Tag: AI
Can AI make art?
Short answer; No, but a person can make art (even quite good art) with AI, just as a few select people can do with a paintbrush or Photoshop. But perhaps AI in art is to be seen more like a musical instrument rather than a ‘tool’, as this long article (converted from a talk) suggests. Another take comes from cartoonist Matthew Inmann, who touches upon some of the same themes and dilemmas, and, funnily enough, also uses the metaphor of music at one point. But he likens AI prompting ‘art’ to pressing the ‘demo’ button on a cheap Casio music keyboard. Both are fairly long, but worth a read if you have time and interest, and both go beyond all the predictable knee-jerk reactions to AI ‘art’ and AI versus Art.
As mentioned earlier, I actively hate 99.9% of AI ‘art’ and visual slop that the net is so hopelessly full of now. But that still leaves 0.1%, and there are some signs of a new genre emerging. Here is a selection of AI-based Instagram accounts that I enjoy: Is it art? You decide..
DDR Mondbasis – Nostalgic German counterfactual reports from the future; love it!
Niceauties – One of the grand old accounts, surreal but with a clear theme and distinct style.
The Strangest Fleamarket – Visually interesting, bizarre creatures and characters.
Moss Carpet – It’s a vibe… Folksy and slightly unnerving tableaus from elsewhere.
Voidstomper – Enter at own risk, bizarre and scary
Aim not here – ‘Excerpts from interdimensional journeys powered by the human mind’
Bonus: Also check out Fellowship a place with many AI-positive artists to explore.
Dreamscape
Observations on AI, July 2025
When people criticize AI in general and LLMs in particular, 90% of the time all I hear is “I’m bad at prompting.” Many laymen still treat ChatGPT as a glorified Google search. Custom instructions and asking directly for what you want will get you far.
Just ask. That is the simple key to a lot of good use of AI. If you don’t like the responses, be more clear about what you want. If you don’t know what to ask, just ask the AI to help with the questions or have it literally question you before answering. Easy. Make sure to give it something to work from; context is everything.
Remember back when we got flashlights in our phones? We suddenly realized how often we actually have a need for a flashlight. It feels a bit like that with AIs; you suddenly realize you have many more questions and topics to explore when you can just ask and get good answers without much friction.
While some of us find great value in AI, there is a clear and growing backlash happening, especially when it comes to visuals. We grow tired of visual gimmicks faster and faster (remember Ghibli-style and cutting up glass fruit?) AI/LLMs are not going away, but the resources needed for running all those servers are vast, and growing like crazy..
When people share long threads where they have a ‘dialogue’ with an AI, more often than not it is just as boring as hearing long and intricate stories about their dreams and nightmares. It’s too specific and personal; without the proper context, it’s close to meaningless. And did I mention it’s also boring?
IBM chairman Thomas Watson is perhaps most famous for the (false) quote I think there is a world market for maybe five computers. Some say half-jokingly; Watson was only off by four, but in an AI world, the five can be said to be the handful of frontier AI models fighting for dominion: ChatGPT, Gemini, Grok, Llama, Claude and perhaps a Chinese wildcard.
If a person claims they know what is going to happen with AI in the next 12-18 months, run away as fast as you can. No one knows this in any meaningful detail; a few may know the broad strokes. And those who know the most are busy building it. Far too often, the predictions remind me of the blind people and the elephant, except AI may very well be the whole room rather than just the elephant in it.
(And yeah, I made this image with ChatGPT, or just Chat as most of the younger people call it now. My guess is that OpenAI will rebrand ChatGPT to Chat later this year, they already own chat.com..)
Me + AI
I believe most of us try to boil AI down to something we can handle mentally; “Well, it’s just advanced autocomplete, it’s just a stochastic parrot, it’s just another tool” etc. etc..”. Sorry to say, it’s much bigger than that. We don’t really know what it is, and people who claim to do so are often dumb or too smart for their own good. Sit down, be humble. AI is not another technology, it’s different. Consider this.
Intelligence may no longer need to be localized within individual minds. Knowledge, once thought to be stored and retrieved like files from memory, now appears capable of being dynamically assembled from vast latent structures. Reasoning, long assumed to be linear and stepwise, is increasingly replaced by parallel, probabilistic processing. And perhaps most provocatively, meaning itself may not require introspection. It can arise in context—without any internal monologue or conscious reflection.
I enjoy using LLMs every single day. For one thing, they completely replace the rather useless 2025 experience of Googling things. I often need to have things explained to me, and LLMs are great at doing so at any given length or complexity; how does this author compare to this one? Make a list of lesser-known movies based on these three unknown movies, give me the five main points from this book, explain this obscure meme. When ChatGPT was upgraded with memory, it was a real game-changer; suddenly, the answers could be based on earlier chats and things you explicitly wanted it to know about you and your way of thinking. And you can have some great conversations on philosophy, personal development, problem-solving, and any other ‘deep’ topics. Just treat it as a dialogue with a (very) smart friend; sometimes they are wrong, sometimes the thinking is flawed, but you will often be a bit smarter afterwards. I think I have some fairly smart friends, and they will quite often disagree on fundamental questions, which is sobering to observe. I also make an effort to use multiple LLMs as a matter of principle, switching between ChatGPT, Claude, Gemini, and Perplexity. Some may say it’s like going from Coke to Pepsi to Jolly Cola, and it may be so. But it often brings slightly different perspectives even if the frontier models feel rather similar. And one of the worst aspects of AI is the feeling that everything becomes average. Just like our physical spaces are evened out by global Internet culture (ever heard of ‘Airspace‘?), it seems like all writing, thinking, art, etc. converges towards some kind of collective global average. I am not a non-AI purist by any means, feels like there will be a solid future demand for people with strong personalities and original perspectives. So much AI slop out here. There is a big Danish AI Facebook group, and it is obvious that much of the writing there is done by machines, peppered with ghastly AI images. And it feels so bad; sometimes I have an actual physical reaction when looking at it. And don’t give me the old “Well, most people can’t tell the difference anyway” argument; I just don’t like machines acting like people, and people acting like machines. And if we are meeting the machines half way, then it’s a full win for the non-human side.
I’m not using much AI when writing posts here. It just feels off to do so. Full disclosure: the WordPress spell checker has AI built in. And I may use AI to brainstorm ideas and topics, but the actual writing is done 100% by me, flaws and all. Many of my notes are collected in NotebookLM now, it’s interesting to chat and enquire into one’s old thoughts. Oh, and I really, really hate 99% of the ‘art’ done by AI image generators, so there won’t be much of that on lovetobuild. And you can explore this site using an AI. Just so you know.
... 