It is no big secret that Artificial Intelligence (AI) is here to stay. While I have not really used it, I have been somewhat intrigued by its uses (in particular, ChatGPT). I recently attended a workshop that was a crash course on dos and don'ts for developing ChatGPT prompts.
ChatGPT pretty much is a large language model (LLM) that uses AI to understand and generate "human" text based on data input and contextual understanding. Basically, one should be able to ask ChatGPT something, and it should spit out an answer for you. Well, the answer may or may not be what you are looking on based on your prompts.
For sh*ts and giggles and to avoid doing housework, I decided to ask ChatGPT to give me a marathon training program. I figured this would be a good way to test ChatGPT because:
I'm running Richmond Marathon in November, and July 1st was Day 1 for my training program. Actually, July 26th is supposed to be my first day of training, but I figured why not start sooner.
This will be my 7th or 8th marathon, so I would know if ChatGPT could give me a "legit" result (well, legit for my needs)
My prompt pretty much asked ChatGPT to develop a marathon training based on the following parameters:
135 days
3-4 running days that include one 10+ mile long run each week
speed workouts (e.g., tempo, interval, and hill runs)
strength training
a rest day
Overall, I was satisfied with what ChatGPT spat out, I did have a couple of minor issues.
ChatGPT did NOT give me a 135-day (~20-week) training program. Instead, it gave me a 16-week training program. I guess that I can deal with this, because (based of my limited understanding) ChatGPT does not actively conduct research. Instead, compiles its answers based on information that is already out there. Maybe it's better to say that it cannot create something out of nothing, instead of saying it cannot actively conduct research. From my experience training for marathons, most marathon training programs typically are 12 to 16 weeks. Although I asked ChatGPT for a 20-week program, I guess the best it could do was provide a 16-week program, which is more common.
I would have liked a bit more variety for the speed workouts, particularly with the interval workouts. For the interval workouts, ChatGPT pretty much increased the number of intervals with the same distance instead of changing the number of intervals and distances. For instance, ChatGPT wants me to run eight, ten, twelve 400m intervals with 200m rest over the course of the training program. However, I would prefer something where the interval distance would also change (e.g., 400m, 800m, 1600m, etc.). This limitation probably is on me because I did not really specify the type of intervals that I wanted to do during the training program. I also ran into the same issue with the hill workouts because ChatGPT did not provide any information about these workouts. Again, that's on me because I did not give it a lot of details in my prompt.
Everything else from my ChatGPT-generated marathon training program looks pretty legit. Actually, I think they are legit because the other things for my program (e.g., HIIT workouts, easy runs, and strength training) are pretty general/generic. I think that I'll play around with ChatGPT and ask it for more specific types of HIIT, easy run, and strength training workouts. Ugh. I hope that I saved the prompt that I used so I will not have to start over because ChatGPT might give me a different answer for the same prompt. Maybe, that might be something for me to test out.
Actually, I did somewhat play with a specific workout. Since I'll be in Spain later this month, I asked ChatGPT to give a 14-mile long run in Santander. Here's what it gave me.
Unfortunately, ChatGPT could not give me a map/route for this route. According to the hubby, I've already been to these places during several of our trips to his hometown and 'should know' where these places are. LOL. To be honest, Playa de los Peligros is the only place that I can remember off the top of my head. I only remember Playa de los Peligros because it translates to Beach of Danger(s). While writing this blog post, I asked ChatGPT the reasons for this beach's name, and this is what it told me:
The origin of this name is historically linked to the navigational challenges that mariners faced in the past. The bay near the beach has a complex system of currents and submerged rocks, which posed significant hazards to ships and boats trying to navigate the waters...Despite its name, Playa de los Peligros is now known for its calm waters and beautiful scenery.
I will say that ChatGPT is right about its beautiful scenery. This is a picture that I snapped when I ran along the Beach of Danger(s) last year.