The Minutes of Dylan Patel on kv cache https://minutesof.com/dylan-patel/on/kv-cache/ Everything Dylan Patel has said on kv cache: 3 verbatim quotes between November 2025 and July 2026, each with a timestamp and a link to the recording it… en Sun, 30 Aug 2026 16:22:19 +0000 Patel observes Asian researchers consistently produce KV cache reduction papers unknown to American researcher https://minutesof.com/q/734c899f-c40e-4627-bc76-4d78dc3ab398/ https://minutesof.com/q/734c899f-c40e-4627-bc76-4d78dc3ab398/ “Some paper that reduces KV cash, and no researcher in America has even heard of this paper.” — SemiAnalysis Weekly Wed, 06 May 2026 00:55:06 +0000 kv cache Patel observes Asian researchers consistently reference KV cache reduction papers unknown to American research https://minutesof.com/q/3c600994-8f55-4267-9e3e-11503f5d67cb/ https://minutesof.com/q/3c600994-8f55-4267-9e3e-11503f5d67cb/ “It's the fucking best thing ever. DeepSeek and TurboQuant were the, like, most precipitous ones that popped up the most,” — SemiAnalysis Weekly Wed, 06 May 2026 00:55:06 +0000 kv cache Patel states avoiding redundant prefill through KV cache can cut inference costs to one-fourth of previous lev https://minutesof.com/q/c93c6e47-fe3d-405d-b28b-2d99a4919dee/ https://minutesof.com/q/c93c6e47-fe3d-405d-b28b-2d99a4919dee/ “So you can cut your cost to one fourth of what it was previously if you just don't do the pre fill. Right?” — Clockwork Fri, 21 Nov 2025 17:18:51 +0000 kv cache