PyTorch, every mention
7 scenes (2026) · ← back to PyTorch
tap a year for its mentions
every year 2026 anyone Ankit Gupta 5Nick Joseph 3Michael Truell 1Jensen Huang 1Francois Chaubard 1
Verbatim, from the transcripts: the passages where PyTorch comes up
Multi-GPU Kernels, Intelligence per Watt, Heterogeneous Inference, and More | YC Paper Club · Y Combinator
- ▶ 2:18 Francois Chaubard And then, uh, we have Mark coming up next, a former PyTorch maintainer, uh, co-founded, uh, GPU mode with my friend Casey Elward, um, and then is co-founding, uh, core automation with, uh, OpenAI VP research, Jerry Twerk.
- ▶ 31:20 unnamed speaker I worked a lot on systems, like, for about, like, five years working on PyTorch. 2 times in the scene
- ▶ 36:33 unnamed speaker Like, you typically have some sort of PyTorch reference, because we sort of all agree in the community that PyTorch is generally more correct than most libraries, but not more performant. 2 times in the scene
- ▶ 41:11 unnamed speaker Um, but, like, one of the reasons why, like, uh, what Shampoo isn't more deployed is because it depends on this algorithm called the QR factorization, which is, like, very, very slow in PyTorch. 2 times in the scene
- ▶ 43:55 unnamed speaker To me, this, like, flywheel is reminiscent, like, when people say, oh, PyTorch is more correct. 4 times in the scene
- ▶ 1:09:16 unnamed speaker These things are very tricky to do in, you know, like a tensor-based programming language, like PyTorch,
Jensen Huang: The Mindset That Built NVIDIA · Y Combinator
- ▶ 40:03 Jensen Huang If not for open, if not for Linux, if not for Kubernetes, if not for all of these, you know, platform, if not for, uh, TensorFlow or more important, uh, PyTorch, right?