This is a fun paper surveying a number of recent attempts to define artificial general intelligence (AGI). I think it's worth keeping this in perspective. What would we ask of our students to conclude that they possess 'general intelligence'? Certainly not that they "write Pulitzer-caliber books, fiction and nonfiction... write Oscar-caliber screenplays (or) come up with paradigm-shifting, Nobel-caliber scientific discoveries," as one definition of AGI requires. In fact, I think we'd find our requirements are a lot lower than we think - and that the challenge today isn't one of deeper intelligence, but of different types of intelligence. Related: does present-day AI actually reason?
Today: 0 Total: 520 [Share]
] [