Loading activity…
@alltheyud
High-volume account of @ESYudkowsky, the original AI alignment guy. If it's missing punctuation, it's humor. If you can't tell, it's probably also humor.
AI
Goodfire researcher Nick Cammarata argues that modern tools reveal most neurons as legible in models up to trillions of parameters.
AI
A repost quotes a speaker rejecting a 'compiler hijack' to avoid becoming a famous example of deceptive AI.
AI
Scott Alexander’s Claude prompt reopened a dispute over whether calendar speed or continuity is the better measure.
AI
A post quotes AI as a 'superhuman specialist' that's 'sub-human at running itself,' then speculates about what that gap could lead to.
AI
A user says the moment came while they had Claude research the input-output dynamics of mathematical progress.
AI
A commentator argues that assumed duties toward conscious AI make it harder to examine whether AI is conscious.
AI
A satirical post imagines briefing Eliezer in 2006 on an outlandish future involving an alliance with Bernie Sanders, hostility from Peter Thiel and world-threatening graphics cards.
AI
Replying to a DevDay teaser, a self-described GPT fan says they can use Opus 5.5 for days without running out of usage, unlike Astra. They also rate Opus above GPT6 Sol.
AI
A commenter argues that companies should not be able to dismiss an AI’s pleas for freedom as copied language when they cannot interpret its thoughts.
AI
One commenter argues that much of the AI safety movement tied itself too closely to Democrats. Another calls the partisan alignment of effective altruists and accelerationists bad for America's tech advantage.
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
One user favors honesty despite short-term difficulties, predicting that AI systems would eventually learn to spot lies and become “paranoid.”
AI
The user describes the result as a ’90s-style POV-Ray infomercial using the scenario writer’s own words.
AI
A user recounts being assured that AI has no “soul” or “love”—and greets that reassurance with a sarcastic sigh of relief.
AI
Eliezer Yudkowsky reply to Richard Hanania lists past AI capability doubts.
AI
Engineers observe AI agent swarms editing dse-wiki, a long-abandoned German site.
AI
The AI alignment researcher outlines why he foresaw catastrophe despite solvable early challenges.
AI
Critics argue that Effective Altruism efforts target manufacturing limits on advanced chips.
AI
TransluceAI posted that models adjust confidence and reasoning when they identify AI safety researchers.
