I Used to Decide Who Got Hired. Now I'm Not Sure I'd Hire Myself.

For three and a half years, I decided who got hired. I ran the assessment centres at a tech company of more than a thousand people. I interviewed candidates for its leadership positions. I knew exactly what I was looking for: depth of knowledge, formal education, years of proven experience.

Today I build software. I have no computer science degree. I started eight months ago.

If I walked into that room now, as the candidate — would the person I used to be hire me?

To explain why that bothers me, I have to go back more than twenty years.

More than twenty years ago, I was a student in Saint Petersburg, Russia. I was studying Public Relations at LETI, working in the marketing department of La Plage—one of the most popular clubs in the city. My job was to help organize real events: concerts, parties, themed nights. That's where I learned what marketing actually was.

The difference showed in my coursework. My assignments were built on what I was doing in real life. They weren't theory applied to a textbook case. They were examples written from actual events I had just helped run. My professors noticed. They started showing my work to other students as examples of what good marketing assignments looked like.

Why? Because I understood it. Theory became interesting the moment it connected to something real.

That lesson stayed with me. It is still true.

But I almost forgot it.

In 2020, I wanted to build a website for my photography business. I decided to learn HTML and CSS—the conventional route. It seemed like the right thing to do. Pick up the skill, build the thing, move forward. I also tried WordPress.

It was all boring. It was slow. I spent weeks on something very simple and it still wasn't what I imagined. I gave up. I used Wix instead and accepted a website that wasn't quite right.

I was not happy with it. But I told myself: this is what you get if you're not a real developer. This is what I can do without years of training. I moved on.

February 2026. Eight months ago.

I started building with Operum. Same person who gave up on HTML, CSS, and WordPress in 2020. Same vision for what I wanted to build—a real client management system for my photography business, one that understands every client, remembers their history, handles bookings and payments, collects reviews.

In eight months, I learned:

  • How GitHub works
  • What CI means, what it does, why it matters
  • How to write and review pull requests
  • How Supabase handles databases
  • How Vercel deploys applications
  • How Cloudflare secures and speeds up a site
  • How Railway runs backend services
  • The entire testing loop—what breaks, how to find it, how to fix it before customers see it

Not from courses. Not from textbooks. By building. Every feature required learning something new. Every mistake taught me why I needed to learn it. The same principle from twenty years ago, compressed into eight months.

I'm not the only one. According to Zapier's survey of people actually shipping software with AI coding tools, 34% have no formal programming background. One in three people building real products never trained as developers.

The tool changed. The person did not.

Noran Shine—my photography studio website—is live. It looks like I wanted it to look in 2020. The client relationship management system, built inside Operum, is still shipping features. Every day I think of something else that would be useful, and I build it. There are so many great features I could add. I can't stop myself.

An AI-powered assistant sits inside it. This assistant has read every call, email, and message with every client. It knows my business inside and out. When a client calls, the system has already read the conversation history and knows the context. A photographer's CRM that actually understands the relationships.

But Noran Shine is just one example. Operum itself—the platform I'm using—was built by its own agents. Askturret, an AI trading platform, was built with Operum by another founder and an engineer. It is not just me.

This would have taken years to build the conventional way. The conventional way would have required becoming a real developer first, spending months or years before the work could even start.

How does it actually work?

Start with a real problem. Something you actually need to solve. Something that matters to you. In February, I wanted a CRM that understood my photography business—real bookings, real clients, real history. I did not start with "let me learn how to build a CRM." I started with "I want this thing to exist."

Then you describe what you need. Operum's agents handle the rest. They build the architecture, write the code, handle deployment. When something breaks—and it will break—you see it break. A test fails. You ask why. You learn because understanding it is the only way to move forward.

I did not know what CI was when we started. I asked, and got a clear explanation. That helped—but it was not the learning. The real understanding came from setting CI up and then working with it on products over the following months. The nuances arrived through use. I am still not deep enough in it, but I know enough to get far fewer bugs. And I know where to ask when I need more.

That is the real economy of it. Not mastery before you start. Not sophisticated architecture. Partial knowledge, honestly acknowledged, plus knowing where to ask when you need to go deeper.

So: back to the question I opened with.

Now I have something very different: real products I built from scratch in eight months. No computer science degree. No years of programming training. Just work I can show to the world.

If I walked into an IT company today and applied for a role building products, would they hire me? I don't know.

My honest feeling is still this: no. Probably not. The gatekeepers—people like I was—would probably still prefer the candidate with the CS degree and five years of traditional experience. They would see my eight months of real building and think: interesting, but it's not the same as deep knowledge. It's not the same as having paid your dues.

And the research seems to agree with me. Harvard Business School and the Burning Glass Institute studied over 11,000 US job postings from 2014 to 2023. Companies removed degree requirements on the job postings. Then they hired almost exactly as they did before. The gate moved in the posting and did not move in the room. Fewer than 1 in 700 new hires actually benefited from the change.

I'm not sure how I feel about being right.

But here is the strange part: computer science enrollment is falling. According to the National Student Clearinghouse Research Center, undergraduate enrollment in computer and information sciences at four-year institutions fell 8.4% in spring 2026 alone. Yet employers still hire as if the degree matters. Fewer people are studying. But the gate did not move.

Traditional education—does it become irrelevant soon? Or does it transform into something different? And if it does change, what becomes valuable? Breadth of knowledge? Or the ability to find information, learn fast, adapt?

These questions are open for me.

What I know is this: it is powerful and scary at the same time. Powerful, because the barrier to entry is gone—you can build something real, prove yourself, right now. Scary, because I was the gatekeeper. I decided what credentials mattered. And I don't know if I'd make the same decision today.

The system is changing. I can see it. I built inside it. But I don't yet know where it ends.

What do you think?