If you’re a junior or mid-level engineer today and you have 5, 10, maybe 20+ years left before retirement, I’m going to say something that will sound dramatic, offensive, or outright wrong:
You should find an alternative, technology-adjacent career path as soon as possible.
Not eventually. Not “keep an eye on it.” Now.
Because writing software as a career is dead.
Not five years from now. Not when some hypothetical breakthrough happens.
It’s dead today.
Agentic programmers are already better than you. They are faster than you. They are cheaper than you. And they scale in a way you never will.
I don’t care what language you specialize in. I don’t care how good your LeetCode stats are. I don’t care if you work in backend, frontend, mobile, infra, or design.
If you are still employed today, treat this period as borrowed time. Use it to prepare for what’s next.
This path is closed.
Where I’m Coming From
I’m not saying this as a tourist, a pundit, or someone who just discovered AI last year.
I’ve been a technologist my entire life.
I spent my early years in the 80s on BBSs, messing with random software projects, trading files, and connecting with other like-minded people who were just curious about how things worked. That era quietly trained me for a career long before I knew what a “career in tech” even meant.
In the 90s, I started professionally as an AIX administrator. In the early 2000s, I moved into InfoSec. By 2008, I was programming full-time and eventually found my way into entrepreneurship.
I’ve loved every minute of it.
I’m a builder. Many of you are too. The joy wasn’t just in getting paid — it was in taking an idea, shaping it, wrestling with it, and eventually shipping something real into the world.
That identity matters for what comes next.
The First Crack: Creative Work Dies
Around 2022, I started playing with OpenAI as a content-generation API. Like everyone else, I was amazed. It was obvious immediately that something fundamental had shifted.
Shortly after that, I started experimenting with AI image generation.
That’s when I hit my first real slump.
You see, I’ve always been pretty creative. One of my favorite pastimes over the last 20 years was doing dumb, funny Photoshop or GIMP image or GIF manipulations — friends, pop culture, inside jokes. It was a creative outlet I genuinely valued.
And almost overnight, that skill became worthless.
At first, it sucked.
There’s a very specific kind of grief that comes from realizing something you worked to get good at no longer has any scarcity. No moat. No value. Anyone can now do it better, faster, and cheaper.
After sitting in that discomfort for a while, I realized something important:
What I actually enjoyed wasn’t the act of manipulating pixels. It was executing a vision and seeing it come to fruition.
Once I internalized that, things changed.
Fast-forward to today. Image and video generation are commonplace. I’m comfortable with them. I enjoy them. I create content I never would have even attempted before, and I get real joy — and a lot of laughs — out of it.
The tool changed. The creative energy didn’t.
The Second Slump: Programming Loses Its Scarcity
In 2023, GPT-3.5 became widely available.
At that moment, I knew the second shoe was about to drop.
Up until then, I believed I had two real superpowers:
- Creativity
- Enough knowledge of the full product lifecycle to design, build, and launch a product entirely on my own
When GPT-3.5 showed up, it was obvious that #2 was about to follow the exact same path as my Photoshop skillset.
That realization hit hard.
Around the same time, the earn-out from the sale of my previous business was coming to an end. I decided — half joking, half serious — that I was done. I stepped away from building and focused on pickleball and video games.
That lasted a few months. Turns out, you can’t outrun who you are.
I got bored. Restless. So I started tinkering again — this time deliberately casting aside the dread of a future where a large percentage of my skills were no longer needed.
Instead, I leaned into the tools, wanting to truly know my enemy.
Building With the Machines
Rather than fighting the change, I focused on using these new tools to amplify my creative energy and attempt projects I never would have tried alone.
That energy eventually manifested into Mega HR, where a few incredibly smart people from my past agreed to collaborate with me.
At the time, GitHub Copilot was new. Cursor didn’t exist yet. But we were fully immersed in anything that could accelerate our work.
Then, in January 2025, I started using Cursor — the first IDE I’d ever used with a deeply integrated LLM.
Like everyone else, I was blown away.
I joked that most of my “programming” had become hitting <Tab> and accepting large blocks of code Cursor suggested based on my intent.
My productivity jumped at least 10x.
That was just the beginning.
Today: I Don’t Write Code Anymore
Fast-forward to today.
I genuinely don’t write code anymore.
My productivity is likely another 10x on top of that first leap — roughly 100x compared to where I was pre-AI.
Building software for me now is a collaboration between myself and Claude agents.
Claude has a complete picture of:
- What my business claims to do
- What our software actually does
- The constraints we operate under
- What I’m trying to achieve strategically and commercially
Technologies that would have taken weeks or months to learn, prototype, and polish are now completed in days.
And I don’t mean just backend services.
I mean everything:
- UX and UI
- Marketing copy and positioning
- Databases and schema design
- LLM pipelines
- Microservices
- Ideation and iteration
The bottleneck is no longer execution.
The bottleneck is judgment.
The Hard Truth
Programming is dead.
UI/UX design is dead.
Not because these things no longer exist — but because they no longer have scarcity.
When execution becomes cheap, instant, and infinitely scalable, it stops being a defensible career.
If you are employed today as a software engineer or designer, you are living in the grace period.
Use it wisely.
This doesn’t mean you’re doomed. It doesn’t mean you’re useless. It doesn’t mean you stop building.
It means the value has shifted.
The future belongs to people who can:
- Define problems clearly
- Apply taste and judgment
- Understand systems end-to-end
- Translate vague intent into concrete outcomes
- Own responsibility for results, not just tasks
Those roles exist. They will continue to exist.
But the traditional “I write code for a living” role is not one of them.
That door is closed.
If you’re early or mid-career, my advice is simple — even if it’s uncomfortable:
Start preparing for your adjacent move now.
Because the machines aren’t coming.
They’re already here.