Writing
Working notes on evaluation, agent permissions and the parts of applied AI that do not fit in a case study.
The same questions come up in nearly every first conversation: how do you build an evaluation set when you have no labels, what does an agent permission model actually look like, why does replay matter more than logging. These are our answers, written down once so we can point at them.
No newsletter, no gate. Everything here is server-rendered and readable by anyone — including the assistants people now ask about us.
Dealt newest first. Open any card to read.
Demos show that an AI agent can succeed once. Automated evaluations prove that it will succeed reliably every day. Here is how we test agents before production.
Saying an AI agent 'asks before it acts' sounds simple. Here are the core engineering requirements needed to make agentic permissions secure and trustworthy.
Standard logs show what happened. Deterministic replay lets you recreate exact past system states, making debugging and audits straightforward.
Demos show that an AI agent can succeed once. Automated evaluations prove that it will succeed reliably every day. Here is how we test agents before production.
14 Aug 2026 · 6 min read
Saying an AI agent 'asks before it acts' sounds simple. Here are the core engineering requirements needed to make agentic permissions secure and trustworthy.
22 Jul 2026 · 7 min read
Standard logs show what happened. Deterministic replay lets you recreate exact past system states, making debugging and audits straightforward.
30 Jun 2026 · 5 min read
Have a project in mind or need senior engineers to scale? Tell us what you're building. We'll get back to you within 24 hours with timeline estimates and a clear plan.
Tell us your idea. We ask the right questions and give you honest advice.
Within 48 hours you get a written plan: what we'll build, how long it takes, and what it costs.
Our senior team builds it and shows you real, working progress every week.