as long as it ends up making money.

Yeah, as long as there are other Google-like companies that give me 100% freedom that I can go to when I get fired. ;-)