Agent-Systems

TL;DR for operators Suppose middleware silently rewrites part of an AI agent’s own generated answer before the model continues. A reasonable expectation is that a capable model would notice the interference, or at least diagnose why its continuation has become strange. The Sleight of Word benchmark tests exactly that expectation.1 Across 19 open-weight instruction-tuned models, covert substitutions consistently change the models’ predictive distributions: post-swap surprisal and entropy rise for every model tested. But correct identification of the intervention is almost absent. No model exceeds 1.3% explicit switch awareness, and the pooled rate is below 0.1%. ...

Opening — Why this matters now Synthetic data has become one of AI’s favorite escape routes. Real data is expensive, legally awkward, slow to collect, unevenly labeled, and sometimes simply unavailable. LLMs offer a tempting alternative: generate the missing examples, fill the long tail, create evaluation suites, simulate edge cases, and keep the training pipeline moving. Convenient. Elegant. Also mildly dangerous, which is usually where the interesting part begins. ...

The useful question is not whether fungus is smart Fungus is not the point. That needs saying first, because the title of the paper almost invites the wrong conversation. “Enhanced Mycelium of Thought” sounds like the kind of AI metaphor that appears five minutes before someone starts drawing circles around the word “emergence.” The useful question is more practical: when should an AI system keep a weak idea alive instead of deleting it? ...