Nathan Helm-Burger comments on What do coherence arguments actually prove about agentic behavior?

[Nathan Helm-Burger]

Also, both human and deep neural net agents are somewhat stochastic, so they may be randomly intermittently exploitable.