I don’t think there’s a contradiction. Anthropic writes this because they want people to believe this, the author argues though that no one believes it.

I believe there’s a small minority of people that truly believes AI is a friend, but I would say it’s a psychological pathology.

I don’t bother trying to guess what large companies really think: a/ they’re made of so many different stakeholders I don’t think it’s possible. And b/ I know money is the most important thing if they are large enough and have lots of anonymous investors, I don’t need to know anything else.