You are confusing individual people working on the project and the corporation that ultimately decides the entire direction of the project.
You are also assuming that all people are inherently good and will immediately leave if something morally questionable is asked of them, and that people can even realise that something is morally questionable.
In no particular order:
- Google is a corporation which is driven by two metrics: advertisement revenue and Monthly Active Users. AMP is driven to improve and increase the two.
- Every single discussion outside of implementation questions have been shut down by the AMP team. The latest round was about AMP4Email where the team has gone out of the way to derail, shut down and outright ignore any direct questions: https://github.com/ampproject/amphtml/issues/13597
- A regular AMP page rarely weighs less than 1 MB when reached from cold cache. The only reason they are fast is because: Google dominates search, and Google aggressively preloads AMP pages from its own huge geographically distributed CDN.
So when you click on an AMP page it's instant only because it's already been preloaded. Meanwhile Google's own page speed tools say that AMP are not fast. And to add insult to injury, AMP is not even valid HTML5. https://ferdychristant.com/amp-the-missing-controversy-3b424...
Of course independent superhackers can't do anything because they don't have Google-level infrastructure. Because Google makes AMP appear to be faster than even http://motherfuckingwebsite.com which is just text.
Just a quote: "The performance benefit of prerendering is enormous, so enormous that prerendered pages can load hundreds of KB of JavaScript and still display “instantly” by the time the user taps on a link." https://redfin.engineering/how-to-fix-googles-amp-without-sl...
- Taking the previous bullet point further: AMP is a hack that has nothing to do with actual website performance or "slimming down" of websites. The only reason AMP exists is because of Facebook Instant Articles. It's not because of walled gardens. It's not because of "open web". It's just the desire to not be outdone by competition and the strive for ad revenue and MAUs.
It's a way to promote a few select websites and penalise every other website, even those who strictly adhere to Google's own performance guidelines: https://developers.google.com/speed/docs/insights/rules
Any and all concerns about AMP where entirely dismissed by Google (and AMP team) until a significant portion of publishers have started voicing their concerns and some have even started pulling out of AMP.
Why publishers you ask? AMP is primarily targeted at publishers and publishers are the only ones even allowed to somewhat participate in AMP. To quote "Standardizing lessons": "constant adjustment to publisher and user feedback." which is half truth at best. The only working group on AMP is amp-news-publishers (https://github.com/ampproject/amphtml/blob/master/contributi...), and any "user feedback" that is allowed is "intents to implement" and comments on how a feature is implemented. All other discussions are shut down.
- The new standards proposed by Google only seek to cement the status quo.
Web Packages is a way to store signed prerendered and preloaded pages on a CDN
iframe promotion explicitly describes how and why AMP pages are "fast": they are preloaded in invisible iframes during search. The only goal of this proposal is to help google make those frames visible and switch the browser URL when the iframe is presented to the user.
This does not actually solve any performance issues. You can still develop your 8 MB site. However, it will load instantly just because Google has preloaded it while you were searching.
Do not be fooled by the use cases. The primary benefactor of aggressive preloading and caching is Google, who already has the entire infrastructure in place, and dominates search. No mater how much you try and adhere to all the performance guidelines, you will never beat a pre-rendered and preloaded page stored on Google's CDN.
- And yes. Ultimately the AMP team in its entirety fails to accept or acknowledge any shortcoming of AMP. Even the "Standardizing lessons" spew the same easily falsifiable myths around AMP heavily sugar-coated in self-aggrandizement.
On top of that, no matter how many times they say that the process is well-lit (they say it three times in the article), the process is entirely opaque, developed by Google, for Google, and in Google's products and then presented post-factum as a "standards proposal".