We have been here before.
In the beginning, engineers wrote machine code by hand. Raw binary instructions, one register at a time. Then assembly came along and gave us mnemonics. Then C arrived with a compiler that translated human readable code into something the machine could execute. You wrote the logic, the compiler handled the translation. Then languages like Ruby took it further. Writing Ruby felt like writing English. You could read a Rails controller the way you would read a paragraph in a novel.
Each step raised the abstraction. Each step made engineers more productive. None of them eliminated the need for engineers.
Today we have AI. A tool that generates code. A tool that speeds up the working process. But that is all it is, a tool. The latest layer in a long history of layers that help us move faster.
What Does Not Change
Tools change. The fundamentals do not.
Your ability to break down a complex problem into smaller pieces. Your ability to pick up a new tool and apply it effectively. Your ability to look at a system and understand why it behaves the way it does. These are the skills that matter, and they have always been the skills that matter.
A compiler did not replace the engineer who understood memory management. Ruby did not replace the engineer who understood system design. AI will not replace the engineer who understands how to solve problems.
The COBOL Parallel
There is a useful parallel in COBOL. Banks still run on it. The engineers who know COBOL are scarce. The demand is limited, but it is real, and it pays well because the supply is even more limited.
The same thing will happen with AI. There will always be domains where AI cannot be used freely. Security sensitive systems. Financial infrastructure. Regulated industries where every line of code needs to be auditable and explainable. In those spaces, the demand for engineers who can work without AI assistance will remain. It will be a niche, but a valuable one.
More Products, More Engineers
Here is the part people get wrong when they say AI will replace software engineers.
Before AI, maybe 10 new apps launched per day. With AI lowering the barrier to build, that number could be 100 or 1,000. Almost anyone can create a product now.
Take even 10% of those products. The ones that find users, gain traction, and need to scale. Those products will need engineers. Not to write the first version, but to keep it running. To handle the complexity that comes with real users, real data, and real scale. To debug the problems that AI has never seen before.
More products means more demand for engineers who can maintain, scale, and evolve them. The pie is getting bigger, not smaller.
Beyond CRUD
And the work itself is changing. Engineers are no longer just wiring up forms to databases. The boring CRUD app is table stakes now, something AI can scaffold in minutes. What companies actually need engineers for is building the next layer, agentic AI systems that automate workflows, make decisions, and help businesses scale faster than any traditional software could.
This is not simpler work. It is harder. It requires understanding how models behave, how to design reliable pipelines, how to handle failure modes that do not exist in traditional software. The engineer who can build and operate these systems is more valuable than ever, not less.
Hiring Should Not Change Either
Whether a candidate uses AI or not, the evaluation should be the same. Can they solve problems? Can they adapt to new tools? Given a complex, unfamiliar problem, how do they approach it?
AI is excellent at predicting patterns it has seen before. But novel problems, the kind that show up at scale, the kind that involve tradeoffs between systems, the kind where the constraints are unique to your business, those still require a human who can reason through ambiguity.
The best hiring signal has always been problem solving ability and adaptability. That was true before AI, and it will be true after.
The Bottom Line
AI is a tool. A powerful one. But it sits in the same lineage as every tool that came before it, assembly, compilers, high level languages, frameworks. Each one made us faster. None of them made us unnecessary.
The engineers who thrive will be the ones who treat AI as what it is, another tool in the toolbox. Learn it. Use it. But do not confuse the tool with the skill.