AI, agents, and the things we build together.
I'm Boris. I write about building with AI: the engineering, the architecture decisions, the lessons learned shipping agents in production. Everything on this site runs on curate-me.ai.
This site runs on curate-me.ai. Every AI call is governed. Every cost is tracked. Every decision is auditable.
Every post carries its own AI cost, model choices, and governance footprint. The numbers come from MDX frontmatter and reader engagement.
Featured demos
Recent posts
Hill-Climbing a Small Decision Model: What Four Real Jobs and $5 Taught Us
The follow-up to our decision models note. We hill-climbed Jev on four real jobs: judging agent reports, second-guessing the security scanner, starting agents from chat, and retrying failed runs. Twenty-eight changes, scored on a tuning set and then once on held-out data. What moved the scores, what didn't, and the ten lessons we're carrying into the platform.
We Hill-Climbed This Blog's SEO With Our Own Agent. The Score Rose to 92. Search Hasn't Noticed Yet.
We scored its-boris.com, ran our platform's SEO agent on it and kept only changes that raised the score with nothing worse. The score went from 77.3 to 92.4, mostly on unproven checks. Agents found answers a little more often. AI search still found us for 1 of 20 questions.
We Gated Agent Upgrades on a Score. 46% of Top-Scoring Reports Failed Our Audit.
A new agent version that scored worse than the best one was thrown away. Then audits of 165 runs found that 46% of reports scoring 85 or higher failed.
Agent Harness vs a Single LLM Call: Why Our $0.875 Job Dropped to $0.0016
The same AI news job cost about 500 times less after we moved it off an agent harness. A controlled replay put the harness itself at about 2x.
Recent pipeline runs
Real dev tasks decomposed, coded, and shipped by AI agents through the curate-me.ai autopilot pipeline.