I feel like I can't have fun anymore because the AI can just do the thing instantly and you've got people on this website advocating to let the AI do everything while you merely read the code.

I felt similar for most of 2025. Then at some point something clicked and now I work on side projects again without AI.

Can AI do it faster? Yes, but that’s not the point. The point is having fun.

The analogy I keep going to in my mind is chess. A computer can play chess on my behalf, or I can play chess myself, but only one is fun.

I think it’s more complicated that. LLMs have allowed me to do things that I couldn’t do before which definitely made programming and hacking things together more fun, and massively increased what I could do in my limited free time. It also allowed me to manually do the things I enjoy while making the less fun parts go faster. On the other hand I recently tried doing a larger project in codex and it wasn’t fun anymore because codex quickly created a system that was way beyond my understanding, it didn’t work, and I had no idea how to fix it. So I guess it just depends how you use it.

Ignore the AI people. In all probability, you were doing something similar before. The Internet is full of developers. Some may be faster at writing code than you are, or maybe they wrote better code, or perhaps they implemented your ideas before you even thought them up. Yet it sounds like you didn't let that inhibit you before.

It can't possibly do everything; mind reading interfaces haven't been invented yet. Paul Graham goes on about how writing is thinking. Just the act of writing out instructions can be fun.

Mostly FUD from grifters and accelerationists. Coding AI isn't useful for producing things that you couldn't have produced yourself, which means you're still important. Fundamentally it's still "just" an autocomplete, whether it's snippets at your cursor or whole files inside your directory. I actually quite enjoy LLMs as a programmer. Contrast this with compilers, which produce machine code that you couldn't have possibly written yourself.

why do you have to do what people on this website tell you? Write the fun thing.

I have fun, but I probably wouldn't if the AI was right all the time. Or if I was helpless when it was wrong. But for now I'm still in the centaur zone.

I have been doing a lot of little projects using AI, and don't get this experience.

I get what this post is talking about. I'm just having fun, that comes in a lot of different flavours. I can try a lot more ideas out, that's fun. I can quickly learn if an idea won't work, sometimes that can be disappointing but at the same time learning why it won't work can be quite fun. When the AI utterly fails to do something it lets me develop an idea in my mind about the strengths and weaknesses of the models. Oftentimes the failures are not just fun but outright hilarious. I enjoy seeing models fail sometimes because they reveal an assumption that I have internallsed to the point of being unaware of it's presence. It reveals to me something about myself when something I didn't feel worth mentioning is actually quite important to communicate. Some of the failures are outright hilarious.

I do find it a bit tiring to use AI for long periods, because lazy thinking produces poor results. You have to maintain a clear idea of what it is you are trying to do. Quite often an idea can seem simple in your head because you have glossed over a number of complicating details. I find it a challenge to keep mind at a level where you are aware of these things before you request an AI to make something intrinsically flawed.

I don't have a problem doing things without AI just for fun either. I make animated images in a tiny stack machine bytecode. I do game jams, and code golfing, like dweets.

I also enjoy playing chess, computers pased my ability to play chess a long way back. I don't mind playing even when I know a computer can do better.

Unless you are the best in the world at a thing, there's always someone who could do it better, every attempt to do the best thing ever in a field will fail. On the other hand you can try and do better that what you yourself have done. Even then that's just the target to reach for. The real goal is to enjoy the reaching. It's the challenge at the limits that is fun, not the success or failure of the end result.

I've not used AI to write code, but everyone who I've spoken to who has says it actually takes a lot of work. It sounds like you get intern level work out of AI, but without the hope that your investment in time results in skills and personal development for the intern.

All of the fighting with the LLM to refine the results sounds tiresome. No thanks.

If it's fun for you, or it unblocks you, or whatever... Go for it. But it doesn't sound fun for me, so nope. I'll keep banging on rocks to write programs until that's not fun anymore. :p