Stuart_Armstrong comments on Agents detecting agents: counterfactual versus influence