
Julia Turc
@juliarturc • 20,466 subscribers
Explaining AI on YouTube • YC S24 Founder • Ex-Google Research • Eastern-European nihilist & American optimist
Videos

"World models" is one of the buzziest yet ambiguous terms in AI right now. I started this video with many questions: - How are they different from video generation? - Can they do more than AI slop? - Can LeCun be trusted given that he wears knee-high white socks? Many thanks to TJ Galda and NVIDIA AI for helping me answer (most) of these questions!
Julia Turc246,917 görüntüleme • 23 gün önce

Early on, Jonathan Ross foresaw the unprecedented demand for compute of the 2020s. After pioneering Google's TPUs, he founded Groq, a "Language" Processing Unit dedicated for inference. After joining NVIDIA, he is betting on the GPU+LPU combo to power the agentic chains of AI calling AI calling AI. This interview is part of my larger series on dedicated AI chips, so drop any suggestions/questions below. Thank you NVIDIA for hosting us at your HQ! 00:00 Intro 01:00 The Google TPU & Groq origin story 02:41 How is Groq different from a GPU? 04:43 Static scheduling makes Groq faster 05:47 Does Groq work with Mixture-of-Experts? 09:27 Are LPUs limited to text models? 11:03 Diffusion models 13:41 NVIDIA Vera Rubin: the GPU+LPU combo 15:26 Will Groq still be sold as a standalone chip? 16:49 How does agentic AI impact inference economics? 19:21 Will AI replace CUDA kernel engineers? 21:08 Will AI democratize hardware design? 26:11 Jevon's paradox: An endless demand for compute 29:04 What should kids learn in the AI age?
Julia Turc43,980 görüntüleme • 6 gün önce

This is why we need to gate-keep science. Two pseudo-intellectuals thinking they discovered something deep, conflating >social attention >transformers attention >quantum physics observer (attention) These have nothing in common, other than the ambiguity of English language. Naked ladies on Instagram have nothing to do with a weighted average followed by softmax. But they're both so mind-blown by their discovery. Dunning–Kruger will only get amplified by AI sycophancy. Please call me out if you see me going beyond my own DK threshold.
Julia Turc119,070 görüntüleme • 27 gün önce

Diffusion models clicked for me when I started seeing them through the lens of particle motion. I built this interactive playground where you too can clickety-clack to understand how drift, noise, and other hyperparams control diffusion. I hereby submit this as penance for the sin of YouTube edu-tainment 😇 Link in the first comment.
Julia Turc32,207 görüntüleme • 3 ay önce

An H100 GPU can run 10^15 FLOPs/s. But somehow, a model like Gemini Flash-Lite is stuck at 200 tokens/s. These numbers show just how much of a bottleneck memory bandwidth is. In this video, we look at how LLMs are evolving to work around this limitation. Thanks Inception for sharing your insights!
Julia Turc25,821 görüntüleme • 3 ay önce

Just published a new Flow Matching tutorial (link below). It's a little unorthodox, so I want to hear your feedback before turning it into a video. On the surface, the FM algorithm looks totally arbitrary: 🫤Pick two arbitrary points 🫤(Arbitrarily) decide to draw a straight line 🫤(Arbitrarily) model particle moving at constant speed 🫤Pick an arbitrary point in time 🫤Evaluate the velocity and use it as training label 🤯Mathematically totally valid Hopefully this tutorial will turn all of these arbitrary-looking things into intuition.
Julia Turc17,970 görüntüleme • 2 ay önce

Diffusion clicked for me when I read about score-based models, a line of work pioneered by Stefano Ermon (et al.) at Stanford. So it was a full-circle moment to collab with him and Inception on a video about training & sampling techniques for making diffusion LLMs faster.
Julia Turc26,941 görüntüleme • 4 ay önce

When reading diffusion papers, my most common reaction is "mkay the math works out but WHY WOULD A SANE PERSON CHOOSE TO DO THIS". This included the (foundational) DDPM formulas. The saving grace is that you can visualize it as 2D particle motion and get a solid intuition.
Julia Turc20,469 görüntüleme • 3 ay önce
Daha fazla içerik yok.