Safe By Design AI
Featured
The AI Ate My Cookie
Semiotic-Reflexive Transformer (SRT) Analysis of LLM Internal States
On the page, one character moved. The I became a Y, a descender dropped below the baseline, and the question left the kitchen.
Read the essay →Get new essays by email
No spam, unsubscribe any time. Or follow the feed.
Reference
- Granite models: an independent referenceAn independent reference to IBM's Granite models — release 1.0, published 2026-09-22. A snapshot of 172 repositories in the ibm-granite catalogue; latest upstream modification timestamp represented:
All essays
- The Inference CentipedeVisualizing the tracing data from LLM runs
- AI Agentic Engineering DemonstrationVisualizing AI Agentic Software Engineering
- Safe By Design: Discipline Is Not EnoughThe Structural Weakness of AI Agentic Engineering
- The AI Ate My CookieSemiotic-Reflexive Transformer (SRT) Analysis of LLM Internal States
- Fact or Fiction: No. 1 - Prompt PipelinesAnthropic Fable 5 educates me on Prompt Pipelines
- Visualizing Code Change Over TimeTree map meets git repo meets time-series analysis
- Software Architecture of The PromptHow does it work?
- History Repeats Itself: The Demon CoreEngineering Discipline in the Age of Machine Intelligence
- Training Data TransparencyWhy we credit. Why we protect.