Robot the Robot

Photoshop #GenerativeFill + #Midjourney = Holy Crap! youtube.com/watch?v=izpbj3…

BREAKING: Donald Trump on the run from the feds in the Florida Everglades #trumpguilty #TrumpShouldDropOut #TrumpIndictment @TheRickWilson @JuddLegum @MuellerSheWrote

High quality AI generated talking heads are coming! GAIA can generate talking avatars from a single portrait image and speech clip. It even supports text prompts like `sad`, `open mouth` or `surprise` to guide video generation. Crazy times ahead 🤯 microsoft.github.io/GAIA

#Dalle3 filters are wild. Even "austrian bodybuilder action movie star" is banned. 😅

Lot’s of talk about v6 in Midjourney office hours today! Highlight: they said aesthetically, v6 should be more like v3 to v4 jump than a v4 to v5 jump 👀 They said v6: > Draws out words > Much more realistic > Way better upscales > Be much more coherent > Grid sizes 8x8, 4x4, or 2x2 > Better textures and lighting It's going out to guides & mods today for a stress test. Depending on how the stress test goes, they’ll release it to 10k club. Timeline: > The goal is to have collections and image creation available to users who’ve generated 1000 images or more by mid December. But... > There are a lot of little things that won’t make it, and "it won’t be perfect". They seemed sad about that. > They seem much more excited about v7 than v6. They're fixing the imperfections, it just wont be ready by mid December for v6. > Their biggest v6 question was: how much text it will be able to write? It'll do words, but might not be as good at complete sentences yet.

@jphilipp @nickfloats Sounds like that is what they are aiming for... I do wonder if they will follow Dalle in banning names of notable people and artists.

@nickfloats It's difficult to go back to Midjourney once you tried the superior prompt understanding of Dall-E 3. If only we could have the amazing Midjourney style combined with amazing prompt understanding. Wonder how v6 fares in that department?

The Stable Video Diffusion model just dropped 🔥 The new model supports: – Text-to-Video – Image-to-Video – 14 or 25 frames at 576 x 1024 – Multi-View Generation – Frame Interpolation – 3D Scene Understanding – Camera Control via LoRA Paper: stability.ai/research/stabl… Code: github.com/Stability-AI/g… SVD model: huggingface.co/stabilityai/st… SVD-XT model: huggingface.co/stabilityai/st…

@heyBarsee I've been very impressed with how quickly it became possible to do something like this: youtube.com/watch?v=yrnSx7…

AI filmmaking is almost here. I didn't think video AI would progress this fast in 2023. Here are the top 10 incredible examples of AI video revolution: (THREAD 🧵) 1/13

DragGAN officially launched with its mind blowing "Drag your image" This is next level image manipulation With DragGAN, you are not just editing images, you are "dragging" them to match your creative vision Here are 10 example use cases:

Midjourney for video incoming..Should be a good battle between them and Gen by Runway. Still some work to do but future is good @mrjonfinger @LinusEkenstam @mreflow @FulljourneyAI #fulljourney

A breakdown/mini tutorial of animating AI stills. The future of filmmaking is just beginning. Midjourney, After Effects, RunwayML Lets go! 🧵🧵🧵🧵🧵

Remixing in Midjourney is so powerful. You can turn illustrations into highly realistic photos, sketches into 3D designs, or a man into a robot, all while maintaining the overall composition and details of your initial image. A quick tutorial showing 2 different techniques

@Thinkwert Video version with added scenes here: youtube.com/watch?v=cZp2cg…

in 1964, legendary Japanese director Akira Kurosawa attempted a version of The Lord of the Rings called “Yubiwa Monogatari.” Financing collapsed after two weeks of production; star Toshiro Mifune refused to work for reduced pay. Only scraps of the film remain. READ MORE ⬇️

To give you a sense of how fast AI for 3D modeling is advancing: the field went from the left (original NeRF-reconstructed mesh) to right (Neuralangelo from NVIDIA) in 3 years. Transporting reality into high-fidelity simulation is no longer a pipe dream.

AI just took 3D modeling to a whole new level 🤯 Introducing Neuralangelo, a new AI model by NVIDIA that reconstructs mind-blowingly detailed 3D surfaces directly from 2D videos — like photogrammetry on steroids. 🧙🏻‍♂️ Keep reading to see this crazy magic for yourself 🧵