Using AI to Get Fit

Several months ago, I made the decision to get in shape. I needed to lose weight, and I wanted to improve my overall health. I’ll talk about my journey in a future post, but I wanted to share how I used AI in a way that you might not expect.

One thing I have done is to be more active. There has not been a day in the last two months or so that I have not engaged in some form of physical activity. I have been doing some strength training. When I play golf and use a cart, I walk more than I would normally have to, and I’ve started walking more around my neighborhood.

I use Strava to track my walks, and I tend to pore over the stats and see to see how I can improve my performance. I’m lucky that I have several different routes I can take, depending on how I’m feeling, the amount of time I have available, and the weather. The route I take most often is approximately 3 miles (5 km). Most days, my older son will join me on my walks. It’s a great way to spend time together and discuss work, sports, and life in general. On days when I walk alone, I listen to music while I walk.

I noticed something interesting about the days I walked by myself—I tended to walk faster whenever I listened to music. Looking at the stats from these walks, I noticed that while my overall times were faster, my pace was all over the place. I would walk fast for a while, then slow down, then speed up again. It was not consistent…at all.

Then it hit me…it was likely because the music I was listening to affected my pace. I started to pay attention to the tempo of the music I was listening to as I walked. Sure enough, when slower songs were playing, I tended to walk slower, and when faster-tempo songs were playing, I walked faster. I decided to create a Spotify playlist with songs I liked that had a consistent tempo. I wanted to see if this would help me maintain a more consistent pace.

However, this was a daunting task. I have over 800 “Liked” songs on Spotify, and other playlists I have created have hundreds of songs. I tried to do this manually, but it was taking too long. I needed a way to speed up the process.

Here is where AI came in. I found a Spotify MCP server that I could use with Claude. I asked Claude to help me create a playlist of songs that had a more consistent tempo. I asked Claude to analyze recent songs, find uptempo ones, and create a playlist that would last at least 55 minutes. (It’d never take me that long to finish the long route, but I wanted to have some extra time in case I wanted to walk longer.)

Claude churned away and created a playlist that met my requirements. I was amazed at how quickly it did this. I tried this new playlist on a walk earlier this week, and I knocked about a minute off my best time for the long route. However, there were some songs that I wouldn’t consider “uptempo,” and my pace was still quite inconsistent.

I went back to the drawing board yesterday and asked Claude to create a new playlist. There was some back and forth with me asking to remove some slower-tempo songs and replace them with more uptempo ones. I also had to ask (a few times) to make sure the playlist did not contain back-to-back songs by the same artist. Finally, Claude got it right and came up with this playlist. Yes, it’s over 55 minutes, but I’m okay with that; it gives me a buffer in case I ever want to expand my route.

I tried this new playlist on my walk today, and not only was my pace much more consistent, but I also knocked another minute off my best time for the long route. I finished that route in about 48 minutes. I’m not setting any records, but it’s quite an improvement for a guy who got winded walking up a flight of stairs just a few months ago.

I have a short-term goal that I want to reach (which could be very soon) before I share more details about all I’ve been doing. However, I was so excited about this that I wanted to share it now. I am sure I will continue to tweak the playlist as I go, but I am happy with the results so far.

Photo by Mohamed Nohassi on Unsplash