HN 精选

Future Computers Will Be Radically Different (Analog Computing)

Veritasium
7 人推荐 · 跨 5 个月 · 2022–2025
计算机体系结构与CPU机器学习与深度学习科学方法与科学传播
Future Computers Will Be Radically Different (Analog Computing)

中文导读

Veritasium 深度解析模拟计算为何或成 AI 救星:从传统模拟计算机讲到现代的 The Analog Thing,再到 Mythic 如何利用闪存芯片直接做神经网络推理,直击数字硬件遭遇的功耗与算力天花板。

为什么 HN 推它

目前把模拟计算与现代神经网络结合讲解得最通俗易懂的硬核科普,尤其是探访 Mythic AI 芯片的部分干货十足。

「Veritasium 讲得通俗透彻,而且还在视频里实际演示了机器。」HN 网友原话:Veritasium explains it really well in general here (and demos the device)
英文原摘要(HackerTube / Claude 整理)

Veritasium explains why analog computing may suit AI workloads, covering old analog computers, The Analog Thing and Mythic's analog chips that use flash memory for neural network inference. It argues digital hardware is hitting limits.

Treated as the accessible reference on analog computing for neural nets, especially the Mythic AI segment.

HN 上的提及

critical bklaasen · 2022-03 · HN 原文
The tech is great, but the (presumably most compelling) use cases this guy can come up with are automated surveillance and efficiencies in junk food production.
strong_rec iKlsR · 2022-03 · HN 原文
suddenly enough a Vertiasium video pops up and about everything just slots into place
rec cultofmetatron · 2022-03 · HN 原文
here's two youtube videos that explain is much better than I could.
reference yazaddaruvala · 2022-03 · HN 原文
Cited as example of improving compute hardware
reference sargstuff · 2022-04 · HN 原文
Neural networks / analog stuff

相关视频