AMP HTML is interesting
I think Google has learned from AngularJS, that the earlier you make public a framework, and ask for partners/collaborations, the better. "Let's do it a community effort". I don't know if it will succeed, but at least is they are doing it open source. That's to be praised. I wouldn't imagine something like this coming from Apple, for example.
After saying that... If AMP works for you, Google, as a smart trick to load faster some news snippets in your news carousel, congrats. But don't call it a framework. There was no need, you can set up a guideline of "best practices", but AMP is like a pretentious joke, creating a "subset" of the web, I don't care if it's done with the best intentions. Only Google has the power to push it down the throat to his media partners. Sad for them.
Aside of that. AMP performance is NOT impressive. It's just "good"
After all the tricks that it's doing, cutting down the expressiveness of HTML and SVG, controlling your JS loading, telling you how to create an equivalent HTML canonical page, no forms allowed, and so
many other restrictions, some of them plain silly... after all that... the content doesn't load that fast anyway. It's a trade off that no one in its sense would accept. Unless you "have to".
It's like creating a m. site at the document level.
I'm glad that AMP is public so we can de-construct it and take the best parts of it, without the crap.
For example, a week ago, I postedhttps://twitter.com/dcorbacho/status/649508751544852481
that the work of Andrea on Web Components is amazing, and "under the radar". I'm sure AMP will give the document-register-element a big boost in popularity, since it's the only dependency that AMP has.https://github.com/WebReflection/document-register-element
Still my opinion on AMP is not final though. But the first impression was quite bad. Let's see how AMP evolves.