via OpenAI Blog [7] — OpenAI introduces CoT-Control and finds reasoning models struggle to control their chains of thought, reinforcing monitorability as an AI safety safeguard.

via Substack Zvi — I’ve been trying to find a slot for this one for a while.

via LessWrong AI [3] — What's the best case scenario regarding OpenAI's contract w/ the Department of War (DoW)?We have access to the full contractIt's airtightOAI's engineers are on top of things in case the DoW breaks the contractThere's actual teeth for violationsBut even then, the DoW can simply switch vendors. Use Ge

via LessWrong AI — Click here if you just want to see the Database I made of all[1] AI safety papers written since 2020 and not read the methodology. To some extent the core idea here is to encode as much info from these papers into something small enough that an AI with a specific problem in mind can take in all

via LessWrong AI [9] — tl;dr We’re incubating an academic journal for AI alignment: rapid peer-review of foundational Alignment research that the current publication ecosystem underserves. Key bets: paid attributed review, reviewer-written synthesis abstracts, and targeted automation. Contact us if…

via Substack Zvi [2] — The attempt on Friday by Secretary of War Pete Hegsted to label Anthropic as a supply chain risk and commit corporate murder had a variety of motivations.

via The Verge AI [2] — Anthropic is making it easier to switch to its Claude AI from other chatbots with an update that brings Claude's memory feature to users on the free plan, along with a new prompt and dedicated tool for importing data from other chatbots. These upgrades could allow users who have been using rivals li

via LessWrong AI [2] — What a weekend. Two new wars in Asia don't qualify as top news. My first reaction to Hegseth's conflict with Anthropic was along the lines of: I expected an attempt at quasi-nationalization of AI, but not this soon. And I expected it to look like it was managed by national security professionals. He

via MIT Technology Review [4] — On February 28, OpenAI announced it had reached a deal that will allow the US military to use its technologies in classified settings. CEO Sam Altman said the negotiations, which the company began pursuing only after the Pentagon’s public reprimand of Anthropic, were “definitely rushed.” In its anno

via The Verge AI [4] — On Friday evening, amidst fallout from a standoff between the Department of Defense and Anthropic, OpenAI CEO Sam Altman announced that his own company had successfully negotiated new terms with the Pentagon. The US government had just moved to blacklist Anthropic for standing firm on two red lines

via Substack Zvi [2] — This is the long version of what happened so far.

via MIT Technology Review [4] — Pull the plug! Pull the plug! Stop the slop! Stop the slop! For a few hours this Saturday, February 28, I watched as a couple hundred anti-AI protesters marched through London’s King’s Cross tech hub, home to the UK headquarters of OpenAI, Meta and Google DeepMind, chanting slogans and waving signs.

via Alignment Forum [5] — Authors: Gerson Kroiz*, Aditya Singh*, Senthooran Rajamanoharan, Neel NandaGerson and Aditya are co-first authors. This work was conducted during MATS 9.0 and was advised by Senthooran Rajamanoharan and Neel Nanda.TL;DRUnderstanding why a model took an action is a key question in AI Safety. It is a

via ArXiv cs.AI [6] — Large language model (LLM) agents typically rely on reactive decision-making paradigms such as ReAct, selecting actions conditioned on growing execution histories. While effective for short tasks, these approaches often lead to redundant tool usage, un

via ArXiv cs.AI [8] — Everyone from AI executives and researchers to doomsayers, politicians, and activists is talking about Artificial General Intelligence (AGI). Yet, they often don't seem to agree on its exact definition. One common definition of AGI is an AI that can do

via ArXiv cs.AI [3] — Synthesizing high-quality training data is crucial for enhancing domain models' reasoning abilities. Existing methods face limitations in long-tail knowledge coverage, effectiveness verification, and interpretability. Knowledge-graph-based approaches s

via LessWrong AI [5] — TL;DRYour base LLM has no examples of superintelligent AI in its training data. When you RL it into superintelligence, it will have to extrapolate to how a superintelligent Claude would behave. The LLM’s extrapolation may not converge optimizing for what humanity would, on…

via Alignment Forum — Also available in markdown at theMultiplicity.ai/blog/schelling-goodness. This post explores a notion I'll call Schelling goodness. Claims of Schelling goodness are not first-order moral verdicts like "X is good" or "X is bad." They are claims about a class of hypothetical coordination games in the

via Substack Zvi [2] — The Department of War gave Anthropic until 5:01pm on Friday the 27th to either give the Pentagon ‘unfettered access’ to Claude for ‘all lawful uses,’ or else.

via LessWrong AI [3] — TLDRThis is a post announcing a lot of new ARENA material I've been working on for a while, which is now available for study here (currently on the alignment-science branch, but planned to be merged into main this Sunday).There's a set of exercises (each one contains about 1-2 days of material) on t