But you won't learn as much about search as you would at Google, or about logistics as you would at Amazon, or about psychological manipulation as you would at Facebook. All that knowledge is locked up in the corporations that profit from it, and if all your software knowledge is based on what's publicly available, you probably wouldn't even know it exists. You might learn how deep learning works and how to apply it to MNIST digits from a web tutorial, but you won't learn anything about how to apply it to other problem domains or even what those other problem domains it can be applied to are from it.
And economic returns are all based on supply & demand, so if you're supplying a service that everybody else can learn too, your economic returns are bounded by what all those other people are willing to work for. The big bucks accumulates to monopolies, those firms and the people who work within them who know things that nobody else does.
But if your goal is to work for next decades google, Apple, etc, then no, Google is not the place to be.
Start working on Bitcoin. Start working on Ethereum. Get to the cutting edge.
Everything you do at google right now is about 10 years out of date (and I'm including machine learning and all their "cutting edge" AI stuff) and targeted at supporting their existing established business.
Take this from someone who has been on the cutting edge for 30+ years-- the challenge for most people is that the cutting edge changes, and you have to be paying attention to follow it... by the time you learn mobile development, surprise, there's bitcoin! (as an example.)
I don't recommend anyone to really go to Google to work on the future (besides perhaps if you want to get an edge in AI). It's just that SV has a lot of people that have accumulated experience while building Google, FB, etc and that helps a lot no matter what comes next. You get a lot of experience and people able and willing to take risks which is a nice basis to build something new.