Why Join Google
Posted to the Overcoming Bias NYC and Bay Area LessWrong mailing lists shortly before Alyssa left MetaMed Research to join Ray Kurzweil’s AI team at Google in Mountain View. Aimed at rationality- and existential-risk-minded software engineers considering where to work.
Next month, I will be joining Ray Kurzweil’s AI team at Google Mountain View. I thought I’d take some time to explain why I’m doing this, and what’s attractive about Google as an employer.
If you’re a software developer, you probably know about Google’s famous perks: the free food, the shuttles, the massage chairs, the dog-friendly offices, the famous co-workers, and so on. Plenty of other people talk about that, so I won’t repeat the usual spiel. Instead, I’ll talk about the things I myself didn’t know when I was interning there (back in 2011), and which may be especially relevant to list members.
Google has the world’s strongest AI team. Or, to be more accurate, Google has many strong AI teams, each with their own areas of focus and many growing rapidly. They recently acquired DeepMind for over $500 million , along with over a dozen other robotics and AI companies in the last six months alone. In addition to Kurzweil and Hassabis, Google’s AI people include Sebastian Thrun (of Udacity and self-driving-car fame), Peter Norvig, Andrew Ng, Andy Rubin of Android, Geoffrey Hinton, Jeff Dean (the man, more than anyone else, responsible for Google’s awe-inspiring massive parallelization toolkit), and on and on and on. Even before the DeepMind acquisition, Norvig estimated that Google employed “less than 50 percent but certainly more than 5 percent” of the world’s machine learning experts.
As part of the DeepMind acquisition, Google has explicitly set up an AI safety board to protect against existential risk from UFAI, with the effort apparently spearheaded by Shane Legg. More than any other company, Google is a place where succeeding at FAI and avoiding UFAI is a serious and practical issue.
Google is the only large company making a serious anti-aging effort. Before Google, anti-aging was a fringe interest, with Aubrey de Grey’s SENS and similar groups running on small donations. Now, thanks to Larry Page, the anti-aging company Calico is being headed by Arthur D. Levinson , the chairman of the board of Apple, with a (presumably) correspondingly large budget. Time Magazine’s cover proudly announced “ Can Google Solve DEATH? That would be crazy — if it weren’t Google.
” Peter Thiel, among others, has (rightly) criticized Silicon Valley for focusing on “little apps” that “don’t matter”. Google has an entire division, the ‘CIA-like’ Google X , focused on ‘real’, breakthrough progress in physical technologies. Most of the projects there are carried out under the utmost secrecy, but we can get a feel for what goes on there by watching the Solve For X ‘moonshot’ talks , posted online.
You’ve probably heard about self-driving cars. They’ll prevent accidents, save time, speed traffic, yadda yadda, a great invention. But hang on a sec. Look out your window. Look at what you see. Some buildings, streets, trees, lawns, parks, and so on. Unless you live in Manhattan, literally everything you see, from the height of the buildings, to the width of the streets, to the lawns and green space, to how long the driveways are, everything , was explicitly built around ensuring that everyone and everything had enough parking (check out your city’s zoning code if you don’t believe me). There are more acres of parking in the US than there are acres of buildings; if you combined it all into one big lot, it would cover the entire state of Massachusetts. The modern US is built around parking, just as much as the antebellum South was built around cotton. Once we have self-driving cars, all of that becomes unnecessary, since the car can drive itself somewhere else when you aren’t using it. Self-driving cars will literally wind up rebuilding the entire physical United States, one year and one red brick at a time, unless the Singularity arrives first.
Many Singularity/rationality folk are already at Google. My own team includes Moshe Looks, Marcello Herreshoff, and Jesse Liptrap, and a quick look at MIRI’s workshops shows many more — Nate Soares (Google), Mihaly Barasz (Google), Nisan Stiennon (just joined Google), and on and on. My guess is that Google probably has more of us than any other for-profit company.
If you want to join Google, the easiest way is to get a current employee (of whom there are many on-list) to refer you; this will greatly increase your chances of getting a phone screen. Having just gone through the interview process, the questions are not generally that hard (although you will make some mistakes, as you’re on a whiteboard and under time pressure); just practice, read through McDowell’s Cracking the Coding Interview , and you’ll probably be fine.
And, of course, the perks are nice.
Disclaimer: this is only my opinion; I don’t speak for Google Inc. Most Googlers get a bonus if someone they refer gets hired, but this does not apply to me, as I haven’t started yet.