What’s changing
AI agents write code, analyze data, draft documents, handle requests. They do it fast, often well, sometimes badly, and always with confidence.
Value is shifting. Producing costs less. Specifying precisely, judging the result, and checking what was produced matter more.
Why the pure manager is disappearing
For years, a technical career led naturally to management: less hands-on work, more coordination. A CTO could run a large team without having opened a code editor in years.
That model no longer holds, for three reasons:
- teams are shrinking. A small team equipped with agents produces far more than before. There are fewer people to coordinate, and more technical decisions to make
- judging takes understanding. A leader who doesn’t understand the work can’t evaluate what agents produce, or the people driving them
- the pace has changed. An idea can be tested in days. A CTO who can’t test it themselves decides on opinions, not results
The CTO who builds
A good CTO today:
- knows the craft hands-on: reads code, knows how an order flows through the systems, knows where the data lives
- drives AI agents themselves to analyze existing code, prototype an idea, build a tool
- sets quality rules: what must be reviewed, tested, and approved by a human, and by whom
- plans the organization around agents: which roles change, which profiles to hire, how to train the team
- is still a leader: decides, owns decisions, talks to the board. Building doesn’t replace leading, but it makes leadership credible
What it means for a small or mid-sized company
It’s an opportunity. A small company could never match a large one’s headcount. With well-driven agents, a small team can deliver what was out of reach not long ago.
On one condition: being led by someone who knows how. A pure-manager CTO in a team of eight is expensive and slows everything down. A CTO who builds speeds up the whole team.
What it means for hiring a CTO
When hiring, or choosing a fractional or interim CTO, a few simple questions:
- when did they last build something themselves?
- how do they use AI agents day to day?
- how do they check what those agents produce?
- how would they evolve your team with these tools?
Monolithic Lab’s position
25+ years of experience, and a daily practice of AI agents. Nicolas Mussat drives them himself to prototype, analyze, and build, and passes that practice on to the teams he leads, technical or not.
See also: AI transformation, and what 25 years of experience brings to the CTO role.