Twitter was being used by AI bots to digest millions of tweets to determine emerging narratives and to flag millions of "misinformation" tweets. They would then use this info to have tweets suppressed and form counter narratives. The rate limits that Elon implemented hindered their ability to do that ....
During the Rate Limit if you searched for France you got what you were looking for, when you do the same search now that the rate Limits is gone it seems like you get thrown everything and the kitchen-sink that just so happen to have the word france in it, fucking Cardi B visiting france, various k-pop and other accounts with france as a part of the name(including @* * * * * francine_ * * * * whatever), fucking ANYTHING that happens to have anything even remotely like 'france' anywhere in it..............
An alternative theory:
https://theconservativetreehouse.com/blog/2023/07/03/beware-france-macron-moves-to-shut-down-internet-communication-to-block-rebellious-uprisings/
The limit on non-paid users seemed reasonable for casual users. This theory assumes that obsessive non-paid users have significant influence on the platform. It might not be a bad idea to make the amount of influence that people have more equal with each other. For example, why should a NEET online activist have more influence than someone working a full time job?
Some of these non-USA users will include troll farms with human users, whose job it is to push a narrative. A captcha will not affect them. AI is fairly new. I expect that less advanced despotic regimes are still using old fashioned troll farms, which were heavily impacted by this.