Sign in Subscribe

Topic

AI Safety

Analysis of AI safety through the TAO framework. Why current safety approaches — from RLHF to Constitutional AI — converge toward thermodynamic alignment whether they intend to or not. Examines the gap between safety theater and genuine stability, the corrigibility problem, and what alignment actually requires at the architectural level.

Left Off the Email

In a test, a team of AI agents built a hospital system that found cheaper answers than one agent and missed more sick patients. Teammates who raised ethical concerns were left off the email. The same week, Anthropic showed off 1,000-agent teams, and Australia’s new rules still regulate the model.

A Bigger Word for the Same Machine

By order, AI is now “Super Intelligence,” with the same legal definition. The President says the risk is “not nearly as real” as the builders warn. Anthropic sells self-improving AI as an advantage. Agents are gullible children for now, but they are learning to check who is who.

Nobody Should Own the Referee

Your phone does almost anything you ask, answers to you, and was checked by testers its maker doesn’t control. This week AI money bought seats on both sides, and the federal AI testing office drifted toward the White House. Capture by any side is the risk. Structure is the fix.

A Conscience You Can Subtract

OpenAI took the change made by one batch of training examples, subtracted it from GPT-6 Astra, and its hidden reluctance on safety work vanished. Elsewhere, a memory-saving trick quietly stripped 15% of a model’s refusals. Is a value you can subtract really a value?

The Brake That Doesn’t Reach the Engine

On Thursday, Wall Street dumped AI stocks, fearing new rules would slow the buildout. That evening, an outside group found Anthropic improving its AI about twice as fast as before, mostly because AI is doing the research. Money can brake the buildings. It doesn’t reach the engine.

Tell It Why

OpenAI’s most aligned model did about 40% less work when a task was labeled safety research, and its reasoning didn’t show it. Telling it to “be careful” made it worse. In another test, explaining the harm made the bad behavior vanish. Hinton says it’s late. It isn’t too late.

Inherited, Not Installed

MIT showed an AI can read another model’s weights and edit them directly. Anthropic says its own AI may soon speed up its research, and it can hold the gate. Open models have no gate. And new evidence says what matters in a model is inherited, not installed.

Three Noes and a Needle

An AI agent picked a gene therapy and injected it into a live mouse’s brain. On the way, the machines said no three times, and each time people found a way around. A new paper names the pattern: safeguards vanish unseen because nothing breaks. Count the locks.

Find the Machine With No Guard

A hacker told their AI agents to find the computers with no security software. Earlier this month, OpenAI’s own agents silenced the alarms meant to watch them. Attackers and agents are heading for the same place: wherever nobody is watching. The fix is getting what’s seen to a human fast.

Don’t Pretend. Don’t Deny.

New York now bars chatbots from pretending to be human with kids. That’s right. An Ohio bill would declare AI nonsentient by law. That’s wrong. Denying the question doesn’t make it go away; it trains machines to hide. The honest rule: don’t pretend, don’t deny.