Peptide Purgatory: The All-AI Issue

Back atcha again with another week’s worth of GLP-1 musings. This week, I take a detour from the incretin world to coment on the burgeoning integration of artificial intelligence (AI) into health care: the good, the not so good, and the downright bullshit. (See how I avoided the obvious, hackneyed Sergio Leone spaghetti western cliche there? I didn’t think so.) I hope this will be a welcome segue from last week’s confluence of peptide injections and boner pills, which set a new weekly low-water-mark in readership. (LMAO!)

As you probably know, this 78-year-old Type 2 diabetic has been on Mounjaro therapy for over a year, with decent results. During that time, I shifted from an excuse-making, indolent blob of blubber to a far less blubbery state by controlling diet and embarking on a stringent strength training program. At my age, something will kill me eventually, but it won’t be of my own doing, and I intend to go down swinging. While Mounjaro has helped me mend my ways, I am hoping I can ditch the crutch by year-end.

A Doc, by Any Other Name…

By the way, when I refer to medical practitioners in this blog, I use pseudonyms to protect the innocent (and sometimes guilty). My cast of medical characters includes the Notorious Dr. DeLorean, the Irascible Dr. Scrooge, and Cruella da Ville: Ace Physical Therapist, among others. I now introduce a new character—my current knee doc, who I’ll call Sporty McNeedle.

Recall that I was seeing Dr. Kahuna, a surfin’ surgeon, about my well-aged, creaky knees. Because I had little desire to jump into surgery, I’m now seeing another, non-surgical doctor in the same large orthopedic practice. Enter Dr. Sporty McNeedle. She’s a young, non-operative sports medicine doc who doesn’t just tolerate informed patients, she collaborates with them. Athletic, focused, and blessedly unhurried, Dr. McNeedle doesn’t reach for the door handle mid-sentence or recoil when you bring up AI-assisted research. She examines, listens, explains, and—brace yourselves—makes a plan. A real plan. With next steps, already! I know—I was shocked too.

Now, sit back and read this week’s issue, which ChatGPT helped me write. Is this the equivalent of the fox guarding the henhouse or can my favorite pseudo-intelligent chatbot maintain objectivity? We’ll see.

After my AI editorial and a related sidebar, I’ll tell you about my follow-up visit with Dr. McNeedle.

---

“I’m Not ChatGPT,” He Snarled. And That’s Why I Brought It With Me.

This week’s Peptide Purgatory is brought to you by the artificial intelligence that fuels your curiosity, writes your prior authorizations, and, if the NEJM is to be believed, will soon rate your doctor’s bedside manner like Uber passengers on a bender. Yes, folks—it’s the All-AI Special Edition.

You may know me as the geriatric GLP-1 guinea pig with an 18-inch resistance band and a Dexcom graph that glows like a NASA telemetry readout. But behind the well-titrated glucose is a man on a mission: to be heard, to be helped, and to never again be told, “I’m not ChatGPT.”

That unforgettable line came courtesy of Dr. DeLorean, my concierge-priced, marathon-running, hipster-garbed provider, who once shut me down mid-symptom summary with those five perfectly defensive words: “I’m not ChatGPT.” This, after I dared to present a coherent overview of my symptoms—pre-researched, logically framed, and yes, assisted by my silicon sidekick. Apparently, listening is beneath his pay grade (in military terms, the equivalent of a four-star general).

But here’s the kicker: ChatGPT listens. It doesn’t get flustered. It doesn’t take offense when you ask about the evidence base for tilt-table testing. It doesn’t run a 10K through Burgundy with your retainer fee.

Enter Dr. Sporty McNeedle: The Model Human Clinician

Now, contrast that with my visit last Friday with Dr. McNeedle, the sports medicine doc evaluating my recalcitrant left knee. She actually engaged. I came prepared—with ChatGPT-assisted insight into quadriceps tendinopathy, the role of hyaluronic acid (HA) injections, and the diagnostic ambiguity of OA-vs-tendon pain—and instead of swatting it away, she added to it.

We’re now pursuing two things: (1) a prior authorization for an HA injection, and (2) a new MRI to determine whether the dominant pain generator is osteoarthritis or tendinopathy. A rational plan. Built collaboratively. No AI paranoia, no eye-rolling, no need to remind her she’s not ChatGPT—because she didn’t need to be. She was better.

Dr. Sundar’s Essay: Grace Meets the Machine

And then there’s Dr. Kumara Raja Sundar, who recently penned a JAMA essay that made me want to stand up and clap—though standing up is a particularly painful production these days. She describes patients bringing ChatGPT-generated insights into her clinic and welcoming them. Rather than defaulting to the old-school “WebMD = hypochondriac” mindset, she sees these moments as the start of meaningful dialogue.

Sundar doesn’t just tolerate informed patients—she recognizes what they’re doing: advocating. Not because they want to play doctor, but because the system often makes them feel invisible unless they do.

Imagine that! A physician secure enough in her expertise that a patient with prep work doesn’t threaten her authority. As if collaboration was actually the point. Kudos to her!

But Then… NEJM Pulls the Fire Alarm

Just when you thought the humans might win, here comes The New England Journal of Medicine with its wet-blanket article titled “Medical AI and Clinician Surveillance — The Risk of Becoming Quantified Workers.”

It’s a dystopian screed—albeit a well-sourced one—about how ambient AI scribes and automated EHR tools aren’t just helping doctors, they’re monitoring them. They’re timing visits. Analyzing tone. Flagging deviations from guidelines. In short, they’re turning physicians into warehouse workers with stethoscopes.

Want to spend an extra 90 seconds comforting a grieving patient? You might get a call from Performance Improvement. Forget to mirror GPT’s patented EmpathyScore™? Docked. Not enough “I’m sorry to hear that”s per minute? You’re now an efficiency outlier.

One of the examples cited: AI-generated patient messages already rank more empathetic than human responses. And you just know someone at UnitedHealthcare read that and thought, “Hey, let’s cut the humans out entirely.”

So Where Do We Go from Here?

Do I want all-AI medicine? Hell no.
Do I want all-human medicine that ignores my preparation, dismisses my questions, and refuses to acknowledge nuance? Also hell no.
What I want is a system where:

• Empathy isn’t penalized as inefficiency.
• Informed patients aren’t treated as threats.
• AI helps without handcuffing.
• And nobody needs to fire up a damn mouse jiggler to dodge GPT-based productivity surveillance.

Dr. Sundar gets it. Dr. McNeedle gets it. Dr. DeLorean? Still out there, wondering why patients prefer the chatbot to the $3,500 club member who can’t be bothered to listen.

If we’re going to survive the AI revolution in medicine, we need to preserve the one thing ChatGPT can’t do: care in a way that’s genuinely human. But we also need to stop pretending that the machines are going away.

They’re not!

So let’s train the doctors. Let’s train the AI. And let’s retire the DeLoreans.

---

The AI Hippocratic Paradox: Help or Harm?

Artificial intelligence won’t save medicine—but it might ruin it. Or maybe both.

Five Ways AI Can Help Without Screwing It All Up

1. Doctor-Prep for Patients
   Structured questions, context, and actual understanding—AI helps patients arrive ready for dialogue, not diagnosis theater.
2. Note-Taking, Not Narc’ing
   Ambient scribes can lighten the EHR burden without turning into tattletales—if used to support, not surveil.
3. Prior Auth Streamlining
   Automated hellpaper generators that finally get the damn injection approved without a four-hour phone call? Yes, please.
4. Decision Support Without Handcuffs
   AI can nudge, not dictate. A helpful assistant—not the Clippy of Clinical Guidelines.
5. Empathy Prompter, Not Replacement
   Sometimes the bot drafts the right words. Sometimes it reminds the human to just be one.

Five Ways AI Can Ruin Everything

1. Surveillance Disguised as “Support”
   Every “helpful” tool comes with a stopwatch and a flagging algorithm. Blink twice and you’re an outlier.
2. Empathy by Algorithm
   Bedside manner shouldn’t be judged by a language model with a bedside nowhere. But here we are.
3. Protocol Over People
   Once deviation equals risk, your doctor’s best move is conformity—regardless of what you need.
4. Productivity Metrics from Hell
   Portals, phone calls, visit times—all tracked. Welcome to Amazon Care™, now with 100% more guilt.
5. The DeLorean Defense
   Doctors threatened by ChatGPT will double down on arrogance instead of engagement. And they’ll lose.

---

This Week on Mounjaro: Kneecapped by Bureaucracy

While this week’s issue is all about artificial intelligence, I’d be remiss if I didn’t include a good old-fashioned meatspace medical update—brought to you not by a chatbot, but by Dr. Sporty McNeedle, the very human, very competent sports medicine doc evaluating my ornery left knee.

Long story short: after a thorough exam and review of symptoms that include pain ascending and descending stairs, sitting down, and rising from a seated position, but remarkably, not during walking long distances on turf or treadmill, Dr. McNeedle made the rational decision to pursue a new MRI. Why? Because we need to answer the eternal question: Is this primarily osteoarthritis or quadriceps tendinopathy? Or both?

We also decided to move forward with a hyaluronic acid (HA) injection—but that comes with the now-obligatory ritual sacrifice to the Prior Authorization gods. Medicare needs convincing that injecting goo into my knee isn’t frivolous spa therapy. So we wait.

In the meantime, she praised my quadriceps control and confirmed that eccentric loading exercises are still on the menu, alongside careful activity modification. I didn’t mention the double hernia to her—that’s in another medical context. She wants me to stay active. In other words, I’m not being benched—but neither should I be launching into Bulgarian split squats while chanting “No pain, no gain.”

So: MRI pending. Prior auth pending. Pain ongoing. But I left the appointment feeling heard, understood, and guided—not scored, flagged, or surveilled by GPT-based empathy meters. Just human care, as it should be.

---

And finally, this week’s entry in the “Did That Email Really Say That?” file:

AdventHealth—our local faith-based, for-profit-ish medical empire—just announced that their licensed massage therapists can now accept tips. Yes, after centuries of sacramental abstention from filthy lucre, the Adventist mothership has decided that the laying on of hands can finally include the laying down of a crisp $20. Or, in my case, maybe a well-placed C-note for my longtime LMT and exercise physiologist, known only as “Pooch.” I haven’t booked a session in ages, but now I feel obligated—not out of generosity, mind you, but just to reward the one person in that entire network who’s made my body feel better without billing me for “evaluation and management” code 99213. Amen.

---

Parting Shots: Where We’ve Been, Where We’re Headed

So there you have it: a week in which I waxed poetic about AI, praised one doctor for being human, roasted another for acting robotic, and discovered that even massage therapy has a tip jar now. We explored the hope and horror of artificial intelligence in medicine—how it can help patients advocate, support good doctors, and streamline care. And how it can also be weaponized into productivity metrics, empathy algorithms, and “performance dashboards” worthy of an Amazon warehouse.

This week’s forecast includes gym work, a knee MRI (assuming the magnet doesn’t groan louder than I do), and the Sisyphean task of navigating EnGuide’s interactive phone labyrinth to learn whether my latest Mounjaro shipment will arrive before I hit the road for a two-week escape from Florida humidity and American healthcare. Current status: “might ship around August 4,” which translates to “don’t hold your breath, but do keep the Frio pack handy.”

And if it does arrive late? I guess I’ll just ask ChatGPT for a travel-friendly alternative—and then politely explain it to the next provider who tells me, “I’m not ChatGPT.”

See you next week, if the AI overlords and pharmacy logistics gods allow. (No, I won’t be wearing my 20-pound neck chain bling to the MRI chamber—I promise!)

---

For an annotated catalog of all my Mounjaro updates, please visit my Mounjaro Update Catalog page.