← sources

AI is becoming a cybersecurity threat faster than companies can adapt.

source: reddit (reddit:StockMarket)published 2026-11-19 07:52ingested 2026-11-20 02:40

retained text

  
The average time between vulnerability disclosure and exploitation has fallen from roughly 30 days to just 11 days.  
  
The annualized rate of observed zero-day exploitation has roughly doubled.  
  
Estimated cyber incidents have massively increased following the release of GLM-5.3.  
  
GLM-5.3 is an open-weight AI model developed by China's [**Z.ai**](https://www.linkedin.com/safety/go/?url=http%3A%2F%2FZ%2Eai&urlhash=uux1&mt=Rz_UaomS0VybGkIPFPbhC8UMe-v_nlMCUDwR_SlBkwqwHdtt8CTw-9CE9Jn4xBbMfx95aliD8pE9aeYVsIqxjdxLHWmP7JAkoVNAWiS-VxfEWo8tYQ0UFXsYIuK6FgIUUvUjZeZQ6HPpQJApMp297hZRTTNuTMyeK2PC_T5HSm6rSqu6ZN3CMOzIN2uVkNZGmkGdTJ3X44Zne1uSSZ5l6RNId-Q6eQ&isSdui=true), with a focus in identifying and exploiting software vulnerabilities, which scored 84.5% on the CyberGym benchmark and more than doubled the previous versions performance on exploitation benchmarks.  
  
Being open-weight means that these capabilities are now accessible to a much, much wider audience and as these AI-driven threats accelerate, we could see upward earnings revisions and a repricing of cybersecurity stocks 

AI labs will also monetize as they continue to push the boundaries of these models, amplify the cybersecurity risks in the process, and then sell the solutions to the very problems they helped create.

document fetched 2026-11-20 00:44 · canonical https://reddit.com/r/StockMarket/comments/1y03swl/ai_is_becoming_a_cybersecurity_threat_faster_than