I perceive AI from an engineering perspective, as I use this tool for specific purposes, the final effect of which is tangible/visible. In addition, such a perspective allows me to define a kind of boundary between human/my own capabilities and the competencies possessed by AI. It is also a kind of calculation of the implementation of the process, which helps to assess: how many things I am able to do on my own, and how many I need support (at different levels) in the context of the time implementation.

There are things:

  • that I can do immediately;
  • which it will take me some time to make;
  • which it will take me a significant amount of time to make;
  • which I could learn “from scratch” by spending a lot of time;
  • which I can’t do on my own.

The time factor in this case is a relative value: does it mean an hour or maybe a few days? Is a lot a month, or maybe a year? It is difficult to establish a clear measure that would fit all people. It is even very unlikely (I do not use the word “impossible” on purpose, because I have already seen for myself several times that “impossible” can turn out to be very “possible”).

By adopting such a structure, it is quite easy for me to adjust the use of a tool that is AI.

During several months of cooperation with Chat GPT, I developed a scheme that, on the one hand, allows me to develop my capabilities, and on the other hand, imparts competencies to AI that I do not have, in order to ensure quite efficient task completion.
Why “quite efficient”, as it turns out, even with the support of such an advanced tool, the human factor still generates delays related to the search for the most appropriate solution. However, on the other hand, the search is exactly what it is all about, in order to find something you need to spend a little more time – by the way, I have defined a kind of “Paradox of Development” (all that remains is to develop a mathematical justification 🙂

So much for the introduction, ad rem.

Instilled with the idea, I decided to develop some kind of visualization for the music, specifically for the album “News From The Greylands”, which is, unlike “Blues From The Couch”, a loose collection of songs, “News…” has the certain hallmarks of a concept album that tells a story.
Anyway, this was the idea from the beginning, to give coherence and chronology to the events described in individual songs. However, after completing it, I noticed a few gaps that require some kind of justification, and that’s why the best way seemed to me to develop a visualization that would give proper meaning to the entire “News from The Greylands” album.
In a situation where shooting music videos for individual songs was completely out of the question, I decided to create a set of illustrations. And this is where the AI tool has proven to be incredibly useful. I don’t have artistic skills, when I try to draw a dog, sometimes a horse comes out :)) In any case, making the illustration was completely beyond my capabilities.

It started quite ordinarily:

– Chat, draw a picture of a car driving at night, in the rain –

And Chat, of course, generated an image, quite neat, basically I couldn’t find any partical drawbacks.
Then the question arose – how much I want to transfer control over the implementation of ideas. It required me to think for a little bit of time… It was then that I defined the division presented at the very beginning and began to act more methodically. I defined detailed descriptions of illustrations, which I “collided” with AI comments on the basis of dialogue – a kind of exchange of views. The subject of discussion was everything, from color, mood, time of day, placement of objects, to small details and inscriptions.

It took three weeks to create 13 illustrations. Is that a lot? That’s right :))

Why illustrations and not, let`s say, an AI-generated music video (which is, after all, a solution that’s becoming quite common and brings some pretty interesting results) ?
I have a bit of a fascination with (still) images. It’s like visiting an art gallery – you can focus on, or even delight in, finding details that are so imperceptible at first glance.
The next step was to bring the illustrations to life by using specific video effects, generating a change of space and a play of light. Adding lighting effects or image distortions is, in a way, a kind of hint on what to pay attention to, or simply a subtle dynamic that allows you to endure until the end of the song.
The Chat immediately offered its help, providing descriptions of specific solutions. I don’t know why, but I decided to do this stage myself (maybe the question about the level of control came back?) based on the free version of VSDC Free Video Editor.
It took another month, and it consisted in getting to know the interface and then the capabilities of the program.

And in the meantime… I improved the musical sound of the songs, i.e. mix & mastering, finding imperfections, eliminating interference, generally working on better quality of recorded music.

And so, after three months of intensive work, I can happily present the final result. As it usually happens, I could still improve something. Let’s call it small imperfections have already been noticed, but as someone taught me “Done is better than Perfect”.

And what about the plans – I’ll probably focus on “Blues From The Couch”, but more as separate visuals of individual songs.

And further… T.B.C.

I invite you to take a longer look, preferably in the evening, when most things can be put aside. Probably one is not enough, so let it happen – a few evenings also sound good 🙂

See You : ))

Tom