Is AI closing the door to entry-level tech jobs?

I know someone who is studying software engineering at university. Not a bootcamp grad, not self-taught. Someone formally trained, doing the work, learning the fundamentals. He spent months applying to internships and junior developer roles alongside his studies. Dozens of applications. A handful of responses. Zero offers.

He wasn't underqualified. He was doing everything right. But every job description he opened asked for things that, two or three years ago, nobody expected from someone just starting out. CI/CD experience, familiarity with AI tools, cloud certifications, production-level thinking. Entry-level roles that read like mid-level requirements.

I'm a cloud software engineer and consultant. I work with different clients across different projects, building with AI and shipping products that have AI baked into them. And watching him go through that process made me start asking a question I couldn't shake: did the bar actually rise, or did the door just close?

The timing is not a coincidence. In the last few years, AI coding tools went from early experiments to everyday practice. Tools like Codex, Claude, Gemini, and Cursor became capable enough to handle the work that junior developers were traditionally hired to do: boilerplate code, documentation, basic debugging, simple feature implementations.

Companies noticed. Why invest in onboarding and training someone to do tasks that a tool now handles automatically? From a pure cost perspective, the math seemed to work. Junior headcount dropped. AI tool subscriptions went up. Efficiency improved, at least on paper.

But something important got lost in that calculation.

And it goes further than just replacing tasks. There are reports that Google has started using Gemini to handle work that would have previously gone to junior engineers. When the companies building these AI tools are also the first to replace entry-level roles with them, that sends a very clear signal to the rest of the industry about what comes next.

Junior roles were never just about output. They were about formation. The entry-level job was where you learned to read a codebase you didn't write, made mistakes on real systems, had a senior engineer help you understand why, and built the instincts that no tutorial, no university course, can fully teach. The ones that come from sitting inside a real problem for hours until something clicks.

Here's the part that struck me hardest: this person is getting a formal software engineering education. He's not cutting corners. He's doing it the way the industry always said you should. And still, the door isn't opening. If a university degree in software engineering is no longer enough to get your foot in the door, something has fundamentally shifted and we should be talking about it.

AI gives you the answer before you've had the chance to struggle with the question. And the struggle is the point.

There is a quote I keep coming back to, from Andy Jassy: "There is no compression algorithm for experience." AI can generate output. It cannot generate the years of context, failure, and iteration that shape a great engineer. I can't help to feel very familiar with this situation, because I graduated as a Software Engineer in December 2025, starting my first job in IT in late 2021 when AI was not a thing yet. So now it feels like I was lucky to work just before the AI boom.

When we remove the entry point, we don't just make it harder for people to get in. We cut off the pipeline that produces the senior engineers, architects, and technical leaders of the next decade.

I'm not arguing against AI. I use it every day. It makes me faster and helps me focus on the work that actually requires judgment. But that judgment had to come from somewhere. It came from years of doing the unglamorous work, the same work we are now automating away before the next generation gets their turn.

If you are involved in hiring, I want to ask you to do one thing. Open a junior or intern job description from your company. Read the requirements. Then ask yourself honestly: does this reflect what someone at the start of their career could realistically bring? Or does it reflect what you would expect from someone with three years of experience?

If it is the second one, you didn't raise the bar. You closed the door.

This person is still looking. He's not alone. There are thousands like him, sitting in university classrooms right now, learning to build the future, wondering if there will be a place for them in it.

The door was never supposed to close. We still have time to open it back up.