Published Jul 3, 2024

Beyond Language: Inside a Hundred-Trillion-Token Video Model

Delve into the evolution of AI with Anjney Midha and Jiaming Song as they explore the groundbreaking transition from GANs to diffusion models, and discover Dream Machine—a revolutionary video generative model poised to redefine video content creation and 3D simulation.

Episode Highlights

Topics covered

Episode Highlights

Related Episodes

The Future of Image Models Is Multimodal
Answers 383 questions
Scaling AI for the Coming Data Deluge
Answers 383 questions
Open Models and Maturation: Assessing the Generative AI Market
Answers 383 questions
ARCHIVE: The Dream of AI Is Alive in AlphaGo
Answers 383 questions
Neural Nets and Nobel Prizes: AI's 40-Year Journey from the Lab to Ubiquity
Answers 383 questions
Scoping the Enterprise LLM Market
Answers 383 questions
From NLP to LLMs: The Quest for a Reliable Chatbot
Answers 383 questions
REPLAY: Scoping the Enterprise LLM Market
Answers 383 questions
ARCHIVE: Open Models (with Arthur Mensch) and Video Models (with Stefano Ermon)
Answers 383 questions
The Researcher to Founder Journey, and the Power of Open Models
Answers 383 questions
ARCHIVE: GPT-3 Hype
Answers 383 questions
Remaking the UI for AI
Answers 383 questions
Security Founders Talk Shop About Generative AI
Answers 383 questions
Developer Tool UX in the Age of Generative AI
Answers 383 questions
DisTrO and the Quest for Community-Trained AI Models
Answers 383 questions

Dexa/AI + a16z

Beyond Language: Inside a Hundred-Trillion-Token Video Model

Topics covered

Popular Clips

Data, Inference Speed

Unlocking 3D Understanding

Evolution of Generative AI

Impactful AI Journeys

Future AI Capabilities

Limitations of Language Models

Image Generation Techniques

Dream Machine Insights

Unraveling Causality

Video Revolutionizes Simulation

Breakthrough in Image Generation

AI Research Evolution

Mind-Blowing 3D Reconstruction

AI Evolution

Evolution of AI

Episode Highlights

Evolution of AI Models

Dream Machine Development

Related Episodes

The Future of Image Models Is Multimodal

Scaling AI for the Coming Data Deluge

Open Models and Maturation: Assessing the Generative AI Market

ARCHIVE: The Dream of AI Is Alive in AlphaGo

Neural Nets and Nobel Prizes: AI's 40-Year Journey from the Lab to Ubiquity

Scoping the Enterprise LLM Market

From NLP to LLMs: The Quest for a Reliable Chatbot

REPLAY: Scoping the Enterprise LLM Market

ARCHIVE: Open Models (with Arthur Mensch) and Video Models (with Stefano Ermon)

The Researcher to Founder Journey, and the Power of Open Models

ARCHIVE: GPT-3 Hype

Remaking the UI for AI

Security Founders Talk Shop About Generative AI

Developer Tool UX in the Age of Generative AI

DisTrO and the Quest for Community-Trained AI Models

Beyond Language: Inside a Hundred-Trillion-Token Video Model

Topics covered

Popular Clips

Episode Highlights

Evolution of AI ModelsJiaming Song and Anjney Midha explore the rise of diffusion models in AI, highlighting their advantages over GANs in image generation. They discuss the transition from GANs to diffusion models, emphasizing the latter's consistency and potential for broader applications.

Evolution of AI Models

Dream Machine DevelopmentJiaming Song introduces Dream Machine, a pioneering video generative model that transforms text and images into video content. He highlights its potential to revolutionize 3D reconstruction and simulation by using video as a reasoning engine.

Dream Machine Development

Related Episodes