I graduated with a Computer Science degree about two months ago. For most of university, I genuinely enjoyed programming: working through difficult problems, debugging, understanding what went wrong, and finally reaching a solution. Recently, though, AI tools have changed that experience. Instead of investigating problems myself, I can often ask an AI for an answer or generated code, and the process sometimes feels more like managing prompts than actually programming. I know these tools can improve productivity, but they have also taken away much of the curiosity and satisfaction I used to feel.
I currently work with one of my professors on paid medical and AI research projects, but the available data is extremely limited. I'm not convinced the project can produce a meaningful result, and some days there isn't much useful work for me to do. That has left me feeling unproductive and uncertain about my future.
Before choosing Computer Science, I considered medicine. Looking back, much of the appeal was the status, financial security, and clear career path rather than a strong interest in medicine itself. I chose Computer Science because I was genuinely more curious about it, but now I'm wondering whether my interests have changed or whether I'm simply burned out and disappointed with my current situation.
How can I tell the difference between losing interest in a field and going through burnout? For experienced developers, how do you keep programming intellectually rewarding while using AI tools? And how do you regain a sense of direction when something you once loved no longer feels the same?
5 Answers
Your research doubts and your feelings about programming may be separate problems. A data-starved project can make you feel useless without proving that you’ve lost interest in Computer Science. Before making a major career decision, try changing your environment or project and give yourself time away from the pressure. You graduated recently, so there’s no need to decide your entire future immediately.
AI doesn’t have to solve the interesting part for you. Use it as a debugging partner or a rubber duck: provide documentation and error messages, ask for possible explanations, and keep control of the design and implementation. Avoid having it generate whole features unless you really understand what it produced. You can also deliberately work on personal projects without AI, or study lower-level topics using books, documentation, and manual pages to rebuild your problem-solving habits.
Passion is useful, but it usually isn’t a reliable permanent state. A job may be worthwhile because it provides stability, useful skills, good colleagues, or opportunities to work on meaningful problems, even when every task isn’t exciting. Keep programming for yourself if that’s where you get the strongest enjoyment, and judge your career after testing different kinds of work rather than judging the entire field from one disappointing research position.
You don’t have to define yourself only as a programmer. Computer Science can lead to systems design, research, automation, technical operations, data work, product development, or roles that combine computing with another discipline. A background in medicine or healthcare could also be explored without immediately committing to medical school—for example, health technology, clinical data, public-sector technology, or process improvement. Try small experiments and conversations before making an expensive, irreversible change.
A lot of professional software work is less like solving challenging puzzles and more like maintaining large codebases, fixing routine bugs, and meeting deadlines. That doesn’t necessarily mean programming has become meaningless; it may mean your current role is a poor fit. Look for areas where the problems are genuinely difficult and where judgment matters, such as embedded systems, graphics, infrastructure, security, performance engineering, or scientific computing. In those areas, AI can help with boilerplate while you still own the reasoning.
The important distinction is whether the team values understanding. If people are only producing and approving generated code as quickly as possible, the work can become demoralizing. A team that expects you to explain designs, maintain documentation, and defend decisions can still make AI-assisted development intellectually engaging.