Week 17: Using AI to Become Brian Johnson

The Experiment

I somewhat dream of being like Brian Johnson, that guy who injects his kids' blood into him and is reversing his age (allegedly). I thought for one week, let me have AI help me be more like Brian Johnson. The goal was to track everything for a week. Take photos of all my food, have AI analyze the nutritional content, monitor my sleep patterns from my watch, log my exercise, and count my steps. Essentially, I wanted to create that quantified-self dream—a comprehensive overview of my input and output that would help me understand my body and habits better.

The Process

Here's how I approached this experiment:

1. Food Tracking

• Took photos of everything I ate throughout the week

• Uploaded these images to ChatGPT to analyze nutritional content

• Had the AI break down protein, carbohydrate, and fat percentages

• Calculated total calorie intake

• (Only forgot once—a very good mortadella, buffalo mozzarella, and pistachio pizza)

2. Sleep Monitoring

• Extracted sleep data from my health app

• Tracked hours, quality, and patterns

3. Exercise and Movement

• Logged all exercise sessions

• Recorded daily step counts

• Imported this data to ChatGPT for analysis

4. AI Analysis

• Fed all this data to ChatGPT to create a comprehensive health profile

• Asked for multiple perspectives: "How would Peter Attia approach this?" "What would Andrew Huberman say?"

• Requested visualizations and actionable insights

The Outcome

The results were enlightening, if not entirely surprising. The AI analysis revealed:

• My protein intake was lower than optimal for muscle building (despite my efforts to increase it)

• My sleep for that week was below my usual average of 7+ hours

• Some interesting insights about basal metabolic rate versus consumption

• Cool graphs that visualized my inputs and outputs

It was particularly interesting to see the data through different lenses, asking what various health experts might recommend based on the same information.

What surprised me most was how quickly ChatGPT could translate food photos into nutritional breakdowns on day one (though some entries needed correction). Seeing the final graphs and getting AI-generated insights made the tedious tracking feel worthwhile.

Key Takeaway

Tracking everything you do creates a feedback loop that naturally modifies your behavior (if you know you’re tracking everything you’re a bit more aware). When you know you're documenting every bite, you become more conscious of your choices—you've put yourself under the microscope. However, there's a significant friction cost to this level of tracking that makes it unsustainable without better tools.

Marc Lou built a more sustainable less MVP version of this which looks cool and could be useful (but you still have to always take pictures of your lunch)

Pro Tips for Beginners:

1. Start Small:

Just track your food before attempting the comprehensive approach

2. Be Consistent:

The value comes from complete data, so commit to logging everything for your chosen timeframe.

3. Use context windows wisely:

ChatGPT sometimes "forgot" earlier data, so I had to remind it of previous entries—organize your information before uploading.

4. Prepare for Tedium:

Taking photos of a handful of nuts feels decidedly un-foodie and can get annoying quickly.

Want to Try It Yourself?

• ChatGPT is surprisingly good at identifying foods and estimating nutritional content from photos

• Most smartphones have built-in health apps that track steps and can sync with sleep data

• For a more streamlined experience, consider specialized apps that integrate with AI (though my experiment was about seeing what could be done with general-purpose AI)

• Set a reasonable timeframe—a week was enough to get insights while still being manageable

This experiment was an interesting benchmark and skirmish into the world of health tracking. While I'm not ready to go full Brian Johnson (keeping my blood to myself, thank you very much), I have made one lasting change: I'm still working on that protein intake.