We could have a better solution to this problem with reduced (I am not naive enough to think Google would be fully altruistic here) lock-in and better engineering, and Google is definitely capable of delivering that solution. Even with internal (past Googler) perspective here I honestly can't come up with an explanation  for why AMP is so bad from an engineering and feature set perspective. It's built with a good understanding of the problems users face, and they've adapted somewhat well to the needs of publishers, but the design is a trainwreck from the perspective of a frontend webdev, browser engineer, or someone delivering content that isn't just news articles.
It's quite right and reasonable to start with something and move forward to a more ideal, standardized solution, but it seems very difficult to evolve AMP's design towards anything that really meets the needs of the web platform in the long term. It's basically a walled garden through and through - just a garden that happens to have some really nice flowers, maintained by world-class gardeners.
 There is an extremely cynical explanation here that applies to many low-quality Google product launches - shipping something ASAP to qualify for a promotion. But the AMP team has demonstrated real commitment so even if that was what originally set things in motion, it doesn't explain what's going on here.
I'm not sure how I feel about AMP as a "standard", but I don't object to it as a technology. Really it's just building on WebComponents and proving some more extreme optimization techniques. At least the tech is open and can be reused by others.
Requiring conformity to a Google-controlled standard to be AMP-validated does seem inherently less "open" to me though somehow. Though I also understand that few other parties seemed to care about performance at all prior to this, so at least the strategy is working.
The biggest request I'd have of Google regarding AMP is to have a clearer separation from AMP Cache. The technologies are separate but people have a hard time differentiating them - even on Hacker News. Most of the complaints people surface here aren't even about AMP at all.
I'd also suggest that any AMP Cache provider should be eligible for the carousel in Google SERPs. Or alternatively it should just be discontinued soon, as I have to imagine it's served its primary purpose by this point. Keeping it around longer just encourages comments that Google is being unfair in search (I understand AMP pages are technically not a factor in organic search, but the carousel still provides a search boost).