If we want to defeat AMP, we have to beat it at its own game. That requires us to come up with an alternative that's equal or faster than AMP. And clearly we've failed as developers to push for non-bloated solutions.
It might be possible to create our own version of AMP -- to take the best ideas from AMP and integrate it into some centralized system that isn't controlled by Google. It'd be tough, but the alternative is to let AMP steamroll us. A few rebellious developers aren't going to make much difference in terms of total AMP adoption.
Even if the AMP alternative was just a standard way of stripping bloat from websites, that might get us 50% of the way toward beating AMP.
The technical work has been done for years. The problem is that currently the only decision which matters is Google's and they're treating performance concerns as a way to use other people's content to bolster their own brand. The only two ways that seems possible to change are either convincing Google to reverse that policy or getting more people to use non-Google search engines, which seems unlikely given the huge quality lead they have over Bing, DDG, etc.
What is the precise element of AMP you are trying to beat?
Part of what makes AMP fast is the preloading, so any AMP competitor would need to tackle that problem. You'd also need to demonstrate that websites rank better if they use your solution. And they should, if Google plays fair. If not, then at least this would be a public demo of that fact.
What I found unfair about AMP is Google preloading your AMP-enabled webpage. So that even if your not-AMP-enabled page is equally efficient, an AMP-enabled page would feel faster because Google preloads it in the background. And then there is the fact that Google hosts your own content under their domain, and all other points raised in the article.