Over the weekend, I attended a family gathering. One of my family members is a software engineer, and he mentioned that he has been looking for work for some time. He mentioned he was taking a Udemy course on Deep Learning to help his cause.

Here’s the reality: Tech companies are either cutting or freezing software engineering jobs. The reason? Tools like Cursor, Windsurf, Github Copilot, Claude Code, and others make one great software engineer more productive.

That means you don’t need as many software engineers. And you know what? Companies will keep the best engineers, not the average ones.

AI coding won’t replace great software engineers, at least not yet.

The Problem: Average Engineers Are Getting Squeezed Out

The market has shifted dramatically. When one engineer can do the work of three with AI assistance, companies don’t need the same headcount. They’re making tough decisions:

1. Keep the top performers who can leverage AI tools effectively
2. Cut the middle tier who struggle to adapt or add unique value
3. Freeze hiring until they understand the new productivity baseline

This isn’t speculation; it’s happening right now across Silicon Valley and beyond.

The Analysis: Why Some Engineers Thrive While Others Struggle

The engineers surviving this transition share common traits. They don’t just code; they architect solutions. They don’t just follow tutorials; they identify problems worth solving.

Great engineers understand this: AI tools amplify your existing capabilities. If you’re already excellent at system design, problem-solving, and understanding business needs, AI makes you superhuman. If you’re just copying code from Stack Overflow, AI makes you redundant.

The gap between great and average engineers isn’t closing, it’s widening.

The Solution: Build Something, Don’t Just Learn Something

As I told my family member: Stop taking courses and start building products.

Here’s your three-step action plan:

1. Use Existing Tools to Find the Gaps

Don’t try to compete with ChatGPT, Claude, Gemini, Cursor, or Windsurf. Instead, use them daily and notice where they fall short. What tasks do you still struggle with, even with AI assistance? What workflows feel clunky? What problems do your friends and colleagues complain about?

The opportunity is in the friction, not in the features.

2. Build Small, Ship Fast

We have a blank canvas where we’re only limited by our imagination and focus. Start with the simplest possible version of your solution. Get it in front of users within weeks, not months.

Your first version will be terrible; ship it anyway. Real user feedback always beats perfect code.

3. Focus on Problems You Actually Have

The best products come from personal pain points. What annoys you enough that you’d pay to make it go away? What tool do you wish existed in your daily workflow?

Build for yourself first. If you won’t use it, why would anyone else?

The New Reality: Builders vs. Job Seekers

The job market for software engineers will remain tight. Companies are learning they can do more with fewer people. But the market for useful AI-first software products has never been bigger.

Every business needs software. Every workflow can be improved. Every manual process can be automated.

The question isn’t whether you can code; AI can help with that. The question is whether you can identify problems worth solving and execute solutions people will pay for.

Stop competing for jobs; start creating them.

Bottom line: My family member could spend months perfecting his deep learning skills and still struggle to find a job. Or he could spend that same time building something people actually want.

The choice is clear: Be a builder, not just a coder.

The best time to start was yesterday. The second-best time is right now.