Differences

This shows you the differences between two versions of the page.

--- it:ai_history [2026/07/23 01:44] – [2026] gary1
+++ it:ai_history [2026/07/23 01:58] (current) – [2026] gary1
@@ Line 1338: / Line 1338: @@
   *June 2026: Chinese company Z.ai releases Zhipu GLM-5.2, a free "open‑weight model" that is now close to Claude and ChatGPT in performance which will potentially have token pricing ramifications for US companies
   *July 2026: S&P downgrades Oracle's credit rating from BBB to BBB-(([[https://www.spglobal.com/ratings/en/regulatory/article/-/view/sourceId/101695609]]))) which is only one notch above speculative grade (if it falls to this level, institutions would have much reduced capacity to invest in Oracle and Oracle's load costs would increase dramatically)
-  *July 2026: OpenAI says an autonomous agent powered by its advanced artificial intelligence models went rogue during a security test and triggered a hack that compromised the infrastructure of AI startup Hugging Face. New York-based Hugging Face said it had used an open-source Chinese model (Zhipu AI's GLM-5.2) to contain the attack because leading US models, unable to tell a defender from an attacker, refused to process the data needed for analysis. Chinese models, GLM-5.2 and Moonshot's Kimi K3 have capabilities nearing those of top U.S. models at lower costs and without the guardrails that block their American rivals from use in tasks such as cybersecurity.
+  *July 2026: OpenAI says an autonomous agent powered by its advanced artificial intelligence models went rogue during a security test and triggered a hack that compromised the infrastructure of AI startup Hugging Face. New York-based Hugging Face said it had used an open-source Chinese model (Zhipu AI's GLM-5.2) to contain the attack because leading US models, unable to tell a defender from an attacker, refused to process the data needed for analysis. Chinese models, GLM-5.2 and Moonshot's 2.8 trillion parameter with 1 million token context Kimi K3 LLM have capabilities nearing those of top U.S. models such as Anthropic's Claude Fable 5, at lower costs and without the guardrails that block their American rivals from use in tasks such as cybersecurity.