I did NOT expect this.
A general-purpose AI just did the unthinkable: Google’s Gemini 3 beat radiology trainees on one of the hardest medical imaging benchmarks on the planet.
51% vs 45%.
A generalist model. Not even a medical one.
This is the first real jaw-dropping moment in medical AI this year.
But here’s my hot take:
AI isn’t coming for radiologists.
AI is coming for leaders who still think they have time.
Gemini 3 didn’t just spot patterns; it reasoned through cases like a junior doctor.
And most enterprises still don’t have the governance, workflows, or strategy to handle technology that moves this fast.
The danger isn’t AI.
The danger is underestimating what happens next.
If your AI roadmap isn’t evolving at this speed, it’s already behind.
DM me if you want to build an AI strategy that won’t get crushed by the next breakthrough.
