Artificial intelligence is driving measurable change in healthcare and government operations through disciplined, narrow applications rather than grand frontier-lab promises, according to journalist and author Josh Tyrangiel during an appearance on The Prof G Pod with Scott Galloway. Discussing themes from his book, AI for Good: How Real People Are Using Artificial Intelligence to Fix Things That Matter, Tyrangiel emphasized that the technology’s real value depends entirely on the values and competence of the institutions deploying it.
The Cleveland Clinic Sepsis Model
According to Tyrangiel, the Cleveland Clinic integrated a sepsis prediction model originally created by Beijian Health, an organization founded by a computer scientist whose nephew died of sepsis. Sepsis kills roughly 300,000 Americans annually, which exceeds the combined toll of breast cancer, colon cancer, and opioid addiction. In 2021, the Cleveland Clinic identified that about 3,000 patients died yearly from sepsis within its care—a standard national rate that leadership considered unacceptable.
The prediction tool was integrated via Epic Health, the electronic health record system utilized by most American hospitals. Initial outcomes required extensive refinement and adaptation between intensive care units and general wards. Following a year of unglamorous iteration, the system achieved a 41% reduction in sepsis mortality, accounting for more than 1,000 lives saved, according to the reporting.
The model did not reach 100% accuracy in the ICU, where patient bodies constantly generate chaotic signals. The health system’s CEO noted that the tool did not need to be flawless to be useful, stating that the objective was to improve existing solutions rather than achieve absolute perfection.
Separating Technology From Tech Companies
Tyrangiel’s perspective on artificial intelligence shifted after a conversation with computing legend Danny Hillis, creator of cloud computing, who suggested separating the underlying technology from the commercial entities selling it. This distinction forms the core argument of AI for Good, which moves away from the binary hype cycle of utopian tech promises and apocalyptic warnings.
Instead of seeking sudden breakthroughs like the regeneration of limbs in laboratories, Tyrangiel expects incremental progress, such as cancer potentially becoming a manageable inconvenience over decades. The reporting indicates that AI succeeds when applied to specific problems with measurable outcomes by motivated personnel, and struggles when treated as a universal fix.
Broader Applications and Institutional Challenges
The conversation between Tyrangiel and Galloway also addressed the broader implications of artificial intelligence across various sectors, including government modernization, tax enforcement, warfare, and labor markets. Tyrangiel noted that technological advancement is progressing at a pace that outstrips the adaptation capabilities of American political and regulatory systems.
Additional topics explored during the podcast episode included potential disruptions to white-collar labor markets, the strategic threat of Chinese AI pricing strategies impacting American frontier labs, and concerns regarding how reliance on automated systems might affect the critical thinking skills of younger generations.