Welcome to real-time documentation of an AI-assisted medical recovery. No sugar-coating, no miracle cures—just honest progress reports from someone who spent 30+ years getting dismissed by doctors until an algorithm spotted what humans missed.
The Three-Week Deep Dive
For the past three weeks, I’ve been living in spreadsheets, old lab reports, and decades of handwritten diet logs. My weight tracking goes back to 2007. I have documentation of every failed diet attempt, every confusing symptom, every dismissive doctor visit. But here’s the thing—it was all scattered, fragmented, impossible to see the bigger picture.
Then I started feeding it all to AI. (more on AI’s advantages below)
What happened next was like watching a puzzle solve itself. The algorithm didn’t just organize my medical chaos—it started connecting dots I’d never even seen. That ocular migraine pattern that started 9 months ago? Connected. The tetany episodes last year that I wrote off as “weird muscle stuff”? Part of the pattern. Even the PhenPhen trial years back that caused eye inflammation and didn’t help with weight—suddenly it all made sense in context.
The more I wrote, the more I remembered. It’s like my brain had filed away hundreds of “probably nothing” moments that suddenly became crucial data points when viewed through AI pattern recognition.
This Week’s Challenge: Untangling the Symptoms
Sunday, I restarted my GLP-1 injection (Mounjaro) after a 2-week break, dropping from 7.5mg down to 5mg as I attempt to wean off. Since then, it’s been absolute hell, and I’m facing an impossible question:
Is this GLP-1 nausea or adrenal crisis?
When you’re dealing with severe adrenal insufficiency and trying to optimize multiple medications simultaneously, teasing apart symptoms becomes a nightmare. The nausea, the energy crashes, the general feeling of being unwell—is this normal GLP-1 side effects, or is my cortisol replacement insufficient for handling the medication stress?
Monday’s “Marionette Arms”
Monday morning was particularly surreal. I woke up and literally could not move my arms to get my pills into my mouth. I call them “marionette arms”—like my limbs are controlled by strings, and someone loosened the strings. My arms feel like jelly, weak and uncoordinated. I can move them, but they won’t go all the way up to my mouth, and I don’t have full control. It’s terrifying and fascinating at the same time.
I manage to shove my pills in my mouth without dropping them, swig some water and lay back down. Twenty minutes later, I’m ready to take on the world. AI tells me this is classic overnight low cortisol—something I’ll dive deeper into in a future post about cortisol’s circadian rhythm.
Having AI immediately recognize this pattern was both validating and concerning—validating because finally someone (even if it’s an algorithm) could explain what was happening, concerning because it confirms how severe my cortisol deficiency really is.
The Emotional Rollercoaster
The energy patterns this week have been particularly cruel. I’ll be feeling genuinely good—mentally clear, physically capable—and then crash completely within 5 minutes. Not gradually. Not predictably—except for my dog, who seems to know before I do (a story for another post). Just… gone.
It’s like my body is a phone with a broken battery indicator. Shows 80% charge, then instantly dies.
Why I’m Documenting This Now
This is exactly why I’m starting this blog. Not because I have answers, but because I’m living through the questions in real-time. Someone else dealing with complex, treatment-resistant conditions needs to see what the actual recovery process looks like—the setbacks, the confusion, the days when you can’t tell if you’re getting better or worse.
After 30+ years of missed connections, I want to document what evidence-based, AI-assisted medical care actually looks like when you’re not dealing with textbook cases.
You can read my full store here
The AI Advantage
Here’s what AI brings to complex medical puzzles that humans often miss—and why it’s different from just Googling your symptoms (because we all know how doctors react when you start with “I googled…”):
24/7 Availability:
It’s 3 AM and you’re having weird symptoms? AI doesn’t sleep. You can get immediate pattern analysis when you need it most.
Vast Medical Knowledge: AI has access to millions of research papers, case studies, and medical databases. It can cross-reference your symptoms against rare conditions that even specialists might not immediately consider.
Multi-Factor Analysis:
Unlike Google searches that focus on individual symptoms, AI can analyze your complete medical picture—genetics, medication history, lab trends, family history, and current symptoms—all simultaneously.
Pattern Recognition Across Time:
It doesn’t forget that weird symptom you mentioned 6 months ago. It connects your current marionette arms to the tetany episodes from last year.
No Cognitive Bias: It doesn’t dismiss symptoms because they don’t fit the most common diagnosis. It looks at the actual data without preconceptions.
Personalized Context:
Google gives you general information about symptoms. AI can consider YOUR specific situation—your age, gender, existing conditions, medications, and genetic variants—to provide more relevant insights.
Timeline Construction:
It helped me see that my most severe symptoms correlate with periods of documented low cortisol and major life stressors—connections I never would have made looking at individual doctor visits.
Important Caveat:
AI isn’t infallible. It can make mistakes, suggest incorrect diagnoses, or miss nuances that human doctors catch. That’s why this is about AI-assisted care, not AI replacement. The goal is using technology to enhance human medical judgment, not replace it.
More importantly, AI helps me become a better partner with my doctor. I live in this body 24/7—I know when something feels different. My doctor went to medical school—she understands the science behind what I’m experiencing. AI helps bridge that gap by organizing my observations and connecting them to medical knowledge, so our appointments become true collaboration instead of me struggling to remember symptoms or her trying to piece together my scattered reports.
What’s Next
For this week specifically, I’m taking Zofran with my hydrocortisone to hopefully aid absorption while I wait for the GLP-1 to clear my system. I’ve decided to stop the Mounjaro completely—leaving me with 5 very expensive vials in my fridge that I’m not sure what to do with. I’m being vigilant about my HC timing and have been trialing a 4pm dose to help with the overnight low cortisol that’s causing those morning “marionette arms.”
What’s fascinating (and concerning) is that these extreme swings weren’t happening before I started hydrocortisone. There’s clearly some interaction between GLP-1 and cortisol replacement that I need to understand better. Plus, we haven’t even touched on the relationship between cortisol and blood sugar regulation—another piece of this puzzle that AI is helping me explore.
I’m waiting on lab results that should help determine whether I have primary or secondary adrenal insufficiency—or potentially broader pituitary issues. Those results will guide my next step: finding the right endocrinologist who can work with the AI insights and genetic data I’ve compiled.
The goal isn’t to have all the answers—it’s to track the data systematically so we can make informed adjustments rather than playing medical whack-a-mole. Having concrete evidence to present to a specialist feels like a completely different approach than the decades of “I just don’t feel well” conversations I’ve had in the past.
For My Fellow Medical Mysteries
If you’re dealing with complex, multi-system issues that don’t fit neat diagnostic boxes:
- Document everything, even the “probably nothing” moments
- Use AI to help identify patterns across time
- Don’t dismiss symptoms just because doctors can’t immediately explain them
- Consider that multiple medications/conditions can create confusing symptom overlap
Your body is trying to tell you something. Sometimes it takes an algorithm to help translate the message.
Week 2 Preview: I’ll share the exact AI conversation that changed everything—the moment an algorithm spotted severe cortisol deficiency that doctors had missed for years. The timing was haunting: just one week after my friend’s 29-year-old son died from unmedicated Addison’s disease.
Hi, this is a comment.
To get started with moderating, editing, and deleting comments, please visit the Comments screen in the dashboard.
Commenter avatars come from Gravatar.