AIAnthropic’s confidence about today’s AI risks draws criticism
A commenter disputes a line they quote from Anthropic’s Opus 5.5 blog, arguing that potentially worse risks ahead don’t mean today’s risks are under control.
TLDR
A commenter quotes Anthropic’s Opus 5.5 blog as saying: “We largely understand the risks today’s models present and are well equipped to manage them.” They call that claim false whether “we” refers to Anthropic or the world. In a follow-up reply, they acknowledge the contrast being drawn between current and future models, but argue that warning about potentially worse risks does not justify suggesting today’s risks are largely under control.
Anthropic’s confidence about today’s AI risks draws criticism
A commenter disputes a line they quote from Anthropic’s Opus 5.5 blog, arguing that potentially worse risks ahead don’t mean today’s risks are under control.
TLDR
A commenter quotes Anthropic’s Opus 5.5 blog as saying: “We largely understand the risks today’s models present and are well equipped to manage them.” They call that claim false whether “we” refers to Anthropic or the world. In a follow-up reply, they acknowledge the contrast being drawn between current and future models, but argue that warning about potentially worse risks does not justify suggesting today’s risks are largely under control.