I myself have come across many such articles, this essay is my view on the matter.
No.
Definitely not in the short term.
In the long term it depends, like many others I too believe the existing Software Engineering jobs will get evolved, but there will be new jobs.
Let me share my reasons in two sections
Short Term
Unreliable for Production
AI agents have not yet got capable enough to maintain a production code base on its own. They can one shot applications, but their ability to continually iterate over existing code is still very much low.
“SlopCodeBench, a benchmark of 36 problems and 196 checkpoints where agents repeatedly extend their own solutions. Evaluating 15 coding agents across open and closed models, we find that no agent fully solves any problem end-to-end, and the best agent passes 14.8% of checkpoints. Quality degrades across checkpoints, with structural erosion rising in 77% of trajectories and verbosity in 75.5%. Compared to 473 open-source Python repositories, agent code is 2.3x more verbose and 2.0x more eroded. Explicit quality guidance reduces initial verbosity and erosion by up to a third, without affecting degradation rates.” SlopCodeBench: Benchmarking How Coding Agents Degrade Over Long-Horizon Iterative Tasks
They may someday overcome that limitation, that day may be in the near future or the distant future I don’t know.
No longer affordable
They have already started becoming expensive. The reason for the wide spread enforcement of AI in development was primarily because it was cheap and it is not the case anymore. Initially AI companies beared the loss to increase the usage but they can no longer take it.
The token cost has decreased, but on the other hand the context window has grown exponentially, Competent models resulted in increased usage, Memory prices have skyrocketed. (I may write a separate Essay on the economic angle)
Can’t take Responsibility
AI agents can’t take the responsibility when something goes wrong, which happen very often. We all know the famous “With great power comes great responsibility” line from Spider-Man. lets flip that line, with great responsibility comes great power.
Responsibility gives you power
A human can tell you when your idea is bad; when the deadline is impossible; when the architecture is over engineered; when something doesn’t make sense; when the priorities don’t make sense. But these things are very much had to get from an AI agent.
Agents navigating messy code revisit files more often and reason longer before acting. A recent controlled study found that resolving SonarQube-flagged issues reduced agent token usage by up to 8.5% and cut file revisitation by 34%. Maintainability is no longer just a quality argument.
https://arc.net/l/quote/kzvmhlsl
What it means for us?
moving from the operator guy to the idea guy.
image of responsibilities of a senior engineer.
Long Term
Why do I believe that?
First a short history to give some context.
The fear that a machine that can do what a human does in a more faster and cheaper would replace the human jobs is not something new. We have all already lived through at least one such cycle. (Note to check for other such instances. The Manure story from industrial revolution)
ATMs and Bank Teller jobs
When ATMs where introduced in the 1970’s people thought that it would replace the slower and costlier Teller jobs. But that didn’t, In fact the bank teller jobs and the ATM count increased simultaneously. The bank teller jobs transformed into relationship management
This pattern is so common that it even has a name Jevons paradox.
Is this a fair comparision?
Probably not. Definitely not. ATMs were a specific narrowed kind of intelligence unlike AI. But many people think that AI will also produce the Jevons effect Jevons paradox - AI, So that is a topic for another essay.