systems: everyone on the record — The Minutes https://minutesof.com/on/systems/ Everyone on the record about systems: 3 verbatim quotes from 2 people, October 2024 to October 2025, in one chronology, each with a timestamp and a link… en Sun, 30 Aug 2026 22:33:16 +0000 Dylan Patel: Patel explains GB200's 72-GPU configuration creates reliability challenges compared to 8-GPU syst https://minutesof.com/q/e442c5ed-ed97-4cb7-8195-b9012b3b0fc8/ https://minutesof.com/q/e442c5ed-ed97-4cb7-8195-b9012b3b0fc8/ “If the reliability of each GPU is the same, then when a single GPU fails in 72 GPUs, you have a much higher chance of something failing, right?” — Dylan Patel, Together AI Fri, 03 Oct 2025 20:23:30 +0000 systems Dylan Patel: Patel explains o1's thinking time creates memory bandwidth issues that prevent batching users at https://minutesof.com/q/b7e7fd8f-2e86-4705-b53e-c249725aab84/ https://minutesof.com/q/b7e7fd8f-2e86-4705-b53e-c249725aab84/ “But if you batch higher, k b cache is not just a memory capacity issue, it's also a memory bandwidth issue.” — Dylan Patel, Scaling Intelligence Tue, 12 Nov 2024 04:17:00 +0000 systems Bill Gurley: Gurley explains NVIDIA's networking, NVLink, and CUDA advantages emerge specifically in the large https://minutesof.com/q/3cd5de34-3e0c-4b75-adef-e5e5e2d29a6d/ https://minutesof.com/q/3cd5de34-3e0c-4b75-adef-e5e5e2d29a6d/ “That's when the networking piece thrives. That's where NVLink thrives. That's where CUDA really comes alive in the biggest systems that are out there.” — Bill Gurley, BG2 Pod Sun, 13 Oct 2024 10:35:00 +0000 systems