Loading activity…
@katjagrace
Thinking about AI destroying the world at http://aiimpacts.org and everything at http://worldspiritsockpuppet.substack.com. DM or email for media requests.
AI
A post likens compute to uranium and describes it as fragile and hard to build.
AI
A researcher says existing models could keep running at scale, limiting the immediate economic hit of a pause.
AI
Katja Grace argues that people are usually mistaken when they excuse racing toward AI by pointing to its upsides.
AI
A user asks the question without explaining what “it” refers to in the post’s text.
AI
One poster says they aren’t a ‘doomer’ but share that camp’s concern about AI and support for regulating AI companies.
AI
A debate participant thinks AI takeover poses the greater risk, but finds human takeover via AI troublingly plausible.
AI
A debate participant believes AI takeover is the greater risk, but also finds human takeover via AI troublingly plausible.
AI
The event site lists the call as open to everyone at 7:30 p.m. ET. A speaker promoting it says he’ll open by discussing the threats he sees and why he considers the issue urgent.
AI
One post argues that plans to replace US workers with AI seem to escape the notice of some people who would be outraged by replacements from another country.
AI
One post argues that AI catastrophe estimates depend on policy, while the chance of decisive government action depends partly on those estimates. A separate argument connects optimism about AI’s potential with caution about how to pursue it.
AI
A user highlights an angle beyond capability and motivation: how AI might kill people even if it didn't want to.
AI
FOX4 reports that President Donald Trump outlined the terminology change at the United Nations General Assembly, saying it would apply to all U.S. documents.
AI
Whitesides wants companies, independent experts and policymakers to establish clear red lines, arguing that AI is advancing faster than it can be understood and regulated.
AI
The post imagines a single AI-fueled global government combining what its author considers the worst regulatory and social policies of the past 20 years.
AI
Amodei says Anthropic will give third-party evaluators permanent, employee-level access to its systems to verify adherence to safety measures, report incidents and assess models’ alignment during training.
AI
Another post predicts AI will become “our interface to the world” and accuses people seeking sweeping controls out of fear of catastrophe of working to create a surveillance state.
AI
CBS Mornings reports that current and former Anthropic employees warned AI could kill all humans. Critics worry exaggerated forecasts could undermine trust, while a defender of numerical forecasts argues they convey the seriousness of the concern.
AI
The linked listing identifies the episode as “We Design the AI That Kills Us All,” from Are We Doomed? Grace describes it as a half-hour NPR discussion.
