AI Loopholes, Health Scans, and Rebuilding Cities

Your face might know more about your health than your doctor does. And your glasses? They might be smarter than your phone.
This week, Jeff and Annie explore some of the wildest—and most useful—ways AI is showing up in the real world. From smart glasses that translate conversations at the dinner table to a selfie tool that predicts cancer risk, AI’s getting more and more physical. They also get into how AI is helping speed up wildfire recovery, what the FDA is automating next, and why surveillance tech is getting way more subtle.

We also discuss:

• Meta’s smart glasses + the University of Washington’s spatial translation
• How OpenAI’s HealthBench and Google’s AlphaEvolve are redefining AI benchmarks
• The FDA’s AI move, LEGO GPT, and facial biomarkers for cancer
• Surveillance, bias, and ethical gray zones

• Spatial Speech Translation by UW
• OpenAI’s HealthBench
• Google Expands Gemini to Android Devices
• Google DeepMind’s AlphaEvolve
• Trump Reverses Semiconductor Export Rules
• LegoGPT from Carnegie Mellon
• FaceAge AI Cancer Prediction
• AI Permitting in Eaton Fire Recovery
• Non-Facial Recognition Surveillance
• FDA’s AI-Assisted Drug Review Pilot