The best AIs today (incl the flirty gpt4o, midjourney and udio) are apparently peppered with hidden lenghty prompting to add to the "limited creativity" human-provided prompt... You can easily take the human out of the loop. The AI can: figure out on its own what's *really* that special about the Mona Lisa, list it as prompt/rules, then make slight variations/amplifications specifically based on other classics (what made *them* so special, apparently) and create a new masterpiece. Human artists did that subconsciously/naturally. AI will do it much better... Anything a human finds "sophisticated" (as in various plot lines etc) an AI will soon do so much better, to the level you will really need to read/watch that movie again and again or listen to that song again to discover another hidden element...
@Fatman305 i think AIs will struggle to come up with original ideas. Their ideas will be combined versions of something that already exists rather than a completely original idea. (I think)
@@kezzaman I think the exact opposite. They will be able to detect minute details/differences on what makes humans prefer plot-line A vs B in existing story-telling (top movies, songs, books vs mid vs low), and also detect what level of variation is tolerable/preferable in every tiny element they've broken down. They can reverse-engineer any top hit, just like professional writers and musicians peel layer after layer when encountering top content, then try to mostly recreate the layers.
if they can do a similar functionality as Udio, but without limiting daily prompts like we have already with Copilot (they just do faster the first 15 prompts of the day), i'll be all for it. been digging into AI music making for some personal projects, and MAN, it is super super fun to finnaly be able to listen to some of the songs that once were just in my head or in a txt file in my computer somewhere.
@@lamsmiley1944 I am aware of every. Little. Version. Upgrade. Of any of the *real*, tangible music generation/manipulation AIs (incl for DAWs). BUT are YOU aware of the amount of recent years Google demos that so far are vaporware?
AI made just out of video concept: Background music: By for all of this I mean if google’s stuff is handled all by The sandbox AI of course. Video structure is from GPT 4 omni or Gemini Advanced and the script to. The videos structure of it also is important because the AI LLM’s that both work on it that are of course the most state of the art ones from both companies you could choose but maybe they have different stuff for their features so you would have to get both I’m just saying. The AI’s will also try to get the exact length and one of the best prompts for each of the videos and lay out the structure and style to pin point accuracy in the video. Also ground styles of the video and audio will be layer out. Along with voice AI from google probably because they’re really good at making stuff like that if they haven’t already that sounds incredible, same as the OpenAI people to and the voices that can hopefully soon be more dynamic and controllable. Also it will prompt its self to make dialogue between different characters. You could make 3D models with the AI that it would plan out and use completely on its own to superimpose onto the videos if it needs to do so same as the AI generated images alike from Imagen or Dalle that can be still frames or could be used to like the 3D models that were also AI generated from before to superimpose them into the videos and the AI would also place them and everything. It would obviously write the script including all dialogue and say it and record in all voices. Then it would control the editing software on your computer with access to the controls like mouse and keyboard and stuff with permission and computer vision to download and play with all the files in a planned out fashion to within some time finish the video and once it understands the editing software just by looking at it inside the editing page. It will search the web to understand everything it would need to know how to use it within a very short amount of time. Then it would save the video and upload it to RU-vid, give it an appropriate title, thumbnail, description, and tags in the video editor it was editing and compositing the video inside of. Then it saves and exports it to RU-vid will all those things. Boom, that’s just how easy it would become to create a whole RU-vid skit episode with AI just working by itself.
First of all, you can't "steal" content that's publicly available on the internet simply by using it to train a neural network. Secondly, everyone with a youtube account has already given Google permission to use their content through the ToS. Finally, like so many others, you have a fundamental misunderstanding of how generative AI works. It is not a clone stamping tool. Search youtube for examples of "Runway gen-3 image to video" and come back and explain how the generated videos are somehow "stolen" content.
There's a huge issue! The content will be created from the uploaded content. When it reaches the point of no return with for example 90% AI content. When will the AI understand the concept of our world with videos made by mankind or see the difference between AI created by videos? I think this subject has not been relative to the speed the competitors have created (or launched) models. It's something we'll have to keep in mind. No Mather what! And this doesn't include the urge to go out and explore.
If the prompts get super precise maybe sci fi and fantasy novels that would otherwise take 100's of millions to adapt as a movie could be done cheaper and quicker. Even if it's just a quick fan service thing to see interest.
@@jacobshaw8825 There's more content out there than Hollywood can adapt. And most projects won't see the light of day because execs don't think they're profitable. Even when they get adapted they get cancelled after one season. It doesn't have to put anyone out of work.
Gemini is very good and popular AI. This is why google chrome market share is skyrocketing, and becoming popular because of Gemini AI. Google chrome and Gemini will be the nail in coffin to crappy edge and Copidiot. Soon, crappy edge and Bing including Copidiot will become history.