~/talk

CT Sun, AIC | theCUBE + NYSE Wired: AI Data Pipeline Forum 2026

Talkat AI Data Pipeline Forum 2026 · on Tuesday, 2026-10-13· [machine-learning]
name:CT Sun
links:none on file
slides:none on file

Argues that memory matters more than compute for modern AI, claiming tiered architectures from HBM to cooler storage keep hot context close while offloading less urgent data, and KV caches preserve inference context so systems avoid costly recompute. Notes practical design tradeoffs: thousandx performance gaps between tiers, necessity of different SSD components, and a move to disaggregated converged storage connected via NVMe over Fabrics so compute and data scale independently.

AI-written summary · about

  1. 0:00ScaleFlux AIC: Memory and Storage for AI
  2. 4:18Human-Inspired AI Memory for Inference and Reasoning
  3. 6:41Tiered Memory and Storage for Persistent KV Caches
  4. 10:22Disaggregated Storage Using NVMe Over Fabrics
~/search
type to search 6,518 conferences, 143,918 speakers and 192,133 talks
~/help
cmdctrl k
search every conference, speaker and series
/
focus the filter search
h
home
c
conferences
s
speakers
e
series
t
light / dark — the status bar button does the same
← →
on a talk: 5 seconds back / forward
0–9
on a talk: jump to 0%–90% of the recording
?
this help
esc
close this, then clear every filter — or just the search, while you are typing in it

Ignored while you are typing. Pages without a filter rail have no search to focus.