Industries often have sayings that turn out to be relevant only in the short-term. It used to be that you needed to tackle a "technically difficult problem"; but many startups have succeeded by building something that people want without creating new technology but "executing" well. He's suggesting we cycle between technology and applications of it.

Obviously we need both tech and apps, but is there evidence in the past of them occurring in cycles?

That said, it's true that everyone is complaining that there's nothing really new, that most apps are trivial. And of course, in the last few years bandwidth/browser/cloud has gotten good enough for web-apps to go mainstream and cause modernization across many industries, which requires domain knowledge and "execution".

I'd love opportunity to be in "enabling technologies and platforms" - but it seems innovators in software infrastructure (libraries, languages, middleware) today often get nothing. Exceptions from the past: databases, postscript. Please prove me wrong!