@m_gharbi : Today's visual generative models are mere stochastic parrots of imagery, much like early language models, which could only statistically mimic short sentences with little reasoning. In contrast, modern large language models (LLMs) can comprehend long documents, keep track of • TwiCopy

Michaël Gharbi

@m_gharbi

+ Follow

Founder @reveimage. Ex-@AdobeResearch
. @MIT_CSAIL Ph.D. @Polytechnique X09

ID: 574636059

linkhttps://mgharbi.com/ calendar_today08-05-2012 15:18:44

20 Tweet

541 Followers

214 Following

Michaël Gharbi

@m_gharbi

3 months ago

Today's visual generative models are mere stochastic parrots of imagery, much like early language models, which could only statistically mimic short sentences with little reasoning. In contrast, modern large language models (LLMs) can comprehend long documents, keep track of

thumb_up_off_alt212

chat_bubble_outline30

repeat31

shareShare