Richard Sutton – Father of RL thinks LLMs are a dead end
Descripción del Episodio
Richard Sutton is the father of reinforcement learning, winner of the 2024 Turing Award, and author of The Bitter Lesson. And he thinks LLMs are a dead end.
After interviewing him, my steel man of Richard’s position is this: LLMs aren’t capable of learning on-the-job, so no matter how much we scale, we’ll need some new architecture to enable continual learning.
And once we have it, we won’t need a special training phase — the agent will just learn on-the-fly, like all humans, and indeed, like all animals.
This new paradigm will render our current approach with LLMs obsolete.
In our interview, I did my best to represent the view that LLMs might function as the foundation on which experiential learning can happen… Some sparks flew.
A big thanks to the Alberta Machine Intelligence Institute for inviting me up to Edmonton and for letting me use their studio and equipment.
Enjoy!
Watch on YouTube; listen on Apple Podcasts or Spotify.
Sponsors
* Labelbox makes it possible to train AI agents in hyperrealistic RL environments. With an experienced team of applied researchers and a massive network of subject-matter experts, Labelbox ensures your training reflects important, real-world nuance. Turn your demo projects into working systems at labelbox.com/dwarkesh
* Gemini Deep Research is designed for thorough exploration of hard topics. For this episode, it helped me trace reinforcement learning from early policy gradients up to current-day methods, combining clear explanations with curated examples. Try it out yourself at gemini.google.com
* Hudson River Trading doesn’t silo their teams. Instead, HRT researchers openly trade ideas and share strategy code in a mono-repo. This means you’re able to learn at incredible speed and your contributions have impact across the entire firm. Find open roles at hudsonrivertrading.com/dwarkesh
Timestamps
(00:00:00) – Are LLMs a dead end?
(00:13:04) – Do humans do imitation learning?
(00:23:10) – The Era of Experience
(00:33:39) – Current architectures generalize poorly out of distribution
(00:41:29) – Surprises in the AI field
(00:46:41) – Will The Bitter Lesson still apply post AGI?
(00:53:48) – Succession to AIs
Get full access to Dwarkesh Podcast at www.dwarkesh.com/subscribe
Episodios Recientes
Más podcasts de Sociedad y Cultura
Ver toda la categoría →
Ad Propositum
By shows
Bienvenidas y bienvenidos al podcast de Adpropositum, mi espacio auditivo para acompañarte a conectarte con tu propósito ayudándote a eliminar los obstáculos para acceder a una vida autentica y con sentido. Aqui reflexionaremos y aprenderemos en torno a la vida, el amor, el sufrimiento, el proposito y lo valioso. Un lugar construido para que lo compartas con otros y para que ademas de acceder a mis podcast, tambien encuentres mis medicinas auditivas para el alma.
Modo Taoísmo
By shows
Obtén inspiración para poseer el poder de alcanzar la grandeza y desbloquear todo tu potencial. Solo necesitas motivación y orientación para superar obstáculos y llevar una vida con propósito.
BAJO LOS PALOS by FLEXICAR
By shows
Bajo los Palos es un podcast presentado por Iker Casillas, donde las conversaciones van más allá del fútbol. En cada episodio, Iker invita a diferentes personalidades para hablar sobre experiencias de vida, aprendizajes y reflexiones, creando un espacio cercano y auténtico. Un viaje lleno de historias inspiradoras, desde dentro y fuera del terreno de juego.
Park Predators
By shows
Explore the dark side of the world’s most beautiful places with investigative journalist and park enthusiast Delia D’Ambra. Each week, Delia guides you deep into national parks and forests across the globe, uncovering stories where nature’s breathtaking beauty has masked sinister secrets. From infamous cases that made headlines to little-known crimes that still need answers, Delia’s relentless pursuit of the truth takes her through archives and remote landscapes to reveal the hidden darkness haunting these natural wonders. Because sometimes, the most beautiful places hide the darkest secrets. This is Park Predators.
Monólogo de Alsina
By shows
Escucha y lee todas las noticias del programa. En directo de L-V de 6 a 12:30
Martes De Misterio
By shows
Casos reales de misterio y horror. Testimonios en primera persona. Entrevistas e investigaciones. Conduce: Martín Echevarría (@martinderadio). Cada Martes un episodio estreno para que puedas oír desde cualquier dispositivo. Si tienes una historia para contarnos éstos son nuestros contactos: +54 9 223 6155802 (Whatsapp Producción) // @martesdemisterio (Instagram) // mail: [email protected]
El Cartel de La Mega
By shows
Dirigido por Daniel Trespalacios. Es reconocido por su formato innovador que mezcla entretenimiento, interacción con los oyentes y temas paranormales.