The Time-Crunched Cyclist Podcast by CTS

Is A.I. Coaching Ready for Prime Time? Hear What TrainingPeaks Says!

• CTS • Season 5 • Episode 271

Share episode

OVERVIEW
Cody Stephenson, the Education Program Manager at TrainingPeaks, is back on the podcast to talk about the current state of artificial intelligence in cycling coaching and how TrainingPeaks' developers are using models to improve the accuracy and usability of athletes' training data.

TOPICS COVERED

  • A.I.'s current and future role in cycling coaching
  • How auto-updates of FTP work and when to pay attention to them
  • About TrainingPeaks' Fueling Insights Tool
  • TrainingPeaks Virtual: Ultimate Indoor Cycling App 
  • What's coming next from TrainingPeaks

ASK A QUESTION FOR A FUTURE PODCAST

Guest Bio – Cody Stephenson

Cody grew up racing mountain bikes in Durango, Colorado where he developed a passion for endurance sports, science, math and technology. He switched to the road and track while racing for Fort Lewis College, where he also managed to get a couple of science degrees. Now he gets to write and talk about his favorite topics every day as Education Program Manager at TrainingPeaks. When he’s not helping coaches learn to leverage technology to reach their goals he’s trying to become as good of a mountain bike racer as he was when he was 13 years old.

Resources:

- Cody Stephenson LinkedIn
- Articles: https://www.trainingpeaks.com/coach-blog/how-to-coach-athletes-who-arent-racing/
- CP W’ vs FTP alone
- Analyzing FTP by Joe Friel
- Power Training with WKO:
- Why Train Submaximally? WKO Case Study - Targeting Specific PDC Improvements
- Learning More about LLM’s and AI

HOST
Adam Pulford has been a CTS Coach for nearly two decades and holds a B.S. in Exercise Physiology. He's participated in and coached hundreds of athletes for endurance events all around the world.

Listen to the episode on Apple Podcasts, Spotify, Stitcher, Google Podcasts, or on your favorite podcast platform

GET FREE TRAINING CONTENT

Join our weekly newsletter

CONNECT WITH CTS

Website: trainright.com
Instagram: @cts_trainright
Twitter: @trainright
Facebook: @CTSAthlete

Speaker 1:

From the team at CTS. This is the Time Crunch Cyclist podcast, our show dedicated to answering your training questions and providing actionable advice to help you improve your performance even if you're strapped for time. I'm your host, coach Adam Pulford, and I'm one of the over 50 professional coaches who make up the team at CTS. In each episode, I draw on our team's collective knowledge, other coaches and experts in the field to provide you with the practical ways to get the most out of your training and ultimately become the best cyclist that you can be. Now on to our show. Now on to our show. Welcome back, time Crunch fans. I'm your host, coach Adam Pulford. Picking up where we left off from last week, I am back with Cody Stevenson of Training Peaks to discuss how best to utilize your data in platforms like Training Peaks and maybe a sneak peek of future enhancements to look at into the future. So, cody, welcome back.

Speaker 2:

Glad to be back.

Speaker 1:

All right, so maybe just like a quick recap from last week is that we compared FTP and FRC to critical power and W prime, explaining that these are similar systems insofar as they both determine high aerobic threshold point and an anaerobic measurement which hinges off of that threshold, so to speak. The former is more of a field and practical based, while the latter is the more like research and lab based, and either system is completely fine to use. You just need to know how you're actually going to use it and use the data within it to make the best decisions for your training decisions. So now, if you're going to rely upon that like some sort of like machine learning or AI populated things, to make those training decisions for you, you need good data, good, clean data, as well as the knowledge of how the system works, because, in my opinion, bad data in equals bad data out, and if your model is bad, so will your training. Cody, would you agree with that?

Speaker 2:

Yeah, yeah, definitely All right.

Speaker 1:

Well, let's, let's pick up right where we left off, which was talking about AI and Cody. I asked you is AI going to take over this endurance coaching space, yes or no? And you said it might take some jobs. Let's talk about that.

Speaker 2:

Yeah, I think. Well, I guess I'll start with the jobs and I'll steal a saying not mine, everybody's probably heard it from every industry is that you're not going to lose your job to ai, but you're. You might lose it to somebody who's using ai better than you. So it's, it's going to be a useful tool. It's going to take care of tedious tasks, repetitive tasks, tasks that um might be prone to human error, and so you can't ignore it. But taking over the whole industry, I, I, I don't think so, mostly because one of the most important lessons I learned when I came from being an athlete to working at training peaks you know, I was a high, I wanted to be a high performing athlete. I cared about results. I cared about and I was okay doing it alone, like I liked training alone. I didn't work with a coach. All that was fine. It was when I started training peaks. It took me a long time to learn that I was in the minority with that.

Speaker 2:

Most people just want to be a part of something, and so one thing they get out of endurance sports, even if they're pretty high performing is relationships, and so that's a really important part of what coaches do, like I mean every coach out there listening knows that you aren't just the guy who writes the training plans or the girl who writes the training plans. You are the therapist, the Sherpa, the tire pressure advisor, everything else that an AI can't. And a lot of times you just become their friend and that's just like a piece that AI can't do. But some of the parts that were pretty important for coaches 10, 15 years ago, even now, are going to be done more and more by AI.

Speaker 1:

Yeah, yeah, no, I completely agree with all that, and I would also say that when I'm talking with my athletes, when I'm, when I'm learning about people and we learn this, even like through COVID, right it's just like the person to person interaction is still so powerful that, yeah, I don't, I don't think it's going to go, yeah, I don't think AI is going to take over and take every you know coach's job and that kind of stuff.

Speaker 1:

But, yeah, I think that the people, the athletes, the listeners here that want the benefit of kind of that data-driven AI sort of you know quicker sort of strategy, or kind of you know kind of more accuracy, more consistency, whatever AI is going to help us with, yeah, they'll want to coach that deals in that category. You know that works in that space, and so, um, there'll probably be a delineation of coaches, um, in that way. But I still know some coaches too that are using Microsoft Excel and they're great coaches, right, and they talk with their athletes and they interact with their athletes Great. So it's like you don't even need the data-driven approach necessarily to do to be a good coach yeah, that that's the thing is, most people mostly amateurs, but even, to some degree, pros.

Speaker 2:

They aren't limited in their performance improvements by a lack of data or lack of exploiting data. Amateurs are mostly limited by time, like, yeah, you would just get a lot better if you could train 14 hours a week instead of seven hours a week, but that's just not practical. They're limited by body weight, like it's just like it can be hard to maintain a good body weight with a modern lifestyle. And they're limited by bad sleep or just a million other things. That isn't necessarily like all the science in the world Well, I shouldn't say all the rest of the science in the world beyond like the science that tells you that sleep is important and that not drinking beer will help your performances. That's where the gains are going to be made, not like having a you know billion parameter neural net telling you how to, you know, win a cat to race better. It's probably not going to move the needle.

Speaker 1:

Yeah, no, that's it entirely. And I think too, the human approach is going to be able to cut through the BS, uh, and say, hey, cody, you know what? Let's not have three beers tonight. You know that kind of thing.

Speaker 1:

Good thing, you're not my coach, but that's right, that's why, that's why I'm not, but yeah, so I think like in that way, I mean I, you know, I, I I don't want the listeners to think that I am just like Mr Data and don't care about anything else I think it's super cool, I really think it's going to enhance this industry and there's there's some great stuff coming out, uh, in training peaks and other platforms too, like, uh, vecta you and I talked about that as well as oh, intervalsicu you and I had talked about that kind of in our, in our pre talk to a lot of this.

Speaker 1:

So you know, there's there's definitely more than just training peaks out there, and I think that that is healthy too, because competition drives innovation and what we're seeing out of, you know a lot of these platforms is being is better because there's more game out there in the field. So, with that said, and kind of with that backdrop of understanding how these systems work and why they work a lot with the automatic generation or auto updates and stuff, one of the things in training peaks is something that you've had for years, which is an auto ftp update. So let's have that as kind of a launching pad of sorts. Why would an auto update of a metric be useful and keep in mind, like the ftp updating that's going to change all the training zones and it's also going to change, you know, ctl, build, ramp rates, um and form, fatigue and all this kind of stuff. So it's an important one, which is why I focus on it. But why would it be helpful to have an auto update and why wouldn't it be helpful?

Speaker 2:

yeah, and that's a perfect example of kind of a pre and post like. If you create a historical point in time for like before ai and after ai. Like that's a good example, because the way it works in training peaks now is very pre-AI. It's been around, like you said, that was probably been around in training peaks.

Speaker 2:

Since training peaks have made it's very algorithmic, it's really simple, it's pretty rudimentary and it'll just look for power data or running data or heart rate data and say, see, like we talked about 95% times your 20 minute peak power.

Speaker 2:

It'll just look for that every workout and if you set a new higher 95% of 20 minute peak power, it'll update your threshold. Or you can set it to notify you that, like this happened and maybe you should update your threshold, which I think is a better way that, like this happened and maybe you should update your threshold, which I think is a better way, like you have to have a lot of faith in a system to issue changes on your behalf without alerting you. So I would say like, yeah, get the alert and double checking and everything, but that's the only thing. It'll look for 60 minute power too. So it'll look for two things and update your threshold based on that. But good coaches know like there's a lot of other things that might indicate threshold changes. So it's really useful for like a coach or an athlete that has like a system in place where they trust their power meter and they know it's kind of going to work correctly for them.

Speaker 1:

Perfect, because I was going to say I mean, there's a lot of things that kind of go into that from the coaching standpoint, which you wouldn't want that to automatic update.

Speaker 1:

I was just having this discussion with another one of my athletes because because he did have like a peak all time, uh, 20 minute power, and I said, well, let's not put it there just yet, let's be conservative with it.

Speaker 1:

Why? Well, because I'm going to guide you through some workouts where we're going to try to repeat a similar power or you know, just under, to see how repeatable you are there. Because if we just go with that, you know the highest power that you've done in your whole life and then create all the zones based on that, you're going to get blown out intensity wise for or you could at least right over intensify the athlete. So you're better off to be a little bit more conservative, because my general approach approach is I would rather under train you by 10% than over train you by 1%, because I will get a better result in the long run. And so now you layer that with you know bad data or somebody that doesn't understand the system that is getting auto updates on something as important as FTP, and if you throw in FRC and all these kind of things, it's like you can start to see pretty clear how spider webby and like confusing and terrible it could be with a training program that is based on bad data.

Speaker 2:

Yeah, so that kind of leads into like what the future of that system would be, is like we didn't even talk about it last time. But ftp we talked about as being like a really simple, practical thing. But ftp and wko is a little more sophisticated because it's modeled and estimated based on a whole range of data. So it's looking at your entire power curve, but especially anywhere from like 12 minutes to 60 minutes, and updating FTP based on that, instead of just a 20 minute power. That's not AI, but it's more you know, kind of like you're talking about. It's more inclusive. It's looking at like oh, did you get really anaerobically strong and that falsely contributed to this 20 minute power reading? It can account for that a lot better.

Speaker 2:

Or even just practical things like oh, you have an athlete who has a perfect 18-minute long stretch of road to do a time trial. Well, that would work better. It can collect race data In the future. If we got something like that into Training Peaks as an automatically updating FTP, a more modeled FTP based on all of the data, that's when it gets really, really useful because you have less need for super precise testing, less need for specific dur races and then having that whole like kind of high dimensional data for everybody sets the groundwork for a more like ai based tool for uh indicating these changes yeah, well, not not to put you on the spot, but I mean, is that, is that a future thing that's coming out in the training peaks uh online platform anytime soon?

Speaker 2:

Anytime soon. Uh, I can't say um, probably, cause I don't know. But um, what I can say is that, like training peaks had other, we've been working on a lot of other stuff for the past few years. You've seen strength like good strength training is more important than ai, and so we knew we had to build a strength builder. First, convenience for coaches efficiency for coaches is more important than ai metric estimation. So we had to, like it would have been very foolish of us to put wko stuff and modeled metrics into training peaks before we let you automatically plan a recurring workout Like that would have been silly.

Speaker 2:

So now that we've got a lot of those convenience things built in, we have a lot of engineers who are really excited and really capable looking at playing around with WKO, and then also a lot of really smart product people looking at like what are the most useful parts of WKO? And some of these like other like kind of cutting edge modeling systems. What parts are kind of like maybe falling flat? For the majority of athletes they only work in specific situations that don't really work at all. It's kind of a testing ground. So what pieces can we pull out?

Speaker 1:

And so, um, we are the most active in looking at bringing those things into training peaks that we've ever been in looking at bringing those things into training peaks that we've ever been, yeah, yeah, and you know, shout out to everybody at training peaks um, because they think in that way, right, rather than just being um, super cool with all the data or overly engineered or something like this, which I mean.

Speaker 1:

Some of these other platforms, like, again, intervalsicu I've really liked it, but there's a lot it's, there's a lot of engineering, ish type stuff that's going on there, and so I don't think an amateur person would be able to just pick that up and, and you know, fly with it as quickly as somebody like Cody or myself or somebody who you know, who knows data well, you know that kind of thing Whereas whereas training piece it's much more like iphone, right, you can pick it up and start using it pretty quickly if you're tech savvy, and so I think that it's super smart on their end and, um, they, they put they don't put the coaches. Well, you can tell me if I'm wrong, but they put the coaches and the athletes kind of like to uh on the same meaning, like how would a coach need to work their business to run quickly with this and how can an athlete take this software and run quickly with it, as opposed to getting bogged down into the data. It seems like that's what you guys do.

Speaker 2:

Correct me if I'm wrong it seems like that's what you guys do. Correct me if I'm wrong. Yeah, that's always been like kind of the driving principle is like let's enable a coach and athlete to do their thing kind of on their own as easily as possible, and it's a good thing. Other people are making those decisions, because I would have fallen on the other side of like. I see intervals, too, or I see other apps and I'm like that's really cool. We should work on that now Exactly. Thankfully, there's more practical, practical people who have coached 40 athletes at once and understand how important like real world stuff is and it pushes us to get those things done fast, so then we can try to move on to like not move on to, but then like spend some time on the cool stuff.

Speaker 1:

Yeah, no, speaking of cool stuff, one of the things that I've been involved with, uh, with y'all over at training peaks and I know that there's other I mean, there's a ton of other coaches that are as well but the fueling insights tool has been something that I got really excited about because, um, uh, working with uh various people to uh build a tool that estimates kilojoule expenditure or calorie expenditure in order for a workout, a race or whatever you're doing in order to get a better fueling plan for the athlete, is essentially what this tool does. Can you tell us a little bit more about that? And I don't even know if it's like, deployed to everybody just yet or how much can we talk about that?

Speaker 2:

It should be. I believe it's in everybody's account. No, it might not be enabled by default, but you should be able to go into your account settings if you're a premium or a coach and look for a data field on the workout card that says fat calories and carbohydrate calories instead of just total calories for a workout, and that's really. It's just two little data fields that sit there. But, like you said, kind of the big thing we all know over the past couple of years has been like the amount of fuel that people take, Like it. It boggles my mind. Like you and I are kind of from similar eras of racing. It was it used to be one gel an hour and that was a kind of like 25 gram gel, a 25 grand Joe.

Speaker 2:

Yeah, like, oh, a hundred calories an hour of gel, you have got glycogen, you'll be fine. And now I'm like reading about how much people are eating now, and it's just a logistical battle, for how do you carry or hand somebody that much food and then how does your gut handle it? So, yeah, when you're talking about 100, 150, 200 grams an hour of carbohydrate, and then on the flip side, on the ultra side, uh, fat at fat adaptation has kind of like fallen out of popular discourse, but I still think it's really important and something to keep in mind is like oh, how much fat are you burning? How much fat could you be burning? How much carbohydrate are you burning? So, like a tool to like show you a better estimate of that is what this tool has been and we wanted to. It took us a long time. This was the kind of thing where if you just came up with like an estimation algorithm, say like what some companies will use for VO2max, you just come up with an equation that's pretty close, you know, see if it's, check a regression against it with real data and then plug that in and it'll give you a number which you could have done with like, oh, everybody, once they get to 60% of VO2 max or 50% of threshold, that's where they're going to be start burning more carbohydrate.

Speaker 2:

We could have done that, but instead Inigo San Milan has been doing a lot of research, collecting a lot of data.

Speaker 2:

He kind of had this inkling that it's not consistent enough among everybody, among everybody, and so his kind of big idea was that and this is part of the configuration and training peaks for it is that it's grouped based on, like kind of your level in the sport and your sex as well.

Speaker 2:

So those are both settings that go into this is, if you're a beginner versus competitive, versus world class, the, the ratio of carbohydrate and fat that you burn at a given intensity is actually pretty different among those groups. So so we have to account for that. So there's a lot more data to put in, a lot more like cleaning and picking that apart and figuring out a good, clean way to implement that with real world data. So that's what it's doing, kind of behind the scenes, and he's written some blog articles for us Like, if you want to get in the nitty gritty, he posts about it on. I think he posts about it on his own social media. He's done a podcast with us about it, so there's a lot more information. He could speak to better about it, but that's what it's doing is like starting to help us understand nutritional, a good estimation of nutritional demands for events.

Speaker 1:

Yeah, yeah, that's absolutely it, and I want to get them on this podcast to talk a little bit more about that once everything is, you know up and running and going. But to kind of recap this for our listeners, it's you know the kind of soup to nuts of it is. It's like we know you're expending a lot of uh calories while training. When you're doing, um, endurance, uh sport with a lot of intensity, it's more carbohydrate. When it's ultra stuff going all day, it's more fat. Right.

Speaker 1:

So we can build the intensity, we can build a model to show how much calories from carbohydrate you're burning, how much from fat you're burning, and then develop a fueling strategy on that carbohydrate in particular, and then we just kind of lump in like some sustenance. That happens with, you know, the fat and some of the protein that you're coming, but it's going to be driven primarily through carbohydrate and it's. It's not like it hasn't changed what I'm doing as a coach, but it's helped me to show it better and I think that that's what's so important, which, which is crazy to me that you guys have probably spent like years developing this and working with one of the greatest like coaches and physiologists out there right now to do this thing, where I'm just like, oh yeah, it doesn't do anything different, like it doesn't change what I do, but it validates what I do, which I think is powerful yeah, yeah, that's I mean.

Speaker 2:

Yeah, putting real numbers on helps. And one thing we've seen where it is changing people a lot is people coaching very new athletes that like lowest group, the most like entry level or like least trained group. It was surprising to all of us, to the point where we kind of doubted it, like we, we kind of double and triple check the data that they hardly burn fat. And that was kind of surprising. Is that even at really low intensities, yeah, so that is starting to change.

Speaker 2:

Like you think like, oh, if you go as a new athlete and weight loss is part of your goal, like, oh, it actually could be harder than you think, because as soon as you're at 30% of threshold, which is really like not a real training level, you're already burning almost entirely carbohydrates. So it's like oh, that can help set expectations or change fueling strategies for that type of athlete. Like oh, you're probably not going to, you're going to have to rethink parts of this. For that type of athlete. Like oh, you're probably not gonna, you're going to have to rethink parts of this for that. And that was that was kind of opening to some of us. But yeah, then at the upper end it was like just putting solid numbers on it and validating it to athletes.

Speaker 1:

Yeah, no, totally agree with that, and so follow up question to that for you, because I don't know. This is, is, is is the feeling insights tool available to self-coached athletes on premium subscription to training things it should be.

Speaker 2:

I believe it is.

Speaker 1:

I'm like 90, sure it is okay okay, we will confirm that and put a link into because there are some articles out there, we'll put a link into it. Um, I just want to be sure that I'm accurate in stating that because, again, this is a new tool and for those who are curious and they're self-coached athletes, I would definitely use this because, like Cody said, it's like if you're new to the sport, this can. This is another tool, another layer to help guide you in the stuff that's super hot right now, which is fueling, fueling your body to meet the demands of, uh, whatever you're doing out there, you know ultras, crit, racing, cross, whatever it is. So, um, cody, we're kind of going long here, but I I definitely want to mention TP virtual. This is another tool that you guys have been rolling out.

Speaker 1:

I was on the original stuff. Uh, I will say it was like cool but buggy and lacking a couple things and like recently, as I've been like utilizing it more, it keeps on blowing my mind because of the enhancements to it, because of how it works, but I think goes back to how you like, how everybody at training peaks kind of thinks with, like the coach and the athlete here they can pick it up and run with it and also customize and let the coach go do their thing with it by building community. Can you talk a little bit about TP Virtual, what it is and why people listening should maybe go try it?

Speaker 2:

Yeah, tp Virtual is like our virtual training and racing platform with kind of, like you said, two or three kind of different pillars that I think make it kind of unique and exciting One. You talked about like community and branding. Other virtual platforms do pretty good with community too. We have lots of like custom branding, so like if you as a coach want to set up your own team kit and club kit, you send us like a. We have a format file, you send us your kit, we set it up. We're moving into more and more like race branding, so like if you set up a race series, the finish banner and the things on the side of the road can be your club or your events company or your coaching company or just the name of the series. So you can like do a lot of customization, get athletes, like I said, a lot of athletes just want to be a part of something so they can feel like, oh, I'm in this group and then like for the coach-athlete relationship, like a spectating and coaching mode, and we're still working on more stuff with that so you can just like watch your athlete race without having to be like try to set up and be riding at the same time, if you don't want to be Easier ways to like watch racing for like, if you have a whole squad or a group of people riding.

Speaker 2:

And then to me, me the coolest part but is it's got a high focus on realism, especially for like, how does road racing work in the real world?

Speaker 2:

Can we make it as realistic as possible in this world?

Speaker 2:

So the physics modeling is real, pretty realistic, as realistic as any, but then also realistic cornering and drafting physics, especially the cornering part.

Speaker 2:

So if you try to pedal through a corner, we can't make you crash, but instead we can. There's a penalty, there's a like a power and speed penalty, so it'll slow you down compared to somebody who coasts, knock you out of the race line and then your trainer will kind of like, create resistance that you have to pedal out of it, create resistance that you have to pedal out of it. So it's like, oh, if you're a coach, coaching tactics, coaching like cornering dynamics, can be hard to do in the real world. Especially, say, you have athletes in the Midwest or someplace cold, you could set up a weekly 45 minute race series with all of your athletes spread around the world and like, watch them do it and it's. You know it's not perfect compared to the real world, but it's 20 hours of practicing, like having people work on their team tactics together, because those aspects of the physics are more realistic yeah, well, I I do want to say there's nothing wrong with staring at a wall and pedaling your bike when you need some me time.

Speaker 2:

Yeah, oh so training peaks virtual also has a zen mode and and I can't remember the name of the other mode, I think it's incognito mode or something zen mode turns off the display, so all you see is like your workout data, and the other one removes all of the other people so you can ride alone.

Speaker 1:

So I'm going to use that anyway. So we don't need to get into the psychology of coach AP right now, but you can see everybody listening, hopefully you can. You can hear um, that there it's evolving, right, like there's, there's cool tech, there's cool stuff, cool tools that are being pumped out and and and again. Like this is one of them, right, so go try it, see if you like it, um, and you know, if it's not your cup of tea, cool, uh, move on. But I think in that way, like people who are creating and innovating, um, you know, jump into that community, jump on that bandwagon because, um, they, they are going to drive um this this market.

Speaker 2:

they're going to drive new things, so new things. Probably the last two questions here for you, Cody, Is there any little?

Speaker 1:

teasers of new things more that we can divulge for our listeners that may be coming out sometime soon from Training.

Speaker 2:

Peaks. I wish I could contribute Without getting you fired, of course, yeah, I wish I could contribute more, but almost most of the really cool things we're working on right now Uh well, there's a cool thing coming out in training peaks virtual soon, so just pay attention. I can't say anything about that, but it's, it's pretty, pretty sweet. Um, most of them are actually just in flight enough that we've talked about them. Fueling insights, you know, not done. There's cool stuff gonna.

Speaker 2:

We don't actually know what those cool things are going to be. We need to start hearing from people using it. But, like, we see this as a beginning. Um, new sport types, again going back to like, kind of fundamentals. If you haven't noticed, we're starting to add sub sport types, so like, oh, you can add yoga or you can specify gravel biking or indoor riding. Um, again, basic, but really, really, I think, important because it's been requested for so long. Um, we talked about, like, we are doing the most research into getting WKO ideas into training peaks that we've ever done. Um, so that's, that's exciting. And, yeah, I guess, tell us what kind of things, what else you want to see cool things that should be coming out.

Speaker 1:

Yeah, I mean and I will put that to the listeners In this way, you can write in to us. I work with Cody, dirk, tony, many others at Training Peaks on various whenever they reach out and ask stuff, but on various projects that come up. So if you're listening to this and you say, man, it would be super cool to have XYZ, feel free to write in to us about it. You can go to trainrightcom backslash podcast and click on ask a training question and you just submit it in to me there. And when I'm talking with Cody or whoever else, um, I will mention that in um, some of the cool stuff that we want to see.

Speaker 1:

I was just actually had a meeting, uh, like three weeks ago, with um two of the engineer, like platform engineers, and it was so cool because, like, I was definitely talking to people who were, you know, supremely smarter than I was, but they were like asking the questions of like. I was definitely talking to people who were, you know, supremely smarter than I was, but they were like asking the questions of like okay, so how do you use the PD model to do this? And I'm like oh, my God this is a great question.

Speaker 1:

Let me tell you all about it, and I don't know what they're going to do with that information, but they said it was really valuable, right? So in that way, listeners um, kind of giving you a voice with somebody who is helping to drive the ship at Training Peaks write in and tell us what you want to hear more about. So and I think that may be a good way to end this, cody is there anything else that you want to tell our listeners? People who are likely using Training Peaks or they might be using something else, are likely using training peaks or they might be using something else. Um, what do you want to say to them as it pertains to data hygiene? Cool techie stuff. I want the ai right now and I'm so confused.

Speaker 2:

I just want to learn more and go fast on my bike uh, I think you can cover all most of the training and coaching stuff. I guess I should say one more thing on AI. We kind of alluded to it. I think the best way to make progress with it if you're already integrating a lot or you are curious about it is to do some basic learning about how they work, Especially like right now.

Speaker 2:

Ai is such a big thing, but what most people are talking about right now is LLMs and so like oh, can you, even if you're not a super technical person, look up the Matchbox neural network or the Matchbox machine learning model. It was like the first analog proof of concept for how a neural network can learn and it was done totally like literally analog. There were no electronics involved. It was about in the 1960s. Go find a podcast or a 10 minute YouTube video about how that works. Read a couple articles about how these things work and that can help put you in such a better frame of reference for what they do, how they work and then how you can realistically use them. It'll help you understand what tools it will help you, what things it can help you do really well and avoid the things that like oh, it's, that's how it works. It's never going to be able to do X now that I understand that, so it's worth learning just a little bit of the basics to like. You're hiring an assistant, so learn how they work.

Speaker 1:

Wow, that is a great. That is a great way to end this podcast. I think that's a really good advice. I'm going to go do that myself because I've been telling my, my athletes as well as like on this podcast like, learn how the system works before you're critical, before you you're you buy it before you jump into it, before you just like learn more of how it works. Right, then go, do and then keep on learning, because I think that is probably the healthiest approach with any system that you're pouring your life into.

Speaker 1:

Because, keep in mind I mean you spend 7, 10, 12 hours doing this training thing. You want to make sure that it's actually going to give you the result that you want. So learn how the system works Exactly. Yeah, cool, cody, I appreciate your time so much. We got two episodes out of you and a ton of knowledge. So thank you for your time and please tell Dirk and everybody over at Training Peaks that thank you for letting us steal you for a little bit and, you know, maybe this turns into more podcasts down the road.

Speaker 2:

Yeah, thanks for having me. It was a great time. Hope to be back.

Speaker 1:

All right, Thanks, Cody. Thanks for joining us on the Time Crunch Cyclist podcast. We hope you enjoyed the show. If you want even more actionable training advice, head over to trainwrightcom backslash newsletter and subscribe to our free weekly publication. Each week you'll get in-depth training content that goes beyond what we covered here on the podcast. That'll help you take your training to the next level. That's all for now. Until next time, train hard, train smart, train right.

People on this episode

Head Shot

Adam Pulford

Host