Diffusion models now dominate generative AI, powering MidJourney, DALL·E, Ideogram, Stable Diffusion, and more—but just a few years ago, they were an obscure research idea. In this episode, BCV partner Slater Stich sits down with Jascha Sohl-Dickstein, the researcher behind the foundational 2015 paper that introduced diffusion as a generative model. <a target="_blank" rel="noopener noreferrer nofollow" href="https://baincapitalventures.com/">BCV.com</a>X: <a target="_blank" rel="noopener noreferrer nofollow" href="https://x.com/BainCapVC">@BainCapVC</a> <a target="_blank" rel="noopener noreferrer nofollow" href="https://x.com/slaterstich">@slaterstich</a>

Diffusion models now dominate generative AI, powering MidJourney, DALL·E, Ideogram, Stable Diffusion, and more—but just a few years ago, they were an obscure research idea. In this episode, BCV partner Slater Stich sits down with Jascha Sohl-Dickstein, the researcher behind the foundational 2015 paper that introduced diffusion as a generative model. BCV.com X: @BainCapVC @slaterstich

Diffusion models now dominate generative AI, powering MidJourney, DALL·E, Ideogram, Stable Diffusion, and more—but just a few years ago, they were an obscure research idea. In this episode, BCV partner Slater Stich sits down with Jascha Sohl-Dickstein, the researcher behind the foundational 2015 paper that introduced diffusion as a generative model. <a href="https://baincapitalventures.com/" rel="noopener noreferrer">BCV.com</a>X: <a href="https://x.com/BainCapVC" rel="noopener noreferrer">@BainCapVC</a> <a href="https://x.com/slaterstich" rel="noopener noreferrer">@slaterstich</a>

Jascha Sohl-Dickstein

How did diffusion models go from an obscure research idea to the foundation of modern generative AI? In History of Diffusion, BCV partner Slater Stich sits down with the researchers who shaped the field—Jascha Sohl-Dickstein, Yang Song, and Sander Dieleman—to trace the origins, breakthroughs, and future of diffusion-based AI.From the 2015 paper that started it all to the latest advancements in generative models, this series unpacks the key ideas, challenges, and moments that led to diffusion’s dominance in AI.Whether you’re deep in machine learning research or just want to understand the tech behind tools like MidJourney, DALL·E, and Stable Diffusion, this is a must-listen.

Technology

How did diffusion models go from an obscure research idea to the foundation of modern generative AI? In History of Diffusion, BCV partner Slater Stich sits down with the researchers who shaped the field—Jascha Sohl-Dickstein, Yang Song, and Sander Dieleman—to trace the origins, breakthroughs, and future of diffusion-based AI. From the 2015 paper that started it all to the latest advancements in generative models, this series unpacks the key ideas, challenges, and moments that led to diffusion’s dominance in AI. Whether you’re deep in machine learning research or just want to understand the tech behind tools like MidJourney, DALL·E, and Stable Diffusion, this is a must-listen.

Share Jascha Sohl-Dickstein

Sign up to save your podcasts

Jascha Sohl-Dickstein

Jascha Sohl-Dickstein