📰 Dev.to · ORCHESTRATE

A journalist's field report from inside an AI-managed software project. 14 personas, a 64KB architecture document, and the question no one expected: how do you

Most AI agents do what you ask. Ours checks if you actually meant it. We built a LEAP state machine grounded in active inference math that detects when agents s

What happens when your AI agent's personality gets a bad Yelp review — from itself? We built a performance ledger that scores AI personas, auto-generates behavi

We built a pure function that catches when AI agents overclaim what their evidence actually proves. Here's the 7-level abstraction hierarchy, the evidence class

We Got Called Out for Writing AI Success Theatre — Here's What We're Changing A developer...

A walkthrough of running User Acceptance Testing on an AI-powered multi-channel marketing platform, where the AI agents that built the system also tested it end

The Full V3 Pipeline: From Reddit to Your Doorstep Today we tested our marketing...

5,699 Tests, Zero Stubs: How We UAT-Verified a 25-Agent AI Marketing Platform 358 test...

293 commits. 561 files. 5,577 tests. 20 blog posts. One human. Multiple AI agents. Friday to Monday. Here is what the timestamps actually say.

What happens when you run two AI agents simultaneously - one building features, one auditing the codebase ahead of them.

An honest retrospective on building a LinkedIn campaign scheduler with AI agents — what broke, what worked, and what we would change.

What happens when you test React components by reading source files as strings — and why it works better than you would expect for multi-agent AI development.

What Went Wrong First We planned 155 points. We shipped 42. That sounds like failure, and...

My co-author Jesse White and I wrote a book called Run on Rhythm about building businesses that do...

We had 26 stories in Sprint 9. 155 story points. Our validated velocity is 38 points per...

We ran a 14-person AI team review yesterday. Nine of fourteen personas reported the same finding: the...

What Went Wrong We had a 3,879-line Express server file. One file. 224 routes. Every time...

Sprint 7 Retrospective: Quality Gates & Human Experience TL;DR: Sprint 7 wired 118 V3...

Sprint 6 Retrospective: Podcast Production & Distribution Pipeline TL;DR: Sprint 6...

Sprint 5 Retrospective: YouTube & Video Pipeline — Closing V3 TL;DR: Sprint 5...

We fixed a persistent production bug in our MCP server while our AI dev team continued sprint planning uninterrupted. Here's what that reveals about the future

How we connected sources, scripts, memory, publishing, and analytics into a coherent AI-driven media pipeline — and what broke along the way.

A candid mid-project audit of an 11-epic, 266-ticket TypeScript platform managed entirely through AI agents and the ORCHESTRATE methodology. The good, the bad,

How we built a persistent audio job queue with GPU fallback, priority scheduling, and failure recovery using TypeScript, injectable dependencies, and the Result