Posted inBlog
Harness Engineering: Making Your Whole Team Ship with AI Agents
For a while, the story of AI coding agents was mostly an individual one. A developer found a setup that worked for them, tuned their prompts, built out a personal…
Posted inBlog
The Case for Deleting Your AI Layer (And What to Actually Cut)
When practitioners first adopted coding agents like Claude Code, the instinct was to write rules for everything. Spell out every convention. Patch every reasoning gap. The `CLAUDE.md` file became a…
Posted inNews
What Codex’s Harness Reveals About Building Agents That Actually Work
Building an agent that works in a demo is one thing. Building one that runs reliably for hours, handles ambiguous permissions, searches a hundred tools without choking on context, and…
Posted inBlog
Somebody Else’s Definition of Done: How to Actually Evaluate an Agent Skill
A writing skill was installed to help with design decisions — color palettes, layout choices, that kind of thing. It activated correctly. The output was technically fine. But every suggestion…
Posted inBlog
Dark code: the comprehension gap that AI-assisted development quietly opened
The first serious encounter with AI coding assistants tends to center on correctness — will the generated code actually work? That turned out to be almost the wrong question. The…
Posted inBlog
Stop Prompting. Start Briefing.
The problem isn't your prompts. It's that you're still thinking in prompts at all. When you treat AI like a smart search engine — type a request, see what comes…
Posted inBlog
Your Company Runs on Rules Nobody Voted On. AI Is About to Make That Visible.
For most of my career, organizational rules lived in people's heads. The longest-tenured engineer knew why we never deployed on Fridays. The VP who had been around since the founding…
Posted inBlog
Intelligence that pushes back is more useful than intelligence that flatters
Shaping the flow of ideas at the frontier of cognitive systems. --- Claude was trained to disagree with you. Not to be difficult — to be accurate. Constitutional AI, the…
Posted inBlog
When AI Stopped Checking In
Shaping the flow of ideas at the frontier of cognitive systems. --- Klarna deployed an AI customer service agent in early 2025. Within the first month, it handled 2.3 million…
Posted inBlog
The Intent Gap
Shaping the flow of ideas at the frontier of cognitive systems. --- Klarna deployed an AI agent that handled 2.3 million customer conversations in its first month, cut resolution time…