How Do I Lose Weight?

Diets suck. I chose a different path and it worked.

I designed a system, not a diet. By partnering with AI and using a little metabolic math, I lost the weight with less struggle then I had in the past.

Here is the five-step system I used.

I decided what an ideal weight was (or at least a better one, just set one, change it later if you want)

Step 1: The “Two-Step” Metabolic Calculation

Instead of picking a random calorie number, I went to Google and searched for a “TDEE Calculator.” Your TDEE (Total Daily Energy Expenditure) is the total number of calories your body burns in a day based on your age, weight, and activity.

I did two specific calculations:

  1. Me Now: To see how many calories I needed to maintain my current weight.
  2. Me Later: I plugged in my target weight. This gave me the “maintenance calories” for my future self.

The Strategy: I didn’t drop my calories to rock bottom. I simply started eating the amount of calories my future self would need. This prevented the “diet crash” because I wasn’t starving; I was just practicing the exact eating habits I’ll need for the rest of my life.

Step 2: Setting Human-Sized Milestones

I stopped trying to lose 100 pounds overnight. Science shows that losing 1 to 2 pounds per week is the sweet spot for keeping the weight off long-term.

I used the 3,500 calorie rule as my guide: since one pound of fat is roughly 3,500 calories, I aimed for a daily deficit of about 500 calories. Instead of one giant goal, I celebrated every 5% of my body weight lost. These small blocks kept me motivated without feeling overwhelmed.

Step 3: AI as My Personal Nutrition Coach

Partnering with generative AI was my “force multiplier.” It handled the mental load of nutrition tracking and meal planning so I didn’t have to go back to the nutrition apps I used before. Sitting in front of your meal searching for which version of “chicken breast” looked the most accurate and estimating how many grams I had on my plate sucked. It was a pain in the ass, I am pretty decent as estimating sizes, but not great at it. I used the AI to outsource that part, and to optimize my macros (protein, carbs, and fats) to make sure I was losing fat, not muscle.

Here is how I used AI for nutrition tracking:

  • I built and refined a custom AI Assistant for nutrition tracking. It did a better job estimating calories and macros than I was, and it did it faster.
  • I told it what I ate, it calculated my macros, and gave me feedback
  • I meal planned and strategized when I should prioritize protein, when I should focus on fiber, how do I stay satisfied…

Step 4: Focus on Body Recomposition (Not Just Weight)

I didn’t want to just be a smaller version of my tired self. I wanted to be “sturdier.” This is called body recomposition; losing fat while building muscle.

The secret is protein. I aimed for 1.6 to 2.2 grams of protein per kilogram of body weight (or roughly 0.7g to 1g per pound). Protein is the most satiating nutrient, meaning it kept me full for hours and protected my metabolism from slowing down.

Step 5: Identity-Based Habits

The biggest shift was psychological. As I stated in the beginning of this post “I decided”. I made a decision, sure I wanted to lose weight, I wanted to feel healthier, but wanting something and deciding to do something are very different things. Motivation is nice, but if I waited for it to move me, I wouldn’t have moved. From my experience, motivation comes after the momentum, and the momentum comes from discipline. But the very first step in that entire equation is the decision.

Because I was eating at my future self’s maintenance level from Day 1, I wasn’t waiting for a “post-diet” life to start. I was already living it. I created a system that models how I will eat forever, making the weight loss a permanent side effect of my new lifestyle.

It doesn’t mean I cannot eat this or that, I can, and occasionally do. But I also make different choices that fit my eating habits of me at my ideal weight. By understanding the math (TDEE current, future, ~3500 calories per pound) I could quantify the mission and I knew if I wanted to be that new me at that goal weight how much I would need to consume on a daily basis once I achieved it. So it wasn’t weight loss by dieting anymore. Dieting was temporary and the weight returns when eating without knowledge of what I was consuming returned.

The Bottom Line

Basic math, online calculators, understanding what you are consuming and adjusting as you go will get you there. The TDEE calculators are general guidance, not precision. After you consistently track your nutrition for a while, use that data to better understand your personal TDEE, how many calories does your body use to maintain your current weight. Use the tools available to you to help. The journey isn’t easy, but you can make it easier by using the right tools.

I started with AI assistants and kept evolving it until I built the solution I still use today. I build MackCrows Track Coach, I use it daily, and I attribute my success losing weight and hitting my body recomposition goals to it. I wanted to keep developing it as a tool for everyone to use.

Why Not Just Use the AI Assistant?

There are some limitations to those AI Assistants, but you can, in fact, if you want my original prompts for the early versions I used, email me, I will share them with you. But here is why they didn’t work for me long term.

There is a limited memory, they would forget what I ate in a couple of days. I had no long term tracking of my weight, my calories, my macros. I started tracking it “offline” by creating my own database, but the AI Assistant didn’t have access to it, so it was a new nutrition partner every few days. Starting over with a new partner every few days is not a good experience, it puts the burden back on me to track and that is what I wanted to offload to the AI.

The Track Coach solves these pain points, I don’t have to spend more than a few minutes per day logging my entire days meals, snacks, and bites along the way. It has a full back end database and it follows a strict rule: any piece of data you input has a purpose and direct benefit to you. If you have ever onboarded to a nutrition app then you already know the oddball questions they ask are not because they need to know that information for your benefit, they want to know to feed their own curiosity, or marketing partners, or whomever they sell data to.

The app is far more than a nutrition tracking app, it does so much more like meal planning, recipe creation, coaching, education, but that is another post for another time. Until then, get out there, you got this!