Well yes, and if you are currently inclined to pay for a coach it is because you recognise that a cookie cutter plan of the web is not going to work for you.
But most people don’t pay for coaching. AI should be able to improve on the one plan fits all off the web, by adding a degree of modification based on the athlete’s data and feedback. And I’d wager there are a fair few online coaches that aren’t doing much more than that already.
absolutely…but that’s just one step up from a book…
absolutely again, but athletes are paying for that…
and so athletes wil have to make a call between two uncertainties…
The AI coach is likely to be complimentary with a larger platform though, Garmin, TP or Strava ect. If the gap between complimentary AI coach & paid for coach is small & likely better in a lot of cases then its good bye paid for coach. Online paid for coach at least.
why would that be the case?
and how would the athlete know?
I guess the platform in question would have access to a lot of athlete data to help it learn.
And that’s how I see it.
What will be interesting is if the hype and buzz around AI means it can actually claim to be a premium product and people will pay because it has AI in the name.
Then there is the 4th option, where a coach uses AI to help analyse the athlete data but applies their experience to any output before feeding back to the athlete.
Because I know how crap some online coaching is & if the AI coach is presented as basically free then the choice to go to an additional paid for service is likely to be a more considered option.
You have got to imagine there’s is a lot of potential for AI to to look at many more data points than regular TSS, fitness, freshness or whatever & pick out some nuggets.
AI is a just tool not a mother, you don’t have to let great be the enemy of the good.
and a lot of crap data.
It’s a results business, ultimately the stopwatch doesn’t lie.
I agree with most of this except I think the value is in the data so the AI will likely be presented as free so it’s fed more data.
The internet is full of crap but it doesn’t stop LLM’s being very effective.
i do
how many data points might it look at when Sean fails his driving test?
i don’t think there are many great coaches out there…certainly there are many good ones…
AI will not be anymore than adequate as a coach…
it’s an experiment of 1…
no one will know if coach a will be better than coach b in that instance…
but loses the learning in the analysis…
Which is the case now, doesn’t change just because coach B is a computer.
Currently coaches mostly trade on reputation, the results they have achieved with their athletes (or maybe personally, but that is a whole other category). AI will be no different, if my mate starts winning races, and he says it’s thanks to the SuperCoach3000 then I’m more likely to check it out myself.
Of course but realistically probably more beneficial than not. It learning to dismiss poor data would be an early lesson I reckon.
Having it understand and recognising progressive overload, seeing where gaps follow too rapid a rise would be clever.
Linking training to race performances would tricky IMO. Garming and TP don’t know when I race.
I agree there but that could help an athlete, rather than a coach.
My customer boss used chatgpt to analyse a load of data and the insights were quick and definitely enlightening. This was how we spend our work time for context. Not preparing for an endurance race.
Depends on how you measure “very effective”. I have been given some incredibly stupid results from ChatGPT when asking development questions, basically repeating “crap” it has read, some of which would leave a company open to hacking.
Correct…
Indeed they do…however, currently more athletes look initially for athleticism rather than coaching prowess in their initial selection…
they are inseparable…or at least hardly distinguishable…
certainly that will give AI some traction, but the same outcomes are not guaranteed…
certainly…but having just taken on a former channel swimmer, how would AI evaluate her skillset, knowledge, experience and training history?