From the StarDevs team

Blog

Notes from the StarDevs team on building AI agent systems, the infrastructure they run on, and measuring what actually got done.

Articles
4
Latest
October 22, 2026
Follow
Atom feed

  1. Ten Marks, One Hand: The StarDevs Diagram Kit

    9 min read Mykola Lobanov

    The glyphs of the StarDevs diagram kit, mint line drawings on a dark ground, one star in white

    The glyphs we draw our systems with, the rules for putting them together, and the files to draw your own.

    • diagrams
    • design system
    • icons
    Read the article

  2. The New AI Benchmark Is Cost per Completed Task

    13 min read Mykola Lobanov

    Glasses resting in front of a darkened screen of code

    The model benchmark tells you what a model can do. The production benchmark tells you what the system can finish—and at what cost.

    • cost per task
    • AI economics
    • LLM cost
    Read the article

  3. The Production AI Agent Stack in 2026

    14 min read Mykola Lobanov

    The Earth at night from orbit, city lights tracing the coastlines

    From model calls to governed systems that can plan, act, recover, and prove their work

    • AI agents
    • agent architecture
    • MCP
    Read the article

  4. Multimodal Agents Are Becoming Creative Teams

    11 min read Mykola Lobanov

    Overhead view of a person typing on a laptop beside a notebook and coffee

    Video, voice, documents, slides, websites, and visual-to-code workflows are converging around a shared brief. The opportunity is to carry creative decisions through an entire project—and make the next revision easier.

    • multimodal AI
    • creative teams
    • generative video
    Read the article

Working on something like this?

Talk to us about your work See the services