Raw LLM Responses

On your point on the economy at scale you are 1000% correct, like without VC funds this is an absolute nonstarter because the power requirements to run the data centers for this are so expensive it would require we basically rework the global electric grid to fuel these things, and while I think that alone should be a disqualifying factor, the real damning thing is that the ROI on AI art / writing is insufferably low, and it will continue to be low because it's pretty clear that the models are actually fairly capped out in terms of their capabilities, because a Machine Learning model has just inherent limits in the nature of the algorithm itself, it will never be 100% because it's literally not possible in the math for it to do so, it will always be a composite. So you have these MASSIVE upfront investments in Data centers, Power, and Water that then need to be offset by the profit, and I guarantee you as soon as these companies start making any real money Disney, et all are going to make it clear that only Disney gets to use Disney images for Art Generation. Sure Disney may in house AI it up, but if your little AI company starts training on their images and you're turning a profit, they're going to eat you alive. The only reason the major companies haven't intervened already is they aren't stupid, they see this as an adorable experiment and want to try it in house with their own stuff to reduce cost, but they don't see these various AI popups as any real threat because they don't have any actual money, it's all investor VC. It's also clear that once serious publication houses got wind of their material being the training material that they will absolutely let the lawyers lose, and laws will start getting written to govern AI. That's why the current budget proposal in the senate is laced with that anti AI clause because AI companies know what the next step in the process is, they know what's coming next is their butts getting legislated into the ground and they want to put that off for as long as humanly possible, because they're already in the red as it is, if they're in the red and having to pay royalties they've made themselves a magpie that just gives them debt. At an economic level there's way to many factors at play for Machine Learning Image and Writing generative AI to be wroth the investment. Now... what I do expect AI to get used for is a ton of in house stuff. I expect you'd see a lot of internal documentation fed to In House Generative AI models to extrapolate company policy on a given subject matter or quickly generate procedures and processes without needing a ton of red tape around it by using the companies existing Procedures and Processes. Stuff that by it's very nature a company doesn't want leaving its protected enclave. On the "Artistic" AI Front I imagine it will be doing a lot of background work in movies / etc. to do spot work that would otherwise take an artist hours for a detail that probably isn't going to be on screen long enough to catch anyone's attention, and again that will be using the studio's own licensed material because they are going to want the background work to have the same overall style as the thing they're actually working on.