Earthgrid的动态

I'm bullish on AI videos. It seems like LLMs are reaching a bit of a ceiling. At least, things aren't progressing as fast as it used to. But AI for image and video generation still has insane untapped potential. Look at this. The physics. The lighting. The subtle movements. You couldn't replicate this in a 3D program. It just wouldn't come out very natural-looking. And this, right here, is as bad as AI-generated videos will ever get. It costs Netflix $100M-$250M (on average) to produce a show. YouTube will soon have an abundance of Netflix-grade clips, videos, series, and shows. By individual creators. At home-budgets. And I couldn't be more pumped about it. Are you going to enter video in 2025?

Creative development with Sora, Veo-2 are totally going to flood the market in 2025 for generating content at scale. This could have an impact in terms of emergence of a lot of copy cat content creators across the web - laying foundation for the cream of authentic content creators becoming even more powerful (and expensive to work with). “YouTube will soon have an abundance of Netflix-grade clips, videos, series, and shows.” Not sure we’re there yet for “series, and shows”, but for “clips, videos” - sure, this is already happening. Entertaining? Yes. Powerful storytelling? Not much yet. I’m yet to see a full blown TV series created by AI (not the animated kind such as the 26-episode series, Qianqiu Shisong) that can deliver on intricate human emotions, complex plots, with method acting prowess shown by top creative talent. But then, according to several movie Directors including award winning Director, James Hawes, we are less than 5 years away from the use of GenAI to create full fledged TV shows (source: https://lnkd.in/gRaaPfp9). Now, that would be a serious leap in use of AI tech for advancing creativity. Happy 2025! ??

It continues to amaze me how quantity is highlighted and trumpeted over quality. All these 'tools' will do is continue the unabated theft of IP, increase energy demands, and produce 99% slop. Cheers? Or Jeers? #technology #ai #genai #creativity #problemsolving

The reality of AI video...

In 2010, Paramount Pictures financed Puss N Boots for $130MM; taking over a year to create for theaters…now an amatuer can produce better quality graphics in days. But the art of good story telling is more essential than ever.

I'm bullish on AI videos. It seems like LLMs are reaching a bit of a ceiling. At least, things aren't progressing as fast as it used to. But AI for image and video generation still has insane untapped potential. Look at this. The physics. The lighting. The subtle movements. You couldn't replicate this in a 3D program. It just wouldn't come out very natural-looking. And this, right here, is as bad as AI-generated videos will ever get. It costs Netflix $100M-$250M (on average) to produce a show. YouTube will soon have an abundance of Netflix-grade clips, videos, series, and shows. By individual creators. At home-budgets. And I couldn't be more pumped about it. Are you going to enter video in 2025?

These new generative AI video capabilities feels a lot like the birth and rise of Adobe InDesign in 1999… One of my first jobs was working as a lithographer for creatives working in QuarkXpress. It took a few years for InDesign to mature, but its powerful RGB & PDF workflow now meant everyone could create proper graphic designs. Mind you, not everyone was good at designing! But at least their artwork was without errors (mostly) and it was printable. Now Canva can do much of the same in the browser thanks to those same RGB/PDF workflows. The AI will (might?) democratize or even replace the video and 3D tools of today in the near future, but it still needs people with a vision to guide it. AI works best if you understand the concept and process of what you are trying to reach. Ask simple questions, get simple results. In my opinion, in order to explain a concept, you still need to understand it (and the capabilities and limits of the tools) to the fullest. Current animators, FX / 3D people have a massive head start. I feel some smart ones will show us the possibilities and the limitations.

Whichever way - Just get things done efficiently according to resources/time. AI led/overseen by humans? Or Humans led/guided by AI? Or both? Read more about the Human AI Collaboration Paradox - https://lnkd.in/d3Y-Bj4k

Hmm ?? llm / #generative #AI for video is absolutely game changing all hail the cat ?? ????videos as we are entering a synthetic reality / uncanny Valley for told by Michael criterion rising sun ??/movie . As we are seeing the democratization of special effects , #deepfake to cat ?? video with chopsticks ?? and pasta ?? only divine Providence might know how much energy ??budget this will consume to our grid ? I believe right now the United States at 3% of the grid is dedicated to data centers and we might be looking at 20 to 30% of the grids energy a few decades out . As old nuclear power ??plants case in point 3 mile island ?? are starting to come back online in order to deliver energy for data centers.

2025 is going to be the year for AI generated video. We're reaching a "chatGPT" moment for AI models that can render convincing video content of whatever you can imagine...like a cat making stir-fry noodles from scratch then chowing down! Between OpenAI's Sora and Google's Veo-2 (among probably many other competitors that I don't know right now) there is a race to become the production tool of the future. It is only a matter of time until we get the first entirely AI-generated show/movie and the ubiquity of AI content explodes even more. This is at once exciting and terrifying. ?? Exciting: The barrier to entry of creating good-looking video content is being lowered considerably if not destroyed altogether. Anyone with an idea can just start typing in their story to the video generation tool and start getting takes as if they had a movie set and production crew at their beck-and-call -- pretty cool! I anticipate many solo creators making fun and attractive content that would otherwise be impossible to make alone. This will create some amazing opportunities for creative folks to share their stories and imaginations in a new and accessible way. It could also further democratize the creation of new content (depending on how costly/available these AI tools are). ?? Terrifying: The film and video production industry will be heavily disrupted. 3D artists, production staff, and maybe even on-screen talent will have to compete with these tools in the future. Their roles likely won't go extinct, but they could see fewer opportunities due to these video production models. This will likely trigger a market shift for the film ecosystem even more disruptive than that brought on by video streaming/the internet. Additionally, there is the risk of malicious actors using these tools to create even more convincing "deepfakes" that sow even more confusion and conflict into our daily life. ?? Regardless of how we feel and what we think will happen with these tools, the cat is out of the bag--and making noodles for now. We have to deal with this technology and its consequences as the cat starts to do more than make noodles. ?? In order to do that the best possible way we need to understand these tools deeply, talk about their consequences earnestly, and stay vigilant for those who would misuse our freedom of innovation and creation against us.