I'm finding it harder to care about coding now that AI can generate so much of the code for me. I used to build small projects to explore ideas, understand how systems worked, and enjoy the programming and system-design process itself. Now I can ask an AI to explain a system, generate an implementation, and answer follow-up questions almost immediately.
I've never been especially motivated by launching products. I mostly enjoyed programming for its own sake, solving problems, and learning how things fit together. It feels like AI has removed the parts I found most interesting and left behind tedious polishing and debugging. I also worry that anything I build will be dismissed as AI-generated and added to an endless pile of mediocre software.
Do other people feel this way? How do you stay motivated, and how do you think about personal projects or presenting your work when AI can produce something similar?
5 Answers
A lot of the value in software was never the keystrokes. It’s choosing the right problem, understanding users, designing a maintainable solution, testing assumptions, and making sure the result is reliable. AI can generate a plausible script, but it doesn’t automatically know whether the problem is worth solving or whether the solution fits the real constraints. The role is changing, so it may help to focus on deeper design, debugging, performance, security, and communication rather than competing with a model at typing speed.
AI doesn’t make personal projects meaningless any more than cameras made painting meaningless or calculators made mathematics worthless. The point can still be the experience of making something and learning from it. If you no longer enjoy coding, that’s worth accepting honestly, but the existence of AI isn’t a reason you have to stop doing things you find satisfying.
If you’re doing this outside work, you don’t need to justify it through productivity or career value. Make something because you’re curious, set yourself constraints, learn a language the slow way, or build a project where understanding the internals is the goal. If none of that feels enjoyable anymore, explore other kinds of creative work instead of forcing yourself to recreate an old motivation.
I still do the important engineering myself and use AI mainly as a search tool or for small bits of boilerplate. I care about understanding the system and becoming a better engineer, not producing the largest possible amount of code. If the model writes something, I make sure I can explain why it works, what its tradeoffs are, and where it might fail.
That seems like the healthiest balance. Avoiding AI completely probably puts you at a disadvantage, but handing it the architecture and accepting whatever it produces means skipping the part that teaches you.
The useful way to work with AI is to keep ownership of the problem and the design. Decide what you’re building, draw the architecture, think through the data flow, and make the major technical choices yourself. Then let the model handle repetitive implementation, documentation, or scaffolding. It becomes a faster keyboard rather than a replacement for your judgment.
This also preserves the discovery process. You can use AI to build an experiment, but first predict what should happen and investigate when the result differs. That keeps the learning instead of outsourcing every question before you’ve thought about it.
I’m not sure every manual skill becomes valuable just because it is rare. Photography is a good counterexample. Still, creating something for your own enjoyment doesn’t require it to have market value.