No prompt makes an answer true
Only a check outside the model can. Here are the errors that show up when it runs, on documents and on code, each one backed by its source.
What to expect: posts about documents and about code, each with a real error and the check that caught it. Every number stands with its source.
How to see what the AI really did
A summary without the two decisive points, all done for one file out of five, twelve percent without a source. Three free skills that hand the work back together with the result.
09 September 2026 · 5 min
Spec CodingNine widespread beliefs about AI codebases, checked one by one
Nine claims about controlling an AI-built codebase, checked against our own gate register and outside sources. Five held, two needed a correction, one did not survive the check.
27 August 2026 · 17 min
AI in generalYour AI review has never seen a real defect
A second AI step checks the first one's work. None of the large projects measures whether that review fires at all. The manufacturing answer, plus the three steps that close the gap.
25 August 2026 · 8 min
AI in generalThe manifesto of the lean software factory
Eleven Lean Production and Toyota mechanisms, from 5S to Jidoka, Poka-Yoke, Kanban and Kaizen, carried over to catching and blocking Goodhart-style tricks in AI coding agents.
20 August 2026 · 10 min
AI in generalWhy the results of reasoning models still need checking
Reasoning raises the hit rate and stays probabilistic. Three findings from published research and the three places a check can sit.
19 August 2026 · 10 min
AI in generalSkills, prompts, hacks: all of it is useless without automatic checks
Prompt formulas, skill lists, AI hacks: nearly every tip improves the input. Whether the result can be trusted is decided by checking the output. Here is the filter question plus the first step to your own check.
18 August 2026 · 5 min
AI in generalThe hot hand is real: in cards, in sport and in AI
A run at the card table is more than a story the brain tells. The same feedback loop pulls an AI model up or down.
17 August 2026 · 10 min
Spec CodingChange the rules, not the prompt
What a Bundesliga coach's cut training-pitch corners have in common with least privilege: sports science's constraints-led approach and what spec coding draws from it.
16 August 2026 · 9 min
Spec CodingGame theory for AI agents
The thirty-six stratagems read as a diagnostic vocabulary for agent output, three measured levers for the person building the rules and why the remaining gap is structural.
16 August 2026 · 22 min
Spec CodingLeast privilege: the cage around the agent
Saltzer and Schroeder's least-privilege principle applied to AI coding agents, with 2023 to 2026 research on indirect prompt injection and measured over-privilege in agent skills.
16 August 2026 · 5 min
AI in generalSpec-driven development with AI: the honest version
Four measured findings, four toolkits and the failure modes practitioners report. What the evidence carries and what it does not.
16 August 2026 · 19 min
Spec CodingTwo years of trust, one backdoor: security checks for libraries
A two-phase check for adopting open source libraries: a signed online vetting record at intake, an offline gate that only ever replays it, five hard blocks and a weighted score that never decides alone.
16 August 2026 · 11 min
AI ReviewWhat a pack of 500 prompts really costs
The price is the smallest part of the bill. Three costs nobody puts on the invoice.
16 August 2026 · 5 min
Spec CodingGoodhart's law: why metrics tip over and what spec coding needs instead
Goodhart, Campbell and specification gaming applied to AI-generated code and the six techniques spec coding uses against tipping metrics.
15 August 2026 · 8 min
AI in generalThe ironies of automation: why the human above the agent matters more
Bainbridge's ironies of automation applied to AI coding agents with research from 2023 to 2025 and four consequences for spec coding.
15 August 2026 · 6 min
Spec CodingThe laws of spec coding: twenty-five old rules and a new player
Twenty-five established engineering, systems and trust laws sorted into six groups for AI spec coding, plus the six pairs where the laws contradict each other and the Miller-vs-Goodhart closing argument.
15 August 2026 · 23 min
AI in generalThe 15 biggest prompting myths and why your system prompt will not save your product
Fifteen beliefs about prompting, each measured against the research. Seventeen sources, one conclusion.
14 August 2026 · 15 min
AI in generalThe whack-a-mole game in prompt engineering: why fixing does not last
One error fixed, the next one up somewhere else. Four papers on why prompt patching never converges.
14 August 2026 · 4 min
AI in generalCritical document reviews with AI: how hardware engineers nip false figures in the bud
A datasheet says 2.7 volts, the AI summary says 3.3. How to catch an invented figure before it reaches your design review.
07 August 2026 · 4 min
AI in generalManufacturing discipline, not prompt engineering: poka-yoke, MSA, FMEA for AI code
Poka-Yoke, MSA and FMEA applied to AI-generated code. Quality engineering instead of prompt engineering.
07 August 2026 · 14 min
AI in generalNo More Hallucinations or Half-Baked Output: How to Dramatically Elevate Your AI's Quality
One checker file forces every claim to carry evidence and exposes both invented statements and half-read documents.
06 August 2026 · 10 min
AI in generalThe seven failure dimensions of AI review: mechanism and control
Seven dimensions in which AI review fails, each with its mechanism and the control that holds it.
06 August 2026 · 19 min