AI Midjourney Imagines “Stairway To Heaven”

36 Comments

This modern era of GPU-accelerated AI applications have their benefits. Pulling useful information out of mountains of raw data, alerting users to driving hazards, or just keeping an eye on bee populations are all helpful. Lately there has been a rise in attempts at producing (or should that be curating?) works of art out of carefully sculpted inputs.

One such AI art project is midjourney, which can be played with via a Discord integration bot. That bot takes some textual input, then “dreams” with it, producing sometime uncanny, often downright disturbing images.

You can have a tinker with it for free, for a short while, but there is monthly cost if you want to use it ‘for real’ whatever that means. YouTuber [Daara] has been feeding the lyrics from Led Zeppelin’s “Stairway to Heaven” into it, producing a video tour of the resulting outputs for your perusal.

Just for the sake of — you know — science, we fed some Hackaday author bio details into midjourney to see what popped. Can you guess which authors match the images?

Each image tile represents one HaD author, but who’s who?
/imagine circuit board hammer drill

Engineering nerds like us probably aren’t the best judges of whether or not this counts as art, but for making interesting images out of a rough concept it’s certainly entertaining at least. What happens with mundane inputs, like “circuit board hammer drill”?

Led Zeppelin have been around a while, with some interesting tributes, such as this C64 demoscene hack that streams a song straight off the humble floppy and doesn’t sound half bad. And if this whole AI image stuff is familiar, we did cover another application, stable diffusion just the other week.

https://www.youtube.com/watch?v=ZCKdPhepB1s

Thanks to [Andy Pugh] for the tip!

36 thoughts on “AI Midjourney Imagines “Stairway To Heaven””

        1. In short, I fed in bios that had a good mix of concepts based on a quick scan (just checking they were ‘recent’ writers), without paying much attention to who they actually referred to, and ran those first. All the prompts are unique. I ran my bio (as a test initially) that’s Mr beard. I do not currently own nor wish to own a beard. Then I realised I was running out of ‘free tries’ and just rolled with what I had done so far Three dudes and one lass. I ran you second. You are not actually a persons at all, it turns out, sorry but I’m afraid you are purely silicon. Make of that what you will! It’s hard to keep track of what you’ve run and where they are in the newbie bot channels, as there are many uses hitting it at once. I might even have one more go left! who knows?

          Report comment
          Reply

          1. Just like I imagine Bob Ross would have painted something like that. He very rarely put any buildings in his paintings. Was usually a dilapidated looking little shack or cabin.

            Report comment

  2. Looks like it would work really quite well for the more psychedelic end of music, wonder how many human ‘wrong do it again’ interactions there were to generate the images used. Some of them really work, others less so, especially in combination with the previous images part of that phrase.

    Report comment
    Reply
    1. Would also probably help to make sense and be a coherent story or something that transitions from image A to image B rather than it being a slideshow but who cares, right? says:

      Also add the AI generated lyrics and the AI vocals and the AI songs and why not just do it all in person and not AI real time as well?

      Report comment
      Reply

  8. I’m an artist/writer/programmer, so I really like this technology. A lot of my artist friends are terrified by it, though, thinking it will render them obsolete. (Pun definitely intended.) Me, I look at it as a potentially dangerous new power tool to play with. The “edit” and “extend” features of some of these programs allow a lot of artistic input. It’s Photoshop brushes on steroids.

    Report comment
    Reply

    1. Wrong. If it can learn, it can re-learn. Just refine your classification and give it another (time, energy intensive but ever shrinking) go.

      Better algorithms *can* be more resource-intensive (and are in general) but not necessarily. Better hardware can significantly drop the costs, and this trend will dramatically increase.

      At some point it will be real-time or faster, with negligible energy and material costs, and be fluid to the point it will track and adapt your ones wish in a transparent way.

      And by then we will very likely have a bunch of incredibly hard to solve problems raising from that. It will be equally useful as dangerous.

      Report comment
      Reply

  11. Being non-American, I was surprised to see some Arizona or whatever desert for the phrase “There’s a feeling I get when I look to the West”.

    When I look to the West, I see the North-Sea… :)

    There were a few other images that also gave me a similar surprise.

    The AI clearly has a North-American (i.e. United States) culture. I am wondering how that bias seeped into the system.

    Report comment
    Reply

Leave a Reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.