Raw LLM Responses

I'll brake this hive mind solidarity in comments with a take that would definitely be unpopular here. Andy, you criticise Shadiversity for contradicting himself with his pro-AI takes at 5:20, and yet this is exactly one of the very first things you do yourself with anti-AI takes. 8:09 You say: _"When you rely to draw the rest of the owl for you, what you do is you're _*_relying on artists who can finish it_*_ without AI..."_ Um, excuse me? What specific *artists* does he rely on in this specific example you've shown? You've mocked it yourself that what he does here is "just" converts a hand-drawn sketch into a photo-like image. So, what are the *artists* the model has "stolen" the skillset from, who can take a line art and turn it into a photo? I'm not talking about ultra-realistic paintings _(they're amazing but you can still tell those are paintings, not photos)_. No, I'm talking about an ability to *draw* a photo, by hand, and it would be indistinduishable from a real one. So who are those masses of artists the model mimicks? Or do we, for the sake of the argument, consider a phone camera to be an artist now?.. And please, spare me from the hypocritical "photographers are artists" take. Nobody debates this specific wording. But the overwhelming majority of photos all the models were trained on are just the usual photo-slop from instagram. I'd be amazed if the professional photos take even 1% of the dataset _(more honest estimate would be 0.1% or even 0.01%)_ , which barely influence the model weights but which would be EXTREMELY hard to find, verify that it's actually a "professional photo" form a photographer who didn't even bother to explicitly mark it as one, and exclude it. The AI learns not from the skillset, but from the visual patterns that people in mass describe as good or bad. As long as it's a real photo, the model basically doesn't care if it's a professional or amature one, the model *CAN* generalize and that's the whole point of it. So this "photographers are artists" take _(and hence, the following ultimatum about compensating photographers or preamptively removing any potentially debatable work from the dataset)_ does nothing more than putting an unreasonable road block intentionally designed to stall any AI progress with bureaucracy - to a complete stop. This is a criticism for just ONE take in your video. You aggresively utilize various forms of manipulations across the whole 1.5 hours of it. Each of those are manipulations and misdirections that can be easily said with 10 words, but the opponent would need 3 paragraphs to debunk it. The best I can say is... well, stay in your echo chamber of circle... "tw"erk. Or don't. Whatever. While I'll learn how to use diffusion models to expand my horizons of possible.