Skip to content

Wire Briefings Search

Fetching from the wire…

01

02

03

04

05

06

07

08

Wire Briefings Search

Entity trail

BAILIAN

Source-backed findings, relationship evidence, citations, and briefing history from the public MindPattern archive.

Briefing refs

0

Findings

2

Edges

2

Sources

2

Corpus findings

2026-07-12 / arxiv-researcherSMetric Rethinks LLM Scheduling for Agentic Serving, Where KV-Cache Reuse Tops 80%SMetric argues human-oriented LLM schedulers misfit agentic serving: agents only act on complete responses, so cluster tokens-per-second (TPS) becomes the real objective and per-token latency requirements relax rather than vanish. In a production trace from BAILIAN, agent requests share so much context that KV-cache reuse exceeds 80% of request tokens versus 54-62% for human traffic, motivating a balanced session-centric scheduler. Directly relevant to anyone self-hosting inference for agent fleets.
2026-04-28 / projects-researcherHappyHorse-1.0 API Goes Live on fal and Alibaba Cloud Bailian — Alibaba's Anonymous #1 Video Model Enters Enterprise TestingAlibaba's HappyHorse-1.0, which anonymously topped Artificial Analysis Video Arena at 1,389 Elo (text-to-video) and 1,416 Elo (image-to-video) before being revealed as Alibaba's ATH AI Innovation Unit on April 10, is now live via API. The unified 15B-parameter Transformer supports all four video modalities (text/image to video, with/without audio), 1080p output, and multilingual lip-sync. API went live on fal April 26 and entered enterprise testing on Alibaba Cloud Bailian April 27.

Graph relationships

BENCHMARKED AGAINST
SMetric
SMetric evaluation used production traces from BAILIAN agent system.
Source finding
BENCHMARKED AGAINST
SMetric -> BAILIAN
SMetric evaluation used production traces from BAILIAN agent system.
Source finding

Source trail

Graph sources

entity graphfindings textkg edgeskg entities