Goodbye Apps, Hello Data
The traditional consumer app experience is vanishing before our eyes
Over the last 10 years we have been hit with app fatigue. Every company from clothing brands, to wellness companies and even my fridge brand created their own “App”. If you look at your phone you probably have 50 apps you have opened once. That world is changing and it’s changing fast (thank god). With the advancement in agentic coding, some (not all) apps are starting to take advantage and tear down their walls, allowing their users to access their data and use it how they want.
Prediction: in the next 18 months you’ll have one app for your entire life with just “connectors” allowing you to “talk” to other platforms (goodbye ugly icons cluttering my home screen, hello simplicity).
The Lead: Connect Everything & Train With Your Data
This week we shipped a full Athlete Guide inside our free intro course for how you can now connect all your wearables (yes Garmin, Whoop, Oura, Coros, Zwift… everything) and create your own customized training platform that’s built exactly how you want. The guide is a walkthrough of Intervals.icu, a free platform (although we encourage you to support them with a $4 donation) that created connections into all these wearables.
What this allows you to do is pull in all your own training, health, and recovery data into one centralized interface that is completely customized to you. From there you can build your own customized training plans for whatever you are training for.
This is a glimpse into the future of applications and something you need to pay attention to.
What’s Hot Around the Industry?
Spotify built a running mode… “again”
High-level: Rolling out yesterday to Premium on iOS in the US, Canada, UK, Ireland, Australia, New Zealand and Sweden. 25 presets you tune by workout type, duration, BPM and genre, then it matches tracks to your tempo (cool in theory). They shipped a running feature in 2015 and killed it a couple years later. Will this be better?
KR Take: This sounds great in theory and if your run lines up to one of the workout types Spotify has pre built (steady, tapered, pyramid, or intervals) it probably is. But with the massive amount of variety in runners’ plans, I think Spotify missed a big opportunity here. They should have embraced this democratization of apps and allowed you to connect your wearable so that you can load your actual workout and then have songs customized to your actual plan and a coach that knows when to push you and when not.
Runna & Endorphins shipped heat adjusted pacing in the same week Badwater 135 (World’s Hottest Race) saw two records fall
High-level: Runna’s Adapt for Heat recalculates your paces so a slower effort in heat delivers the same training stimulus as the original target on a cool day (80% of their users report skipping or changing runs because of heat). Endorphins shipped the same idea inside a free data layer. Runna’s is subscribers only. Endorphins’ release included larger updates: a promise to free data (forever), shoe tracking, a Claude integration, and a public roadmap.
KR Take: Max Jolliffe’s win at Badwater 135 and the release of the same heat feature by two leading companies in the same week firmly cements two key ideas: 1) heat training is here to stay and you better learn it if you want to perform at your best, and 2) using data to make informed decisions about your training has never been easier. Looking at how Runna (Now a Strava subsidiary) and Endorphins each approached the second point, you can clearly see two very different approaches to the same problem. On one hand you have Runna, who not only gated everything behind a paywall, but also had its parent company launch a very public MCP, but then restrict access to the majority of users. And on the other you have Endorphins, which has publicly come out and said that their users’ data is free… forever. Oh, and also openly launched a Claude integration in the same breath. The companies that will win this next phase of consumer fitness apps will not be the ones that build stronger walls, but will be ones that remove their walls and allow consumers to build and customize what they want, when they want.
Garmin’s “Whoop Killer” already a flop?
High-level: If you ask the answer would be a resounding yes, as they put it to the test in an open water swim and saw comically bad results. After swimming two 15-minute efforts with CIRQA, an Apple Watch Ultra 3, a Polar Sense and a WHOOP MG… CIRQA came last, by a lot. But if you ask Nick at Tom’s Guide, his tests comparing CIRQA to the Garmin HRM were actually fairly close. He did two interval sessions on the track, an easy run, and a strength workout, and only the strength workout showed a real discrepancy.
KR Take: If you are looking for a leading edge HRM that is accurate down to the beat, then CIRQA is not it. But I don’t think Garmin ever intended it to be this. This device is for someone who is slowly getting into tracking their data and wants something that looks clean and is simple to use. I do not think this “kills” Whoop and I don’t think this is for a serious athlete who wants the latest tech (it’s even running an older gen 4 sensor). But if you are new to data tracking and want something that looks nice, then this is a great entry point.
ChatGPT launched Health… and our healthcare system will never be the same
High-level: Live to every US user on 7/23, web and iOS, free plans included. It connects Apple Health alongside records from Epic, Oracle Health, One Medical and Function Health, then pulls medications, labs, recent visits, sleep and activity into any chat. Ask it what changed since your last panel, how your sleep tracked against your build, or to build you a proper diet plan.
KR Take: This is one of the biggest changes to our healthcare system in recent history. For years, one of the biggest gaps in healthcare has been medical records from multiple providers and none of them having the full picture. Not only does this solve that issue, but it now gives you access to leading medical advice directly in your pocket (consult your primary doctor before doing anything… this stuff is still AI after all). People are sleeping on this, but this will forever change healthcare. Out of the stale brick and mortar and into the consumers’ hands, where it should be.
Let’s Get Technical Dev Notes
Biggest MCP (Model Context Protocol) update since its release 2 years ago
An MCP is what lets an AI actually work inside a service you already use. Instead of you having to know exactly which endpoint in Garmin to call, you connect through MCP and the model knows where to look and how to get you an answer. Two real changes and one tradeoff:
Connections used to open with a handshake that set up a session the server remembered. That’s gone. Every request now carries what it needs and stands on its own. Scales better under load, so these models can churn through your data quicker and more reliably.
Not free, though. A dropped request used to pick up where it left off. Now it just starts over.
Long jobs finally have a real design. Instead of a request sitting there blocking until it finishes, the server hands you a ticket and you check back. That’s what you need the day you want an agent that goes off, works for twenty minutes, and reports back on its own.
Opus 5 launched and it basically killed Fable 5
After all that Mythos hype it’s kind of hysterical to see how quickly Fable 5 just got replaced. Opus 5 matches or beats Fable 5 on most coding and knowledge work benchmarks, but costs half the price ($5/$25 per million input/output tokens versus Fable 5's $10/$50). Fable 5 was great for the couple minutes it existed, but man am I glad to have a better, cheaper, faster alternative again.
Let’s Wrap It Up
It is an incredibly exciting time to be a consumer in the fitness world. All of these companies, which we’ve created such allegiances with, are starting to tear down their walls and allow us to connect our beloved data across platforms. Gone to the days of being tied to one platform, and hello to the world of customization. Now it’s time to harness it.
I can’t wait to see what you build.
Kevin
Train with your own data. Build your own AI coach.
Athlete AI helps athletes and coaches take control of their training with AI. No code required. Join the free community for the 7-Day Challenge and the full classroom, or get the news in your inbox every week.
READY TO BUILD YOUR OWN APP OR AI AGENT? SEE ATHLETE AI LABS →