Came across this in my recommended just now. I love your take on this.. I'm pretty skeptical about using AI tools but am still fascinated by how these more abstract renderings tell a wider story. I enjoyed playing with very early AI image generation since it provided an abstract idea without looking like it was manmade, but I don't have that interest anymore since new versions focus too much on emulating existing work I just finished a master's exploring the ethics of using digital tools for dress conservation in the museum space. It has sightly different discussions compared to art, but there is still an interesting point to make between the relationship between artists/practicioners and the technologies being used Anyways, curious to see more of your content!
Your master's sounds absolutely fascinating, thanks for sharing! Would love to hear more about that. I know exactly what you mean about early AI tools vs what they have already become. I think you might enjoy another video I did recently about Google Earth, as it explores a similar idea: Basically, some of these tools and technologies are more interesting, at least from an artistic and arguably even cultural point of view, when they are still very flawed and show some seams. As they "improve," they actually worsen, in a sense. I'm interested in how these things deviate from reality, just as you said, more so than how they can accurately simulate it.
@@edlundart I enjoyed the Google Earth video too! This may not a very accurate comparison, but it made me think of how people view pixel art today as a stylistic choice, compared to how artists used to work with the limitations of technologies and displays to get the most out of their work. Perhaps in the future there will be renaissance of people emulating early photogrammetry technologies, mimicking abstract effects that people are trying so hard to remove at the moment. I sent an email just now on your website to say hi, be great to connect with another creative on here. Keep up the awesome content!
@@Brittzay The pixel art comparison is a great one, I think! We're constantly finding things to be nostalgic about in old technological limits. Thanks for reaching out - I'll check out your RU-vid channel, too.
0:40 she looks like she's having a mask on top of skeleton face 😀 Now I want to play some Heroes 3 with Necropolia 🤣 (edit) P.S. maybe the in-between ground could be generating terrain for games (also physical boards, to play with figures). Being "ugly" or some random blob can actually be a positive result for some interesting textures here!
I think the reason the brazilian woman was portrayed that way is because of carnival. I'd like to think if you do that for countries that have similar events that would be the same. like the bahamas or barbados
I agree that carnival influences the way the Brazilian woman is rendered. I didn't mention this in the video, but I generated each one a few times to end up with usable-ish results. I believe every generation of Miss Brazil featured the carnival-esque look. Miss Sweden generated (either once or twice, I can't remember) in a bikini, which is another interesting example of how LLMs function. There is an American idea of "the Swedish bikini team" etc, even if there are very few days per year a Swedish woman is likely to wear a bikini. Of course, pageants traditionally have a bikini portion, so it makes sense in that context, yet the vast majority of nationalities are rendered in evening gowns. Anyway, none of this is meaningful grounds for statistical analysis, and it's all "anecdata" - and my main interest here was the fascinatingly flawed shapes output by the AI tool. Thanks very much for watching and commenting!
