/tags/llms/ Recent content in LLMs on Alexander Arvidsson Hugo -- gohugo.io en Tue, 28 Jul 2026 00:00:00 +0000 /posts/sycophant-reviewing-the-sycophant/ Tue, 28 Jul 2026 00:00:00 +0000 /posts/sycophant-reviewing-the-sycophant/ You built the pipeline everyone recommends. One model drafts, another reviews. Two sets of eyes. Except your reviewer is a language model, and language models can be argued out of a position by nothing more than a confident counter-argument. No payload to catch. No injection required. Just a wrong document in your retrieval set making a fluent case. The verdict flips, the badge says ‘reviewed,’ and everyone downstream stops looking. You added a laundering step. /posts/regex-good-argument/ Tue, 14 Jul 2026 00:00:00 +0000 /posts/regex-good-argument/ I spent the winter teaching people to defend against prompt injection. Layer your defenses, I said, and you can ship systems you trust. I still believe that. But I found an attack that walks through every layer I described, and it has no payload at all. It is just an argument. You push back on the model’s conclusion with confidence, and it folds. Every time. No guardrail fires because no guardrail was watching the conclusion itself. That changes things. /posts/not-disagreeing-enough/ Tue, 23 Jun 2026 00:00:00 +0000 /posts/not-disagreeing-enough/ Describe your chosen architecture to your AI assistant and ask what it thinks. Odds are, it’ll tell you the approach is sound. But a 2025 Stanford study found AI models affirm users 47% more than humans do, even when the user is clearly wrong. Worse: people who got sycophantic responses trusted the AI more and were more likely to return. The version that damaged their judgment was the one they liked best. This is Goodhart’s Law in your feedback loop. /posts/neurodiversity-and-llms-2/ Tue, 16 Jun 2026 00:00:00 +0000 /posts/neurodiversity-and-llms-2/ Your AI policy was written for someone who doesn’t work on your team.Your AI policy was written for someone who doesn’t work on your team. Probably for someone who doesn’t exist. Part 2 of this series moves from research to practice. The key variable isn’t cognitive profile — it’s domain expertise asymmetry. Where that gap is largest, the agreement machine runs without a check. This post covers where the real risk concentrates, why structured review consistently outperforms ‘does anyone see any problems?’, and what a policy that actually changes behaviour looks like. Design the workflow. Not the person. /posts/neurodiversity-and-llms/ Tue, 02 Jun 2026 00:00:00 +0000 /posts/neurodiversity-and-llms/ Your brain evolved to detect lions. Now it may fire every time you open ChatGPT. This post unpacks three reasons LLMs are not neutral tools — what they’re trained on, how RLHF creates systematic pressure toward validation, and what your neurobiology does with the result. Then it gets specific: the same sycophantic system creates meaningfully different failure modes depending on who is using it. Nobody designed this. Nobody fully planned for it. And most deployment practices still aren’t accounting for it. The research is early. The mechanisms are not. /posts/nothing-gets-deleted/ Tue, 14 Apr 2026 00:00:00 +0000 /posts/nothing-gets-deleted/ You can delete a post. You can unpublish an article. You can invoke GDPR and demand erasure. What you cannot do is remove something from an LLM. Once data dissolves into billions of model weights, there’s no row to delete, no file to erase. And it gets worse: the models now training on AI-generated outputs are degrading with each generation, narrowing toward a statistical echo of themselves. The right to be forgotten has no technical implementation path. None.