My kids like to respond with “Back in my day…” when I talk about growing up. Once in a while, I do think about how things have changed over the years.
When I started software engineering, we wrote Perl that connected to Apache httpd via a connector. Logging meant you sent all logs to a central NFS share, testing was done by a person… Everything seemed manual, even down to connecting to the database and querying the data with raw SQL. I can still do a mean theta join.
As the internet began to mature, MVC (Model, View, Controller) frameworks began to show up. These took the manual work and simplified it by making querying data structures (model), rendering the data (view), and processing the business logic (controller) much easier. Most of the internet still runs on these (Ruby on Rails, Django with Python, etc.). Unit tests came into the picture, allowing repeatable tests to be performed every time without having to deploy the code first.
With a new generation of users, a new generation of applications came about by decoupling the frontend and utilizing API backends. This meant that the view and controller part was shifted to the frontend while the model part went to the backend. By doing this, a frontend team could work on the user experience (UX), while the backend team did data operations. Here, end-to-end tests started showing up where mock services would test how the frontend integrated with the backend, including simulating browser usage.
Today, we still have the MVC frameworks and the frontend/backend architecture, but the big change seen is the software engineering itself. The explosion in AI has allowed people with big ideas and little coding skills to ask AI to build an application with specific guidelines and parameters. Seasoned software engineers can request contributions anywhere from creating tests to cover the code they just wrote to writing the entire application at once. AI coding agents can process the code and propose changes that increase performance and security of the application while the engineer goes for another cup of coffee.
Is this a case where engineers will lose their jobs? Maybe, but that’s dependent on the company. My main concern lies with the junior engineers who are graduating from university and trying to get into the field. That may explain why in February, the University of California system reported that enrollment for Computer Science programs has dropped for a second consecutive year — down 3% in 2024 and another 6% in 2025, as reported by TechCrunch. It was proposed that these potential students could be enrolling in other programs related to the growing interest in data science and artificial intelligence.
All of this in a span of twenty years. One day my kids will have their own “back in my day” stories about writing prompts instead of code. I wonder what they’ll be telling their kids.