Learning and Coherence
Joscha discusses the unique loss function in human learning, emphasizing how we filter information to enhance coherence. He suggests that while current transformer models are effective, they may not be the most efficient path to achieving AGI. Instead, he envisions future systems that utilize fundamentally different algorithms, potentially leading to more elegant solutions than the brute-force approach seen in today's language models.In this clip
From this podcast
Lex Fridman Podcast
Joscha Bach: Life, Intelligence, Consciousness, AI & the Future of Humans | Lex Fridman Podcast #392
Related Questions
Will large language models scale all the way to artificial general intelligence (AGI) as discussed in the episode Joscha Bach: Life, Intelligence, Consciousness, AI & the Future of Humans | Lex Fridman Podcast #392 and the clip Real-Time Learning?
Will large language models scale all the way to artificial general intelligence (AGI) as discussed in the episode Joscha Bach: Life, Intelligence, Consciousness, AI & the Future of Humans | Lex Fridman Podcast #392 and the clip Real-Time Learning?
Will large language models scale all the way to artificial general intelligence (AGI) as discussed in the episode Joscha Bach: Life, Intelligence, Consciousness, AI & the Future of Humans | Lex Fridman Podcast #392?