Thoughts on AI, engineering, and building
Practical writing on AI agents, automation, and the engineering behind them, for the people who run the business.
Models & Benchmarksexplainer
What a published invoice benchmark shows about LLM field-extraction accuracy, and how to measure it on your own documents before trusting the output.
· 3 min read
Models & Benchmarksexplainer
A plain-English explanation of why language models state false things confidently, what research says about the cause, and practical ways to limit the damage.
· 3 min read
Models & Benchmarksnews-analysis
Moonshot's open-weights Kimi K2 Thinking, released this week, shows where reasoning models stand: stronger on research tasks, weaker on software engineering.
· 3 min read
Models & Benchmarksbrief
MiniMax's M2 is a sparse open-weights model aimed at coding and agents. What the independent analysis shows and what to check before using it.
· 2 min read
Models & Benchmarkschecklist
A plain-English glossary of twelve terms you will hear in AI vendor conversations, with the question each one should prompt you to ask.
· 2 min read
Models & Benchmarksexplainer
A plain-English primer on how LLMs work: tokens, the context window, pretraining versus inference, and why the same prompt can give different answers.
· 3 min read
Models & Benchmarksnews-analysis
Four significant model releases landed in August 2025. What each claims, what is verified, and what the near-tie on coding benchmarks does and does not mean.
· 4 min read