This was pretty interesting at the time, but AFAIK it never got any followups or provided detailed benchmarks of humans/AIs, which was too bad. It was an intere

This was pretty interesting at the time, but AFAIK it never got any followups or provided detailed benchmarks of humans/AIs, which was too bad. It was an interesting approach, and training DRL agents with that as a curriculum would be even more interesting these days.