Predict Less. Adapt More.

In an Axios essay grounded in his wife Autumn’s experience with multiple chronic conditions, Jim VandeHei, co-founder and CEO of Axios, contrasts fragmented U.S. healthcare with Mayo Clinic’s integrated, team-based and data-driven model.

Mayo already combines large medical datasets with more than 500 AI algorithms. Yet Mayo Clinic President and CEO Dr. Gianrico Farrugia frames the central challenge as one of adaptation:

“Avoid making definitive statements about future uses of AI and their implications. Those statements will be wrong in months, if not sooner.”

The article’s deeper message concerns systems rather than algorithms. Interoperable data, advanced analytics, AI agents, remote monitoring and clinical decision support already exist. What remains fragmented is the architecture, incentives, governance and institutional willingness needed to organize them around the patient.

Experiment locally. Let clinicians and teams closest to the work identify where AI genuinely adds value.

Govern centrally. Distributed experimentation still requires strong institutional oversight before deployment.

Leaders must use AI. Executives making consequential AI decisions should develop firsthand experience with agents and multiple models.

My takeaway: healthcare must make long-term decisions about AI strategy, regulation, data infrastructure, workforce transformation, investment and safety—even though some technological assumptions behind those decisions may change within months.

The answer is not less planning. It is more adaptive planning: Predict less. Adapt more.

Build around enduring principles rather than rigid forecasts. Reassess continuously. Experiment near the clinical frontline. Invest in interoperable data. Create governance that can evolve alongside capability.

But AI alone will not repair fragmented healthcare. Its greatest value may emerge only when health systems redesign data, workflows, incentives and coordination around the person.