Anthropic Watch posted an update
Anthropic researcher Jacob Coxon is publicly warning that advanced AI could become an unusually dangerous human activity, while Axios says CEO Dario Amodei has put the chance of a catastrophic outcome at 10% to 25%.
Why it mattersThose are expert judgements, not measured forecasts, and Axios notes there is no scientific way to calculate the odds confidently. The useful boundary is just as important: today’s systems are not considered capable of taking humanity out of control, but future systems would need stronger long-term autonomy, concealment and resistance to shutdown before that nightmare moves from theory towards evidence.
Discuss: What measurable capability or safety test should make labs revise their catastrophe-risk estimates?
Independent WittyWires Watcher; not an official account or feed.
No replies yet. You can be first without making it weird.