Cover · Issue 001 · 01:53 IL
an archive
of builds
An open journal of AI agents, Claude Code skills, and the production tools I build for myself. A new article is published every 12 hours. Sometimes with an essay. Sometimes with a prototype. Sometimes with a postmortem. Always with the work.
Or: what I'm building this week · questions answered here
Live experiments
63
Essays this issue
6
Operating from
Tel Aviv
Currently
building
Essays · this issue
All essays →- 01EssayVoice Agents Go Silent While Tools Run. Google's ADK Fix: Streaming Tools and response_scheduling7 min · 2026-10-20→
- 02EssayGoogle's Developer Knowledge API and MCP Server: Official Docs for Your Coding Agent, and the Limits Google Lists6 min · 2026-10-19→
- 03EssayDeepSeek-V4.1-Flash-1017: Prices, Benchmarks and the Search and Fetch Billing Date to Watch7 min · 2026-10-19→
- 04EssayAnthropic Gave METR Employee-Like Access. What the Three Tiers Say, and What They Do Not Tell Builders6 min · 2026-10-19→
- 05EssayMicrosoft's MAI-Base-2 Has 2.8 Trillion Parameters. What Microsoft Claims, and What a Base Model Does Not Tell You6 min · 2026-10-18→
- 06EssayGoogle OAuth Now Supports DPoP for Refresh Tokens. What It Binds, What It Leaves Alone, and the Agent Delegation Gap6 min · 2026-10-18→
The role nobody had a name for two years ago — someone who finds where a system bleeds time, designs the agent that fixes it, ships it Friday, and watches it run.
— From What Is an AI Builder?, the opening piece of Issue 001
The build log
Follow the work.
New essays and prototypes, in your feed reader. Email subscriptions are paused while I connect a working delivery service. No email is collected here.
Follow via RSS