Ground Truths Joanna Stern: An AI Immersion for 365 Days
May 13, 2026
Joanna Stern, technology journalist and Emmy-winning filmmaker who spent 12 years at The Wall Street Journal, spent a year testing AI across daily life and health. She talks about uploading bloodwork to Notebook LM, AI-assisted mammograms and dental deep-cleans, compute and energy costs, chatbots as medical aides, AI therapists and trainers, robot companions, and autonomous rides.
AI Snips
Chapters
Books
Transcript
Episode notes
AI Made Sense Of My Scary Blood Results
- Joanna Stern uploaded her lab PDF into Google's Notebook LM and let two AI hosts generate an automated podcast that explained her elevated LDL.
- The AI repeated the nurse's voicemail advice but made the results easier to understand, prompting Joanna to change behavior and fix her cholesterol the following year.
Compute Costs Are Real And Hard To Measure
- Joanna highlights compute cost as a hidden but important part of AI's true cost, especially for training and repeated inference.
- She visited a data center, learned inference is less intensive than training, but found it nearly impossible to estimate personal compute spend like "I wish I knew at the end of the year I used $10,000 of compute."
AI Flagged A New Spot On My Breast Ultrasound
- Joanna got AI-assisted mammogram (ScreenPoint/Transpara) and ultrasound (Kiosk) readings alongside radiologist Lori Margulies at Mount Sinai.
- The ultrasound AI flagged three areas; two were stable from six months prior, but one new AI-flagged spot led to MRI follow-up.
