The way I understand John Mearsheimer's "offensive realism" it doesn't seem compatible with Mustafa Suleyman's views on how the coming wave of AI can be kept benevolent through liberal means and democratic cooperation.

That's a bit scary to be honest.