Below are some talks that have been recorded, listed reverse-choronologically, which means, as you go down the list, it deteriorates in quality. My best talks are the ones that have not been recorded.

• Pitfalls of next-token prediction @ Simons, 2024 [1h talk]
• Out-of-distribution generalization @ CMU AI lunch, 2025 [1h talk]
• Failure of uniform convergence bounds @ Princeton, 2020 [10 min talk]
• Algorithm configuration @ Simons, 2016 [30 min talk]

Below are slides I’ve designed:

• Creative limits of next-token prediction [Oral Slides][1h Talk Slides]
• Pitfalls of next-token prediction [Slides]
• Out-of-distribution generalization [Slides]
• GAN optimization [1hr talk - slides] [NeurIPS Oral slides]
• Algorithm configuration @ Simons, 2016[Slides]
• Downscaling LLM slides[Slides]

These are some of my cleaner posters:

• Creativity ICML’25 poster
• Downscaling LLMs ICLR’24 poster (A poster done in collaboration with Tian)
• Pitfalls of NTP ICML’24 poster
• Interpretability theory ICLR’21 poster
• Out-of-distribution generalization ICLR’21 poster

My earlier posters (like uniform convergence and GAN optimization) are too dense for my liking in hindsight, a sentiment that applies to my taste in research too e.g., when I compare my older paper on generalization bounds with any of my later results.