Of course! Is this actually news though? They also did things like this before: they paid for a patent on drop out regularization. Math isn’t patentable, but a corporation will turn as many screws as it needs to make money.

A for-profit company doing research is always in the interest of improving the bottom line. I’ve always read papers from corporate labs with a big grain of salt. They won’t publish something that puts them in a negative light. They have an incentive to exaggerate and bend the truth. A companying publishing something is always marketing.

This is why it’s vital to have academics and non-profit foundations leading the frontier of research. It’s why the White House illegally pulling congressionally appropriated funding for universities is such an utter tragedy (and extreme national embarrassment). Everyone has their bias and their agenda they’re trying to advance, but publicly funded research is the better option when compared to for-profit corporate research.

I'm sure they'll start publishing everything again as soon as OpenAI do.

How could they do that? Don't they know they owe it to us to make the world a better place? /s