What I think would be useful is if you got some feedback from the AI on how it thought you were progressing. Like if it was evaluating things like heartrate and power and how well you were holding on to reps and gave feedback like “you need to increase/decrease intensity on your solo rides if you’re going to hit your target”.
2 Likes
Also, if there were a way to evaluate day by day if can get away with more training in a given zone within the analogy of training plan. For example, will adding volume in endurance zone kn recovery days hurt my progress? Or what specific progression I can expect to gain from a given workout.
2 Likes
On the most recent podcast @NateP mentioned the new ML model was scary accurate at predicting your time to freshness or some similar metric. @Jonathan gave an example where he moved a workout to later in the day and the system made the workout harder because it knew he now had 5 more hours of recovery before training.
Not that this answers your original question, though it does suggest we’ll be getting a lot more ‘predictive’ insight soon. Hopes shouldn’t be too high though, a lot of this is difficult to actually present. I wouldn’t be surprised if the AI is just ‘updated’ and we see workouts behave differently without really understanding the ‘why’
3 Likes
But would the AI know that I get kinda sleepy later in the day and might actually be less fresh 
You should go listen to the segment. I had the EXACT same thought as they were discussing. I heard something to the effect of: “we know how you perform at different points in the day and after other sessions, so we’ll tune for that”
It certainly sounds too good to be true, and i’m sure there are scenarios where it’s just wrong… I think it’ll be missing a lot of input to help make some recommendations and thats where you need to still use a bit of your brain when it comes to training.
If they got to the point of partnering with a food tracking service (myfitnesspal, loseit, etc.) I think RLGL could be next level and help give warnings about under/over fueling.
1 Like
I use AI Endurance and while I like their approach to using AI to calculate progress/plans, I really don’t like the AI generated verbal feedback that it gives. It’s unnecessary wordy and repetitive. And basically regurgitates what the dashboard already showed you.
If you have ever used CoPilot to listen in to and summarize your conference call, you’ll know what I mean.
In TRs example it would be akin to generating 2 paragraphs worth of text explaining to you that you should take an easy day if you see a yellow light. Which you already knew if it happened to you.
I would have thought AI pulled all the Garmin health data like your HRV, heart rate, BP etc and put that in the mix