I’ve been in medicine for over twenty years, which means I’ve had a front-row seat to a remarkable amount of change.
When I was in medical school, smartphones were still relatively new. Electronic medical records were far less common than they are today. Telemedicine occupied a small corner of healthcare rather than becoming a routine part of clinical practice. Even the way physicians accessed medical literature looked dramatically different.
Medicine has never been a static profession. New drugs emerge, surgical techniques improve and diagnostic tools become more sophisticated. Every generation of physicians inherits a version of medicine that looks different from the one that came before it.
For most of my career, though, those changes followed a pace that felt manageable. Technology advanced, institutions adapted and eventually a new equilibrium emerged. Artificial intelligence is the first development I’ve encountered that makes me wonder whether that sequence is about to break down.
Most conversations about AI focus on what the technology can do. Every week seems to bring another impressive capability, another model release, or another example of software performing a task that previously required specialized expertise. The technology itself is fascinating, but I’ve found myself becoming increasingly interested in a different question.
What happens when technology starts moving faster than the institutions built around it?
That may sound like an abstract concern, but medicine provides a useful example.
How Medicine Traditionally Adapts
One reason medicine has remained remarkably resilient over time is that it tends to move carefully.
That can be frustrating when you’re dealing with outdated software, cumbersome regulations, or administrative systems that seem frozen in time. But caution exists for a reason. Medical decisions affect real people, and the consequences of getting things wrong can be significant. As a result, healthcare institutions are generally designed to prioritize stability and safety over speed.
Medical school curricula also evolve gradually. Residency requirements change slowly. Licensing examinations undergo extensive review before major revisions occur. Hospitals often spend years evaluating new technologies before implementing them broadly.
The basic question that always has to be answered is this: is the new technology demonstrably superior in some way to the status quo? If not, the technology will generally fall by the wayside and be passed over for adoption.
For most of modern medical history, that approach has worked reasonably well. Technological progress occurred, but it typically unfolded on a timeline that allowed institutions to adapt. New imaging technologies, minimally invasive surgery, robotic platforms and advances in pharmaceuticals all changed medicine for the better. Yet those changes generally occurred over years or decades rather than months.
There was time to study them.
Time to debate them.
Time to determine where they fit.
The institutions responsible for training physicians were never expected to transform overnight because the world around them rarely required it. That assumption is now worth revisiting.
Why This Feels Different
Artificial intelligence is hardly the first technology to generate excitement. Medicine has seen plenty of innovations arrive with promises that ultimately proved exaggerated. What’s different about AI is not simply its capability. It’s the pace of improvement.
ChatGPT was based on GPT-3.5 and was released in 2022. GPT-4o arrived in 2024. GPT-5 followed in 2025. Since then, increasingly sophisticated multimodal and reasoning models are emerging every few months! Tasks that seemed impossible for AI systems just a few months ago are now routine.
The exact trajectory remains uncertain, and predictions in technology are notoriously difficult. History is full of experts who underestimated change and others who overestimated it. Still, it’s difficult to ignore how quickly these systems have improved.
Research published in journals such as Nature from 2023 and 2025 has demonstrated increasingly strong performance of large language models across a range of medical tasks, including clinical reasoning, diagnostic support, evidence synthesis, and clinical documentation assistance. At the same time, health systems around the country are already adopting AI-powered tools for ambient clinical documentation, administrative workflows and decision support.
The conversation is no longer theoretical. The technology is already entering everyday practice.
Whether it ultimately transforms medicine as profoundly as some believe remains to be seen. What seems increasingly difficult to dispute is that the timeline feels different from previous waves of innovation.
Historically, major changes in medicine unfolded over decades. Artificial intelligence appears to be evolving on a timeline measured in months.
Medicine Is a Pattern Recognition Profession
Part of the reason this development feels particularly significant is that medicine itself is fundamentally a pattern-recognition profession.
Medical students and residents spend years learning which details matter and which do not. We learn to recognize presentations, connect seemingly unrelated information, identify abnormalities and narrow a nearly infinite set of possibilities into a manageable diagnosis. It’s the reason why physicians eventually make such a big salary when we are practicing attendings.
A patient walks into the room with fatigue, weight loss and abdominal pain. Most people hear a collection of symptoms. A physician immediately begins sorting through patterns.
- What fits?
- What doesn’t?
- What needs further investigation?
- What represents the greatest risk?
That process becomes increasingly refined through years of training and experience. Much of medical expertise ultimately rests on this ability to recognize patterns and apply them in context. This is where AI becomes particularly interesting.
The strengths of modern AI systems overlap with many of the same cognitive tasks physicians spend years developing. They can process enormous amounts of information, identify patterns across large datasets, summarize complex literature and generate surprisingly sophisticated analyses from incomplete information.
I experienced this the other day, when a medical LLM helped me diagnose a rare case of viral orchitis when provided with the clinical history and ultrasound image.
None of this means physicians are becoming obsolete. Medicine involves far more than diagnosis alone. Communication, trust, judgment, ethics, procedural skills, leadership and the ability to navigate uncertainty remain deeply human qualities.
Still, it’s difficult to ignore the fact that some of the activities traditionally associated with expertise are becoming increasingly accessible through software. That possibility deserves serious attention.
The Institutional Challenge
The question that keeps pulling my attention back is not whether AI will improve. I suspect it will. The evidence for this becomes more irrefutable every month that passes.
The difficult question is whether the institutions surrounding medicine can adapt quickly enough when it does.
Medical schools cannot redesign curricula every six months. Residency programs cannot restructure training pathways every year. Licensing organizations, accreditation bodies, hospitals, insurers, and professional societies all operate within systems that were built for stability.
Again, that’s usually a feature rather than a flaw. The challenge arises when the external environment begins changing faster than the system’s ability to respond.
Imagine a student entering medical school today. By the time they complete four years of medical school, residency and potentially a fellowship, the world they enter could look dramatically different from the one that existed when they began training.
That has always been true to some extent. But now, with AI, it’s a certainty.
Looking Ahead
The question that keeps coming back to me isn’t whether artificial intelligence will continue improving. It certainly will, and soon.
Every indication suggests that these systems will become more capable, more accessible and more deeply integrated into everyday life over the coming years. Exactly how far they’ll go remains uncertain, but the direction seems increasingly clear. What feels less clear is whether the institutions surrounding medicine can adapt at the same pace.
Medical schools, residency programs, licensing organizations, hospitals and professional societies all play an important role in maintaining standards and protecting patients. Most were built to evolve carefully, and for good reason. Healthcare is not a field where rapid, untested change is usually desirable.
But there is a meaningful difference between moving carefully and moving slowly. Historically, those two things have often looked similar because technological change itself tended to unfold gradually. Institutions had time to observe, evaluate, debate and eventually adapt.
I’m no longer sure that assumption holds. For most of my career, it felt reasonable to believe that the future would arrive incrementally. New technologies would emerge, medicine would adapt and training pathways would evolve along the way.
Artificial intelligence is the first development that has made me question whether that process will continue to work as it always has.
Not because physicians are disappearing and neither because medicine is becoming obsolete, but because the timeline itself is changing. And if that’s true, then the most important questions over the next decade may have less to do with technology than with adaptation.
- How should physicians be trained in a world where AI will beat the best human diagnostician?
- What skills become more valuable as AI becomes more capable?
- How should medical education evolve?
- What responsibilities do professional organizations have in helping physicians navigate these changes?
The day may soon come when physicians are accused of negligence not for using AI, but for failing to use it.
I don’t pretend to have the answers, but I do think we need to start asking the hard questions. Because whether we feel ready or not, this conversation is already here.
In Part 2, I’ll explore what the next decade might look like if current trends continue and how medicine could change as a result.
Sources
- Stanford Institute for Human-Centered Artificial Intelligence (HAI): AI Index Report — https://hai.stanford.edu/ai-index-report
- NEJM AI — https://ai.nejm.org
- Goldman Sachs Research: The Potentially Large Effects of Artificial Intelligence on Economic Growth — https://www.goldmansachs.com/insights/articles/generative-ai-could-raise-global-gdp-by-7-percent
- McKinsey Global Institute: The Economic Potential of Generative AI — https://www.mckinsey.com/capabilities/quantumblack/our-insights/the-economic-potential-of-generative-ai-the-next-productivity-frontier
Want to support the blog?
- Join our investor club at Cereus Real Estate
- Visit my Recommendations page
- Check out my wife’s food blog: Eat Dessert First
- Check out my TikTok channel
- Follow me on Instagram
- Follow me on YouTube
- Contact me with questions
Experience the financial benefits of real estate without dealing with the headache!
