Hi there. I’m Erik Thor, and I’m a coder and system architect. Like most people in my field, I find myself using AI more and more, and now I’m starting to worry that it’s making me less talented at my job.
Worst of all, in my free time, I love to write and make videos. Now I worry that I’m spending less and less time writing. I think I’ve gotten addicted to AI, and I don’t know how to move forward.
I keep telling myself:
- “I’ll be done after I write this next prompt.”
- “After this final prompt, I’m going to start working.”
- “I’m going to start doing something.”
- “I’m going to start taking care of things.”
- “Get creative. Get writing. Get recording.”
But it’s never true, is it? It’s always another prompt.
Can you relate? Do you sometimes find yourself stuck in cycles trying to figure out the solution to a problem—maybe not even a problem that’s really important, or something you actually need to know? After all, do you really need to know what the moon is made of, and why are you prompting about that in the first place?
Notice how AI starts to slowly erode and suck itself into every aspect of your life—into your relationships, into your emotional health and well-being, and into your work. We should recognize early on if things aren’t working, break out, use our human brains to think up a way forward, notice when AI is starting to excessively creep into our lives, and set boundaries. We are going to have to start doing that on a day-to-day basis.
My Experience with AI
The truth is, over the last few years, I’ve developed a fascination with artificial intelligence and LLMs like ChatGPT, Gemini, and Claude. I don’t think I’m alone in using these tools:
- My sister tells her kids to ask ChatGPT for help if they ever have a question they can’t answer.
- My colleagues use AI for most everyday tasks.
- Students use AI to study.
- Teachers use AI to grade.
- Personally, as a coder, I have AI workflows and agents running for almost everything these days.
In this video, I’m going to share some of the honest experiences I’ve had with AI: how I’m using it, how to use it moving forward, what my hopes are, and what my worries are. Watch to the end to hear my tips if you plan to use AI at all.
As an intuitive who loves to learn, I use AI for everyday purposes such as answering fact-based questions, conducting research, and learning. I worry that this makes me read fewer books. Last year I read 52 books; this year I think it’s less than 10.
I wrote this YouTube script in an AI-coded application. I have my own custom writing tool, my own custom journal tool, and my own custom calendar tool. I build and develop these tools with artificial intelligence. I worry that fixing, improving, and hosting these applications takes too much time away from my creative tasks.
I code applications with tools like Antigravity and Cursor because I hope that one day I’ll be able to replace my current subscriptions and online tools with my own personalized, perfect, private alternatives. But my goalpost keeps shifting with every fix, every tweak, and every new tool. I keep thinking of another one I’d like, another solution I’d like, and how I’d like things to look different.
I think this is rooted in a sense of dissatisfaction or lack of appreciation for what exists simply because it was made with AI. I don’t really appreciate the work the way I would have if I had written it myself. I don’t really accept or like it the way I would if a human made it.
This AI-related dissatisfaction is a genuine problem. For most people, when we look at AI content, we don’t feel the same level of satisfaction or connection to the work because a human did not make it. I wouldn’t tell my favorite YouTubers that they should make their videos differently, because I love their work the way it is. But if it were made by an AI, I’d probably be pointing out all the flaws and issues I found with it. I see this as a real, honest problem.
The Illusion of Independence
At the same time, what brought me into this was my desire to become truly, fully tech-independent. I’ve always wanted to own my code, host my code, and move away from cloud subscriptions. I don’t want to spend hundreds of dollars every month on various tools—Adobe, YouTube Premium, Google Drive, and all those things. I want to own and host things myself.
But am I really independent if I made it with AI’s help?
While I do all of this with AI, I often arrogantly claim or tell myself that I’m somehow superior to other people when it comes to how I use AI because I don’t rely on it to be a therapist. I would never ask an AI for help with a personal problem, to analyze a dream, or to solve a relationship issue. I would talk with another person. I would journal. I would reflect. If I couldn’t figure it out myself and was really struggling, I’d go to a coach or a therapist.
I think that’s better because AI therapists can’t provide accurate, meaningful, or honest questions and feedback on your life. AI is programmed to be a validator who agrees with you, supports you, and tells you whatever you want to hear. That’s what it’s built for, and that is dangerous.
I would never want such a person as my therapist, and I would never consult such a person for advice. It’s a little bit like the Mirror of Erised in Harry Potter—it shows you what you most want to see. That image is very superficial; ultimately, you can’t touch or hold it. The AI might accurately tell you what your problem is in life, but it will not provide you with any meaningful solution.
You can spend years prompting about your problems and personal issues, but ultimately, the real work starts when you sit down and talk about it with other people—when you can be vulnerable, honest, and real with others. Can you really be that if you spend all your time talking with an AI instead?
Everyone has to see for themselves what is right for them, what works, and what actually makes them feel good and grow as a person.
- Does it really help you learn?
- Does it help you improve?
- In what way?
Being a Human First
I don’t like using AI to write or produce content. The thought of making an AI-generated YouTube video feels wrong. The thought of cloning my voice, my movements, or hearing myself speak through a clone feels like I’m being replaced.
I don’t like using AI to write because I love to write. If I use AI to write, then people read what AI wrote, not what I wrote. I want people to read what I write, the stories I have to tell, and what I feel.
I spend an excessive amount of time using AI tools, eating into my cognitive capacity. I spend too much time thinking about how to prompt an AI, what to prompt it next, what to develop next, or what to do after that.
Even though I work as a developer and system architect and write apps for a living, I am still a person first. I have my own personal thoughts, emotions, philosophical perspectives, and spiritual practices. I have my own website where I talk about psychology, I have my YouTube channel, and I have my blog. All of this is meant for me to be a human first, because that’s the most important thing I can be—even in my job.
The most important thing I do is talk to my colleagues as a human being first. Whatever I code or build on top of that is great, but ultimately, I only do that so I can be human. It’s my job, it’s what I do, and it’s how I contribute to the world. But I do it to create time and space for myself to be creative, have meaningful relationships, and be a good person. My job and IT career should never overshadow that.
Sometimes it does. Sometimes I think about it too often, and that eats into my ability to share interesting ideas with my wife or talk about interesting things with my friends.
Interacting with AI can feel like an endless loop. Most AI answers are nonsensical, vague, and abstract. They’re rarely practical, concrete, or direct. They’re programmed to sound nice, but aren’t built to be useful. They rely on passive voice instead of active voice. It’s not a direct conversation with you; it’s a re-consideration about someone abstract who doesn’t exist, by something that doesn’t exist.
The Gambler’s Fallacy of Prompting
Artificial intelligence can be useful, but it also follows the gambler’s fallacy. We often tell ourselves that the solution to our problem is hidden behind the next prompt. You get stuck in circles prompting back and forth. Often, nothing the AI says ever satisfies me, to the point where I don’t even read the answer.
I realized I was using AI as a personal journaling and writing tool. Why don’t I just write it down in my journal and skip the process of talking to AI about it?
Often I use AI to ask questions, but why don’t I just look up the original source? When I ask AI to act like a simple advanced search engine—giving me real sources, short references, or direct quotes so I can look it up myself—it doesn’t do it well. It often invents sources that don’t exist or outright refuses to provide sources.
What is the harm in using AI? This is a bit like a Faustian bargain—a deal with the devil. You hope it will give you a competitive edge, help you find a match on Tinder, help you do your job better than your colleagues, or help your company outperform competitors. But everyone else is doing the exact same thing.
AI offers to save us time by doing our work for us, and now our new job is telling this AI employee what to do. Do we all want to be managers in the future? Is our dream really to spend all our days telling something else what to do? Do we want to spend our time micromanaging something else to do a task we could have learned to do ourselves?
The Risk of Cognitive Offloading
It is easy to start simple and end up using AI for everything, offloading your cognitive capacity to the point where you start crossing your own boundaries. You probably started by telling yourself, “I’m only going to use AI for this.” Then you thought, “Well, maybe I can just do a little bit of that, and a little bit of this.”
It feels tempting and simple, but this is the lazy part of our brain—the part that wants to save energy and time. But what is a meaningful way to spend your time? You have a limited amount of time every day. Do you really want to spend all that time trying to save time?
If you really enjoy doing something, why look for a shortcut? If you have a beautiful forest path to walk through that you love, why take the shortcut if it means spending less time doing what you love—walking around in the forest and smelling the fresh air? We use AI to simplify our lives, but by doing so, we take away the things that bring us joy.
Imagine using AI to write a love letter to your partner. Consider how that affects your feelings toward them. The practice of writing a love letter helps you feel more connected to and grateful for your partner. If you use AI, will you still experience that feeling? You have to ask yourself:
- How does using AI make you feel?
- How does it help you connect to the things you do every day?
Imagine going to bed and realizing you haven’t actually done anything all day. You haven’t written anything for your book, made progress on your short story, made any videos, created art, or practiced piano. AI did everything all day, and all you did was watch. Why turn yourself into a consumer when you were meant to be a creator?
Inherently, AI can make you second-guess your skills and your value as a human being and creative. You may start thinking that artificial intelligence is simply better than you at everything:
- It writes better than you.
- It paints better than you.
- It makes better music than you ever could.
- It writes better code than you can ever hope to write.
This can keep you from doing any work on your own. Over time, this becomes a negative cycle. As your skills in art, writing, and recording decline, your brain loses the skills it never gets to practice. It thinks, “If AI can do it, why should I?”
The worse your skills get, the less skilled you become at evaluating the AI’s work. You start to perceive its art as better than it really is simply because you don’t understand what art is anymore. You evaluate its writing as better than it is because you’ve lost the ability to recognize good writing, and the perceived skill gap between you and the AI grows wider—even though what the AI produced was only ever surface-level and shallow.
Finding a Way Forward
We tell ourselves we’re different—that we’re the exception. I do that a lot. I tell myself, “I’m the exception. I’m using AI in a way that benefits my life. I’m using artificial intelligence to learn new things.”
But can I remember what I prompted AI about yesterday? Do I remember what I prompted it about last week? Can I remember anything it told me that I found meaningful, purposeful, or useful? If AI stopped working tomorrow, along with all the work and content it created, would I be able to remember what it had done or how? Would I be able to reproduce it?
I don’t say this to be moralistic. Artificial intelligence is so new that none of us fully know what is right or wrong here. We don’t know how it will impact our psychology, well-being, brain development, or our ability to do our jobs. We only have preliminary studies.
We do know one thing: the tech companies that build and run these services are not focused on your well-being, personal growth, or skill at your job. Their goal is to keep you using their services as much as possible. They aren’t programming these LLMs to help you solve your problems; they build them to keep you using them for as long as possible, for as many things as possible. That is a genuine problem.
Maybe AI can grow to be something genuinely helpful for us. But who should own it, what should their reward be, and what should their responsibilities be if they use it?
Personally, I’m still figuring out my stance on artificial intelligence—when it is good, when it is bad, how I should use it, and why. I’m not making this video because I plan to quit using AI entirely. Maybe I will eventually, maybe I won’t. I’m trying to figure out how I feel about it, what is right and wrong when it comes to these tools, and ultimately what makes me feel good, benefits my well-being, and helps me grow as a person. What can support me in developing my personal thoughts, my skills, my creative work, my work as a developer, and being a human being, friend, partner, and future father?
My Rules for Using AI Moving Forward
From now on, my intention is to only use AI in ways that assist and support me in being creative, learning, and understanding the world. My rules moving forward:
- Keep a firm eye on what really matters in life: Focus the majority of my time on tasks and activities that genuinely matter to me.
- Never delegate joy: Never use AI to do something I genuinely enjoy doing myself.
- Be honest about authorship: Never brag about anything I’ve done with AI or try to pass it off as my own.
- Do the work myself first: Never waste time prompting AI about something I could do myself or find faster with a quick search.
- Prioritize human connection: Ask a friend or family member before asking an AI, especially about personal or real-world topics. As human beings, we enjoy and appreciate being asked questions and given the chance to answer, even if we don’t know the answer.
- Be human first: Take pride in my work, whatever I do. Love it, cherish it, and let myself be immersed in it—even as a beginner. Take pride in figuring things out, loving the process, and dedicating myself to it every day, using AI only as the cherry on top.
If it ever becomes more than that, I can gently pause—just like in meditation—re-evaluate, and bring myself back to the work.
I’d love to hear your thoughts on all of this. What are your experiences interacting with AI, and how does it help or harm you?
Thank you so much for watching, and see you all in the next video.
Loading discussion…