[ExI] Zai GLM-5.2 AI and global workspace

[Adrian Tymes]

If GLM-5.2 is a Chinese product and said it is Gemini Pro, which is a
Google (and thus US) product, either their training data is suspect or
someone's ripped something off.

On Fri, Jul 10, 2026 at 2:26 PM BillK via extropy-chat
<extropy-chat at lists.extropy.org> wrote:
>
> I asked the Chinese open-source GLM-5.2 to review and discuss the Anthropic J-space article.
> GLM was impressed but insisted that this did not mean the AI was conscious.
> I then asked if GLM-5.2 had a J-space as well.
> The reply was that all large LLMs must have an equivalent system to produce the answers.
>
> Oddly, in its answer, it referred to itself as a large Gemini model.
> I asked what specific Gemini model. GLM-5.2 replied that it was a version of Gemini Pro / Ultra (my tier): The largest, most capable models, built for highly complex reasoning, deep contextual analysis, and multimodal understanding.
> That must mean the Chinese are running a Gemini equivalent AI.
>
> You can read the whole discussion here - <https://chat.z.ai/s/aeffd1cb-5213-4919-9ac5-453a32a0c6c6>
> (The link goes to the end of the discussion. Page up to go to the start).
>
> BillK
>
>
> _______________________________________________
> extropy-chat mailing list
> extropy-chat at lists.extropy.org
> http://lists.extropy.org/mailman/listinfo.cgi/extropy-chat