Memo: Tang Jie, the founder of Z.ai, the Chinese lab behind GLM models, argues frontier AI capabilities should stay “as open and widely accessible as possible”
Bloomberg
Context & Ripple Effects
Z.ai’s prior coverage shows a consistent open-model strategy: GLM-4.5 was released as open source and positioned as lower-cost than DeepSeek, while GLM-5.1 was made available under an MIT license. The company has also described Chinese AI models’ open-source approach as a route to global users.
Tang Jie’s memo therefore puts an explicit policy rationale behind an existing distribution model rather than signaling a standalone product change. It matters because GLM is being positioned around broad accessibility at a time when Chinese open models are already prompting responses from US companies and policymakers.
First-order effects
Z.ai publicly strengthens its commitment to making frontier GLM capabilities broadly accessible, reinforcing expectations that its model releases will remain open rather than restricted to a proprietary API or platform.
Developers and organizations evaluating GLM receive a clearer signal about the lab’s intended access model, alongside its prior open-source and MIT-licensed releases.
Second-order effects
The stance raises competitive pressure on other model providers to justify closed access through superior performance, tooling, safety controls, or commercial support rather than model availability alone.
Broader access to capable Chinese models can expand their international developer footprint, extending the policy and competitive concerns already associated with open-source releases from DeepSeek, Qwen, and peers.
Third-order effects
If leading frontier labs continue pairing high-end performance with permissive distribution, AI competition may shift further from exclusive model access toward deployment, integration, compute, and ecosystem strength.
The resulting tension—between global diffusion of open models and policymakers’ concerns about their reach—could make access rules and governance a more consequential competitive variable, though this memo itself does not establish a regulatory outcome.
The trend: Frontier AI competition is increasingly becoming a contest between open-model distribution designed to maximize adoption and more controlled approaches designed to monetize and govern access.
i really don't get why anthropic communicates like this it feels like it'd be much better PR-wise for them to “Fable is included in the paid plans until X date, which may be extended based on capacity” the other thing is announcing this 1-2 days in advance, rather than day of
As before, you can use up to half of your weekly usage limit on Fable 5. After that, you can continue using Fable 5 with usage credits, or switch to another model to keep working within your remaining limits. More details here: https://support.claude.com/...
This is of course in line with DeepSeek's vision, which has an entirely different corporate lineage. Zhipu did it earlier and equally authentically. We are seeing a convergence. The Chinese AI community might well build God and call him open public infrastructure. [image]
I think the impact of this document has not yet been understood. This is not some silly mumbling about “application scenarios”. Jie Tang is a Tsinghua professor, one of the patriarchs of Chinese AI, leader of the GLM project. He says: yes, I know what AGI means. It must be open. …
Got a chance to read the Geopolitechs translation in full yesterday (🔗below). Three things stand out to me: 1. We should be reading Prof Tang Jie's letter primarily as a mission statement for https://z.ai/, not as a bold claim that AGI is imminent. He argues that the