Embracing AI to Train for an Ultra Marathon

About a week ago, I embarked on a journey to conquer the OSR thirty, a thirty-mile ultra marathon around Manhattan. With the race just fifty days away and minimal training under my belt, I decided to enlist the help of an AI personal trainer. Using Clay and OpenAI integrations, I set out to program a trainer that would offer effective running advice and automate my training schedule.

Setting Up the AI Trainer

Within Clay, I established several columns to guide the AI: the date, day of the week, and day of training. My training regimen is structured by the week, with interval workouts on Tuesdays, tempo runs on Fridays, and long runs on Sundays. Making sure the AI understood this schedule was crucial for accurate advice.

To prepare for the race, I created a column to track the days until the event. This information was vital for planning my taper period, which begins about ten days before the race to reduce mileage and intensity, ensuring I feel good on race day.

Prompting the AI

The real challenge was programming the AI to provide solid training recommendations. Using OpenAI's API, I crafted a prompt detailing my current fitness level, personal bests, and training preferences. I emphasized a conservative approach to avoid injury and burnout, aiming to peak at seventy miles a week with a longer long run.

The AI's recommendations were surprisingly spot-on. For instance, it suggested a rest day on a Saturday, aligning perfectly with my schedule. It advised a sixteen-mile long run on Sunday at a challenging pace, which was ideal for fifty days out from the race. The AI even accounted for the taper, recommending shorter runs as the race approached.

Enhancing the Experience with Automation

To make training more engaging, I added creative run titles to each day's plan. These fun titles, like "Sunday Stride" and "Mellow Monday Mileage Builder," were integrated into my Google Calendar using Zapier, providing daily reminders of my training schedule.

Connecting Clay to Google Calendar involved using an HTTP API integration, allowing the seamless transfer of data from Clay to the calendar. This setup ensured that every morning, my run details were ready and waiting, making it easy to stay on track.

Conclusion

This AI-driven approach to marathon training not only provided effective guidance but also added a layer of fun and organization to the process. As I prepare for the OSR thirty, I feel confident that this innovative method will help me achieve my goal. If you're interested in trying Clay, the link is in the video description. Happy running!