HC2023-K2: Hardware for Deep Learning
What it is
Bill Dally of NVIDIA gives a Hot Chips 2023 keynote on the challenges of building hardware for deep learning. He traces how GPUs have been adapted for machine learning, especially training.
Why Hacker News recommends it
HN points to it for an insider view of why NVIDIA's AI gains came from design rather than process, and for data-movement power costs.
- 2 people recommended it
- 2 different months
- 1 strong recommendations
- first 2023, last 2025
What HN commenters said
“The whole talk is brilliant and worth watching.”
Dally's keynote dismisses neuromorphic hype; whole talk praised.
Footpost, 2025-06 · Strong recommendation · in “Neuromorphic computing”
“Bill Daly gave a great presentation at Hot Chips a few weeks ago”
Bill Dally explains Nvidia's AI gains came from smarter design, not just process.
tverbeure, 2023-10 · Recommendation · in “Nvidia, one of tech's hottest companies, is fine with remote work”
Excerpts from Hacker News comments, linked to the originals. Summaries and labels were written by a language model from those comments.
Topics: nvidia, gpu architecture, hot chips, ai accelerators, bill dally
Also recommended by the same people
talkSaving Voyager 1! - Bruce Waggoner at !!Con 2024!!Con · 4 HN recommenders · 2024–2026
explainerThe Fast Fourier Transform (FFT): Most Ingenious Algorithm Ever?Reducible · 7 HN recommenders · 2021–2026
other8-bit computer updateBen Eater · 14 HN recommenders · 2017–2023
lectureLight Years Ahead | The 1969 Apollo Guidance ComputerTNMoC · 11 HN recommenders · 2021–2026
tutorial“Hello, world” from scratch on a 6502 — Part 1Ben Eater · 11 HN recommenders · 2019–2026
talkFinding the Missing Memristor - R. Stanley WilliamsThe Qualcomm Institute · 12 HN recommenders · 2010–2024
talkThe Ultimate Game Boy Talk (33c3)media.ccc.de · 10 HN recommenders · 2017–2026
talkThe Thirty Million Line ProblemMolly Rocket · 11 HN recommenders · 2018–2024