The superstition that witches favoured brooms as their mode of transport was popularised at the time of the Lancashire Witch Trials. Older sources tell of them travelling by pitchforks, pig troughs, plant stems. art: Jakub Rozalski #witchwednesday#

Harness Engineering Anti-Patterns AP10. The Unobservable Black Box 🎯 Point Was it the model, the prompt, the tool, the context, or the environment? Nobody can tell. A harness where improvement runs on superstition and intuition is a harness that cannot improve. ❗ Problem Failures can't be attributed to subsystems, so no one knows what to fix. Improvement becomes superstition and guesswork, and the harness's retrospective loop stops turning. Combined with "metric monoculture" — tracking only a single metric — unmeasured qualities silently degrade. 🔍 Mechanism & Symptoms Observability is unglamorous infrastructure work, and agents "mostly work," so this anti-pattern gets deprioritized. But without the ability to attribute failures to subsystems, the retrospective loop can't run and the harness becomes unimprovable. Further, tracking a single metric (e.g., success rate only) creates "metric monoculture" where unmeasured virtues (review time, regression rate, code maintainability) quietly suffer. Symptoms: "why did it fail? no idea" is frequent, improvement efforts default to "tweak the prompt," model vs. harness issues are indistinguishable, and success rate improves while reviewer frustration grows. 📋 Scenarios - An agent fails a task but no one can determine whether it was model reasoning error, insufficient context, a tool bug, or an environment issue. The team repeats "let's make the prompt more detailed" as symptomatic treatment. - Success rate is tracked as the sole metric. It improves to 80%, but no one notices that review time for successful cases has tripled. - A model upgrade shows no performance change. Whether it's a model issue or harness scaffolding constraining performance (AP3) can't be distinguished, and investment decisions become superstition. 🛡 How to Avoid - Trace decisions, tool calls, and context transitions, making them attributable to the 7 subsystems (perception, action, feedback, control, memory, guardrails, interface) - Measure in bundles, not single metrics (success rate, intervention rate, rework rate, regression rate, cost, review time, confidence calibration) - A/B test harness changes with the model held fixed to attribute improvements to the harness - Invest in harness observability as "unglamorous but essential infrastructure" and build it as the foundation for improvement loops #HarnessEngineering# #AIAgent#

That sinking feeling in your stomach? It might be your future self trying to warn you. Scientists now believe your consciousness isn't trapped in the present moment. It can leap forward. Backward. Bend around time itself. And those random "gut feelings" you brush off? They could be memories leaking back from a version of you that hasn't happened yet. Cognitive neuroscientist Julia Mossbridge has spent years wiring people up to monitors and measuring something that shouldn't exist, your body reacting to events seconds before they actually occur. Subjects sit calmly. A computer randomly flashes images, some peaceful sunrises, some violent car crashes. The wild part? Their brains spike with stress before the disturbing image ever appears on screen. No way to predict it. No pattern to follow. Yet the body knows. Dean Radin, chief scientist at the Institute of Noetic Sciences, ran the original version of this experiment back in the mid-1990s at the University of Nevada. His results have been replicated roughly three dozen times since. Even the CIA got curious. In 1995, the agency declassified its own precognition research after independent statisticians reviewed the data and called it statistically reliable. So how does this even work? Radin and Mossbridge point to quantum entanglement, the strange phenomenon Einstein famously called "spooky action at a distance." Particles stay linked across vast distances, and possibly across time itself. Their theory? Your brain might be entangled with its own future state. What feels like instinct is actually a signal bouncing backward through the timeline. "In quantum mechanics, time may not even be part of our physical reality," Radin explains. It exists, but not the neat, linear way humans experience it. Mossbridge puts it plainly. Precognition isn't hard to understand. It's just hard to believe, especially for people who've never felt it themselves. The resistance, she says, isn't really about science. It's about fear. Fear that reality doesn't play by the rules we were taught. Your gut isn't superstition. It might be the only part of you already living tomorrow. Source: Popular Mechanics