We talk with Boston University researcher Nicole Spartano, PhD about what continuous glucose monitoring can reveal in people without diabetes and why those patterns may matter for future diabetes and cardiovascular risk. We dig into Framingham Heart Study findings (https://diabetesjournals.org/care/article/doi/10.2337/dc26-1197/172437/Beyond-Traditional-Glycemic-Measures-CGM-Glycemic), why A1C may not match CGM in non-diabetes ranges, and how to avoid overreacting to glucose spikes.

An interview with Shannon Lantzy, PhD about why diabetes technology should be judged by how many “spoons” it gives back, not only by A1C and time in range. We dig into FDA decision-making, cybersecurity, faster software validation, and where AI can truly help without creating unsafe risk.

• Shannon’s path from math and NASA consulting into FDA-focused decision science
• Why patient preferences can change how benefit risk gets weighed
• Cybersecurity realities for connected body-worn diabetes devices
• Spoon Theory as a practical model for chronic disease burden
• Spoonshot as a push to speed up safe medtech software throughput
• The real bottlenecks: validation, integration, interoperability testing
• Why medtech cannot ship updates like consumer apps
• Four ways to think about AI in medical devices
• Unmet needs: personalization for pregnancy, young children, alarm tolerance, and affordability
• Why explicit tolerances and clear validation define AI’s limits

We talk with Amir Hayeri, founder and CEO of BioConscious Technologies, about using machine learning to predict glucose trends and turn CGM streams into actionable clinical foresight. We dig into accuracy, trust, liability, and why clustering dysglycemia may reveal risk that HbA1c can miss.

• Amir’s origin story and the prevention-first philosophy behind BioConscious
• Predicting glucose 60 minutes ahead using CGM data and how accuracy is measured
• Why an app without a clinical workflow has limited impact
• Population triage for clinics using clustering to surface high-risk patients fast
• Clinician concerns about liability and the fear of replacing doctors
• Building trust through explainable AI and a clear user interface
• Plans for studies and publication areas including gestational diabetes and dysglycemia
• The “glucose atlas” idea and what CGM can show in non-diabetic users
• Limits of mixing CGM with wearables data and why model complexity can hurt performance