Skip to the card

Card 363 of 9762026-09-14 issue

Observed arrival · 2026-09-14

Horizon RunMap — MoE architecture by Lucas Ribeiro

horizonrunmap.com Observed source
Editorial interest 82/100 Selection signal · not a rating of the site

An independent technical preview explaining how mixture-of-experts models can keep compact expert weights in RAM while moving selected computations to the GPU.

Landing page captured for the 2026-09-14 issue.

Field notes

The architecture keeps the complete immutable expert-weight source in RAM and uses GPU memory as a working area for the experts selected during inference. Its GLM-5.3 example contrasts roughly 1.49 TB of BF16 weight memory with 744 GB in FP8, then relates those figures to B200 capacity, server headroom, context length, and concurrency. The page also exposes audit material, JSON data, and links to model specifications and deployment references.

Observed signals

Read the marks

Editorial observations of this landing page, not a rating.

OpenPublic substance visible
PrettyNotable craft visible
ProPolished or operationally mature
NicheUnusually specific use
HumanPersonal, local, civic, or handmade

One card from the complete issue

Pull Over for the Index

263,868 arrived 1,000 judged 976 catalogued Enter the complete issue
horizonrunmap.com

Landing page observed 2026-09-14. The live site may have changed.