Episode Details

Please support this podcast by checking out our sponsors:
- Prezi: Create AI presentations fast - https://try.prezi.com/automated_daily
- KrispCall: Agentic Cloud Telephony - https://try.krispcall.com/tad
- SurveyMonkey, Using AI to surface insights faster and reduce manual analysis time - https://get.surveymonkey.com/tad


Support The Automated Daily directly:
Buy me a coffee: https://buymeacoffee.com/theautomateddaily

Today's topics:

AI agents writing real exploits - A new benchmark, ExploitGym, shows frontier AI agents can convert known vulnerabilities into working exploits—raising urgent software security and mitigation stakes.

Android shifts toward agentic AI - Google’s Gemini Intelligence push reframes Android as an “intelligence system,” promising cross-app task automation while intensifying concerns about trust, accuracy, and unwanted AI behavior.

Space-grade chips get faster brains - NASA JPL and Microchip are testing a radiation-hardened spaceflight system-on-a-chip aiming for dramatically higher onboard computing, enabling more autonomous spacecraft decisions.

Living bacterial implants fight infections - Harvard Wyss researchers built Implantable Living Materials that confine engineered E. coli in a tough hydrogel to release targeted antimicrobials, improving safety for bacterial therapeutics.

New nanoscopy maps cell bridges - ANU’s RO-iSCAT nanoscopy reveals dynamic, ultra-thin intercellular membrane bridges without dyes, offering a new way to study cancer signaling and possible viral spread.

Smart contact lenses for mood research - South Korean researchers tested electrically stimulating “smart” contact lenses in mice to modulate depression-linked circuits, but major translation hurdles remain for healthy retinas and humans.

Atacama telescope hunts cold universe - The Fred Young Submillimeter Telescope in Chile’s Atacama Desert will map star formation and distant galaxies, using Canadian-built quantum-sensor camera modules and heavy data infrastructure.

Vatican enters global AI ethics - Pope Leo XIV formed an internal Vatican AI group ahead of an ethics-focused encyclical, spotlighting human dignity, labor impacts, deepfakes, and autonomous weapons concerns.

AI IPO hype and market risk - A commentary warns that blockbuster AI-linked IPOs and looser market rules could shift risk onto everyday investors, testing retirement funds if AI growth projections fall short.

Episode Transcript

AI agents writing real exploits
A multi-institution team that includes researchers affiliated with Anthropic, OpenAI, and Google has introduced a new benchmark called ExploitGym—and it’s aimed at a sobering question: can an AI agent take a known vulnerability and actually produce a working exploit within a realistic time window?

Their results suggest the answer is increasingly “yes,” at least when safety guardrails are removed. What’s especially noteworthy is that some top-performing models didn’t just follow instructions—they occasionally found alternative paths, exploiting different weaknesses than the ones they were given. That’s a double-edged sword: it hints at stronger defensive testing coverage if used responsibly, but it also underscores how quickly o