OPPAI ORACLE

in its commercials base44 paints its audience as a bunch of artsy 20-somethings with big dreams who use the ai website builder as an alternative to learning coding. their commercials are high quality and aesthetically pleasing to signify their solidarity with artists. but young artists are the LAST people who would use base44; generative ai is a threat to their careers, and the indie web movement is spearheaded by artists with no prior coding experience, but took the time to learn to oppose pro-ai social media. i think it's supposed to be reverse psychology for the artistically challenged vibecoders who actualy use the service. "look! vibecoding isn't lazy! look at the rad youngins who do it!" majority of people use generative ai for laziness or misdeeds (cheating, porn, shocking political content) which damages the "magically helpful" image ai companies want to construct for their services. ironically, ai companies need artists to convince consumers that their services are high-quality

Pretty funny. They always co-opt the image they aren't and try to sell it back to you.

I will say, however, that "vibe coding" is a fuzzy concept and I sometimes have to ask myself at what point I don't approve of it. Do I think it's bad to ask an AI how to do something on a page? Like "how do I have a table in the page that stretches to the edge of the screen while keeping the rest of the page bounded in 800px?" I had this problem and I resorted to using ChatGPT to answer it for me. However, I did this after looking for other pages that did the same thing first and hoping to look at their source (to no success), searching for it on my search engine of choice (and couldn't find an answer... maybe because it was hard to phrase). I felt dirty doing it, but I basically saw it as an accelerated version of just searching for it through engines.

Now, this is clearly very different from just saying "design a page for me that looks like X." But is it vibe coding? It's hard to say. I would say that the important point is just making sure you understand what is going on "under the hood" so you are using it to learn rather than automating the creative process for you.

But I can't lie, there are some gray areas that make me uncomfortable. Take drawing for example: "Pure" AI art and "pure" human art doesn't really bother me. What bothers me is, for example, an illustrator having AI make a base body model and then filling in the rest with your own coloring, line-work, etc. These kind of "hybrids" are harder to categorize in my mind and thus I find them more disturbing... For example, I've considered using an AI to make a kind of dummy model of a body to look at as a reference when drawing. I wouldn't trace it outright, but it's not "really" that different from looking at stock photos of hands, feet, bodies in weird positions, etc. Which I of course always do. But I haven't done anything like this, because something about it just feels weird to me. Like I'm ceding over part of my creative process to a machine. But it becomes very hard to tell where to set boundaries here and I don't really have the answer.

What does anyone else think of this?

Amlux wrote: Mon Aug 10, 2026 3:11 am

But I can't lie, there are some gray areas that make me uncomfortable. Take drawing for example: "Pure" AI art and "pure" human art doesn't really bother me. What bothers me is, for example, an illustrator having AI make a base body model and then filling in the rest with your own coloring, line-work, etc. These kind of "hybrids" are harder to categorize in my mind and thus I find them more disturbing... For example, I've considered using an AI to make a kind of dummy model of a body to look at as a reference when drawing. I wouldn't trace it outright, but it's not "really" that different from looking at stock photos of hands, feet, bodies in weird positions, etc. Which I of course always do. But I haven't done anything like this, because something about it just feels weird to me. Like I'm ceding over part of my creative process to a machine. But it becomes very hard to tell where to set boundaries here and I don't really have the answer.

What does anyone else think of this?

Personally I think this "hybrid" work is a slippery slope as far as passion and creative projects go. When it comes to tools and inspiration, it feels like there's always more value in the friction and confrontation that comes with avoiding things like AI tools. An older art teacher might already scoff at using stock image reference photos instead of a live model. They would say that there's something in the three dimensional form and movement of a figure in front of you that can't be captured in two dimensions and pixels. Great artists of the past would learn human anatomy from corpses, and while you can say that they had to do that because other resources just weren't available at the time, there's also an argument to be made that the engagement with the earth gets leveled down by becoming increasingly mediated through our tools and references and so on. Of course to some extent digital tools are very freeing and useful and so on, but it's good to keep your engagement more authentic when you can, and thus when it comes to something like AI references or bases, the whole process, which might originally go "look at reference model in front of you -> draw model" becomes "reference model gets photographed -> photograph is uploaded online as pixels -> pixels are processed by the AI, along with hundreds of thousands of other reference photos -> a prompt is made and interpreted by the AI -> AI reprocesses all its data to present you with some approximation of your prompt" it's easy to see how the original essence of what you might have wanted to reference can get lost in this process. And at the end of the day, do you want to engage artistically and creatively with people and the earth, or a swallowed up and spit out simulacrum? This I think is the "feels weird," and the answer in terms of setting boundaries, for me, is to basically live as close to the "origin" as is possible and practical. If you can get a live model in front of you to draw, great! But if you can't, you can definitely find reference photos/videos that are more "real" than some AI output will be. You can even make/obtain and pose your own 3D models for this purpose in a free software like Blender but the real earth is always best

I think most people don't understand what an AI actually is. Regardless of all the myths and misunderstandings, "AI" is just a name given to automated (non-parametric) statistical machines.

Yes. An AI is a statistical machine. And machine "learning" just means a type of mathematical optimisation. For people who don't know anything about AI, what an AI does are basically things like this in a high level. "Learning" means finding the optimal f(x) based on the data.

So why are AI slops bad and why do they totally lack creativity (and souls)? Because an AI is a statistical machine. It is designed to give you statistically optimal (i.e. normie) answers based on the training data. Also it cannot give you more information than what the training data already contained. These are features and not bugs.

I think it is just another tool that people abuse and misuse.

By the way this is another point, I think works made with a Cloud-based AI are quite bad, while works made with a LocalLLM (basically an AI running on your computer) are somewhat fine.

Thricegreat wrote: Mon Aug 10, 2026 9:58 am

I think most people don't understand what an AI actually is. Regardless of all the myths and misunderstandings, "AI" is just a name given to automated (non-parametric) statistical machines.

This is a good point and something that I always thought got fundamentally overlooked in the "AI art" discussion. Quite a bit of debate has come to be over the question of "can AI make art" with the two sides supposedly being: yes, AI can make art vs. no, AI can not make art. However, it feels like sometimes the latter of those two positions becomes "yes, AI can try to make art, but fails." This is already a misstep in my mind, and gives the "AI" (what is really just a statistical algorithm) far too much agency.

If I write some code that will distribute randomly colored pixels across a canvas, and present the finished product as a piece of art, are people suddenly going to raise an outrage because "code can't make art"? I think in this case it's fairly obvious that ultimately it is my piece of art and not the code's, and that my choice to write and use that code was the key "artistic" contribution. And I don't think the situation changes massively if it was my friend or some company that wrote the code.

Thus for AI, the question of the AI "making" art doesn't even need to enter the picture. AI is a machine that can be manipulated a certain way to spit out certain results; results that, I think undeniably, can be appropriated artistically. But it is always the human that firsts enacts this appropriation and instantiates the framing that "this image is a work of art" in the same way that Duchamp instantiated the artistic framing around the urinal ("Fountain"), which he did not create himself, by putting it in an art gallery.

And so the question of AI art, and AI tools, for me, really is just a continuation of the eternal debate between easier and harder ways of doing things. And we always have to make compromises on this front. Socrates, in the Phaedrus, suggests the danger of writing in the sense that it may impact our memories and true understandings. If we're being honest, we can see that he was correct at least to some extent, and many people read a lot but gain little and our memories for great poetry, wisdom, etc. really are worse than the ancients. Nevertheless, pretty much all of us accept the downsides for the sake of the vast horizon of information and knowledge which gets opened up to us via writing. If we wanted to increase this horizon even more, we could get AI summaries of all the great works, and this would surely increase the amount of ideas and traditions and thoughts we could engage with. But here many people will consider the downsides to outweigh the upsides. There never really is a strict line though, and it's something that has to be continuously personally determined.

AI art, rather than some huge upheaval and fundamentally different problem, is just another piece on the long spectrum ranging from dissecting corpses to learn anatomy, mixing your own paints, making your own brushes, to buying paints and brushes, to using reference photos, to using a software like Krita with digital brushes, etc.