I am assuming he is referring to cryptocurrencies & blockchain tech. There are plenty of businesses doing well but that's irrelevant. Waiting for established businesses in a growing space before jumping in isn't very innovative.
On the other hand, in the same time period (actually in a shorter time period!), machine learning driven technologies like voice and image recognition have gone from sci-fi to commonplace. I ask: which is more innovative?
If you wait until it's obvious to everyone that something's a good idea, you're not being innovative, you're just an opportunist. Similarly, if it turns out that it's not actually a good idea, you're not being innovative, you've just wasted a whole lot of time.
Innovation happens at the intersection of contrarian & right. You have to be willing to stake your time, energy, and ego on an idea that many people believe is bad, impossible, or not worth it (otherwise it would've been done already), but you also have to be right in your judgment.
1. That Google didn't research crypto internally and purposely decide not to pursue it.
2. That crypto is worth staking Google's time, energy, and ego on.
3. That crypto involvement is a direct reason a company is or is not innovative.
Internal research is useless.
The only way to know whether a new idea will work is to try it: put some minimal form of it in front of people and see if they actually get use it. People are complex; sometimes they do things that all rational thought says they shouldn't (buying Bitcoin and investing in ICOs would both fit into that category, IMHO), and many times they don't do things that all rational analysis says they should (I participated in research projects while at Google for both micropayments for content and for labeling trustworthy sources on the results, both of which most people would agree are good goals but neither of which has ever worked, and I founded a startup afterwards to provide career guidance to undergrads, which every adult we talked to said "I love what you're doing, and totally wish that existed when I was a student" and every student said "I love that idea and would totally use it", and then promptly never looked at it again once we built it.)
There's no guarantee that external research works either, but at least you've learned something applicable to the next idea.
And then get skewered left and right for shutting down products? You can't have it both ways.
Fast forward to today and 10x to 1000x the number of us wonks are getting an overlapping amount of Reader's value (plus new value) from Twitter. The purging of Reader is a distant memory, but it doesn't mean it's necessarily bad to try a bunch of things and kill the non-viral losers (in a relative sense).
> The only way to know whether a new idea will work is to try it: put some minimal form of it in front of people and see if they actually get use it.
And in the face of a universe of potential ideas how do you determine whether you invest time into prototyping pre-sneezed tissues? Internal research. You’ve just moved the goal post.
b) 'Wasting time' is a part of research. If you never fail, you're missing opportunities. That said, you pick your failures with the hope of getting the right kinds of wins... which brings us to...
c) Giant centralized cloud-based businesses have giant centralized and for-the-most-part trusted datastores. There's no reason to use a blockchain if a standard distributed database suits your needs. From where I sit, blockchains are mainly a solution in search of a problem.
(and just imagine the press cycles if a FAANG company rolled out a blockchain, and thus burnt god-knows how much coal bringing it mainstream...)
I'm equally convinced - and I suspect I disagree with most HNers here - that there are also some problems for which blockchains are the only solution - they just don't work with centralized data stores, mostly because nobody would trust a single entity to manage that data store. Google isn't even looking for those problems. In general, Google does not look for problems, it takes problems that everybody knows about and looks for solutions. There are thousands of entrepreneurs in the cryptoverse who are looking for those problems.
There was a time, early in its history, where Google was willing to take a solution - download the web and keep only the links - and then find the problem (search) for which it was the solution. And then they leveraged that solution into all sorts of other problems - webmail, navigation & directions, news, academic papers, video, etc. But the thing is that this is a risky business: most solutions in search of a problem don't actually solve anything, so the time spent solving them is wasted.
Name a single actual problem that is only solved by the blockchain.
I'm not sure if you're being facetious, but assuming Amazon is a member of "FAANG": https://aws.amazon.com/managed-blockchain/. Does that count?