Your body’s check engine light is flashing

AI is the new doctor’s waiting room

My best friend’s four-year-old daughter was just diagnosed with an aggressive brain tumor.

The doctors say she has 12-18 months to live.

Heartbreaking news like this tests my rational optimism. My wife says the scariest part is little Rayne felt fine - up to the moment she wasn’t.

How can we feel okay on the outside while something lethal grows inside? We track asteroids in outer space better than we track trouble inside our own bodies.

As a Rational Optimist, you know AI is making incredible advances in medicine. We’re curing previously incurable diseases with it. AI is even starting to come up with new drug mixtures on the spot to beat diseases we used to be helpless against.

37-year-old Joseph Coates, for example, has a rare blood disorder called POEMS syndrome. He was told to choose where he wanted to die.

But a rare-disease doctor in Philadelphia used a custom-built AI system to search for new cures hidden inside old drugs. The model trawled the medical universe and surfaced a cocktail of drugs no one had ever tried.

Joseph tried it, and he’s now in remission.

Treating diseases is one thing. But what if we could use AI to prevent them altogether?

We can. The vast majority of human health problems are fixable if we know about them early enough.

AI is giving medicine a new superpower: early detection.

I hope we’ll look back at 2025 as our emergence from the medical dark ages. That we relegate a whole class of diseases – those that can slowly kill us without us even knowing until it’s too late – into the “defeated” category, like polio.

My Whoop health tracker bracelet gives me a tiny preview of this future. My skin temperature often spikes 1-2 days before I feel sick.

Now imagine receiving a warning like that years in advance for serious health problems. Not just “you slept badly,” but a data-driven wake-up call: 

Your heart attack risk is rising.

Book a scan now.

Cardiovascular disease, aka heart attacks, is the No. 1 killer worldwide. It is mostly preventable if we identify the warning signs early enough. Innovators are racing to do just that.

London’s Imperial College created an AI-enabled stethoscope. In a trial of 12,000 patients, it doubled diagnoses of heart failure and tripled detection of atrial fibrillation… from just 15 seconds of listening to each patient’s heartbeat!

Researchers at Mass General Brigham built an AI tool that sifts through coronary CT scans and accurately estimates your 10-year heart attack risk. 

Cancer is the other big, often silent killer. Catching cancers early, when they’re still fixable, is the closest thing to a “cure” we’re going to get in the 2020s. 

Pancreatic cancer almost always hides until it’s too late. A new AI-assisted blood test called PAC-MANN reached about 98% specificity in early trials, giving patients a fighting chance to catch it while surgery can still help.

MIT and Mass General built an AI that forecasts lung cancer up to six years ahead of time, long before a radiologist would flag it.

At Harvard, the CHIEF AI model showed 94% accuracy in identifying 11 different cancer types.

See where this is going? Most of these tools are in trials, but they’re advancing fast.

It’s a matter of time until AI offers us a comprehensive radar that spots health “bogeys” far in advance enough for us to fight back and win.

 The medical robo-prophet

Delphi-2M is a new AI model that can forecast your odds of developing 1,200+ illnesses decades in advance.

Instead of predicting the next word in a sentence like ChatGPT, Delphi “reads” your health record. Then it maps your chances of developing everything from heart disease to cancers to skin and immune conditions, stretching 20 years into the future.

Delphi can tell you your chances of developing diabetes by age 70 based on your diet.

Or how the next 10 years of your life might go if you keep smoking, vs. if you quit.

In medicine, AI illuminates the invisible. It’s a sequence reader with perfect memory. It spots subtle patterns that humans miss, then can translate them into plain English next steps.

And unlike even the sharpest specialist, AI never forgets. It can recall the odd footnote buried on page 200 of a 1995 paper, and connect it to a scan you had last week. That’s its No. 1 superpower.

AI already has the knowledge level of a world-class doctor. By “world-class,” I mean in the top 1% of physicians worldwide.

OpenEvidence is the first AI system to score a perfect 100% on the US Medical Licensing Exam, the grueling test every doctor must pass.

OpenEvidence is already used by 40% of US physicians across 10,000+ hospitals.

AI’s No. 2 medical superpower is access.

With no more than a smartphone and an internet connection, you can now “chat” with a world-class doctor, 24/7/365. You don’t need a special AI model, or to book an appointment 3 months in advance, or to pay an insurance deductible.

ChatGPT already has the knowledge level of an expert doctor. It’s open to everyone. If you’re not using it yet, you’re leaving a life-improving opportunity on the table.

You know one of our mottos at Rational Optimist Society:

This is the worst AI doctors will ever be.

They’ll learn to spot more diseases, further in advance, and with higher accuracy.

Follow the curve out a few decades, and most people in the future need not die of disease. They’ll die of old age.

AI will save lives, money, and years of human potential. At the end of the day, it means more extra years with the people we love most. And isn’t that the point?

It’s still up to you to act though.

As I mentioned, most of the breakthroughs I’ve shared are still in trials. But you don’t need to wait years to benefit. Regular ol’ GPT-5 scored 97% on the US Medical Exam. And you can access it for free.

Wearable health trackers are a big part of this story. They too get better every year.

My Whoop wristband already tracks my heart rate, sleep quality, workouts, blood oxygen, and more. Apple and Stanford recently trained an AI model on 2.5 billion hours of wearable data like that.

It was able to spot infections or injuries and monitor sleep…

Better than hospital equipment!

Curious about my own body, I exported an entire year of my Whoop data and plugged it into ChatGPT. I asked: “What’s really going on with my body?”

It told me my resting heart rate (53 bpm) was solid and improving. But I only average about 7.1 hours of sleep a night, and my body is screaming for 8.8 hours.

That’s a problem AI cannot fix. It’s called “having a 10-month-old son.”

We still track our cars’ health better than our own hearts. AI is starting to give us a check engine light for our bodies. The “dumb” numbers from a wristband are already becoming early warning signals you can act on.

The future of health care isn’t about hospitals buying bigger machines. It’s about individuals, you and me, choosing to use the tools already in our hands.

Here’s how.

Start by showing AI your health. Gather your test results and scans. Dump out your medicine cabinet and snap photos of every bottle.

Then tell the AI about you: your conditions, surgeries, meds and doses, allergies and family history. I use “voice mode” for this, which lets you talk to the AI instead of typing.

Once you’ve got the basics in, start asking questions: “Summarize my health in plain English. Tell me what I should be watching.”

Then you can go deeper. If you’ve had a coronary CT, ask it to analyze the image and give you a risk score.

This isn’t something you do once and forget. Create a “project” on ChatGPT with all your data you can come back to.

To be clear, AI doesn’t replace human doctors. But it is a new first line of defense that’s easy to access for everyone.

Going to a human doctor is a pain. You book an appointment, sit in a waiting room among sick people, deal with insurance, and agree to pay whatever it costs, which is often a mystery.

Then, you get 15 minutes with the doctor. They might be rushed and stressed. They may or may not appreciate you asking questions. And that’s if you live in a place with easy access to doctors, which billions of people do not.

AI solves all of that. The best waiting room is no longer in a hospital - it’s your living room. You should be asking GPT questions like: “What should I double-check with my doctor?”

Don’t sit back and wait for the healthcare system to take care of you.

Take ownership of your health.

It is extremely sad that not much can be done in situations like my friend is facing with his little girl. No parent should ever hear that sentence, and no child should ever carry that weight.

How can we be rational optimists when tragedy can strike any of us out of nowhere?

For me, the answer is: we can make things better. With another decade of progress, AI will catch most tumors when they’re still small and curable. Although we cannot eliminate tragedy, we can vastly reduce it.

Now picture 2040. You’re in the clinic. A test flags the same disease that took your dad.

But this time it’s early. Your smartwatch picked up the first signs. The AI tells your doctor to start a therapy that didn’t exist a decade earlier.

You go home shaken but alive, and hug your kids.

Rational Optimists, our job is to help this future arrive sooner.

Let’s be the generation that built an early warning system for our bodies.

Say a prayer for Rayne tonight, and hug your loved ones a little tighter.

Hey, you’re one of the first 20,000 members of Rational Optimist Society

Since starting ROS less than one year ago, we have amassed 20,000 members, all through word of mouth.

We have big plans that include:

• Creating educational material for children.
  

• Building the Innovator’s League to connect top innovators and entrepreneurs.

We’re also considering ways to better foster a community.

As one of our original 20,000 members, we value your ideas. Let us know what you think in the comments. 

We suspect there are at least 1 million rational optimists out there, just waiting to be awakened. Sometimes one illuminating story is all it takes. Please forward this letter to your friends and family.

Onward and upward!