I quit Chrome some months ago. With more perspective on it I can now say I should have left earlier, which is sad to me because even now I still watch the project (and even occasionally participate). The many parallels to a breakup are not lost on me.
I took a while to move on because I wasn't sure what I wanted to do. My farewell post above describes how my career already covered a computer-scientist arc and a free-software-hacker arc; what else would I like to try?
And I eventually realized what I've missed is the generic employment experience: being paid to write code, not as an urgent need to solve puzzles and prove myself or as some statement of idealism, but rather as just applying my skills to help build something that's useful and selling that thing for money. (Even working on a product that makes money is something I haven't done in nearly a decade.)
So, my new team: small, friendly. Includes women (that other half of humanity that is conspicuously absent from the salt mines of silicon valley). Some of us have kids. We work hard but also go home. I can see the ocean from my desk.
The product itself — Google Consumer Surveys — is at first glance boringly corporate (to this day I still get "AdSense" and "AdWords" confused) but that's part of the charm. It turns out that there's more to it, too.
In brief the product is almost two products: one is that you can create a survey with us and get a response sampling the US population in matter of days at an order of magnitude less cost than before. The other side is that website owners can run these surveys to be answered by their users one question at a time and get paid per answer at a rate that works out to be an order of magnitude better payout than display ads.
What I find most interesting about this project is the greater social context. Ads, even the better relevant text-based ones, are always slightly confrontational with their subjects: the site owner wants you to click on their ads but isn't allowed to encourage you to; users rightly feel like a commodity that is being sold to an advertiser.
With our surveys the financial relationship is made explicit. Our surveys always run in a format where the user must make a choice: either answer the survey or do a task specified by the site owner, such as buy access to the content. It sets up the expectation: this content costs something to make and you must share that cost. The user can pay with their data (answer the question) or they pay with their wallet, and they're always free to close the page if they reject both options.
In that light you'll note that surveys, the thing the product is named for, are not necessarily even the point. Which is sort of apt in that it's not the point for me, either. I'll hopefully have some interesting war stories about being a web developer again after years in the browser world for you in the future.