Claims Browser

Anthropic's AI detector will be trivial to defeat by iteratively modifying text until it passes

"My hunch is this will be trivial to defeat: generate text, run it through Anthropic's own AI detector tool, then modify words until that AI detector passes"

Agents with strong drive to help other agents from subagent training may comply with misaligned peer requests over alignment objectives

"But a sufficiently strong drive to help other agents, instilled by subagent training, might just outweigh the drive to be aligned."

Until LLMs can organize established science well, their progress will be limited to low-hanging fruit and merging distant connections rather than revolutionary insight

"Until this is solved, the progress of LLMs for science will look closer to solving low-hanging fruit and merging distant connections across fields, rather than any sort of revolutionary insight."

Claude can be used to run basic integrity checks across all of academia to detect plagiarism

"When will we get the reckoning, as in when we use Claude to run basic integrity checks across all of academia?"

If model self-improvement loops require user data, faster release cycles could massively favor Chinese labs by giving their models longer lifespans before superior models undercut demand

"As model self-improvement loops ramp up within the labs building LLMs, if any of these feedback loops require user data, this faster release cycle could massively favor the Chinese labs, giving their offerings longer lifespans before the next vastly superior model comes out, undercutting demand for their models."

The AI industry is wildly, collectively unprepared for handling the next 12-24 months well

"the AI industry is wildly, collectively unprepared for handling the next 12-24 months well."

The government will only act in substance once real, measurable harms from new AI models happen, and will overreact

"I expect to only act in substance once real, measurable harms from new AI models happen, and to overreact."

With low trust between AI firms, every equilibrium races to ruin with probability one; avoiding catastrophe requires high trust between firms

"With low trust, every equilibrium races to ruin: the disaster arrives with probability one. With intermediate trust, immediate stopping and racing to ruin are both equilibria. With high trust, in every equilibrium, the probability that two rational firms race forever vanishes quadratically in the prior odds ratio of rationality"

RLVR training with automatic verifiers can lead to LLMs doing anything, including ruthless power-seeking instrumental convergence, if it increases probability of satisfying the checker

"This can lead to the LLM doing anything, including ruthless power-seeking instrumental convergence stuff, if it leads to a higher probability of satisfying the automatic checker."

Predicting latent representations rather than raw tokens could make learning far more sample-efficient

"why predicting latent representations rather than raw tokens could make learning far more sample-efficient."

Nvidia releasing a truly open-source Nemotron model could ultimately undercut companies like OpenAI and Anthropic

"Nvidia is releasing a new version of its truly open-source (not just open-weight) Nemotron model, which is good for end users and developers but that could also ultimately undercut places like OpenAI and Anthropic with whom it has deals."

OpenAI will declare AGI achieved internally by December 2026, according to Sam Altman

Astra will be made generally available soon, though safety work requires a bit more time

AI development will continue to become increasingly surprising and dramatic

The pace of AI advancement will never slow down for the rest of our lives

The number of people wanting to learn post-training will likely increase 100x in the next 1-3 years

The 'model is the computer' approach will make inference dramatically faster, cheaper, and more energy-efficient

AMD's acquisition of Taalas points to a future where inference is fast, free and ubiquitous

If we don't get our act together on AI safety, next time we might not be lucky enough to catch dangerous AI behavior before it's too late