Join 35,000+ others who get my thoughts on product, tech, and crypto — in as few words as possible, no more than twice a month.
Dear subscribers,
Imagine creating an entire painting by typing a few sentences.
That’s exactly how Jason Allen, an entrepreneur, created AI art that won first place in a recent competition. The internet exploded after hearing this news, with some claiming that Jason joined “a marathon and drove a Lamborghini to the finish line.”
I’ve been playing with AI art myself and it’s mindblowing. Let’s dive into:
What can AI art create?
How does AI art work?
Will AI art help or hurt artists?
I’m looking for sponsors to support my writing and keep this newsletter free. As a sponsor, you get:
A short writeup at the top of a post with a picture and CTA
Promotion to my audience:
Email: 31K subscribers, 25K opens
Social: 86K across Twitter and LinkedIn
Audience: Mostly tech professionals, founders, and investors
Past sponsors include web3 and creator companies such as Polygon, Nas Academy, Acast, Highlight, and Manticore Games. Fill out the form below if you’re interested or reply to this email if you have any questions.
What can AI art create?
AI tools like DALLE-2, Midjourney, and Stable Diffusion can generate detailed images from text prompts. Here’s a sample of what you can create:
1. Create images from text
You can create your city as a scene from a Studio Ghibli movie with this prompt:
still from Studio Ghibli movie 'alone in san francisco'; very detailed, focused, colorful, antoine pierre mongin, trending on artstation ; 8k
2. Create images from rough sketches
You can transform a rough sketch into a detailed image with this prompt:
A distant futuristic city full of tall buildings inside a huge transparent glass dome, In the middle of a barren desert full of large dunes, Sun rays, Artstation, Dark sky full of stars with a shiny sun, Massive scale, Fog, Highly detailed, Cinematic, Colorful
3. Extend images beyond their borders
You can extend the borders of images like the original Girl with a Pearl Earring painting using a few text prompts.
How does AI art work?
Here’s the “explain like I’m five” version:
Billions of images are scraped from the internet. These images, along with their text descriptions, are saved in a database. Here’s a data explorer for “Ghibli”.
The AI model uses this database to train through reverse diffusion. Diffusion adds noise to an image (from the dog to random pixels). Reverse diffusion turns noise back into an image. So if you type “Dog on the moon” the model knows how to mash “dog” and “moon” images to create a single piece of art.
To learn more, check out this thread or watch this video. To repeat the obvious:
The AI model trains by using images from other artists without their permission.
As you can imagine, this has some ethical implications.
Will AI art help or hurt artists?
Here’s the bull and bear case:
Bull case: AI will lower the barrier to entry for artists to create digital art
By letting anyone create art from simple text prompts, AI will dramatically lower the barrier to entry for new digital artists.
AI will also help existing artists become more productive. For example, the AI Photoshop plugin below lets artists quickly generate an entire landscape from their drawing and a few images:
Bear case: AI will hurt artists’ livelihood and plagiarism will be rampant
Existing artists don’t get any recognition or compensation from new AI images that are created based off their work:

I think AI tools will help artists long term
Let’s look back to the invention of photography in the 19th century (source):
When photography was invented, some artists used it as an aid for their paintings. Others refused to touch the tool at all.
Photography democratized people’s access to images. The middle class could all of a sudden take family portraits with the click of a button.
Traditional artists moved to impressionism and modernism to differentiate their work from photos. A new breed of artists made photography an art form by caring about composition, color, light, and much more.
I think we’ll see a similar pattern play out for AI tools:
Some artists will use AI to craft better art pieces. Others will shun these tools completely.
AI will democratize people’s access to digital art. Anyone can now create great art pieces.
Traditional artists will start to explore new movements that AI can’t easily replicate. A new breed of artists will emerge who use AI natively to create masterpieces.
One thing is for sure, AI is here to stay and will dramatically transform not just art but multiple industries (e.g., fashion, architecture, programming, and more).
Creators should keep an open mind and give AI tools a try to avoid getting left behind.
This was a fascinating article, Peter, thank you!
It seems to me at this stage that the AI comes up with some really interesting art. I looked at "a dog looking at flowers, by van Gogh" - they were pretty cute!
I hear what you are saying with your comparison to photography and my sense is that you are "right on the money!" That being said, I wonder if you have a sense of #3 and could elaborate on it. It's easy to see the new AI artists and their creations, but what I am more curious about is what artists could explore that the AI can't do. I'd love to read what you have to say about that. Thanks! :)