Research-backed pieces on AI agent deployment
I Took My Framework to Harvard and Oxford
The post-Harvard V6.2 update: what held, what changed, and what Oxford will add next.
3 July 2026 · 7 min read
I Mapped 40+ AI Adoption Frameworks to Stress-Test My Own
I built a framework from experience, then mapped the landscape to challenge my assumptions. What I learned, what I'm incorporating, and what was genuinely absent.
25 March 2026 · 9 min read
Why AI Agents Fail (And It's Usually Not the AI)
Berkeley researchers analysed 1,600+ AI agent failures. The biggest category was specification and design problems.
20 March 2026 · 5 min read
Autonomy Is a Design Decision, Not a Capability Outcome
The same agent can operate at any autonomy level. The choice is governance and design, not technology.
21 March 2026 · 5 min read
The 80/20 Rule for AI Deployment
Less than 20% of effort goes to the AI itself. Over 80% is data integration, stakeholder alignment, and governance.
22 March 2026 · 5 min read
Your Best AI Agent Is Your Biggest Risk
The more reliable an agent becomes, the less capable your team is at catching its mistakes. Bainbridge's paradox explains why.
19 March 2026 · 5 min read
How I Built This Site with Claude
A $10/year website built entirely with Claude Cowork. Static HTML, Cloudflare Pages, a subscriber system, and no code editor.
24 March 2026 · 9 min read

Get notified

New article notifications, nothing else. Unsubscribe any time.