ChatGPT as Coach - Part 1 : Halfway Point

I thought I was doing well on my weight.  After all in 2020, I lost 30 lbs and had kept the weight off.  That is why I was surprised when my Doctor on my last visit said I was borderline obese.  What?  How could that be?  I was running 4 miles a day and going for hikes on weekends.  The doctor mentioned if you were just a little more overweight he could prescribe a GPL1 drug.  I joked about going on a binge at Feast Buffett.  I didn’t really want to take the drug, but I was confused on what I might have to do differently.

Then I decided to turn to Chatgpt for help. If you’d told me a year ago that an AI would help me lose weight, I’d have laughed. But here I am, ten pounds lighter, running farther than ever, and finally understanding why my weight used to jump up and down like a yo-yo on the scale.

When I started treating ChatGPT like a data-driven coach I was able to break through my plateau.

📊 Logging, Not Guessing

Every day I entered my stats into a shared fitness database:

* Weight

* Exercise (mostly treadmill runs and long weekend miles)

* Every meal, drink, and even the “refuel” IPAs

Instead of vague advice like “eat less,” the AI calculated my true calorie target based on exercise load, recovery needs, and my goal of losing about a pound per week. The numbers weren’t emotional—they were math.

But the real breakthrough wasn’t calories. It was sodium.

🧂 The Sodium Surprise

For years, I assumed weight spikes meant I’d blown my diet. One night of pizza or a couple of beers and—boom—two pounds up on the scale.

ChatGPT caught the pattern. It pulled data from my log and highlighted how my high-sodium days (anything above ~2,300 mg) consistently produced next-day weight jumps. Not fat gain—just water retention.

The insight was almost embarrassingly simple:

> “You didn’t gain two pounds of fat overnight,” it explained. “You just stored extra water with yesterday’s salt.”

That small realization changed everything. I stopped panicking at daily fluctuations and started tracking trends instead of snapshots.

⚙️ Rebuilding My Diet

Sodium was everywhere. It showed up in surprising locations like my breakfast burito where the tortilla contributed to unacceptibly high amount of the recommended amount.

I swapped high-sodium sauces for fresh spices and vinegar.

I learned that spicy V8 and pickles were stealth salt bombs.

I added more fiber (beans, oats, berries) to smooth digestion and stabilize energy.

And I kept plenty of water and electrolytes around to offset the days I did indulge.

Once the water swings settled, my true fat-loss trend appeared—steady, linear, and motivating.

🏃‍♂️ Running Lighter, Eating Smarter

I didn’t crash-diet. I fueled my runs, lifted weights, and enjoyed the occasional burger. ChatGPT helped me see the big picture: a week’s worth of choices mattered more than any single meal.

It also reminded me to recover. On long-run days, it bumped my calorie target automatically. On rest days, it cut them back. The consistency—and the data—did the rest.

💡 What I Learned

1. Water weight isn’t failure. It’s feedback.

2. Fiber and protein keep hunger predictable; salt keeps water unpredictable.

3. Weight loss is mostly pattern recognition. AI happens to be very good at that.

4. When you understand the “why,” motivation takes care of itself.

🧭 The Bigger Picture

Ten pounds later, I feel like I've broken through the stagnation. And while ChatGPT didn’t run the miles or cook the meals, it became the world’s most consistent accountability partner—patient, analytical, and always ready to connect the dots between my habits and my results.

In the end, it wasn’t about dieting. It was about learning my body’s language—one sodium reading, one run, one log entry at a time. I still have 10 lbs to go. I still have more to learn.