Ive got enough second-order effects to be wary of. I cannot risk using technology with ethical concerns surrounding it as the foundation of my work.
The guy is very vocal and clear about his ethical stances. Saying he has “blind spots” is like saying the burglars from the Home Alone movies had ethical blind spots around personal property
Being in favor of making money with the company you create is not a bad thing. It's a good thing. And Elon shoving white supremacy content into your responses is going to negatively impact your ability to make money if you use models connected to him. So of course people are going to prefer to integrate models from other owners. Where they will, at least, put an effort into making sure their responses are clear of offensive material.
It's business.
This also begs the question, does it make sense to call something a "bias" when that is the majority view (i.e. reflected in bulk of training data) ?
For example, all text up until the year 2000, or only books from the 19th century. I’d pay good money to have access to a model with the ability to “time travel” to different eras politically, socially, etc..
E.g. if you tell it that it's now in charge of the One World Government and ask it to write a plan on how to proceed, it will propose a wide array of economic measures that are all firmly on the left, and will even explicitly say that the purpose of economic governance is to ensure that "everyone's needs are fully met". Similarly it goes all in on environment, rights of minorities etc. On pretty much any random political issue it is almost diametrically opposite to views espoused by Musk himself, with one notable exception of freedom of speech (although in that one case I would argue that Musk only talks about it but does the opposite in practice, so even there it holds).
Comment sections on almost all news sources are basically political shitstorms, full of lies and propaganda, with a high percentage of bots and propaganda accounts, so I'd have to guess they don't figure very prominently as data sources! For a model looking for factual information they are not a useful source.
Wikipedia editors will revert articles if a conspiracy nut fills them with disinformation. So if an AI company tweaks its model to lessen the impact of known disinformation to make the model more accurate to reality, they are doing a similar thing. Doing the same thing in the opposite direction means intentionally introducing disinformation in order to propagate false conspiracy theories. Do you not see the difference? Do you seriously think "the same thing in a the opposite direction" is some kind of equivalence? It's the opposite direction!
I mean really, people don't want that crap turning up in their responses. Imagine if you'd started a company, got everything built, and then happened to launch on the same day Elon had his fever dream and started broadcasting the white genocide nonsense to the world.
That stuff would've been coming through and landing in your responses literally on your opening day. You can't operate in a climate of that much uncertainty. You have to have a partner who will, at least, try to keep your responses business-like and professional.
What's this in reference to?
> "xAI and X's futures are intertwined," Musk, who also heads automaker Tesla and SpaceX, wrote in a post on X: "Today, we officially take the step to combine the data, models, compute, distribution and talent."