Dr. Marcus's latest NYT piece. I think core point is building AI systems requires a hybrid approach; Purely data-driven merging with symbolic-algorithmic methods.

This is perfectly logical --- but it may be based on a flawed assumption.

This assumes that intelligence is *really* the objective.

What if the *real* objective is simply to suck trillions of $ out of the investment system as quickly as possible before the whole thing collapses?