My AI Dilemma

Via Patreon on January 2nd 2026:

Hey you!

I hope you are doing good!

In recent months and years, I’ve found myself entangling myself in a dilemma, and because I know that this is a quesiton which is extremely relevant for a wide range of people, I’ve decided to make this public to start a conversation about it.

As you may know, since a few years, I passionately advocate against AI and the biases it creates. Of course I argue from the perspective of a design educator here. But to build my own infrastructure, the website for example. It started with translations from German to English with Deepl, then I’ve made a subscription to Mistral “Le Chat” to ask all kinds of technical questions. And then, I’ve discovered “Cursor”, which is extremely helpful to refactor, explain and simplify code of my websites. It helps me running my platform.

The speed and precision with which Cursor translates my thoughts into code is breathtaking. Cursor is different than every other tool I have worked with, because it “knows” the whole codebase and can work on files effectively that depend on each other. It’s an incredibly powerful tool for prototyping: I can simply tell the “agent” what to do.

This all drove me into a… let’s call it… a tangled up situation. Philosophically. Is this still integral? Can I advocate for “downgrading,” for technological sovereignty, repairing old laptops, Open Source software, Creative Coding, and Linux, while I myself use the most powerful AI machinery to build and refine my own systems? Doesn’t it feel like a betrayal of my own philosophy?

For fun, I did something completely paradoxical, maybe even absurd and started a conversation with another very powerful AI: Google’s Gemini. I’ve used it for the first time for this conversation. Here’s the original prompt with its answer.

A conversation with Gemini

My prompt:

“I would like to have a conversation with you about AI. I find myself in a bit of a dilemma. Briefly about me: I am a designer, Creative Coding teacher, and media artist. I am also currently writing a book titled _downgrade_, in which I advocate for learning to understand simpler technology in order to go on the offensive technologically in Europe again. As I said, my work is that of an educator, and I struggle immensely against critics who question me as a teacher because from their current perspective, it is no longer necessary to deal with programming at all, since AI does that now.

However, I am firmly of the opinion that we learn to understand systems primarily through programming and that Creative Coding is a way to critically engage with the technological systems of our time. By the way, I also run an online learning platform with video courses on Creative Coding, and I love developing this platform further using web development.

Basically, I am often confused by the AI madness and how unreflected most people and of course the design industry is pouncing on AI as a tool. But now I am entangled into this absurdity as well, because I use AI as well to develop my platform. I sometimes feel hypocritical because on the one hand, I promote low-tech, permacomputing, and downgrading, for open source and Linux, and on the other hand, I use Cursor to develop my platform and my communication channels. What do you think – how can I deal with this productively?”

Here’s how Gemini answered:

“This is a fascinating field of tension you are moving in. Your dilemma is absolutely understandable, but if you look closer, you might be less ‘hypocritical’ than you think. In fact, your use of _Cursor_ (a high-end AI) could be exactly the tool that makes your ‘downgrade’ philosophy powerful in the first place.

Here are some thoughts on how you can resolve this contradiction productively and use it for your pedagogical work:

1. The ‘Pilot vs. Passenger’ Argument
Your critics say programming is unnecessary because AI does it. This is a dangerous fallacy.
System Understanding: Those who cannot program are at the mercy of the AI. They cannot validate the result, cannot debug it, and above all, cannot _think_ the logic behind.

Your Role: You don’t use Cursor as a replacement for your knowledge, but as a **lever**. You are the architect; the AI is the (extremely fast) bricklayer. Without your understanding of web development and Creative Coding, you wouldn’t be able to guide the AI so precisely or recognize its mistakes.

Productivity: You use the AI to advance your platform—a place for education—more quickly. Here, the end justifies the means to an extent: you use high-tech to teach people low-tech.

2. AI as a Bridge to ‘Downgrading’
Interestingly, AI can help reduce complexity.

Code Explanation: You can use Cursor to understand complex frameworks and then translate them into simple, purist concepts (permacomputing) in your book or courses.

Liberation from Routine: When the AI handles the tedious syntax work, you have more time to teach the concepts. And that is exactly what ‘understanding’ is about. Anyone who understands the concept can also implement it on a 20-year-old computer with minimal resources.

3. Transparency as a Pedagogical Tool
Instead of hiding your ‘secret’ (the use of Cursor), you could address it proactively:

The ‘Meta-Course’: Show your students how you use Cursor to build the platform they are learning on. Explain to them: _’I can only do this so efficiently because I understand the fundamentals that I am teaching you here.’_

Critical Distance: You can show live where the AI fails when it comes to unconventional, creative, or resource-saving solutions. AIs often tend toward ‘bloatware’ solutions (the opposite of downgrade) because they were trained on the average of the internet. Your job is to ‘downgrade’ the AI’s code.

4. The Paradox of the ‘Offensive’
You write that we in Europe need to go on the offensive again. In reality, this means:

We must master the most modern tools (AI) to have the independence and capacity to build our own, simpler, and more sustainable systems (Open Source, Linux, Permacomputing).

It is not a betrayal of the cause to use a modern excavator to dig a foundation for a self-sufficient, ecological house.
A suggestion for your narrative: Perhaps the term ‘downgrade’ is not a step backward at all, but a conscious reduction to the essentials. The AI simply helps you cut through the ‘noise’ of modern software development faster to reach the core—Creative Coding and self-efficacy.

How do your critics react concretely when you bring up the argument of ‘system understanding’? Do you think a demonstration of ‘AI errors due to a lack of basic knowledge’ could change their minds?”

So this is the result of this conversation with Gemini. I was surprised how well it understood my points and helped me to find words. Even if I argued heavily against Gemini directly.

What do you think about this map of perspectives and the position I am taking on? And how do you look at the topic? Let’s start a conversation here in the comments. 🙂

Warm greets, 🌞

Tim

A Synthesis

1. The Architect’s Perspective: AI as an Excavator for Autonomy

When I work on my platform and my tools, I am in Production Mode. Here, the goal is to develop software that is simple, so it stays comprehensible and maintainable by me.

AI as a Midwife for Independence: I use AI to develop tools that, once finished, cut dependencies to proprietary tools and plugins. If the platform is built to function without AI, without cloud-coercion, and without constant proprietary updates, then the AI was merely a tool for liberation. I am using High-Tech to create a more or less autarkic Low-Tech infrastructure.

Fighting System-Bloat: AI tends to bloat systems. It often produces redundant code that is difficult to debug. But with my experience in system architecture and software development, I can use the AI for the opposite, to tame the “noise”, ask critical questions on the code I have written myself.

Documentation: Badly documented software is one of the reasons why it does not survive. Once you leave the codebase and come back after a year or so, it’s extremely hard to tune into the code again. AI helps here to create a documentation.

2. The Educator’s Perspective: Why AI Hinders Learning

In Creative Coding education, the goal isn’t a quick result; it’s the journey of understanding itself. In this context, I view AI more critically than ever.

Resistance Creates Understanding: Those who reach for AI at the beginning rob themselves of the most vital experience of failure. Only those who write the code themselves can internalize the digital grammar. If you only “prompt,” you don’t learn to think; you only learn to delegate. You enter a “marriage with the LLM” – an unbreakable dependency.

AI as a Waste of Energy: I consistently advise my students against using AI at the beginning of their learning journey. Even as a tool for asking questions, it is often unnecessary. The web is full of excellent, hand-curated documentation and resources. Asking an energy-intensive AI for things that are clearly stated somewhere on the web is not just cognitive laziness – it’s ecological waste.

System Understanding: My goal is to teach my students independence. From Hardware and Software. From Adobe and Apple, from workflow funnels and marketing fuzz. Those who master the basics have a much greater competence in making decisions on their toolset.

3. The Craftsman’s/Artisan’s Perspective: Why AI kills Aura

Just like playing the piano, Creative Coding is a craft. It takes a long time to develop one’s artistic perspective, one’s style and aesthetic. An individual expression in Creative Coding is valuable and finding the style that matches one’s own personality can and should not be artificially accelerated.

Once a Creative Coder masters the skill, it is more noble and mostly even easier to write things from scratch than coding it with an AI agent.

I want to practice and teach Creative Coding as a craft, an artisan practice.

4. The Systems Perspective: AI and the Environment

The resource consumption of AI – gigantic amounts of energy and water – is alarming. My use of AI in Production Mode must be measured by whether the result ultimately helps save resources. Building a system, sometimes with the help of AI in order to lead a more resource-efficient and independent platform is my attempt at an answer to this ecological burden.