Tyler Cowen Urges Estimates of AI Cybersecurity Costs

Tyler Cowen posted urging the AI safety community to supply scientific estimates of added cybersecurity costs from hypothetical AI break-ins, along with numerical forecasts and tracking of shorted assets. Rohit Krishnan shared the post on X. Jeff Huber quoted it to stress rigorous methods. Albert Wenger replied that it overlooks tail risks. Taylor Lorenz noted comments criticizing the shorting suggestion but called the first two ideas reasonable. Roon said cyber crimes are not the main concern. Brad Carson answered two questions with a paper and website produced using Claude and twelve agents.