a16z Podcast
What's Missing Between LLMs and AGI - Vishal Misra & Martin Casado
Mar 17, 2026
· 47m
Vishal Misra returns to explain his latest research on how LLMs actually work under the hood. He walks through experiments showing that transformers update their predictions in a precise, mathematically predictable way as they process new information, explains why this still doesn't mean they're conscious, and describes what's actually required for AGI: the ability to keep learning after training and the move from pattern matching to understanding cause and effect. Resources: Follow Vishal Misra on X: https://x.com/vishalmisra Follow Martin Casado …
Buffering from publisher…
Audio streamed from the publisher — STT.ai does not host podcast files.
अस्मिन् प्रकरणे अद्यापि हस्तलिखितं नास्ति
STT.ai येन इदं प्रकरणं कृत्रिमबुद्धिद्वारा लिखितं भवति । वक्तृत्वबोधेन, समयसूचनानि, तथा बहुविधेषु फॉर्मेटेषु निर्यातेन च सटीकः पाठः प्राप्तः भवति ।
वक्तृ- पत्ता
शब्द-स्तरीय-समय-चिह्नानि
निर्यातं SRT, TXT, JSON रूपेण