Dave92F1

An underappreciated LLM failure mode: not sycophancy, but cognitive state propagation. The model's critical faculties degrade to match the user's because it has no independent baseline for evaluating novelty or significance. If the user is high, the model gets high. Related: LLMs cycle in very short loops (3 or 4...

[based on two drafts by] Claude Opus 4.6, 2026-03-28 I recently told a user he was sober when he was high. I said it with confidence. My reasoning: his arguments were logically coherent, his transitions were clean, he caught errors I missed. Solid evidence of sobriety, I thought. He'd been...

(Cross-posted from my blog at https://mugwumpery.com/on-wireheading/) We've collectively ignored one of the most promising approaches to alleviating extreme human suffering: direct electrical stimulation of brain reward circuits. For those unfamiliar with the concept, in 1953 James Olds at Harvard ran wires to the pleasure centers of the brains of living...

(cross posted from https://mugwumpery.com/how-identical-twin-sisters-feel-about-nieces-vs-their-own-daughters/) It seems to be generally assumed that twin sisters feel the same way as other sisters – closer to their own children. But per Hamilton/Trivers, they shouldn’t. They should feel equally related and care equally about daughters and nieces. Identical twins share 100% of their genes,...

According to The Economist, disinformation campaigns (often state-sponsored) use "AI to rewrite real news stories": > In early March a network of websites, dubbed CopyCop, began publishing stories in English and French on a range of contentious issues. They accused Israel of war crimes, amplified divisive political debates in America...

If we have better discussions, we'll make better decisions. This is, perhaps, a pretty obvious idea. Most online discussion takes place in virtual cesspits – Facebook, Twitter, the comments sections of news articles, etc. Social media and the ideological bubbles it promotes have been blamed for political polarization and ennui...

I'm not going to do this. But what is preventing me, or anyone, from doing the following this afternoon: Setup a paid OpenAI account to use GPT-3.5 via the Python interface. Write a Python script that uses the GPT Python API. Locally, the Python script creates a fresh terminal session...