Massive Gains in the Gym with This AI-Generated Strength Plan – Part 4
April 8, 2023
Unless you've been hiding under a rock the last few months, you've witnessed AI taking over our news feeds (among other things). While there is a lot of hype around AI, we've been taking a more practical approach by experimenting with real-life use-cases. In this article, we will show you how we used ChatGPT to create a strength training plan that after 3 months resulted in 6kg (13.3lbs) of muscle gains.
Admittedly, ChatGPT did lose it’s “train of thought” at some point, but it was able to provide some really concrete and understandable examples, that were pretty close to our desired result.
For the overall phases, we didn’t want a ‘peaking and tapering’ phase as we planned to transition into more cycling and endurance training, rather than competition. Next, we populated sets and reps per the recommendation, but added more granularity for each week within a phase, always including a reduced-volume week the last week of the block. Rest between reps was never explicitly defined, so we added rest intervals between sets with increasing rest as the weight increased. After training a week or two, we made some adjustments to the rest because it felt too long between sets (shown below).
And here is the final plan. When you’re doing your actual workout, you’ll refer most often to this and your maxes.
And the last, practical and necessary step to get started is creating the maxes sheet where you capture your 1RPM. With this number, all of your workouts will be catered to your abilities. You can always modify these on any given day if you find your struggling to complete a full week, or recovering from an illness/injury.
After using this plan for 3 months, I’m well on my way to gaining 10kg in one year. After 3 months, I successfully gained 6kg of muscle, only 4kg to go! Stay tuned to see if I reach the goal before the end of the year.
Here are our closing thoughts:
ChatGPT isn’t perfect. In order to get to a quality end-result, we need fair amount of knowledge to know what to ask. Knowing what to prompt, how to prompt, and in which order is the name of the game with ChatGPT.
We could make a good plan much faster with ChatGPT. A novice could have put this together themselves and gotten at least 80% of the way there on their own. There could have been some misdirections, but with some common sense and iteration, you could end up with a good plan.
This plan was made using ChatGPT 3.5. ChatGPT 4 likely would provide better results.