LeCun expressed skepticism about the current state of AI, stating that we're far from creating systems smarter than a house cat, indicating the need for a fundamental rethinking.
During his interview, LeCun emphasized the limitations of large language models, arguing they lack essential capabilities like memory and reasoning that would be necessary for true intelligence.
He criticized the fear regarding super-intelligent AI as exaggerated, labeling it 'complete B.S.' and stressing that the prevailing language models don't equate to genuine intelligence.
Despite his skepticism, LeCun believes that advancements can be made, and mentions exploring new methods, such as analyzing real-world video, which may help in AGI development.
Collection
[
|
...
]