Sounds like just the next calculated step. 1. Introduce $badThing with claims about page load performance. 2. tell sites they need to use $badThing to get into the “carousel” 3. Scale back the requirement of #2 just soon enough that you can argue against any claims of market manipulation via your monopoly. 4. (Continue to) Profit, because 80% of the sites that implemented it aren’t going to just rip it out now that i…
i disagree with 4. Sites need maintenance and evolution. i don't know a single dev that won't eagerly drop something as impacting as amp as soon as it's an option.
Even for those who acknowledge Google may do shitty things, but are more interested in pulling their dick or what have you, rather than doing the work to make a great $thing, AMP gives them an easy out. No interpretation required, no need to worry about timings or what a "core web vital" is. Just follow this bouncing ball. Monkey see, monkey do.
(b) that assumes the dev is allowed to. Plenty of middle and upper management will always defer to "the industry" over any amount developer insight - particularly when the pitch is essentially "lets remove this thing so we can do some more work ourselves to achieve the same result".
(c) even if the dev wants to, and management OK it - maintenance is literally that: maintaining the status quo. Evolution, particularly on the scale of large sites, is a long term thing.
Big news sites are some of the worst offenders for junking up a page. Google could have "fixed" that by using the metrics discussed in the article to determine "carousel" results, from the start. But then they'd just have made a bunch of websites load faster, and wouldn't have added yet another tentacle around the web.