▼Chinese Open-Weight Models Emerge as the Non-US AI Stack

Score history

Daily conviction score, 0 to 100. Higher means the thesis is more strongly corroborated.

Now 84 · -2 since Oct 9 · ranged 84 to 86

Showing the last few days. Unlock full score history.

Why the conviction moved

• Oct 7
  Strengthened +3

  AWS and Ai2 shipped specialist models built on Alibaba's Qwen weights: Strands Decider 2B on Qwen3.5-2B and AstaBrief 8B on Qwen3-8B. Even US hyperscalers and research labs now use Chinese open weights as default base models. The thesis expected a split into rival supply blocs; this shows Qwen spreading inside the US stack itself.

• Oct 6
  Weakened

  Reflection AI positions Beam as a US-built open-weight rival to China's GLM-5.2 and claims it needs three to four times less inference compute. Users cut off from closed US models may get a downloadable US option, which weakens the idea that the non-US bloc has to standardize on Chinese weights.

Showing the last 2 days. Unlock the full record.

Source trail

• Supporting · October 7, 2026

  AWS and Ai2 ship small open specialist models built on Alibaba's Qwen base weights

  AWS and Ai2 shipped specialist models built on Alibaba's Qwen weights: Strands Decider 2B on Qwen3.5-2B and AstaBrief 8B on Qwen3-8B. Even US hyperscalers and research labs now use Chinese open weights as default base models. The thesis expected a split into rival supply blocs; this shows Qwen spreading inside the US stack itself.

  Strands Agents (via Hacker News)

• Contradicting · October 6, 2026

  Reflection AI unveils Beam, a 501-billion-parameter open-weight model built to compete with China's GLM-5.2

  Reflection AI positions Beam as a US-built open-weight rival to China's GLM-5.2 and claims it needs three to four times less inference compute. Users cut off from closed US models may get a downloadable US option, which weakens the idea that the non-US bloc has to standardize on Chinese weights.

  Reflection AI (via Hacker News)