TL;DR: A major Silicon Valley payments giant just laid off 40% of its workforce overnight because AI has automated code maintenance. People are panicking, blaming internet influencers who pushed kids to study Computer Science. But blaming influencers misses the point entirely. Learning to "code" was never a career; it was just a temporary mechanical trade. The engineers who are being fired are the ones who refused to understand business, office politics, and human intent. The only way to survive the AI era is to stop being a "Coder" and start being an "Architect."
James here, CEO of Mercury Technology Solutions. Tokyo - March 3, 2026
A reader recently sent me a panicked message. He pointed to the news of a Silicon Valley payments giant slashing 40% of its workforce overnight because AI Agents had taken over code maintenance. He was furious at the internet influencers who, for the last decade, preached that "Computer Science is the ultimate safe career."
He asked: "Are these influencers just the grim reapers of industries? They pushed civil engineering, and it collapsed. They pushed computer science, and now it's collapsing. What are these kids supposed to do?"
Here is my brutal, unvarnished perspective as someone who builds software systems for a living. Stop romanticizing Computer Science. And stop treating "coding" like it is a magical, irreplaceable superpower.
1. Coding is Just Typing. Math is the Core.
At its core, Computer Science is just Applied Mathematics. But over the last 20 years, we degraded it. We told an entire generation that Computer Science simply meant "Learning to Code."
Twenty years ago, knowing how to translate English was a highly paid, specialized job because most bosses didn't speak English. Today, everyone has translation apps. The need didn't disappear; it was just democratized.
Programming is exactly the same. Learning the syntax of Python or C++ is like learning the rules of chess. It takes a few hours. I studied Information Engineering in the late 90s. I learned how to build a website in a single day. When I needed to learn Linux programming for an internship, it took me a week.
Nobody goes to college to learn the "trade" of programming. You go to learn the mathematical logic. The syntax changes every five years. If you anchor your entire career value on knowing a specific programming language, you will not survive past the age of 35.
2. The Illusion of the "Executioner"
Even without AI, the software industry has always been obsessed with automation. When I was a Systems Architect in 2009, my entire job was building frameworks and instantiating virtual functions so that 10 engineers could do the work of 100. We were constantly building tools to make low-level coders obsolete.
AI just accelerated the inevitable.
The engineers getting laid off right now are the ones who thought their job was to sit in a dark room and wait for a Product Manager to hand them a ticket. They refused to understand the Why. They only cared about the How.
When you only know How to execute a well-defined task, you are competing against an AI that does it perfectly for $20 a month.
3. The Survival Guide: Become an Architect (The Art of BS)
If coding is dead, what is valuable? The Art of "BS" (Business Strategy and Human Negotiation).
A true Architect rarely writes code. Their job is to navigate the messy, illogical reality of human business.
- The Software Architect (The Translator): You don't just write a program. You figure out how to solve a business problem within a strict budget. If the client’s problem is only worth $5,000, but your elegant code requires $10,000 in server costs, you are a failure. Your job is to translate human needs into machine constraints.
- The Industry Architect (The Politician): This is where the real money is. You have to figure out the unspoken needs of the client's leadership. You have to keep their legacy systems running so the VP doesn't look bad. You have to integrate with their preferred vendors to maintain their political capital. You have to navigate the human ego.
No AI can navigate office politics, budget constraints, and executive ego.
Conclusion: Knowing What to Do is Worth 10,000x More Than Doing It
Years ago, a programmer wrote a brilliant script to help our office snag hard-to-get train tickets. I saw his code, realized he was talented, and hired him to rewrite a high-frequency trading algorithm for me in a different language.
He wrote the code perfectly. That code made me a massive amount of money in the financial markets. Did I give him a cut of the profits? No. I paid him his hourly rate.
Why? Because he just typed the code. He had no idea what the code was actually doing or the financial strategy behind it. If he hadn't typed it, I would have paid someone else to do it, or I would have done it myself.
The value is not in the execution. The value is in the intent. The algorithm (the business logic) is worth millions. The programming (the typing) is a commodity.
If you are a Computer Science graduate facing the 2026 AI layoffs, hear me clearly: Stop learning new programming languages. Start learning business, psychology, and systems architecture. Stop being the person who writes the code, and start being the person who tells the AI what code to write.
Mercury Technology Solutions: Accelerate Digitality.
