tailcalled comments on Many arguments for AI x-risk are wrong

[tailcalled]

Do you have a reference to the problematic argument that Yoshua Bengio makes?