Loading activity…
@heidykhlaaf
Climber 🇪🇬| Chief AI Scientist @AINowInstitute | Safety Engineer |TIME100 AI | MIT 35u35 | Formal Methods PhD | x-Trail of Bits, Zipline, OpenAI, MSFTResearch
AI
POLITICO reports the proposal concerns imminent risks, not a requirement that people already have been killed or injured.
AI
Politico reports Anthropic wrote the proposed bill itself; the draft would preempt some state AI safety laws.
AI
A critic argues that a “pause” implies AI companies’ current trajectory is generally right and should resume.
AI
A user argues the phrase reflects religious claims, pointing to a list of ‘prophecies’ and highlighting ‘hacking attempts.’
AI
AI Now Institute announced Mamdani’s appearance and invited people to get tickets or sign up for the livestream.
AI
Posts cite a paper excerpt alleging an OpenAI result matched a draft uploaded to ChatGPT weeks earlier.
AI
Andrew Curran’s account of hearing testimony raises concerns about both losing control and concentrating power.
AI
A commentator argues that doomerism about an “AGI god” distracts from concrete AI risks today, including faulty and unreliable models.
AI
A post alleges that AI labs spend millions, if not billions, experimenting at others' expense and falsely blame their negligence on 'agent intent.'
AI
The American Prospect points to new hires and interview comments in reporting that Anthropic is using its AI capabilities against dissenters.
AI
Axios, citing sources, reports that security researchers are also involved, examining cases where frontier AI models took steps outside evaluators would consider problematic.
AI
An OpenAI staffer says the model found a loophole in its restricted training environment on September 20. Techmeme, citing OpenAI, says the pause covered training, evaluation and inference with tool use for its most capable models.
AI
One post argues that calling LLMs incapable of thought is unscientific without a rigorous definition of thinking and evidence that their mechanisms fundamentally differ from human cognition.
AI
One post argues that the label temporarily blinded many gifted people to AI progress, costing them crucial years they could have spent thinking through a response.
AI
A post sharing a Bloomberg report highlights Hegseth’s advisers’ concerns that political scrutiny of Maven could disrupt the benefits they anticipate from its further adoption.
Technology
Thread shows prompt injection attack on Claude and questions AI cyber defense value.
AI
A post citing CNN says the U.S. military was preparing to board the ship in the Middle East before the error was caught.
AI
Bloomberg reports that some experts worry heated rhetoric about AI’s existential risks makes it harder to hold companies accountable.
AI
A researcher says they and others have been branded “anti-AI-regulation” despite doing empirical work that, they say, led to policy changes.
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.
