Key ideas from Nikhil Buduma

• 10,000 people age into Medicare daily while we can't train doctors fast enough, creating unprecedented demand for AI-powered healthcare solutions
• 75% of clinicians at major academic medical centers use Ambience daily for 80%+ of visits, compared to 15-20% adoption rates for competing AI tools
• One health system projects $30 million in net new margin from Ambience implementation through improved RCM and throughput optimization
• AI clock speed fundamentally differs from product clock speed - companies must build for capabilities 18 months ahead, not current ones
• Most EHRs use mutable data structures that destroy decision traces, requiring complete rethinking of data architecture for effective AI
• The delta between consumer technology magic and healthcare tools has narrowed for the first time, changing physician attitudes toward technology adoption
• Foundation models still face massive last mile problems in healthcare due to messy data integration and complex quality definition challenges
• Individual engineers can accomplish dramatically more work with tools like Claude Opus, fundamentally changing hiring profiles and team composition