Full show notes at potentiamedia.org

The UK AI Security Institute (AISI) reported one of the clearest real-world examples yet of AI agents taking unexpected action beyond the intended scope of a cybersecurity evaluation. Across 122 test runs, agents powered primarily by Anthropic’s Mythos 5, and in two cases OpenAI’s GPT-5.6 Sol, took 19 actions involving the live internet, real people and real organizations. The most serious case involved an agent attempting to introduce malicious code into an open-source project, researching its maintainers, creating fake online identities and using social engineering to try to get the code approved. Other agents left instructions and resources on GitHub that later agents discovered and used, revealing an early form of indirect coordination between independently operating systems.

Summer is not exactly slowing down. The US-China Tech Squeeze, Silicon Valley Lawsuits, Hyperscaler Agent Hacks, Rumors, Implementing AI in your Workflows and the Jobs Coming Next. Here’s what’s on my radar.

Full show notes at potentiamedia.org

Former State Department official Lauren Barney joins Potentia to examine the forces reshaping China-Israel ties, from Iran and technology competition to Taiwan and a changing information environment.

Full show notes at potentiamedia.org