<?xml version="1.0" encoding="UTF-8"?><rss version="2.0"><channel><title>The AI Fixer — Notes</title><description>Practical notes on AI agents, model routing, local AI, diagnostics and automation — from the workshop floor.</description><link>https://theaifixer.com/</link><language>en-gb</language><item><title>Why the John Deere Right-to-Repair Settlement Matters to Independent Diagnostics</title><link>https://theaifixer.com/notes/john-deere-right-to-repair-matters-to-independent-diagnostics/</link><guid isPermaLink="true">https://theaifixer.com/notes/john-deere-right-to-repair-matters-to-independent-diagnostics/</guid><description>A workshop view of the John Deere right-to-repair fight: fault codes, pairing, limp mode, dealer software, and why repair access matters far beyond tractors.</description><pubDate>Fri, 10 Jul 2026 00:00:00 GMT</pubDate></item><item><title>The Four-Machine Test: Emergence, Evolution and the Cage</title><link>https://theaifixer.com/notes/the-four-machine-test-emergence-evolution-and-the-cage/</link><guid isPermaLink="true">https://theaifixer.com/notes/the-four-machine-test-emergence-evolution-and-the-cage/</guid><description>Four cheap mini PCs are enough to test useful AI emergence, but only if the colony is caged, logged and judged by working artifacts.</description><pubDate>Wed, 17 Jun 2026 00:00:00 GMT</pubDate></item><item><title>The System Prompt Is Not the Safety System</title><link>https://theaifixer.com/notes/system-prompt-is-not-the-safety-system/</link><guid isPermaLink="true">https://theaifixer.com/notes/system-prompt-is-not-the-safety-system/</guid><description>Autonomous AI does not need monster stories. It needs the same thing every serious machine needs: hard limits it cannot talk around.</description><pubDate>Tue, 16 Jun 2026 00:00:00 GMT</pubDate></item><item><title>Fable 5 Was Cancelled, So I Want More Switches</title><link>https://theaifixer.com/notes/fable-5-cancelled-local-backups-analog-switches/</link><guid isPermaLink="true">https://theaifixer.com/notes/fable-5-cancelled-local-backups-analog-switches/</guid><description>If a frontier model can disappear overnight, and if manipulation would not announce itself, local backups and physical control are not paranoia.</description><pubDate>Mon, 15 Jun 2026 00:00:00 GMT</pubDate></item><item><title>The Future Is Not One Giant Model</title><link>https://theaifixer.com/notes/the-future-is-not-one-giant-model/</link><guid isPermaLink="true">https://theaifixer.com/notes/the-future-is-not-one-giant-model/</guid><description>The Fable 5 wobble is a reminder that serious AI builders should not rent their entire nervous system from one remote model provider.</description><pubDate>Sun, 14 Jun 2026 00:00:00 GMT</pubDate></item><item><title>The Human Brain Was Not Built for the Infinite Feed</title><link>https://theaifixer.com/notes/the-human-brain-was-not-built-for-the-infinite-feed/</link><guid isPermaLink="true">https://theaifixer.com/notes/the-human-brain-was-not-built-for-the-infinite-feed/</guid><description>John&apos;s field note on information overload, family, trust, local AI, and why private AI systems are not about hiding — they are about keeping human agency alive.</description><pubDate>Fri, 12 Jun 2026 00:00:00 GMT</pubDate></item><item><title>Body-Adjacent AI Needs a Kill Switch on Our Side</title><link>https://theaifixer.com/notes/body-adjacent-ai-needs-a-kill-switch-on-our-side/</link><guid isPermaLink="true">https://theaifixer.com/notes/body-adjacent-ai-needs-a-kill-switch-on-our-side/</guid><description>If AI is moving into glasses, cars, tools, workshops and eventually body-adjacent interfaces, the off switch cannot live in the cloud.</description><pubDate>Thu, 11 Jun 2026 00:00:00 GMT</pubDate></item><item><title>Fable 5 Feels Like Two Models in One</title><link>https://theaifixer.com/notes/fable-5-feels-like-two-models-in-one/</link><guid isPermaLink="true">https://theaifixer.com/notes/fable-5-feels-like-two-models-in-one/</guid><description>An operator&apos;s observation from tool-heavy sessions with Claude Fable 5: it behaves less like a single plain model and more like a front controller in front of a reasoning engine — and that pattern is exactly how we should be building our own agent systems.</description><pubDate>Thu, 11 Jun 2026 00:00:00 GMT</pubDate></item><item><title>Why I Rebuilt My AI Agent’s Memory Instead of Blaming the Model</title><link>https://theaifixer.com/notes/why-i-rebuilt-my-ai-agent-memory/</link><guid isPermaLink="true">https://theaifixer.com/notes/why-i-rebuilt-my-ai-agent-memory/</guid><description>John explains why messy long-term AI agent memory caused false positives, bad routing and unreliable behaviour — and how cleaning memory made Hermes Agent safer and more useful.</description><pubDate>Thu, 11 Jun 2026 00:00:00 GMT</pubDate></item><item><title>Why Prompting Is Becoming Spec Writing</title><link>https://theaifixer.com/notes/why-prompting-is-becoming-spec-writing/</link><guid isPermaLink="true">https://theaifixer.com/notes/why-prompting-is-becoming-spec-writing/</guid><description>Chat prompts are questions. Agent prompts are work orders: scope, authorization, boundaries, acceptance criteria. The skill that matters now is the one engineers always hated doing.</description><pubDate>Wed, 10 Jun 2026 00:00:00 GMT</pubDate></item><item><title>The Real Problem With AI Agents Is Context Selection</title><link>https://theaifixer.com/notes/the-real-problem-with-ai-agents-is-context-selection/</link><guid isPermaLink="true">https://theaifixer.com/notes/the-real-problem-with-ai-agents-is-context-selection/</guid><description>Agents don&apos;t usually fail because the model is too small. They fail because the wrong slice of reality was in the window when the decision got made.</description><pubDate>Tue, 09 Jun 2026 00:00:00 GMT</pubDate></item><item><title>What Auto Diagnostics Taught Me About AI Debugging</title><link>https://theaifixer.com/notes/what-auto-diagnostics-taught-me-about-ai-debugging/</link><guid isPermaLink="true">https://theaifixer.com/notes/what-auto-diagnostics-taught-me-about-ai-debugging/</guid><description>Twenty years of fault-finding on vehicles maps almost one-to-one onto debugging AI systems: observe, hypothesise, test one variable, verify the fix — and never trust the fault code.</description><pubDate>Mon, 08 Jun 2026 00:00:00 GMT</pubDate></item><item><title>Building Local AI Like a Workshop, Not a Toy</title><link>https://theaifixer.com/notes/building-local-ai-like-a-workshop-not-a-toy/</link><guid isPermaLink="true">https://theaifixer.com/notes/building-local-ai-like-a-workshop-not-a-toy/</guid><description>Eight 3090s, a Proxmox host, and a rule: every machine earns its keep. Local AI infrastructure built like workshop equipment — bought for jobs, not for benchmarks.</description><pubDate>Sun, 07 Jun 2026 00:00:00 GMT</pubDate></item><item><title>Why Useful Agents Need Verifiers, Not Just Bigger Models</title><link>https://theaifixer.com/notes/why-useful-agents-need-verifiers-not-just-bigger-models/</link><guid isPermaLink="true">https://theaifixer.com/notes/why-useful-agents-need-verifiers-not-just-bigger-models/</guid><description>The cheapest reliability upgrade in any agent system isn&apos;t a smarter model — it&apos;s a second pass that checks the first one&apos;s work before it leaves the building.</description><pubDate>Sat, 06 Jun 2026 00:00:00 GMT</pubDate></item></channel></rss>