The 17th Company

Six months ago this blog got me a job. Then the job killed the blog.

Quick recap for anyone new. Last December I tested something: build and ship real products using AI on pure pay-per-token. $750 and one month later, I wrote up what I found. Partly economics. Partly proof — show how I work, in public, instead of telling recruiters about it.

31 days after the first post — job offer.

And then nothing. Half a year of silence. The job I got through this blog consumed the exact resource the blog runs on. Focus.

So this is a post about what happened. About a pattern I finally named. And about the seventeenth company.

What the blog did

The first post went live on December 17th. The offer arrived January 17th. One month. A recruiter read my posts, saw how I think, and reached out.

Every engineer I knew was grinding through the same funnel — three hundred applicants per opening, ghosted after four rounds, HR asking for "passion" at a salary from 2019. And I just walked around the line. The blog didn't describe my skills — it was my skills, running in public, checkable by anyone with a link.

I remember thinking: no way it's actually because of blog.

The job

Solo frontend engineer. Not the lead of a frontend team, but the entire frontend team. I took it from an empty repository to production release.

I came in believing the hype like everyone else: prompt goes in, product comes out. Make a smart face, collect a salary. I learned fast that it doesn't work that way.

The real work looked like this: architecture in the morning, a "quick" UI tweak before lunch, and a bug from three contexts ago after lunch. Repeat until spring.

The volume was heavy but manageable. What wore me down was the switching. When you switch away, your focus doesn't pause; it drains.

There's also a new idea spreading across the whole industry: that any feature is "a couple of prompts" away. It's an understandable belief, since the demos make it look that way. But anyone who ships AI-written code to production knows the prompts are the cheap part. The real work comes after: reading what came back, catching what's subtly wrong, and testing what it confidently broke.

I was more productive than I'd ever been, and at the same time further behind than I'd ever been.

We shipped, the app became stable, and the engagement ran its course.

By spring there was nothing left of the part of my brain this blog runs on. The blog didn't decline. It just stopped.

The count

Sixteen companies. Do you really want to work in one company until your retirement?

That was never my story. Not once, not by accident, not because I couldn't hold a seat. I got hired sixteen times. I shipped sixteen times. Every engagement had the same shape: come in, build the thing, ship the thing, move on. Some companies call that a contract. Some call it employment. The word changes. The shape doesn't. I call it a finished contract.

Here's what matters more than any of those words: I adapted sixteen times. New codebase, new team, new domain, new stack — every time. The shortest lasted two weeks. The average was nine months. The longest stretched a year and a half. Each one a cold start. Each one shipped — even the two-week one.

Today that's my advantage. I was switching contexts long before it became mainstream. I can read what a company actually needs almost as fast as AI ships a feature. And after seeing sixteen different ways of working from the inside, I can tell what's going to work, where the bottlenecks are, and what should stay exactly as it is.

That's a C-level skillset, built over 10+ years of cold starts. I just never stayed anywhere long enough to hold the title. Yet.

What is actually going on

The industry is not reframing. It already reframed. Every big restructuring is a crisis — and the job market right now is exactly that. Awful, and it won't improve, because AI is genuinely replacing the effort that nobody in a corporate office ever cared about protecting. Every corporate worker knows exactly which meetings, which tickets, which "processes" I mean.

You have to react. Not adapt, not wait it out, not hope the cycle turns. React.

My software developer career is not dying — it's already dead. And not because of this blog post: check your LinkedIn — it's all green with "#opentowork".

It's hard to admit, but for me this picture is already here. Not coming — here. And the blog, the public building, the whole push for visible, verifiable experience — that wasn't strategy. It was intuition. Sixteen contracts of watching the ground shift, and my gut said: make your work undeniable before someone asks you to prove it to a filter.

I keep going — or I start learning proper Czech and switch to the service industry. Programming was always just the one thing I figured out how to get paid for. The question now is whether I can figure out the next one before this one finishes closing.

The 17th company

Sixteen companies hired me. The seventeenth is me.

It's a line on my CV — current position, start date, shipped work under it. Where there used to be a gap between jobs, there's an employer. Four projects running, three consultations active, this blog as the public ledger.

I'm not done with jobs. I'd take an eighteenth — a good team with a real problem beats solo work, and I've never pretended otherwise. What changed is the direction of checking. I'm not in a line of three hundred hoping a filter likes my keywords. I'm here with a URL.

If you're hiring, you can see how I think before the first call instead of after the fourth.

It worked once in 31 days.

What's next

The blog is running again. Not a comeback announcement — just work resuming.

December 17th — first post. January 17th — the offer. Today - The 17th company.

You have your own number. Mine is seventeen.

Role, project, audit — my email is always open. The 17th company is accepting both clients and offers.

— The Founder