LeCun is more accomplished and smarter than I will ever be, but his thoughts on the term 'AGI' just seem like dumb pedantry regarding word definitions to me.

Sounds like it is more like an wishful thinking from a person who spent his life mastering deep nets than a nitpick. Like "Sure deep nets can never be an AGI, but Humans are not AGI's so we can reach human level performance with deep nets, it is not a dead end! IT IS NOT A DEAD END!".

I think the saying "Science Advances One Funeral at a Time" applies here.

Perhaps you should examine your seemings then.

Sounds exactly like what I would say if my life's work revolved around the idea that AI boils down to how many neurons we can simulate