
How Much Hacking Can One Person Automate With AI?
How much hacking can one person automate with AI? This episode examines the difference between speeding up individual tasks and running a dependable cyberattack campaign. The central question is how much human work remains once software begins taking actions and interpreting the results.
We examine Anthropic’s published accounts of misuse, including a data-extortion operation and a separate espionage campaign, alongside Britain’s National Cyber Security Centre assessment. The discussion explains why target counts, automation estimates and confirmed breaches measure different things. It also…
The skinny
The skinny isn't ready yet — notes appear once the transcript is processed.