I don't think its a fair characterization. If the average understanding last week was "we have no idea how to get to 1-minute photorealistic video models", and that understanding massively changed with a single work, it is a breakthrough.
The focus on "novelty is only when you
MULTIMODAL AI
-
Breakthrough in Photorealistic Video Generation Models Debate
By
–
-
Public reaction breakdown to OpenAI Sora video generation release
By
–
I read HUNDREDS of TikTok, IG, and X comments about the @OpenAI Sora release. The rough breakdown is: 40% dread/world is over/nothing is real
30% how do I get access to it/I want it 20% wow/cool/insane
8% why do i need this/what’s the point 2% it still messes up hands -
OpenAI Video Model, Gemini 1.5, and New AI Jobs
By
–
Top stories in AI today: -OpenAI drops AI video generation model
-Google unveils Gemini 1.5 with massive context window
-How to verify Gemini’s responses
-Michigan sells student data for AI training
-6 new AI tools & 4 new AI jobs Read more: http://
therundown.ai/p/openai-chang
es-the-ai-video-world
… -
Midjourney Intelligence Critique Scaling Model Improvement Debate
By
–
Thank you for documenting, I respect that. I felt that your claim was unsound, striking a pose based on an unsurprising improvement in scaling a diffusion model (while you actually know that Midjourney has less intelligence than an autistic tadpole). I decided to throw an egg at
-
OpenAI Introduces Sora Text-to-Video Generation Feature
By
–
🚨 BREAKING NEWS: @ChatGPTapp introduces Sora, the groundbreaking text-to-video feature!
— God of Prompt (@godofprompt) 16 février 2024
🎥 Transform text into engaging videos effortlessly.
🔐 Full control: Customize visuals, style, and pacing.
🤖 Sora's innovative technology extends GPT capabilities to visual storytelling.… pic.twitter.com/mLQlsLUm43BREAKING NEWS: @ChatGPTapp introduces Sora, the groundbreaking text-to-video feature! Transform text into engaging videos effortlessly. Full control: Customize visuals, style, and pacing. Sora's innovative technology extends GPT capabilities to visual storytelling.
-
Caution About GPT5 and Its Multimodal Capabilities
By
–
You don’t even know GPT5 yet. And regarding it’s multimodality I would be cautious with my words
-
AI Attention Mechanisms and Hidden Complexity Recognition
By
–
It works incredibly well. During the first few views, my attention follows the most salient parts of the scene, such as the paw repeatedly touching the nose, and only after I saturate the attention draw for high significance features do I realize that kitty is a shoggoth
-
Unable to Distinguish Real Minecraft from AI Generated Content
By
–
I honestly wasn’t able to differentiate between real Minecraft and AI generated
-
Synthetic Data and Unreal Engine Video Generation Concerns
By
–
Likely synthetic data, unreal engine. A few of the videos felt that way
-
Waking Up to a New Text to Video Model
By
–
i went to sleep, i wake up and there's a new text to video model?