@ylecun 오늘의 포스팅 - 2025-12-27
📌 오늘의 포스팅
🔁 [RT] BREAKING: New CNN Poll says Trump is the worst Pre...
BREAKING: New CNN Poll says Trump is the worst President in history.
🔁 [RT] LeCun's JEPA has evolved into a vision-language mo...
LeCun's JEPA has evolved into a vision-language model, with 1.6B parameters rivaling the 72B Qwen-VL. Instead of predicting words directly, the proposed VL-JEPA learns to predict the core "meaning" of a text in an abstract space, ignoring surface-level wording variations. This
🔁 [RT] Lol no normal American can afford this And you’re...
Lol no normal American can afford this And you’re more poor Shall we call you Muripoor
🔁 [RT] We don’t simulate every atom to model intelligence...
We don’t simulate every atom to model intelligence. We don’t use quantum field theory to model road traffic. JEPA taught me the importance of learning physics at the right level of abstraction. Thank you @ylecun and the JEPA team— it was a privilege to work with you.
Why current AI models are bad at planning @soniajoseph_ explains the problem: They don't understand cause and effect. The fix? JEPA. Instead of trying to predict every single pixel (or atom), it predicts concepts. This allows the AI to abstract away the useless details
— @readsail
#뉴스 #트렌드
@ylecun | 2025-12-28