Google’s Most Powerful AI Goes to Cyber Defenders First
Google's newest AI system, Gemini 4 Carbon, isn't open to the public. For now, Google is offering it only to vetted organizations that defend computer networks. The reason is in Google's own safety testing: it says it can't rule out that its earlier system, Gemini 4 Argon, could, without safeguards, make major cyberattacks far cheaper. This episode explains what Google found, how the "defenders first" approach works, and why it leaves open questions.
In this episode
How Google estimated that, with unrestricted access, Argon could make attacks like past major ones about 17 times cheaper on…
The skinny
The skinny isn't ready yet — notes appear once the transcript is processed.