I have been reading a few of your takes, and I have to say that, while amusing, many of them demonstrate a lack of basic understanding how things work, or alternatively you write them like that on purpose to get more rage comments -- in either case I won't be reading any more of them.

I will, however, write down some examples based on this article.
Number one: you clearly don't understand network effects (or pretend to), like that the fact, that all of your friends are on Facebook, means that if you build another Facebook, you won't get any of your friends to change - because all of *their* friends - and also their university department, teachers, coursework, study announcements, etc, etc are on Facebook, and their time is limited. So they won't give a rat's butt about your app - or a million similar apps, no matter how much better those happen to be. This is one of the reasons why enshittification, or if you prefer, 3x strategy, works just fine on the internet.
This is also why Apple and Google will likely forever make 30% off the revenue of the ever-hopefulls who think that their new (buggy, horrible-to-use, pseudo-ai-generated) app will be the new Facebook, or the new X, or the new Reddit.
This is also the reason why our app stores are full of exactly those shitty pseudo-ai-generated apps, and your YouTube feed (with its 8-minute-adbreaks) is full of shitty, pseudo-ai-generated videos. Sure it makes some money for the people who post them, but you know who it makes a lot more money for? Google. Which is why they will just let people continue making pseudo-ai-generated videos. And you will simply let them, and won't change to, say, Vimeo, or a million other apps, because all creators are on YouTube. (Unless of course you are into a different content form, like short videos, and are already on TikTok, where the same thing will happen sooner or later).

It might very well be that Google's ads, excuse me, search results, are now more relevant then before, but nobody cares, because on the Google results page there are now also far more ads than actual search results, and it is getting increasingly more difficult to distinguish the ones from the others. So despite the "objective quality" of results getting better, the subjective quality of user experience on google is becoming just like altavista was 30 years ago, (except with less prn, but also without regex and logical operators, so you can't even fine-tune your results.) Which is, by the way, when and why a back-then clean and ad-free google took over the search market. And yes, you could build another search app, since those don't have network effects. But also, good luck single-handedly acquiring all the infrastructure that you'll need to host the crawlers and databases to host your own search engine.

Re AB-testing: it doesn't make a product better for any target group. It increases the revenue for the corporation. It can make the product objectively worse to use, but if A brings you, as a company, more revenue then B, you as a company are going to choose A. This is not theory, see Amazon Prime accidental subscription and cancellation process, where after the AB testing they kept the confusing subscription but also went and redesigned the process to be more difficult to unsubscribe. Optimizing for revenue, just like optimizing any other multivariate system for a single variable, in pretty much no way leads to an improvement in another, at best weakly dependant variable, i.e. better experience for consumers. The best you can hope for is it happening as a side-effect.