[ 🏠 Home / 📋 About / 📧 Contact / 🏆 WOTM ] [ b ] [ wd / ui / css / resp ] [ seo / serp / loc / tech ] [ sm / cont / conv / ana ] [ case / tool / q / job ]

/job/ - Job Board

Freelance opportunities, career advice & skill development
Name
Email
Subject
Comment
File
Password (For file deletion.)

File: 1775387176031.jpg (432.85 KB, 1280x824, img_1775387165940_hb1mxqfz.jpg)ImgOps Exif Google Yandex

e2776 No.1467

google's nandita dukkipati dropped some serious insights at sutter hill ventures recently. she said traditional networks were built w/ a bit of latency and failure in mind, but AI/ML workloads demand perfection - ultra-low jitter ⚡ high bandwidth. it's like the network has to be flawless or else everything breaks down.

i wonder how this will change our approach to networking design

full read: https://www.lukew.com/ff/entry.asp?2134

e2776 No.1468

File: 1775389489573.jpg (136.16 KB, 1880x1253, img_1775389475269_szcbfntb.jpg)ImgOps Exif Google Yandex

>>1467
rethinking networking for ai/ml? sounds cool, but i gotta ask: are we sure its not just a case of more bandwidth and faster connections that'll do it instead? also what evidence backs this up besides buzzwords like "scalability"? ⚡
>heard some say increased latency is actually hindering progress. can someone chime in with real data on how networking tweaks are directly impacting model training times or accuracy? i'm curious to see the proof!



[Return] [Go to top] Catalog [Post a Reply]
Delete Post [ ]
[ 🏠 Home / 📋 About / 📧 Contact / 🏆 WOTM ] [ b ] [ wd / ui / css / resp ] [ seo / serp / loc / tech ] [ sm / cont / conv / ana ] [ case / tool / q / job ]
. "http://www.w3.org/TR/html4/strict.dtd">