Ambitious projects and optimistic predictions have seemed not-unusual among big-name AI people, since the start.

Marvin Minsky predicted emergent behavior and recursive effects much sooner. (I think this was in all honesty: not realizing how complex the human brain, and not knowing how hardware would develop.)

Douglas Lenat apparently thought commonsense knowledge and reasoning would be easier.

Ray Kurzweil was often making predictions about hard-AI accomplishments just around the corner, which we're starting to see much later.