Raw LLM Responses

@8:16 LLM stands for Large Language Model, not Large Learning Model. And that's for the text ones, visual generative models are usually diffusion based, which is a different thing. As for 'just being trained to generate stuff like what's on the internet', no, not quite. That's pre-training, which is a large part of the process, but not all of it - after that you have various kinds of reinforcement learning, from human feedback or otherwise, which add additional capabilities on top of that. In short, no offense, but if you want to have a really good take, do more research, you missed some basic stuff. -_- As for using AI to do art... I'm of pretty mixed opinions about it. I'm not a huge fan of the current generative style, which is all about spamming prompts until the machine spits out something somebody likes... I do think there's a possibility of these tools being useful, depending on how they're made and how they're used? I'm not a huge fan of the major corporation using them though. Looking forwards to more stuff like angel engine there, though. And, on the topic of AI doing art for itself... The AI's would need a far more robust self loop. Which would go a big chunk of the way to making them be people, which is a terrifying concept. Though I'll admit, I am hella curious what kind of art a fully recursive and self aware AI would make for itself...