ai is coming for your running shoes (and i kind of love it)

i want to be upfront about something: i am not a tech guy. i’m a guy who runs three times a week, checks his splits too obsessively, and once spent 45 minutes googling whether sore calves are normal. so when i tell you i built an app that connects to my strava data, i want you to understand the full weight of that sentence.

i used lovable — an AI app builder — to pull together a simple tool that looks at my recent workouts and gives me actual feedback on what i’m doing. not generic “great job champ” feedback. real stuff. am i running too fast on easy days? are my weekly kilometres all over the place? is there a pattern to when i’m dragging? it took me less than an afternoon to build something that would have cost a developer real money a few years ago.

and that’s kind of the point.

what’s actually changing

AI has been creeping into running for a while now. strava has pacing insights. garmin has coaching recommendations. your watch probably knows more about your sleep than you do. but most of these tools are still fairly surface level — they tell you what happened, not what to do about it.

what’s shifting now is that the connections between your data and actual intelligence are getting easier to make. strava now has an MCP (model context protocol) integration, which means AI assistants can potentially talk directly to your running data. ask claude or chatgpt a real question — “why do my tuesday runs feel so much harder than my thursdays?” — and instead of getting a generic answer about rest days, it could actually look at your data and answer it specifically for you.

that’s not science fiction. that’s kind of where this is heading right now.

what this means for average runners (like us)

here’s what i think is underrated about all this: it’s not really about elite performance. elite runners already have coaches, physiologists, and probably a person whose whole job is their nutrition. AI coaching tools don’t move the needle much for people who already have a support system.

but for a returning runner in their 30s who doesn’t want to pay $200/hour for a coach and is mostly just trying not to get injured again? this stuff is actually useful. ask a specific question, get a specific answer, based on your actual runs. not a blog post written for the average of everyone.

i’m not saying AI is going to replace running communities, or the weird satisfaction of figuring something out the hard way, or the fact that sometimes you just need to run more. but it’s a genuinely useful tool for people who want to think a bit more carefully about what they’re doing without turning it into a part-time job.

the honest caveat

i should be clear: i don’t have strava premium, which limits what the MCP integration can actually do right now. so some of this is still more potential than reality for people like me. but i’ve used what i can access, built what i could build, and the direction is pretty obvious.

the barrier to making sense of your own running data is collapsing. and i think that’s a good thing.

so — are you using any of this?

i’m genuinely curious what tools other runners are actually using. garmin coaching? strava AI features? something you built yourself? or are you still running on vibes and hope like a normal person?

drop it in the comments. i want to know what’s actually useful out there versus what’s just marketing copy with a “powered by AI” badge slapped on it.

2 responses to “ai is coming for your running shoes (and i kind of love it)”

  1. Hi. Random person here. I’ve been using Claude AI for a while and I’m pretty pleased so far. I have unconventional goals and crazy races, so it’s worked OK. I worry that Claude goes along with stuff too easily, and he’s never run more than a 10K or so and can’t relate all that well, but seems better than cookie cutter. I used Runna for about a year and was constantly frustrated by its arbitrary limitations. And Garmin Coach was just totally crazy unrealistic pacing and such. I have Strava Premium, whose AI does little more than constantly remind me that I am smart and cool (which is an obvious tipoff that it doesn’t work, since I am neither.) I look forward to you and others who will take what is a limitless potential that hasn’t been used yet and adapt it to the masses.

    Like

    1. Hi! Thanks for stopping by and sharing your experience. You’ve perfectly summarized the current landscape of AI running tools.

      The contrast you mentioned is exactly what makes this space so interesting right now. On one side, you have traditional apps like Runna and Garmin Coach. They are built on static, cookie-cutter logic that breaks down as soon as you step outside standard distances or conventional pacing.

      On the other side, you have generalized AI like Claude. While it’s incredibly adaptable for your “unconventional goals,” it suffers from being a bit of a sycophant. Without strict, specialized prompting, it will agree to just about any training load because it doesn’t have the real-world context of what tired legs actually feel like after a 10K. And as you noted, Strava’s AI is currently just skimming the surface with generic encouragement rather than actionable insights.

      The real magic is going to happen when we can ground that limitless, conversational potential of tools like Claude with actual, physiological constraints and running science. That’s exactly what I’m hoping to explore here!

      Liked by 1 person

Leave a comment