https://winbuzzer.com/2026/06/01/nvidia-launches-cosmos-3-with-openmdw-for-physical-ai-xcxwbn/
NVIDIA has launched Cosmos 3 as a physical-AI model that combines scene reasoning, multimodal generation and action output, tying the release to a new OpenMDW licensing framework.
#AI #NVIDIA #PhysicalAI #AIModels #MultimodalAI #WorldModels #Robotics
Can a self-supervised model learn good visual representations without ever reconstructing pixels? JEPA, the program from FAIR now continued at AMI Labs, says yes by training the model to predict embeddings of missing data instead. This primer walks you through where JEPA came from, how it works, what's been demonstrated, and where it's headed.
https://benjaminhan.net/posts/20260523-what-is-jepa/?utm_source=mastodon&utm_medium=social
A primer on Joint-Embedding Predictive Architectures (JEPA) — the self-supervised representation-learning program developed at FAIR and now continued at AMI Labs. What it is, where it came from, how it works, what has been demonstrated to date, and how it relates to other self-supervised and world-model research.
LeCun Launches Broadside Against LLMs, Questions Hinton's Stance
AI pioneer Yann LeCun questions Large Language Models and promotes world models for AI's future. Learn why he left Meta and started AMI.
#YannLeCun, #LLM, #AIMachineLearning, #WorldModels, #AIFuture
https://newsletter.tf/yann-lecun-criticizes-llms-world-models-future/
AI expert Yann LeCun says current AI models like LLMs are not on the path to true intelligence. He believes 'world models' will be the future of AI within five years.
#YannLeCun, #LLM, #AIMachineLearning, #WorldModels, #AIFuture
https://newsletter.tf/yann-lecun-criticizes-llms-world-models-future/
https://winbuzzer.com/2026/05/18/starchild-1-the-first-real-time-multimodal-world-m-xcxwbn/
Odyssey has launched Starchild-1 as a real-time multimodal world model that simulates visuals and sound together.
- DAMN Anthropic... Claude Design is ominous (better than Stitch by Google?): https://www.anthropic.com/news/claude-design-anthropic-labs https://techcrunch.com/2026/04/17/anthropic-launches-claude-design-a-new-product-for-creating-quick-visuals/ after Anthropic CPO leaves Figma board: https://techcrunch.com/2026/04/16/anthropic-cpo-leaves-figmas-board-after-reports-he-will-offer-a-competing-product/
- Claude for Word: https://claude.com/claude-for-word
- HY-World 2.0: A Multi-Modal World Model for Reconstructing, Generating, and Simulating 3D Worlds (check the video): https://huggingface.co/papers/2604.14268 https://github.com/Tencent-Hunyuan/HY-World-2.0
#AI #AINews #anthropic #claude #design #microsoftword #tencent #hunyuan #opensource #worldmodels
Winbuzzer
Benjamin Han
NewsletterTF
AIagent.at 🤖 AI News
Tim
Tim Green
Nicole Hennig
Nicole Hennig