BREAKER DELTA
Breaker Delta
Jensen Huang / Nvidia turmoilOpen Source 🔥 Must-read
161
Current wave
#8
Board rank (2026-11-16)
Wave -43Rank →
How is wave height calculated?
Wave height = base + resonance + evidence + first-appearance bonus, uncapped. Base by level: 🔥 Must-see 60, 👀 Watch 30, 🌀 Echo 10; +20 per independent source (multiple links from one source group count once); +8 per original link; +25 on a topic's first appearance. A topic needs at least 2 independent sources to count as a hotspot; a single source stays a echo. Topics with no new development are carried forward at their last wave height, and deltas compare against the previous time point's full board.

🌊 Wave chart

0591171762352012026-11-12 18:002042026-11-13 18:001882026-11-14 06:001682026-11-152042026-11-161612026-11-16
EditionWaveRankLevelChangeTitleSourcesLinks
2026-11-12 18:00201#1🔥Must-readFirst seenNvidia Board Set to Oust Jensen HuangWSJ, NYT, BBC, Hacker News, Redditwww.wsj.com · news.ycombinator.com
2026-11-13 18:00204#1🔥Must-readWave +3 · Rank →Aftermath of Jensen Huang's Resignation: Employees Petition for His Return, Trump Nominates Tim Cook as SuccessorHacker News, WIRED, Reuters, TechCrunch, Bloomberg, Google Newsnews.ycombinator.com · www.wired.com · techcrunch.com
2026-11-14 06:00188#2🔥Must-readWave -16 · Rank ▼1Jensen Huang Resignation: Petition, Subpoena, and a $3 Billion China Project HitHacker News, Reuters, BBC, WIRED, TechCrunch, Wikipedia Most Readwww.wired.com
2026-11-15168#6🔥Must-readWave -20 · Rank ▼4Jensen Huang Resignation: Execs Sign Petition, WSJ Long Read and New Faces on WikipediaNYT, TechCrunch, WSJ, The Guardian, Wikipedia Most Readtechcrunch.com
2026-11-16204#8🔥Must-readWave +36 · Rank ▼2Jensen Huang's resignation: return rumors heat up as allies form a “war room” to lobby the boardFinancial Times, The Verge, TechCrunch, Hacker News, Reddit, Wikipedia Most Readwww.jpost.com · www.theverge.com · techcrunch.com
2026-11-16161#8🔥Must-readWave -43 · Rank →Meituan open-sources LongCat-2.5 weights: a 1.6T agent model trained entirely without Nvidia GPUsThe Decoder, Hugging Face, Redditthe-decoder.com · huggingface.co

Meituan's LongCat team has released the LongCat-2.5 weights: a 1.6-trillion-parameter MoE agent model activating about 48 billion parameters per token, open-sourced under the MIT license; it self-reports 88.4 on Terminal-Bench v2.1, close to Claude Opus 5's 89.1 (both Meituan's own tests, not independently verified); trained without Nvidia GPUs at any point, and its self-hosting guide gives only a deployment plan for 194 Huawei Ascend Atlas A2 units (2 handling images, 64 handling input, 128 generating), with no Nvidia deployment instructions; 1M context, 128k output; API compatible with the OpenAI/Anthropic formats.

Suggested action:Write a tweet / dig deeper — an open-source agent model plus a de-Nvidia training path is worth a full piece.

✍️ For creatorsFrameworks from public writing skills: tweet-writer (hook formulas / PAS), Hormozi's four parts, Dan Koe's rhythm
Short-post outline
Three hook options (the hook decides 80–90% — write it first)
  1. Specific result:“Jensen Huang / Nvidia turmoil” — wave height 161, 16 sources reporting at once.
  2. Curiosity gap:Meituan's LongCat team has released the LongCat-2. But the thing worth watching is not the event itself.
  3. Signal:The signal is not where the noise is: “Jensen Huang / Nvidia turmoil” is already at #8 and still developing.
Body skeleton · PAS (Problem → Agitate → Facts)
  1. Problem:Meituan's LongCat team has released the LongCat-2.
  2. Agitate:Write a tweet / dig deeper — an open-source agent model plus a de-Nvidia training path is worth a full piece.
  3. Facts:Meituan's LongCat team has released the LongCat-2.5 weights: a 1.6-trillion-parameter MoE agent model activating about 48 billion parameters per token, open-sourced under the MIT license; it self-reports 88.4 on Terminal-Bench v2.1, close to Claude Opus 5's 89.1 (both Meituan's own tests, not independently verified); trained without Nvidia GPUs at any point, and its self-hosting guide gives only a deployment plan for 194 Huawei Ascend Atlas A2 units (2 handling images, 64 handling input, 128 generating), with no Nvidia deployment instructions; 1M context, 128k output; API compatible with the OpenAI/Anthropic formats.
    Wave 161 · rank #8 · 16 sources in resonance · Sources:www.wsj.com · news.ycombinator.com · www.wired.com · techcrunch.com · www.jpost.com · www.theverge.com · the-decoder.com · huggingface.co
Short-post tips
Long-form outline
Skeleton · Hormozi's four parts (Idea → Story → Reasons → Soft CTA)
  1. One idea:State your claim in one sentence — distill it from the suggested action; if you can't, the angle isn't clear yet
  2. Story & evidence:Nvidia Board Set to Oust Jensen Huang → Meituan open-sources LongCat-2.5 weights: a 1.6T agent model trained entirely without Nvidia GPUs(Wave trail:201 → 204 → 188 → 168 → 204 → 161)
  3. Reasons, not steps:One reason each from the tech, business and user angles — argue why it is true, not how to do it
  4. Soft CTA:End lightly: I will keep tracking its wave height — bookmark this page to follow along
Long-form tips