Yeah I have the same feeling. I also think it’s weird to say he’s clearly wrong. I mean, it’s a very subtle question exactly where you cross the line that comes with an obligation to credit others. All ideas build on each other and are to some extent mashups of ideas that came before them. It’s not something you can be clearly wrong about in an objective sense, and Jürgen’s view seems consistent.
Looking at how credit and attribution works in science today (google ”citation rings” for example) I can honestly say that I’d much prefer to live in a world where Jürgen did invent most of AI, rather than the one we’re in now.
> it’s a very subtle question exactly where you cross the line that comes with an obligation to credit others
What he means is there are clusters that systematically avoid citing the true origins of important concepts related to computational learning, and prefer to only cite their own "cluster members", a phenomenon that in scientometrics is known as a citation cartel:
Looking at how credit and attribution works in science today (google ”citation rings” for example) I can honestly say that I’d much prefer to live in a world where Jürgen did invent most of AI, rather than the one we’re in now.