Latest from the build
The CoMavenAI blog
Notes from
the build.
What we're building. What we're learning. Practical perspectives from engineers working alongside real teams.
Inside the work
Embedded engineering. Applied AI. Lessons from production.
Explore the thinking
In the Arena
9 articles in this topic
In the Arena #9: when the agent asks, who answers?
Four reads on what AI can do, what it costs, and the people who keep the work useful. Start by checking what happens when an agent needs approval.
In the Arena #8: two of 66 runs caught the $25.7 million
The best agent finished about a quarter of the finance assignments. Four reads on where agent work actually breaks, and where it is already cheap.
In the Arena #7: same model, twice the bill
The software around a model can double what it costs to run. Four reads on what AI really costs, who is already using it, and what agents do when nobody checks.
Get new posts by email
When we publish, you get it. No drip sequences, no sales follow-ups. Just the articles. Unsubscribe anytime.
Loading the form… You can also email contact@comavenai.com with the subject “Subscribe”.
In the Arena #6: your business isn't the frontier
Sam Altman and Dario Amodei want to pace the frontier. Your next useful software project needs the right model, clear rules, and economics that work.
In the Arena #5: when should the agent ask a human?
Where AI tools become operational change, when agents should hand work back to a person, and how to check whether one recurring task is worth a small test.
In the Arena #4: ask for the twentieth run
The agent that passes once and fails the other nineteen, what companies actually bought in July, the one metric that survives cheap code, and the first of 33 questions executives asked.
In the Arena #3: the model is the cheap part
The build that beat the buy, five verbs that moved a benchmark 59 points, who is actually shipping with AI, and the agent that found the answer key.
In the Arena #2: two agents are not two opinions
Correlated blind spots in agent swarms, the OpenAI number worth reading, Agent Plugins, and who funds the compute.
In the Arena #1: what small teams actually do with AI
Small teams stretching jobs with AI, the sixteen agent ceiling, and the EU's new disclosure rules.