I'm sure that's true, but it is also noteworthy that any and every consideration that goes into the initial inclusion of the data before it is fed into the model introduces intended and unintended consequences on the training.
Furthermore, the proliferation of the LLMs themselves is putting negative pressure on survival of the places where all the good data is sourced from in the first place. When traffic to a place like stackoverflow is way down because everyone's reading LLM answers (that the LLM training dataset got from stack overflow), there are less good conversations on stackoverflow to read. Some of these sources of training data may even be caused to cease to exist entirely.
That may be the plan, but that amount of money is not significant to even the poor for more than perhaps a month or so.
A bribe is a transaction--I pay you this, and you let me get away with that. If I want to continue getting away with things, I have to continue to pay, since I am the one continually putting myself in a compromising position. The second I stop paying, the ride ends.
Also, promising payments at some point in the future may fool 3rd party onlookers into thinking you're helping them out, but I can't imagine the people who were promised the money are going to just not notice that they were welched on, even if they love the person that stiffed them.