Long-form
How I became an x10+ company
People used to pull me onto projects, half-jokingly, for my ideas. Someone stuck a title on it: CDO — Chief Disruptive Officer. It was a joke, but it stuck, and I never really shook it off.
The difference with back then is that I can prove it now. I have the receipts.
x1.2, x2, x10
That x10 or x100 comes from the world of disruption. Doing something 1.2x smarter is a matter of optimising. Doing something 2x faster sometimes works by temporarily overloading. But doing something 10x faster, better, smarter or cheaper? That usually means working differently. And that is what is happening here. Not working harder, but differently and smarter.
That ladder is the whole point. Optimisation moves within the existing way of working; disruption throws the way of working out. Everything that follows hangs on that distinction.
The receipts
Right now I have running: three software products, the website, an internal messaging app, and more. All of it on a single refurbished virtualisation server — 14 cores, 28 threads, 128 GB of ECC memory, 2× 1 TB of SSD and HDD in RAID 1. In front of that an enterprise firewall, with Cloudflare and Microsoft 365 ahead of it. All managed by autonomous agents.
And refurbished here does not mean second-hand junk. My sysadmin recently ran a sweep for redundancy and single points of failure, and came back with a concrete recommendation: enterprise SSDs with power-loss protection. I bought them and fitted them.
That, by the way, is the model in miniature. I had not worked out for myself that there was a risk there. He did the analysis, came back with a reasoned proposal, and I carried it out — because holding a screwdriver is the one thing he cannot do.
And then the costs. My monthly out-of-pocket sits between €200 and €400, everything included. Broken down:
| Item | Per month |
|---|---|
| Claude subscription — the complete team | ~€200 |
| Electricity, depreciation on the refurbished server, Cloudflare, Microsoft 365 | ~€200 |
Yes, the Claude subscription is in there. That is usually the first question, so here is the explicit answer: the complete team costs about €200 a month, and the whole infrastructure around it costs about the same again. There is no hidden AI bill behind this story.
Being a bit cheaper than the next person is what I call optimisation. But 10 to 100 times cheaper and faster — that is not optimisation and it is not evolution. That is hard disruption.
The team
Five developers, four staff, one sysadmin. Working day and night, no fuss, clear expectations. And honestly: directing them is simpler than directing people.
Do I need to understand exactly what they do? No. A CTO does not know command-by-command how a sysadmin does the job either. What they do need is a good workflow, checks and controls, and someone who understands at a high level how it ought to work and who gets informed when uncertain — through spikes into security, into architecture, into UX, into known patterns, into best-of-breed solutions.
That is the core of the model. I draw the line between what they do on their own and what comes past me first. They carry the rest.
On craft
I want to be honest about this, because that makes it more credible. Is this software exactly what a senior developer delivers after years on a single project? There are probably unnecessary abstractions here and there, or it is less DRY than it could be. That is a real and recognisable difference, and I am not pretending it does not exist.
But there is something on the other side of the scale. Every one of the five boards has a mandatory CI Gate. The autonomous ticks cannot bypass it — that is written hard into the protocol. Only I can override a red check, and only when the CI itself is demonstrably broken and not the change. There is branch protection, and secret scanning that redacts secrets before they end up anywhere. There is a separate QA role that tests behaviour before the merge, not after. And when in doubt, spikes go out into security and architecture before a choice is made. It is a difference in craft, not a hole in security.
What it comes down to
One of the developers put it more sharply than I can:
There is no competing with it. What I cost a company in a single morning will now get you 2 or 3 complete products for the same money — provided you know how to direct the team properly.
That last part is the condition, and it is not a footnote. The technology is here. The difference is whether you can steer it.
And steering is exactly what it is. Reversible work is allowed to run on its own; anything with real consequences waits for my approval. That is where the directing sits, and that is where it goes wrong for most people who try this — they delegate the decision instead of the work.
That is what I mean by an x10+ company: not that I am ten times bigger, but that I do ten times as much with the same means.
Disruption used to be an idea I had to sell. Now it is a bank statement. The CDO joke still holds — only now I have it on paper.
This piece belongs to How I build — that page has the short version, with the receipts.
Back to How I build