Cycling art using AI

Playing around with Midjourney; thought some of you might find this interesting.

Prompts: bike, alpe d’huez, sunset, Van Gogh


I’ve been experimenting with Midjourney. So far my prompts have returned nothing but rubbish. I’ve even tried copy > pasting prompts from other users. Not sure what I’m missing…

Your image is brilliant. Have to say that Midjourney has caused quite a stir in my line of work.

Yes, my other attempts fell pretty flat, but this one hit the mark.

this is amazing. So does this make you the artist that you put together the prompts that generated the piece? Is the artist the person using the tools or is the artist the paint or the paintbrush or the canvas or the company that formulated the paint?


Yes, I didn’t poke around IP side of things before signing up, but it raises interesting questions.

Here are some of the other “variations.”


Would you mind elaborating on this a bit? Trying to understand how people are learning to adapt (or not) to some of these new tools is often more interesting than the tools themselves.

I work in the creative industry. Graphic design / visualisation / concept creation.

This tool is causing quite a few questions to be asked and a few people are sitting uncomfortably.

Why? On the ‘face’ of it, this tool is able to generate image-based concepts in 60 seconds. You can ask for variations on a single theme and upscale that concept. If you’re a client who has previously paid/employed a visualiser or conceptual artist, this has the potential to be a revelation. What might have taken hours or days, for a single concept, is now 60 seconds away…

or is it?

Whilst Midjourney is good fun, you have little control. Yes, you provide prompts but you can’t finesse. A visualiser or concept artist can tweak the smallest aspect to your exacting specifications. Midjourney doesn’t currently have this ability.

From a purely technical standpoint, the artwork it generates is fun to look at but, to the trained eye, you quickly realise that it’s not really there. Your brain is filling in for the short falling. Even the images on the community showcase forum and those doing the rounds on LinkedIn or Deviant Art are lose, to say the least. Yes, I could present them to you as a concept but, if you then wanted to use them, I’d have to find a way to create them. This is doing a job backwards and more often than not, it’s not time or cost effective.

I could rattle on but you’re probably bored already.

Personally, I’d love to love Midjourney. Problem is, I can’t get it to work. It clearly does work, so I’m the broken link.

As a creative tool, in the right hands, it’s going to be awesome. There will however be a backlash from clients who believe this will prove to be a money saving tool.

At the end of the day, a human will ask you questions:

  • What type of bike?
  • What time of day?
  • What is the weather like?
  • Is the rider a man, woman, child, etc…
  • Are they racing, overweight, is this meant to be funny, etc…

Simple things like this. They’ll be invested in producing an asset specifically for your needs.

Midjourney will, I have no doubt, achieve deeper questioning in the future. What it can’t do is add that human aspect. Has it ever ridden up Alpe d’Huez in early morning rain? Can it relate to how that might feel? Has it ever felt achievement? These are the soft skills that get overlooked which a human being brings to a project.


I also work as creative director in design studio and lecturer at academy of design. And in 2 months I am preparing the exhibition focused on AI images with main focus on bikes. I have dozen of images already. It was pretty fun small project.

The Ai, especially midjourney is pretty amazing to ping the ideas and give you spark. You can generate some amazing concepts, and if you work as a concept artist for games or movies - this is a winner that can speed up concept generation phase tremendously. When it comes to bikes itself - it generates great “emotional” images but it has a huge problem with form of the bicycle as an object.


I’m sure you guys saw this posted by cycling tips. It’s kind of interesting to see things created that don’t make sense as we know the world, but look like they could.


That looks like it hurts.

Thanks for the deep dive here.

This is great. It makes it sound like its not the job killer some folks may think it is but rather a tool to use as a muse when you are stuck with a blank piece of paper or 20 ideas that are all too related and need something totally from left field.

1 Like


It is unnerving to know that this ‘tool’ really is out there and it is going to get better at what it does. At the end of the day, it’s just a tool.

Much like the computer and programs such as Photoshop, they marked progress, not the end of something.

Designers by their very nature are inquiring. I’m willing to bet that Midjourney won’t even end up being used for the purpose it’s being used for now. People will dig into it and other factors and options will come to light.

The designers who taught me used to count off line copy. They used Letraset, Cow Gum, Magic Markers and blotting pads. The idea of a computer must have chilled some of them to the bone. As with most things, humans adapt and we find a new way.

Will it be a better way? Ultimately, the client will decide. I do however stand by my point that it won’t be a money saver. Those people who can really make this ‘tool’ work will create their own markets and charge accordingly.

1 Like

Cool - I would be interested to see that gallery when finished.

Not all of the prompts landed, but the emotional resonance of the image I posted is definitely what surprised me most.

Uploading: 81E953A7-3DBB-4187-B430-8372F59E7C31.jpeg…

I made a few using dalle. “Male cyclist rides strong up a paved path next to a scenic lake, oil painting”

Those are some examples done yesterday as a small sneak-peak. I am focusing on the variability of the images and visual exploration in different styles.


Have to admit, those images are really interesting. I really like numbers 1 and 5.

Do you run the images through Photoshop or are the straight out of Midjourney?

Those are raw images out of midjourney. No human touch, other than prompt.

I have also examples from Dall-E but do not like the results. When it comes to quality of an image (bicycle looks like bicycle) dall-e is great but the results are completely “soulless”. It’s like a stock images generator for me.

Ocean road cycling Dali

1 Like

I would buy that cat-bike as an art print.