There is a big divide between people using LLMs and AI and people who object. We humans always seem to have this kind of phase that we have to go through with every change, every new technology. And it's very often, that technology gets overloaded with the interest or assumed interest of the people providing it.
This is an important and necessary situation where we as human can have a moment of consciousness between the two extremes. Usually this situation will result in a crisis, that will have a cleaning character and bring us down to a more solid and reliable state (of technology AND opinion).
So - despite of the absolutely stunning technology we have access to today - it appears to be an almost religious-type quarrel. One side projects imminent doom, the other propagates the end of all our troubles, an heaven-like future - if we only adapt. Today, in summer 2026, this makes it very difficult to navigate the tech industry.
Codeberg
A few month ago I was very happy to move my hobby code projects away from Github. I wanted to get away from the big tech companies and support more independent, smaller organisations - and Codeberg.org was the perfect match for me. Today, I am a hobbyist developer, but I have been developing software almost all my life. My work today was reduced to a fraction of the development past since the nature of my job changed from engineering software to leading software engineers and products. LLMs gave me a second wind and another chance to learn and to do things I would otherwise not be able to. It was fun.
Now, the Codeberg organization and its members have decided to ban LLM-based projects on their platform. They went even as far as to see LLM-driven development as threat to Free and Open Source Software.
I was on vacation and I only saw the messages go past me on Mastodon. I have not been a member of the Codeberg.org project, so I was not part of the discussion nor vote. And now I have to live with the decision. My own opinion does not matter in this case. I will respect their decision, but since I fall heavily into the sweet spot of their argumentation, I will have to move my projects away from there again. I will keep on using LLMs to create code and solutions, I will continue to experiment with ways to improve pipelines, ensure quality and put all my knowledge and experience into developing systems and harnesses that make good solutions possible. The bubble may burst, but the tech is going to stay.
If you want to read an outstanding description at where we are today, read Kris' beautiful article on LLM Driven Development.
Using AI on my blog
This blog is a personal pastime. There is nothing special about it. I concept, write and publish everything myself here. Sometimes, LLMs help me do some research or an LLM result inspires me to write about something I would not have written about otherwise. AI helps me to code the minimal structure around Kirby. But the texts and the photos are written and produced by me.
If I really use LLMs to produce content here, it must be for a special reason (maybe to showcase something) and I will explicitly mark it as such. You will not miss that mark. If I use LLMs to research knowledge or construct guidance for one of the articles, I will explicitly put a disclaimer on top of the post.
Using AI to create other content
Very often today you already have LLMs in your pre- or post-production processes, transcriptions, filters on images and process steps on videos often make use of LLM tech without you having a chance to change that except drop it completely. If I use LLMs in any such way, I will mark and disclaim it.
I use tools like Riverside and Auphonic. While the content I produce is LLM-free, the production chain is not anymore - at least not reliably.
I will heavily avoid the production of AI-based images or music. I am a good sketcher, a musician, call myself an artist so I will generally have no use of LLMs on this behalf. Sometimes it makes sense to use LLMs to create a showcase or an argument for a discussion. As said above, I will then mark the production with an obvious flag.
Using AI responsibly
The subheadline sounds almost ridiculous as I am writing it. We are overflown with offerings of LLMs helping us out in our lives and since we are in an economy, people try to make a living out of the new tech. Okay, some people even try to exploit technology harmfully and with lots of assumptions about the future. Storage and chip prices thrive, data centers consume lots of resources. I am convinced, the technology will not go away again and we will find ways to source the technology without the mass consumption of energy and resources.
I am experimenting with local models, running on ordinary computers with ordinary performance. I try to never state the same prompt twice, I try to limit my usage as much as I can, either by degradation (using more light-weight models for simpler tasks) or by not using it at all. I use frontline models, too.
Models are a controversy - training them seems to ignore copyrights and the work of authors. IMHO we do not have the necessary mechanisms in place to cope with that. On the other hand, humans have always been using other people's work to get educated and then create their own stuff. We all are standing on the shoulders of giants.
Closing words
So there you have it. My AI/LLM disclaimer. Probably I am going to post a short form of this on a static page (not a blog post). If you want to hear what more inspiring people have to say, here's a YouTube video of The Spirited Man and his AI Manifesto.