Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
After almost a decade of laptop tests, I have raised a truly terrifying amount of health and physical condition. And although I like to turn my daily data, there is a part that I hate: the summaries of the AI.
Over the past two years, a deluge of summaries generated by AI has been sprinkled in each fitness, well-being and laptop application. Strava presented A functionality called Athlete IntelligencePresented as AA, taking your raw training data and relaying it in “Plain English”. Whoop has Whoop coachAn AI chatbot which gives you a report of “daily perspectives” summarizing the weather, your recent activity and recovery measures and training suggestions. Oura added Oura AdvisorAnother chatbot that sums up the data and removes long -term trends. Even My bed Selves me with summaries every morning on how his AI helped fall asleep every night.
The AI of each platform has its nuances, but the typical morning summary goes a bit like this:
Good morning! You slept 7 hours last night with a heart rate at 60 bpm. This corresponds to your weekly average, but your slightly high heart rate suggests that you cannot be completely recovered. If you feel tired, try going to bed earlier tonight. Health is a matter of balance!
Who could seem Useful, but these summaries are generally placed next to a graph with the same data. It’s worse for training. Here is one that the intelligence of Strava athletes has generated for a recent race:
An intense race with high heart rate zones, pushing in an anaerobic territory and recording a relative effort well above your typical scope.
THANKS? I can ask the information of athletes to “say more”, but it regurgitates my effort, my heart rate area and my rhythm measures that I can see in the graphics in the summary of the training. If you did not know anything about my sports history or circumstances surrounding this race, this summary could be read as perceptive. Here is what the summary left aside:
- It was dangerous to triple my mileage in my second round of the year, given the high humidity, more than 85 degrees, and my unequal training history in the last two months compared to the previous six months. Strava has access to weather data And Each training I have done in the past five years.
- I had to interrupt this race because I fell and I shredd my hand and knees. This is information to which Strava has access, because I have downloaded a knotty image in addition to text notes. After adding the said notes, the updated summary only reflected that I cut the short race. My injury has not changed anything in his ideas, even if it is the most important thing that happened during this race.
A more useful overview may have been: “You have run during the record heat of your region. Although you have maintained a coherent and regular rhythm, you have the bad habit of raising the mileage too quickly after prolonged breaks, causing several self -declared injuries in the past five years. A safer alternative would be a mileage lower than lower races to hang on to temperatures. healed.
Runna, an application in the progress of popular execution which also includes Insights, has generated a slightly More useful summary. He said that my next race should be “easy”, the one that is perfectly timed for me to recharge. I’m sorry, but 48 hours are not enough time for my knees to heal safely without the risk of reopening my injuries.
Integrated chatbots are not much better. Yesterday morning, I asked Whoop Coach if I had to run today because I injured myself during my last race. He said to me, “Whoop is unable to respond to the message you sent. Please try to send a different message.” I tried to crop my prompt, saying: “I am injured and I have a soft. I was invited to contact the membership services in Whoop to continue the conversation.
Oura Advisor was more useful, noting in my daily summary: “With your preparation of preparation and recent stress factors such as heat, injury and higher glucose, your body can be more tired than usual today.” It suggested that I prioritize rest. When asked: “What types of movements are correct when you have a wounded knee and a slight box?” He responded with common sense responses like a short and easy walk if there is no pain, light stretching and a reminder to rest completely if I feel a high discomfort. This is closer to an ideal answer, but I had to guide it to the type of answer I wanted. Ideas are SO In general, which they benefit from that they benefit from self -quantification beginners – and even then, only if they are allergic to Google research.
My sloppy race is exactly the type of scenario where technological CEOs say that information on AI could be the most useful. In theory, I agree! He would be Be kind to have a competent and integrated chatbot that I could ask more nuanced questions.
1/ /3
For example, I had an irregular sleep schedule this month. I asked our advisor to sleep if my sleep and preparation trends showed signs of a high risk of injury. I also asked if I had abnormally high sleep debt levels this month. In both cases, he said no – he said I was improvement.
What resulted was a one -hour debate with a chatbot that made me question my own lived experience. When I tried to ask him to dive into a particularly stressful week earlier this month, he told me that his ideas were “limited to (my) week the last and current trends”. This kind of defeat against having six years of data from our data.
After months of realization of Reddit and other community forums, I know I am not the only person to find these characteristics of the AI to be laughable. And yet, Holly Shelton, director of products for our Oura, tells me that the response to Oura Advisor was “extremely positive”, with 60% of users using it several times a week and 20% using it daily. “Beyond the frequency,” says Shelton, “it offers a real impact: 60% say that the advisor helped them better understand the metrics or the health concepts they have previously found confusing.”
Meanwhile, Strava’s spokesperson Brian Bell tells me that athlete information was intended to help beginners athletes and that “the response to functionality remains strong” with around “80% of those who oppose comments by finding the” very useful “functionality to” useful “.
A Whoop spokesperson could not respond by publication.
These Milquetoast summaries? They are probably the best compromise between speed, cost, usefulness, confidentiality of data and legal responsibility
I understand that my frustrations arise from the inherent limits of the LLM Private health data disorder. Strava could be a de facto fitness data center, but it is missing all The health data necessary to create holistic, useful and personalized information. It would take a long time to advise a year of sleep data for trends. This latency is guaranteed to offer a bad user experience. Without forgetting, they should probably increase their subscription from $ 5.99 per month to add this type of computing power. I’m not sure, but Whoop’s coach may have refused my requirements related to injuries to protect themselves from responsibility if something has come to me to follow his suggestions. These Milquetoast summaries are probably the best compromise between speed, cost, usefulness, data confidentiality and legal responsibility. But if so, then let’s be honest. The current AI features are reconditioned data, much like the books of books written by a fourth year student which relies on a summary of Wikipedia instead of reading the book. It is a functionality nailed with adhesive tape and a dream because AI is the zeitgeist. Maybe A Day, these ideas of AI will create a useful and personalized experience with usable information. This day is not today, and it is not worth paying an supplement.
(Tagstranslate) Analysis
Source link
