Notes
Design notes on deterministic AI engineering.
All notes
Rewriting tik4net for Modern .NET
A one‑afternoon rewrite that reduced a long‑lived library to a small, modern, ~40‑file SOC‑style codebase now used in new production code.
When Newer Isn’t Smarter
Newer models are optimized for speed, steerability, and pattern retrieval, but these priorities do not help on narrow, specification‑driven engineering tasks.
A Three‑Layer Terminal Architecture
How AiStudio helped me design and build a terminal engine that’s more capable than I could have done alone
Designer and Programmer Working in Sync
A structured two‑mind workflow from brief to green gates
I Believe — AI, Engineering, and the Future
Why disciplined AI augments engineers instead of replacing them.
I Wrote This in February 2025 and Haven’t Changed a Comma Since
How AiStudio’s Tool System Became Effortlessly Extensible
MacTelnet, .NET, and AiStudio
How a long‑standing problem finally cracked — with help from modern AI
RBW, Stateful Staging, Atomic Writes, and Green Gates
How AiStudio generates dependable code through deterministic rails.
Tile‑Based Context: A Hybrid Approach to LLM Memory
Context is the most misunderstood part of LLM engineering.