The Perceived Value of Software Engineers Is Going Down

A few years ago, if you told someone you were a software developer, they said something like "oh, that must be complicated, I have no idea how that works." Now they say "oh, you do software? I do too, I made a tool yesterday while watching TV." The perceived value of software engineers is going down, and I think it's starting to show up in what companies offer them.

Software was seen as a hard job for specialists. Most people couldn't write a small script, and they knew it. They had no way to judge your work, so they assumed it was complicated and respected it.

Now anyone can describe what they want in Claude Code or Codex and get software that runs. That's easier than using Microsoft Office, where you have to find the right menu.

Anthropic looked at about 400,000 Claude Code sessions and checked how often people clearly got what they set out to do. In the sessions that produced code, people in software jobs got there 34% of the time, and people in other jobs 29%. Managers did best, slightly ahead of software engineers. Anthropic says that could partly come from how it measured this, because managers may be more likely to say so when they get what they asked for. Each group also picked its own tasks. But a manager who reads this won't think about any of that. They'll think they are as good at this as their developers.

Most senior developers, me included, have plenty of arguments for why we are still needed. I wrote about one of them in Vibe Code Is Legacy Code Before It Ships. Even if all of them are right, someone who made a tool last night and watched it work won't see your job as hard anymore.

On October 29, the National Student Clearinghouse reported that enrollment in computer and information science at American four-year colleges fell 12.4% this fall. These are early numbers, and the announcement doesn't say why. But if you're 18 and your uncle tells you he built an app in front of the TV, why would you spend four years on a computer science degree?

I see the effect most in software jobs at companies outside of tech. Say a company makes kitchen cabinets and runs on an order system that someone built for it years ago. Before AI, the owner had no idea how that system worked and was scared of it breaking. When the company needed a developer, it wanted the best person it could get. It offered $120,000 and was very happy when someone competent showed up.

Now the head of sales has built a quote calculator with Claude Code on a Sunday evening. It took two hours, and the whole office has seen it. So when the same job opens up, the owner thinks anyone can do it. True or not, the offer is $85,000, take it or leave it.

Executives act on what they believe AI will do, whether it does that or not. Harvard Business Review reported on a December 2025 survey of 1,006 executives, covering all jobs. Of their companies, 21% had already made large staff cuts because of what they expected from AI. Only 2% had made large cuts because of AI they had actually put in place.

What I'm building

Delegate tasks. Get software.

Give Vroni a GitHub issue, bug report, spec, or rough idea. It reads the repo, plans the change, writes code, runs checks, and works toward a review-ready pull request.

Take a look at vroni.com