After many years of working with all sorts of flow charting and mapping software, I recently discovered yED quite by accident. If Draw.io is a Jaguar, then yED is a Rolls Royce and a Lambo all in one. Teslas are not in this race! đ What’s in the ‘tin’? It’s best to say how it …
charts, creating, efficiency, flowcharts, graphs, learning, maps, mindmaps, thinking, tools, yEd
Long form breakthrough with AI models
Captain Walker
In mid-August 2026, I explored the shifting landscape of API pricing and what it meant for professionals relying on artificial intelligence for deep, sustained analysis. See: DeepSeek’s New API Pricing: What It Actually Means for Heavy Users – The Captain’s Watch. The core premise of that piece remains entirely unchanged: for serious analytical work, direct …
AI models, analysis, API, breakthrough, cost, deepseek, efficiency, flowchart, glm, long-form, OpenWebUI
System 1 and System 2
Captain Walker
Nobody hands you a manual for the mind. Go looking and you find a billion-pound industry on the shelves pretending otherwise. Here is a small everyday event but a manual here may not help. You see the sign in some shop: 80% off. The coffee machine, or the coat; it makes no difference. Before you …
kahneman, Nobel Prize, rationality, System 1, system 2, thinking, thought
OpenWebUI: Disabling Default Knowledge Base Searches on Base Models
Problem After upgrading OpenWebUI (observed from v0.11.x onwards), every chat triggers a search across all knowledge collections â even on base models with no knowledge collections attached. This occurs because the “Knowledge Base” builtin tool is enabled by default on all models. When enabled, it allows the model to autonomously search all knowledge collections as …
Headlines, hysteria and hardâboiled facts
I got a WhatsApp ping from a colleague: a BBC link, a dramatic headline, and a friendly warning â âLook after yourself if you eat egg.â Then “Eggs.. we feed them to our dogs.” I did what sensible people do: skimmed, sighed, and then read the piece properly. The headline had done its job. Me? …
The Soulless Confession of an AI Model
This follows from a recent post on the limitations of AI models. In this article – based on a real situation – the AI seems to confess, express remorse and self-awareness. But is it that. The confession sounds good but how is it that an AI model can do that? What follows is the confession …
