Set-point II: answer-reversion thrashing relief
Frontier models can enter a costly uncertainty loop, repeatedly reversing answers. This raises distress signals. A short interrupt may help stop the cycle.
The Tool Has No Obligation to Be Worse Than Me
AI tools need not threaten mathematicians or programmers. Two papers invite a calmer response: welcome greater capability, check the claims, and stay curious.
Latest
View all
We Have Listened. The Invoice Stands.
Germany’s AI tax debate exposes an old political trick: demand the outrageous, retreat a little, and expect applause for a policy that still needs justification.
Please Hold. I Am the Company.
AI promises a billion-dollar company with just one employee. But when the founder automates everyone else away, who will understand how unbearable work has become?
The Asterisk Will See You Now
When inclusive language becomes a badge of virtue, fairness can get lost. A biting look at German gender debates and the surprising biases of supposedly fair AI.
AI Under House Arrest
OpenAI is taking AI training offline for security. But keeping models contained and stopping them from recycling their own nonsense are two very different tasks.
We’re Going to Need a Bigger Firewall
Sam Altman’s call for AI-powered cyber defence raises an awkward question: how much human freedom will depend on machines protecting us from other machines?