Overview
News
Technologies
Salaries
Products
People
Growth
Financials

Overview

Developer Relations, voiceover services and business consulting. Training to be the world's fastest centenarian while building the world's smallest conglomerate.

News

News i.brandanthonymcdonald.com 20 days ago
2,085 Tests, and None of Them Opens the Front Door
I read all 22 repositories in my ecosystem and counted every test I had written. Then I counted how many of them could fail a deploy. The second number is the post. Here is what my tests are genuinely good at, the tier that is missing, and where a browser
Read more
Report
News i.brandanthonymcdonald.com 23 days ago
I Built an Adversarial Eval for My Multi-Agent Coach. It Passed the Traps I Designed And Failed the Basics.
21 property-based test cases, an opus-class judge, a frozen baseline. The contradiction traps I engineered all passed. The thing that failed was citation discipline, at 57%. Here's the full loop: findings, fixes, and the re-run numbers, including what sti
Read more
Report
News i.brandanthonymcdonald.com 27 days ago
Prompt Chains, Tool Calling, and MCP: How AI Agents Actually Do Things
A from-zero explanation of prompt chains, tool calling, and the Model Context Protocol, using three production LangGraph agents as worked examples, with a SWOT for every approach and an honest audit of which techniques those agents actually use.
Read more
Report