How does this differ from testing an AI in a simulated universe before letting it out into the real world?
It depends on what you define “simulated universe” to be. Here it uses the AIs own prediction module to gauge the future; if you want to call that a simulated universe, the position is arguably defensible.
How does this differ from testing an AI in a simulated universe before letting it out into the real world?
It depends on what you define “simulated universe” to be. Here it uses the AIs own prediction module to gauge the future; if you want to call that a simulated universe, the position is arguably defensible.