Stefano Ermon
Stefano Ermon is an associate professor of computer science at Stanford University and the co-founder and CEO of Inception Labs, a company developing diffusion-based large language models. He is known for his pioneering research in diffusion models, including extending them from images to text, and for co-inventing core generative AI techniques such as Direct Preference Optimization and Flash Attention. Through Inception Labs, he commercializes diffusion language models under the Mercury family, which aim to significantly outperform autoregressive models in inference speed.
“If you think about inference, now not training, inference generation, autoregressive models are still sequential. The computation is one left to right, one token at a time... That kind of workload does not map well to GPUs. That kind of workload is extremely memory bound.”
Source→“There was an inflection point in 2017 when people switched from RNNs to transformers... But if you think about inference... the equivalent at inference time is a diffusion based model.”
Source→“Our Mercury models are on par with the Haiku models, Flash models, Mini Nano models from OpenAI, if you look at benchmarks, while being significantly faster.”
Source→“They were previously using, serving their LLMs on Cerebras... And then they switched over to our diffusion-based LLMs because they can essentially get the same speed as what you would get if you were to run an autoregressive model on custom hardware.”
Source→“Open Router has this very nice way of kind of like looking at all the different use cases... they have like a nice hard taxonomy basically of tasks.”
Source→“All of that started from ideas that were developed in academia in my lab. But that's not the only one.”
Source→“We kind of like came up with this idea of let's train a neural network to denoise images... And that basically became the kind of like underlying technology of modern diffusion models... with my PhD student, Yang Song.”
Source→AI-extracted from podcast / newsletter / paper summaries. May contain errors.