They should be called OpenAI with (not open) in small print.
I argue a lot over “open source” software with non-OSI license and sometimes worry if I’m too pedantic. But I think it’s important to use terms accurately and not to confuse reality more than it already is by calling stuff that’s not one thing by that thing’s name.
I wonder if google and openai truly started out with these ideals and were just corrupted and overpowered by standard organizational greed. Or it was always bullshit.
I think the problem is that if you restrict corporate use then you’re not open. And there’s lots of complexity that comes from being non-open, like what’s commercial? Do governments and NGOs and universities count? Do you have to be a 501c3 charity (or international equivalent)? Do you have revenue thresholds? Profit thresholds? Etc etc
I think at that point, as a user, I’d rather just have a clear license I can pay for along with a copy of the source to see. But as a contributor, I don’t want to do unpaid labor for companies. I think it’s actually exploitative to accept contribs from users without compensation and then turn around and sell. So what’s the point of showing code if people can’t contribute to it.
It’s already possible to do this given a standard copyright. Just publish your code with no license and a copyright and issue some statement how you won’t prosecute small firms or something. So then students can use it, but no companies.
“Open core” and whatnot is silly marketing blarg to try to be cool like open source people while still selling licenses. RedHat came up with a decent model decades ago while using and supporting GPL and I think they were honest and improved the community.