Hey — I’ve been building fitness apps with small teams for many years.
We just launched a new version of GAIN after noticing more people using ChatGPT to generate workouts. That works well in some cases, but we kept running into a few fundamental problems with that approach:
1.) Data quality & safety. A lot of fitness content online is “bro science,” and LLMs inevitably absorb that. In training, bad info isn’t just wrong—it can be unsafe.
2.) Fitness-specific UX. A gym workout is hands-free and time-sensitive. Things like audio cues, rest timing, load tracking, pacing, and exercise sequencing depend on context and capability ChatGPT doesn’t really have (e.g. squat vs curl timing, user goal, fitness level, exercise preferences, injuries, fatigue, etc.).
3.) Missing location/equipment context. Over time, we want to know what equipment is actually in a given gym or apartment setup, incorporate geo-location, and generate around that. Vs. asking users to re-prompt every time.
Tying together a bunch of expert fitness knowledge, we built a rules-based + data–driven workout generator for personalized, adaptive training. It pushes one clear, daily workout and adjusts when you miss days, travel, or change constraints—without requiring you to “start over” each time.
This is early v1. We haven’t turned on many feedback loops yet (tracking, soreness, preferences, etc.), but those are coming. We’ve already mapped out generation for cardio, yoga, mobility, and recovery, including how to balance them over time and according to various fitness goals, user prefs and usage.
I hear people trying to get CGPT to do weight tracking and progressions, but there's so much bogus data to sort through and not much published, so we think we'll do that much better, too, collecting real-world data with a purpose-built tracking UX.
Would love feedback on the approach and the app itself. App is live on iOS and free for now.
Agree that you must find what works best for you in terms of preferences, enjoyment, and practicality.
It's also important to keep in mind timing and the training cycle (break down, recover, supercompensation) as the author points out. If you really want to "optimize" your training, or at least take steps in that direction, doing the same thing over long periods of time will result in a longtime plateau. I think the way most people manage their fitness - do the same lifts at the gym or the same 45 minute routine on the treadmill, for years in a row - results in many people treading in plateau zones and generally underwhelmed by their fitness despite putting in the time.
Stretching is best done after a strength workout to dissipate blood pooling and lactic acid build up in the muscles. Doing so will reduce muscle soreness and facilitate the muscle recovery process.
Of course, stretching in and of itself can (and should) be done as an excellent standalone workout!
My thoughts on "diet" vs. "exercise": they're two pieces of the same puzzle, and I prefer to manage the "exercise" side of the equation more proactively. Reason is because eating in (esp. in America) today is HARD. There's a lot of crap out there, portion sizes are whack, fat, sodium and empty carbs are everywhere. And you must eat multiple times every day, so that means you make many DECISIONS about eating every single day (hundreds?). Fighting against the grain (pun intended) of what food is readily available and easily accessible is an uphill battle, filled with guilty, uncertain, and constrained decision making.
Don't get me wrong, I still try to eat healthy (and I'm lucky to live in San Francisco where there are more healthful options). But, the nice thing about exercise is that you only have to make ONE DECISION every day. And that is to do it. Once it's done, it's done. And it's not too hard to find a fitness activity you really enjoy doing, plus the endorphin rush you get from a good workout is amazing. When you build exercise into your lifestyle as a healthy habit, it's not even a decision anymore, it's automatic.
I'd much rather live a healthy lifestyle with a daily (or almost daily) fitness routine than to painstakingly count calories or awkwardly ask the waiter to 'hold the bun, please, and remove the excess butters, sugars, fats, and salts..' every time I go out to eat. That just makes it awkward for you and everybody else.
We just launched a new version of GAIN after noticing more people using ChatGPT to generate workouts. That works well in some cases, but we kept running into a few fundamental problems with that approach:
1.) Data quality & safety. A lot of fitness content online is “bro science,” and LLMs inevitably absorb that. In training, bad info isn’t just wrong—it can be unsafe.
2.) Fitness-specific UX. A gym workout is hands-free and time-sensitive. Things like audio cues, rest timing, load tracking, pacing, and exercise sequencing depend on context and capability ChatGPT doesn’t really have (e.g. squat vs curl timing, user goal, fitness level, exercise preferences, injuries, fatigue, etc.).
3.) Missing location/equipment context. Over time, we want to know what equipment is actually in a given gym or apartment setup, incorporate geo-location, and generate around that. Vs. asking users to re-prompt every time.
Tying together a bunch of expert fitness knowledge, we built a rules-based + data–driven workout generator for personalized, adaptive training. It pushes one clear, daily workout and adjusts when you miss days, travel, or change constraints—without requiring you to “start over” each time.
This is early v1. We haven’t turned on many feedback loops yet (tracking, soreness, preferences, etc.), but those are coming. We’ve already mapped out generation for cardio, yoga, mobility, and recovery, including how to balance them over time and according to various fitness goals, user prefs and usage.
I hear people trying to get CGPT to do weight tracking and progressions, but there's so much bogus data to sort through and not much published, so we think we'll do that much better, too, collecting real-world data with a purpose-built tracking UX.
Would love feedback on the approach and the app itself. App is live on iOS and free for now.