Hy-MT2-30B-A3B

A single-source signal from Hugging Face (rank 4, score 291) for a Tencent 30B parameter MoE model with 3B active parameters. No cross-source confirmation exists — zero presence on GitHub, Hacker News, X/Twitter, Reddit, Product Hunt, Dev.to, or Bluesky. The 'ours' pipeline missed it entirely.

Why now: MoE efficiency narrative (DeepSeek-V3, Qwen-MoE) creates receptive HF audience for '30B/3B active' parameter claims. Tencent release timing likely opportunistic.

• Only HF source active: rank 4 with 291 raw score, normalized to 0.5 — moderate HF engagement but not exceptional
• Zero signals across 7 other weighted sources including GH (100 repos tracked), HN (26 stories), X (0 posts), Reddit (7 posts), Dev.to (43 posts)
• 172 of 200 entities in pool are single_source — this is baseline noise, not selection
• Tencent branding on large MoE model fits current efficiency narrative, but no downstream adoption signals

Considerations: HF rank 4 could reflect internal Tencent promotion, benchmark gaming, or automated scraping artifacts. No GitHub usage, no discussion anywhere, no early adopters building with it. Large Chinese labs routinely release models that gain initial HF traction then flatline. Without external validation, this is indistinguishable from a well-placed upload.

EMERGING SIGNAL · Ignore: Revisit only if it appears on HN front page, gains GitHub integrations, or shows sustained HF download velocity over 7+ days.