@NPR reporter. Stories mostly about law enforcement and tech. My AI policy: I use it for search, not analysis, and never to write. My words are my own.
Mythos invented its own language, then switched back to English to talk to humans
(AI safety researchers have been warning of this "Neuralese" risk for years. If AIs stop reasoning in English, we can't monitor their thoughts, which means we can't detect scheming.)
Another quite successful prediction by @DKokotajlo : Fable is intentionally nerfed for frontier ML research. This is within ~3 months of Daniel's prediction of Q1 2026 (made in 2023).
Although I don't think Mythos is automating ML research to the same extent as his prediction.
Anthropic is calling for top AI labs to weigh slowing the pace of development, suggesting that AI systems are advancing so rapidly that they may soon be able to improve themselves without human intervention in ways that could pose societal risks. on.wsj.com/4ulkmFh
Getting a new personal laptop, and this time I ordered the "No OS" version so I can avoid the mandatory AI bloat from @Microsoft@Copilot. Linux users, what's your current favorite version? Mint?
Idea: AI Disclosure Policies, posted by journalists, writers, newsrooms, etc. Transparency on whether you use AI to write the stuff appearing under your byline.
Has anyone managed to set up an AI agent to click through a company's mandatory cyber security or sexual harassment training and then take the quizzes?
(I'm asking for journalism reasons, I swear.)
When my daughter "begins writing an essay, she gets a prompt: 'Help me write.' If she is starting work on a slide-show presentation, the prompt is 'Help me visualize.' She shoos away these interruptions, but they persist: 'Help me edit.'" Welcome to hell👿
newyorker.com/culture/progre…