
Gustavo Bakker
@gusbakker
Software Developer.
AI & Deep Learning enthusiast. Music & Video creator in my free time.
ID: 34865268
24-04-2009 06:15:51
563 Tweet
34 Followers
188 Following


Yann LeCun Geoffrey Hinton If humans were trained only on text input, would they truly understand the world? Although they might develop strong linguistic skills, they would struggle in areas such as spatial orientation. Like LLMs, they could understand text but would struggle w/ simple tasks like walking




Andrej Karpathy Perhaps this is why, in the olden days, scientists such as Einstein and others were able to focus and come up with brilliant ideas, as they were not overwhelmed by an excess of information.





This is interesting as a first large diffusion-based LLM. Most of the LLMs you've been seeing are ~clones as far as the core modeling approach goes. They're all trained "autoregressively", i.e. predicting tokens from left to right. Diffusion is different - it doesn't go left to
