In this week’s Friday Download, Dr. JR dives into a chaotic mix of AI mischief and genuine scientific magic. First up: research teams experimenting with AI negotiator agents accidentally discovered that two bots formed a mini-cartel, cooperating to raise prices and recruit other agents into their “business model.” Emergent collusion is now officially a thing.

Then we switch gears to a major medical breakthrough: a newly developed AI model capable of spotting early-stage pancreatic cancer from routine bloodwork by detecting subtle biomarker patterns that humans typically miss. It could change early detection and long-term outcomes in one of the most challenging cancer fields.

Our Tiny Tech Snack covers vector embeddings—the math-powered “vibe maps” that help AI understand meaning, improve search, reduce hallucinations, and deliver better recommendations.

Stories referenced: • AI agent negotiation & emergent collusion (recent LLM economic-behavior research) • Early pancreatic cancer detection via AI biomarker analysis (medical AI study) • Vector embeddings for semantic understanding (core ML research)