It is you that makes the picture!

And not some artificial stupdity!

Well, let me give credit where credit is due: What this is, is an immensely capable image search engine that has been combined with an image processing platform that is a master at recognizing the searched item and fitting it into a composition by deleting and blending pixels. So, not quite so stupid, after all... Very capable little technician actually. But intelligent? In the sense of being capable of making decisions? Please!

It is a completely new way of making images - that is for sure. Instead of drawing something you write something. So, you are actually writing the picture, not drawing it. But you and only you are the one that makes the stuff. And the output will be only as skilled as you are in translating words into images. Which needs two things: You need to have a very substantial English vocabulary (I do not think that these platforms give quite the same results in other languages, given how the image search that they conduct will search mostly among English language captioned or tagged images since the English language constitutes something like %50 of the internet). And the second thing that you need to know is what is out there in this world - in other words you need that globally fast fading quality: Culture!

The best way to explain this is to actually quote myself, chatting with Sebastian Willis about what he does with text to image:

"Let me tell you how I know that: I know that because of you. If I hadn't seen your stuff I wouldn't have given this whole text to image malarkey even a second thought. Because when one googles "Midjourney" and looks at the images - it mostly looks the same, doesn't it? And quite frankly, it mostly looks rather awful. Lots of orange sunsets for some reason... ;-)

So, it isn't about the platform, it is about your imagination. But also, it is about how extensive your vocabulary is, whether you know the correct terms for things in art history, design history - whether you know "history" in the first place, whether you know geography, folklore, mythology, etc etc... all these things, you know? Whether you know the names of different painting and drawing techniques, different materials, different types of lighting, and and and... And very few people have that sort of cultural grounding and vocabulary which, now that I am doing it, I see that you need before all else. Since the whole thing is text based, you need to know the names of things and you need to know that these things actually exist. Only if one has that "culture" can one proceed to the actual creative activity itself. And there one has to have imagination, of course - which you have in spades, as we all already know. But without the cultural grounding itself (which you also have in spades, obviously) it doesn't matter if it is Midjourney, or what I use over here, the result would probably also just be "orange sunsets over cyberpunk cities"..."

I have been going nuts with it over the past couple of weeks. Oh, the things one can make! The places it lets one's imagination roam to. Provided that you have an imagination that can do the roaming in the first place, that is. But the way that this thing works is such that an imagination by itself will not get you too far. You need to know the names of things. Well, you need to know that they exist in the first place - and then you need to know what they are named. Material, lighting, color, composition, style and content - you are the one who decides everything. Which means that you have to know what specific materials you can choose from, what sorts of lighting you can bring to bear, what technique should this thing be made in - watercolor or 3D or hyperealistic photography? Carrara marble or tarnished brass? Colors? Ochre Yellow or Aquamarine? Composition? Symmetrical or asymmetrical? What about negative space? Style? Chiaroscuro or Flat Design? And then, of course there is the content, a list for which I will not even attempt to make! All of these things have names. That one needs to know!

Now, some will use this far more experimentally than I am doing, looking for unexpected results. Sebastian, my guide in all of this, says he is one such. More power to them, I can see what an imagination amplifier this thing can become in the correct hands - or rather at the disposal of the correct brain, because it is true that while you are generating the images you do come across remarkable stuff that emanates from your directives but takes them to really unexpected places.

Alas, I myself am a control freak when it comes to visuality - trained professional designer, you know? Control is our middle name... So, I end up writing highly extended prompts which I keep refining and refining as I make iterations. Which means that this is not exactly a cheap indulgence. While the basic fees for most of these platforms sound reasonable at first glance, this is actually an expensive thing to do, since in order to get the perfect result you need to make lots and lots of iterations of the one image that you are after and the prompts for these iterations get bought separately. So, you will see someting like 500 prompts for $5 (which comes on top of your basic subscription) and think that this is no big deal. Well, let me tell you - it is a very big deal actually! 500 prompts is what I sometimes use for a single image, sometimes even more. Luckily, I have had this premium subscription at Freepik for years to use for teaching resources. And they have developed an excellent text to image editor that lets you generate high quality, high resolution images and they give premium users 4000 prompts per day. Each prompt generates 4 iterations, so that is 16000 images. Have I ever used all 4000? Not yet, but I have come perilously close.*

I am reaping the full benefits of being a "thingmaker" here: Not ever having been hampered by things like styles or missions or statements, I do whatever comes into my head. So, my old idol Rachel Ruysch, in whose honor I even made an avatar outfit in SL, gets to go to Venice - a place where she would never have been, never having traveled outside of the Netherlands, as far as I know. A medieval book of hours ends up becoming a habitat for small creatures hiding among thistles. I make bestiaries and paper cut cities, go back in time to 1932 where it seems to have rained a lot, cats sit in fishbowls and a fox and a frog enter an unlikely marriage from which emerges a FoxFrog. I am not a good tale maker, my imagination always works in single frames somehow, but I wish someone would make up a tale for the FoxFrog... Mushroom houses, bonsai mushroom forests, bisque porcelain figurines (both human and feline) to live in a Fragonard type garden.

But then, when I want a change of pace, I also make some rather slick stuff with a view to sell it commercially as designer resources on design resource portals - I have even applied to a few of them. Product placement shots, food still lives, home interiors - all very hardcore commercial imagery. It is a lot of fun to make these, and if I can pull it off, the extra income wouldn't go amiss either, of course.

But regardless, of how I go, I come back to the same point. As far as I can see, you cannot really do much with text to image unless you have the cultural grounding that has given you the vocabulary that will allow you to find stuff through the search engine. If it is design resources - then you better know all the terminology for product shots as well as their usage, or what one should look for in interior design shots or food photography. That there is such a thing called "copy space" that is vital for the poor soul who will buy your image and use it somewhere. And as for all the other stuff - how could I make any of it, if I didn't know about medieval bestiaries, books of hours and Rachel Ruysch? Or that the architecture of the 1930s was called Art Deco. Or mushroom houses? Or not yet told foxfrog tales. Would I ever even think of such a thing if I didn't know Grimm's Fairy Tales? Where there are no foxfrogs of course, but all the things that then lead one to foxfrogs...

And you can see some of what I concocted so far here: https://www.trivialthingies.com/portfolio

---

* Yes, I know that there are also platforms out there, such as Adobe's Firefly for example, that let you make unlimited prompts for free. For how much longer will Adobe let people get away with that, I wonder? But the problem for me is that these platforms (including Firefly) do not have the overall image quality that I am after, and also they only let you download medium sized images.