Paying Kids To Do Schoolwork
I think that the standard schooling system could be a lot better. This is for two main reasons: It’s slow. [1] It limits agency. [2] This isn’t to blame the people who work in…
Our Work is Low Skill Expression
crosspost from substack The amount of skill behind an outcome is often impossible to discern by scrutinizing the outcome itself, and our goal of making AI go well may be an…
Impressions at the Extremity of Civilization
Content note: this is part of a challenge of writing a blogpost per day for a week. Epistemic status: this is a series of vignettes written as-though diary entries. While…
Cada Tarde host says "these 15 weeks have not been victorious" in Iran
Anthropic Is Taking AI Welfare Seriously. I’m Not Sure It Knows What It’s Measuring.
Claude is a Constitutional AI, this means, in theory, that it operates from a set of principles as opposed to hard rule sets. This is achieved in a somewhat convoluted fashion…
A cheap specialist judge gets used by agents but fails to reduce alignment audit costs
TL;DR I gave AuditBench's investigator agents a lightweight (Gemma 2-2B) EM-toxicity-scorer (judge) as an additional audit tool, targeting a proof-of-concept for misalignment…
American Government Takes Down Claude Fable
No good policy gets announced shortly after 5pm eastern on a Friday. Here we go again. The Once And Future Fable The United States Department of Commerce, as per a letter from…
Not telling is lying
TL;DR: I don't have a clear guideline for when lying is right or wrong, but I have one against ontological obfuscation. The Behavioral Intentional Stance (BIS) A man meets a…
A simple argument for trying less hard
People often make arguments against “trying hard” (working very hard, pushing yourself to the brink, being intensely goal-directed, and so on) by pointing to the risks of burnout…
How might continual learning affect safety and alignment?
This is the third post in our sequence Implications of Continual Learning for LLM Agents . Summary We argue that continual learning (CL) has two major potential safety…
Rubio Signs Sports Diplomacy Deal With UFC Ahead of Weekend Fight at White House
Partnerships between the White House and UFC have raised a bevy of ethical concerns and legal questions.
How to Suffer Less
(Based on a talk I gave at LessOnline 2026 titled “How to get Enlightened” that had similar content but a different framing. I think this is a better framing for the Internet…
On Deadline: White House, Angelo Carusone discusses the slur about Michelle Obama at the UFC event and how rhetoric like it went from the right-wing media ecosystem to the White House lawn
Euro-Office 1.0 Arrives To Open-Source Infighting: 'Compatibility Is Not Sovereignty'
An anonymous reader quotes a report from ZDNet: If digital sovereignty is important to you, and it certainly is in the European Union (EU), then you'll be pleased to know that…
Coinbase Launches Tool To Let AI Agents Manage Trading and Payments
Coinbase has launched Coinbase for Agents, a tool that lets AI agents like ChatGPT or Claude execute crypto trades and manage payments on a user's behalf. "For example, customers…
ACLU Sues After Facial Recognition Falsely Identifies Florida Man As a Child Abductor
fjo3 shares a report from Reason: Police arrested a man in Florida for attempted child abduction in a town he had never visited, and the only evidence linking him to the crime was…