So far I have built this website on JouwWeb because they have a good and simple website editor - that makes it possible for anyone who has never programmed to build a beautiful website. Ideal to try out if blogging is something for me.
I already had a long-running fascination with Japanese art forms such as bonsai and Japanese gardens, but generative AI has led me to delve into the more graphic art forms as well. Due to the isolated location of the Japanese islands and the resulting unique culture, Japanese art has its own signature. While influences from China played an important role in the earlier centuries - and even influences from India via Buddhism - Western influences can also be recognized in Japanese art in recent centuries. On the other hand, Western art has also been influenced by Japanese art on several occasions in recent centuries. A good reason to see what generative AI can do with it.
Sometimes I feel like I have too many hobbies - or too little energy, but the former is worded more positively. Now that spring is in full swing, my garden needs my attention again, at the expense of the time I can devote to this website. That's why today: gardens. No pretensions, no rankings, and just for fun.
Before we start writing a story, let's do some small exercises to see if we understand how to communicate with a chatbot. I always admire people who know and can do Latin. I would love to use this in stories with spellcasting . Can the chatbot help? (I have translated the Dutch text to English for clarity.)
So far I've been practicing generating images for a story I already have lying around. The downside to that is that during the writing I've already figured out what my main characters look like and the world looks like; Conjuring that out of an AI is quite a job and may not even be possible.
In the previous blog I started a list of Printing on Demand websites. These platforms allow your customers to place orders, forward them to their manufacturer(s) after payment and have the product shipped directly to your customer. My list of Printing on Demand platforms was not finished yet, so this week the other well-known providers. There are a number of striking differences:
Although my first reason to generate images with AI was motivated by my desire to illustrate my own stories, it was soon so much fun that I also started creating for the sake of creating. It only costs money, so it would be nice if it also brings in money. Unfortunately I'm not entrepreneurial: I don't scrape together my money to have all my creations printed and then sell them on a market. Fortunately, the internet offers a solution. There they invented 'printing on demand': printing something only when it has been purchased.
Winter is a nice time to curl up behind the computer. Inside it is pleasantly warm, especially after a brisk morning walk. The garden is quiet and requires little maintenance. All the time to generate image after image on my favorite AI platform NightCafe. Sometimes fanciful prompts that take you to faraway places in the past, future, film or literature. Sometimes realistic prompts to recreate the here and now.
The rules on copyright and plagiarism are much older than the current rise of generative AI. We define generative AI as artificial intelligence that can produce texts, images or music, whether or not based on instructions from people. As of 2022, there is controversy over three themes, namely copyright to the writer of the prompt, copyright to the creators of the images / lyrics / music the AI are trained on, and plagiarism of existing works by the AI.
Of all places in the world there are pictures on the internet, but of some more than others. Think first of all of the places where many tourists come, and also of the places that play an important role in films and series on TV. What's on the internet can also end up in the AI, but I've already found that only the truly globally famous places are accurately represented. Below is a sample from very well known to apparently not so well known. I have been guided by songs and events. There is no conscious pattern to its distribution around the world; I scraped together the examples from five months of experimentation.
What art is seems to mean something different in every century. Nowadays, even a toilet bowl can be art, if only it is placed with that purpose. New genres and other art forms are emerging, changing the idea of the function and nature of art. Relevant here is both the rise of digital art and the rise of AI in making art.
If you do it on purpose, it often becomes less beautiful than if it happens by accident: creating strange animals. They sometimes seem to come from another planet, or to be a cross between two familiar animals. It is possible to create this 'accident'. I personally note three options:
With all those options of AIs, choosing the right prompt can become quite complex. Overall, they yield the same results, but it's the details that differ. I didn't have the time to try them all myself. I searched for sources for the best prompts, and those sources had sources too... so here are a few sources that provide tips on how to create the best prompts.
Meanwhile, the market for text-to-image AI has exploded. More and more groups are trying to get their share. I actually wanted to write an article about tips for making good text prompts, but I couldn't see the wood for the trees anymore. So here's an attempt to sort things out.
The big advantage of using AI when creating images is - firstly - the large amount of details, which are - secondly - drawn at high speed. Of course there are differences between the AI. For example, the first Dall-e was simpler than Dall-e2, and could generate less detail. Midjourney and the Stable Diffusion algorithm raised the bar even higher than those two, both in terms of detail and quality of the image. That also means that you can go for an older AI if you consciously choose to make a simple image.
Music is an endless source of inspiration for me and many with me. I regularly prompt the AI in my favorite style to turn this inspiration into images.
To pick up the thread of my previous blog: I have continued making portraits of inspiring older faces. These ladies here were generated in Stable Diffusion version 1.5 with the text:
With my first AI-generated images, I mainly tried to find out what it knew and what it could do. Knowing that the different AIs have been trained with images from all over the internet gives you high expectations, but you really don't know anything yet. My traditional Northern Dutch name 'Harmanna' immediately turned out to be difficult, because it is not given much, even in the Netherlands, and is therefore hardly known on the internet. The AI thought of everything (dishes, villages, inscriptions) but not a woman's name, so I had to help a bit.
A lot can be said about copyright related to AI. As far as I know, no lawsuits have been filed in the Netherlands, but they have in the US, where some artists have sued Stability AI, Midjourney and DeviantArt. In the UK the same threatens to happen against Stability AI by Getty Images. In the meantime I've delved into the subject matter a little deeper, starting with the help of YouTube lawyer LegalEagle who made an informative video about it.
I can never stop making up stories, but it's a lot of work to also write down those stories, or draw them on paper, or convert them into a working game. The AI that are now emerging have the potential to accelerate this process. I started with a lot of questions: what is possible, can it generate my own style, can it make all the details I want, how controllable is it? More importantly: what is my own style? And: which AI do I want to use?
Those who read or watch a lot of science fiction may be shocked by all the recent reports about AI: artificial intelligence that can play chess, that makes art, understands text, can write by itself…. Let me reassure you: for the time being, AI cannot yet take over the world. My own experience with AI generated images makes it clear that AI does not give perfect results. The various text-to-image algorithms available remind me of a three-year-old child. It understands words, but the connection between the words quickly becomes diluted. If I ask for a black circle and a blue square, the chance of getting a blue circle is just as big as the chance of getting two black squares against a blue background. These AIs can't count either!
To start with the elephant in the room, the big problem with AI-generated art is the potential for plagiarism. Let me start with an obvious example: Vincent van Gogh. His unique style is recognizable out of thousands. His self portrait is world famous. I am quite certain that any self-respecting AI algorithm that wants to generate 'art in general' is also trained with Vincent van Gogh's paintings.