July 7, 2026
I spend a lot of time thinking about "AI" - more than I would like. This question continues to bother me, so I'd like to get some thoughts out.
Putting moral issues aside (copyright, environmental, economic, etc), I have far fewer disagreements the idea of using a chatbot than I do vibe coding. Obviously, vibe coding does not take nearly the same amount of cognitive effort as writing software manually. For many years prior to "AI", the best technique to find the information programmers needed was to search the web for our problem. Often you'd find a Stack Overflow post with 5 answers. 5 answers which maybe did not solve your problem specifically, but were close enough to lead you to the solution (forcing you to think about the problem at hand).
The first coding problem I ever undertook (as far as I can recall) was making a JavaScript "web app" which allowed me to time meals depending on when I woke up. (I was a starving artist with a horrible sleep schedule back then) It took only an hour or two and set the ball rolling on learning to code.
Shortly after, I made a little app in Swift which let me record audio on my iPhone and upload it to Dropbox directly. That took an evening and most of the code was found on Stack Overflow. Was it the most cognitively intensive project? No, but I left it feeling empowered, capable, and curious.
You may argue that prompting is a skill; that directing an agent (or swarm of agents) to succeed doing a task is a lot like management of employees, which does take cognitive effort. I'd argue instructing employees who have feelings, opinions, family, friends, community, and personalities requires tact and real people skills; something a good manager has.
That being said, I don't believe using a chatbot in search of an answer is that dissimilar to Googling a problem (disregarding the moral baggage). This was especially true back in the early days of "AI" when the answers were often somewhat vacant or just barely scratched the surface. Now, even open source models will provide suitable answers most of the time.
I struggle with controlling my impulses. If there are cookies in the kitchen cabinet, I will eat them; often all of them in one sitting. To me, using "AI" (or copying+pasting from a Stack Overflow post) feels similar to giving into an urge like the aforementioned cookies.
On my best days using a chatbot, I would read the code and type it in to my editor line-by-line. A relatively small effort which helped cement new knowledge in my brain, building the scaffolding and giving it shape so I can recall it again in the future.
Shamefully, on my worst days I would copy and paste the function verbatim, try it and if it works I change the code style to match my own and move on. It's an odd experience later going through my codebase only to see a function which feels foreign and that which I do not understand despite it having my style.
"So what?", you may be thinking, "why not just take the output verbatim and move on, there by finishing the task?" I tried that and hated it. Or rather, I hated how I felt afterward. I got things done, but at the end of the day I felt empty. I would leave my office, head downstairs, and tell my fiancee that work kind of sucked and was a slog. The work, by the way, is my own software crafted by my own hands over the course of 6+ years. That was a slog? The code where I control the graphics stack, the audio stack, the platform layer. Nobody else touches my code except me. It's a dream! I am so incredibly lucky. How could that be a slog?
In early 2025, I began to realize these tools were not good for my mind. It didn't matter if I was copying it line-by-line or as a whole. I was just prompting, waiting for a response, and then copying it or doing my line-by-line thing. I was not thinking critically, I was not pushing myself, and I was not gaining new skills. I began to feel the inkling in the back of my head that these tools felt a distraction from real work or "the easy way out."
It began to shape of an addiction in my mind. Not unlike the other temptations in my life: weed, porn, YouTube, Instagram Reels, caffeine, cigarettes, and sugar. Whenever I'd find myself hitting a wall a voice in my head would appear and say "why struggle on this problem? It's not fun! Get the answer and move on."
In late 2025, shortly after a move from Ohio to Chicago. I needed to take time to adjust. I decided to take a month off from using a chatbot at all. This was right around the 2025 Advent of Code so it was a perfect time to try… hard (in Odin). So I did! I began to feel my brain light up a bit and started to realize how much it atrophied over the year prior. The AoC problems were taking quite some time each day, but I didn't open any AI assistant and felt proud. When I got back to work on MiniMeters I didn't feel much slower. In fact I could feel the scaffolding being built in my brain: places where knowledge could be stored and ideas could come from. I felt a lot of joy struggling and overcoming whatever issue faced. At the end of the day I felt far more accomplished and was much more excited to start the day.
Coming to now - mid 2026 - I still use chatbots, but with much less frequency. Never with a harness such as Claude Code, never as the first attempt at a problem, but only after searching the web, the docs, and GitHub fails me. And if it is more theoretical I prompt it to not give me code, but a guide or provide pseudocode. In fact, when ChatGPT first came about I would ask it for code in a language I was not using so I was forced to think through the problem more. The downside is that this technique still requires me to have self-control, which is not my strong suit. I believe I am taking a far healthier approach.
It's not lost on me that by using these tools there are skills I am not building. I am not blind to the fact these tools have serious ethical issues. We all need to draw a line in the sand at some point. Some may say they never touch "AI" and never will (I have a lot of respect for these people). Some don't care. I see far less people talking about moderate use of these tools and I feel I must be honest.
Despite all this, I do not care about vibe coded projects! AT ALL! I get the ick when I see them being published, disclosure present or not, and I lose respect for the person who created them. Why? I'm not sure.
Perhaps it's the lack of effort: That the creator didn't care enough to learn what was needed to create what they wanted. *snarky voice* "I know I struggled when working on my project; they just prompted the AI."
Maybe it's the fact they say "I made this thing" when really the "AI" did the work. (I saw someone comment on a Hacker News post saying something along the lines of "AI'm making…" then explaining their project. I felt that refreshingly honest.)
It could be the fact the project could just disappear on a whim.
Likely it is all the above, but it's time for the main question:
It has been happening more often. From friends I've been close with since high school, former teachers, recent mentors, and people who I love dearly. I do not want to push these people away. Not intentionally. Not by debating them on the ethics (some are more willing than others). And definitely not by demeaning them for using these tools. I don't want responses to feel cold to the point they no longer show me things they are excited about, "AI" or not. Plus, who am I to demean when I also use the same tools with the same ethical issues.
I have been telling myself to look at this through the lens of love, try to find the good parts of what they are showing me. Find the parts of their personality peaking through despite the design screaming "AI" to me. Maybe the underlying idea is useful and inventive, captivating and unique. I love that they are creative people! I love that they wanted to show me something they made!
The feeling I most want to avoid is the "You're not a real…" feeling. When I learned to make electronic music I felt and heard it many times. "You're not a real musician. You don't play an instrument." It gave me the drive to push myself, sure, but it hurt for many years. It made me feel I was incapable of learning or somehow less than. I don't want to hurt my friends when they are excited to show me something.
Perhaps I have it all wrong. Maybe I should be honest and that will be the best outcome. Maybe they are on their own journey and vibe coding is the first step to a fulfilling passion. Maybe I am just a hypocrite and a dick who needs some humility. I don't know. Is that the point of this post? I don't know that either.
I don't even think anyone will read this, but whether you think I'm a hypocrite who is full of myself, or share similar feelings I would love to know your opinion:
Drop a reply on Bluesky.
- Joe