Earlier on, the guy seemed more like a karen than a shill, but as he went along, his arguments overwhelmingly leaned towards attacking Facebook's competitors and being friendlier towards Facebook.
QT: https://qoto.org/@olives/112709003324867944
He dislikes transparent training sets, even though it makes it easier for someone to figure out if something shouldn't be in there. Why? Because, Facebook has a non-transparent training set (whose contents are unknown).
His arguments about E2EE were also very specific, so that he could oppose it for Messenger (until Facebook actually implemented it) but not for Whatsapp. Presumably, because he is supporting whatever it is that Facebook is doing.
His arguments about AI are also overly specific, intended to give Facebook a way out. It focuses on very specific types of offensive content but otherwise upholds their model of collecting data.
If he was a genuinely concerned person, he would have seized on more pertinent concerns about web scraping or other things. Instead, it provides a veneer of such scrutiny while doing no such thing.
Essentially, his goal is to trip competitors up on whatever he can come up with, while giving a pass to Facebook.
Even in the case of end-to-end encryption. He was opposed to it because Facebook was opposed to it and wrote grandiose remarks on that.
Then, when Facebook swung towards implementing it (due to European pressure), he doesn't care in the slightest about it.