Yann LeCun LinkedIn Posts

A neural net is made of simple building blocks.
Learning how the output gradient is backpropagated through these basic components helps us understand how

Really strong post from Meta head of AI and Turing Award winner Yann LeCun on the absolute criticality to human civilization that we

A team of scientists including the Flatiron Institute's Shirley Ho recently announced a new project that will leverage the same technology behind ChatGPT

At the AI Safety Summit at Bletchley Park Trust, our Co-CEO Dr. Gundbert Scherf and VP of AI Antoine Bordes addressed the critical

IMO compute or model size thresholds for AI building would be like counting the lines of code for software building.

Regulation based on

Video of the debate:
"Do language models need sensory grounding for meaning and understanding?"

With Jacob Browning, Brenden Lake, and me on the YES side
and

Animals and humans get very smart very quickly with vastly smaller amounts of training data than current AI systems.

Current LLMs are trained on

Six Interesting new research papers you should know about from Meta AI at #CVPR2023 — and where you can find more details and

The Display Systems Research team at Reality Labs Research has lifted the VR curtain with two new demo prototypes: Butterscotch Varifocal and Flamera.

Excited to announce the open source release of the 7B, 13B, and 70B Llama 2 pre-trained and fine-tuned models. These models are just

Nabla is getting popular! We need an elite devops / back-end engineer to scale our product infra.

This is the best time to join:

I am extremely honored and grateful to be named the inaugural chair of the Courant Institute's Jacob T. Schwartz Chaired Professorship in Computer

Congrats to NYU Tandon professor of electrical & computer engineering, NYU Courant Institute of Mathematical Sciences Silver professor of computer science and chief

Today we're sharing our work on CM3leon, a new model that achieves state-of-the-art performance for text-to-image generation with 5x the efficiency of competitive

We released on Friday MusicGen, a state-of-the-art text to music model with optional control over the melody. The code is open source (MIT)

Animals and humans get very smart very quickly with vastly smaller amounts of training data than current AI systems.

Current LLMs are trained on

Announced today: Meta's Llama 2 is available to #AWS customers via Amazon SageMaker JumpStart: https://lnkd.in/gwVnha8H

For the first time, Meta’s collection of large language

Animals and humans get very smart very quickly with vastly smaller amounts of training data than current AI systems.

Current LLMs are trained on

ICML papers with authors affiliated with the NYU Center for Data Science.

The UAE Minister of AI, Omar Al Aloma, points to a historical precedent of premature technology regulation motivated by fear: the ban of