new this weekConstitutional AIessay · 18 min
Raising AI: From Rules to Reasoning
How Anthropic's new 23,000-word constitution signals a fundamental shift in how we guide artificial intelligence
Jawhny Cooke
02/26/26
jawhny.cookevol. 04 — agentic systems
past
the prompt.
what software looks like when the prompt isn’t the whole product. agents, evals, retrieval, and the rest of the stack underneath.
agentic systems— eval-driven dev— context engineering— inference graphs— tool calling— sub-agents— mcp— retrieval— self-verifying codeagentic systems— eval-driven dev— context engineering— inference graphs— tool calling— sub-agents— mcp— retrieval— self-verifying code
※ featured
what i’m
thinking about
new this week— latest writing
AI ResearchRecursive Language Models: Why Teaching AI to Take Notes Beats Giving It a Bigger Brain
01/24/26 · 11 min
AI PromptingDials, Not Dice: A Precision Prompting Guide for Nano Banana Pro and Claude Code
12/07/25 · 12 min
AI DocumentationThe Anti-Slop Framework: How to Create AI Documents That Drive Decisions (Not Noise)
11/30/25 · 16 min
AI SecurityAI-Powered Cyberattacks: How the GTG-1002 Campaign Changed Cybersecurity Forever
11/20/25 · 13 min
— themes
what i write about
※ by post — not by schedule
one email when
the next one’s ready.
no drips, no automations, no “5 things i learned this week.” only the actual post, when it’s actually done. ~once a month.
join 4,200+ readers · unsubscribe in one click · no tracking