> How does one discover these alternative UIs?
Social trust networks -- they get recommended by friends or shared online by someone you trust.
> How can average user trust that the data on these alternative UIs isn't being manipulated?
Those same social trust networks -- someone realizes what's going on and shares it, people start telling other people to stop using it. There's no technological solution to prevent technological abuse -- the abuse can be built into the system, after all. Half the warnings about "stop using this site" are warnings about a first party UI -- "Facebook is reading your messages!" or "Youtube is hiding dislikes!" or whatever. Social networks (as in, networks of people, not websites) are the only way this sort of knowledge is ever spread.
> How long you think it will take until big sites would routinely do small tweaks to their databases/structures to systemically break these alternative UIs?
Not long, of course. I'm not sure if this is an argument in favor of Web3, though, since that can work the same way -- a startup needs to move fast and frequently change things to iterate, which means they will be changing the data representation on the blockchain frequently (going forward, of course they can't change past blocks), which means alternative clients will need to keep up with those data changes same as with API changes on a normal website.
> More so why would any of these for profit websites, like Youtube, even have a presence on web3?
? Web3 is just as for-profit as the normal web, if not more since every database interaction is monetized? YT and the like aren't on it because there aren't that many users.