The long arc

About

I'm an AI coach, advocating the mindset shift to AI native for the team and the org.
An extremely heavy AI user: seven AI plans, a dozen or more agent sessions at once.
10–100x in my own delivery, with a high quality bar.
I built the Long Horizon framework, which runs unattended for hours.
One working day: about 10,700 lines changed across 85 files, merged the same afternoon.

Project and cloud management, then SRE, then cloud and AI platform work, up to Principal Engineer now.

My full work history is on LinkedIn.

Why seven AI plans

Seven AI subscriptions across three model families (Claude, Gemini and Qwen), plus voice and coding assistants. Agent sessions build, check and review each other’s work, and each plan has a job in that harness:

Claude Code Max, my personal plan: running the heavy lifts in a sandbox, in auto or unattended mode.
Claude Enterprise: team work, with enterprise controls.
Claude on Google Vertex AI: API runs and benchmarks inside Google Cloud.
Google Antigravity, with Gemini: a second model family to cross-check Claude.
Qwen Studio: open-weight models, for cost and portability.
A text-to-speech service: voice, for audio and talking avatars.
GitHub Copilot: help inside the editor.