I've learned there's a BlueSky account using the handle @seaspinner. That is NOT me - I have no idea who they are - and to avoid confusion I do want people here, especially my followers, to know that isn't my account. I'll let you know if/when I get an account there.

🚨 Holy shit… Stanford just published the most uncomfortable paper on LLM reasoning I’ve read in a long time. This isn’t a flashy new model or a leaderboard win. It’s a systematic teardown of how and why large language models keep failing at reasoning even when benchmarks say they’re doing great. The paper does one very smart thing upfront: it introduces a clean taxonomy instead of more anecdotes. The authors split reasoning into non-embodied and embodied. Non-embodied reasoning is what most benchmarks test and it’s further divided into informal reasoning (intuition, social judgment, commonsense heuristics) and formal reasoning (logic, math, code, symbolic manipulation). Embodied reasoning is where models must reason about the physical world, space, causality, and action under real constraints. Across all three, the same failure patterns keep showing up. > First are fundamental failures baked into current architectures. Models generate answers that look coherent but collapse under light logical pressure. They shortcut, pattern-match, or hallucinate steps instead of executing a consistent reasoning process. > Second are application-specific failures. A model that looks strong on math benchmarks can quietly fall apart in scientific reasoning, planning, or multi-step decision making. Performance does not transfer nearly as well as leaderboards imply. > Third are robustness failures. Tiny changes in wording, ordering, or context can flip an answer entirely. The reasoning wasn’t stable to begin with; it just happened to work for that phrasing. One of the most disturbing findings is how often models produce unfaithful reasoning. They give the correct final answer while providing explanations that are logically wrong, incomplete, or fabricated. This is worse than being wrong, because it trains users to trust explanations that don’t correspond to the actual decision process. Embodied reasoning is where things really fall apart. LLMs systematically fail at physical commonsense, spatial reasoning, and basic physics because they have no grounded experience. Even in text-only settings, as soon as a task implicitly depends on real-world dynamics, failures become predictable and repeatable. The authors don’t just criticize. They outline mitigation paths: inference-time scaling, analogical memory, external verification, and evaluations that deliberately inject known failure cases instead of optimizing for leaderboard performance. But they’re very clear that none of these are silver bullets yet. The takeaway isn’t that LLMs can’t reason. It’s more uncomfortable than that. LLMs reason just enough to sound convincing, but not enough to be reliable. And unless we start measuring how models fail not just how often they succeed we’ll keep deploying systems that pass benchmarks, fail silently in production, and explain themselves with total confidence while doing the wrong thing. That’s the real warning shot in this paper. Paper: Large Language Model Reasoning Failures

America wasn’t born in a moment. Before it was a nation, it was an idea—and in 1776, nothing about it was inevitable. ‘On This Day…1776’ is a new TIME Studios series executive produced by @DarrenAronofsky that tells the story of the year that shaped a nation. Released in short weekly episodes around the dates events occurred. Premiering on TIME’s YouTube channel. Watch the series: time-magazine.visitlink.me/I…

New article on AI and fascism link.springer.com/article/10…

We used to open-source libraries in order help other humans. But now I don't care if Anthropic saves 1,000 tokens thanks to my library. The only way I see this working is some sort of marketplace where developers offer AI-optimized libraries and the clankers choose to pay when they predict it'll save them tokens.

OpenAI allows users to directly adjust ChatGPT’s enthusiasm level techcrunch.com/2025/12/20/op…

Pool report from @joeygarrison — new plaques installed outside the Oval Office that mock past Democrat presidents. The one that names Barack Hussein Obama says he presided over the "highly ineffective 'Unaffordable' Care Act," among other critiques.

The president’s TruthSocial post about the Reiners’ horrific deaths is quite a contrast from how Rob Reiner reacted to the murder of Charlie Kirk.

This is both disgusting and expected. The Scumbag-in-Chief