This study is currently going viral, perhaps because it is one of few that are statistically extremely strong. But as I read it, I realized we have one BIG blind spot in how we measure exercise. The study, published in British Journal of Sports Medicine (doi; 10.1136/bjsports-2025-111351), suggests that for achieving a 30% reduction in cardiovascular disease requires roughly 560 to 610 minutes of moderate to vigorous activity per week. This is around 80 minutes every day! So movement is crucial: Running, swimming, cycling, playing football, tennis, etc. The study shows statistically that movement – undoubtedly – matters enormously! The good thing: We became very good at measuring it. Smartwatches now measure how long we run, it calculates distance, pace, heart rate, HRV – some can even estimate VO2 max. These are affordable! You can track your cardiorespiratory activitry and fitness reliably enough for 20 to 40€ with a Xiaomi MiBand. But the study made me think about something else. If all of the sudden, everyone is exercising cardio for 80 minutes a day: What happens during those 80 minutes? Imagine two people who both exercise 80 minutes every day. Their sophisticated Garmin Fenix smartwatch will essentially tell the same story. But mechanically, their bodies could experience something completely different. How much force are they producing? How quickly can the do so? What happens when they land? How do they decelerate? How are loads distributed between the left and right leg? Can they stabilize themselves after a rapid movement? What I want to say: Two people doing the same exercises for 80 minutes a day can arrive at a similar cardiovascular health, while ending up at entirely different muscular or join health. One becomes stronger and more resilient, the other accumulates weakness, asymmetry or excessive joint loading. In other words: They can have the same VO2-max, run the same distance at the same pace – while still having completely different strength, stability, force absorbtion and joint loading profiles. We measure the “dose” of movement very well, with every new smartwatch generation we can measure the cardiovascular response better. But outside of elite sport, biomechanics science labs, or rehabilitation (when it is too late), people do not have access to measuring their mechanical interactions with the ground. Why does it matter? When you run, every single step means you apply force to the ground and then absorb force again. Can you tell me, how you absorb forces when you run? Or Whether legs behave differently between left and right? Or how it changes after training, an injury, or simply as you get older? Unless you go to a specialized lab, you don’t know the answers. I think this matters enormously, because health is not only a cardiovascular event. And I’m very happy that with Movetrics.com we are working on making these measurments available to everyone on earth. This is a real photo of our Movetrics F1-3D force plate: If you find this interesting, we are currently looking for 20 pilot customers who want to help us make this kind of force measurement more accessible outside elite sport and biomechanics labs. For example: gyms, amateur sports clubs, physiotherapy, applied science. Our idea is simple: you use the system with real clients, in real world settings, give us honest feedback, and help us shape what the product needs to become, in order to reach every human being on our beautiful planet. If this sounds relevant for you or your organization, send me a message or email: ms@movetrics.com Link to the mentioned study: https://bjsm.bmj.com/content/early/2026/05/28/bjsports-2025-111351.long

The problem is not slop, most AI written analysis and texts of today's top models are quite honestly of higher quality than of the average human. The problem is the fake authority it creates when humans publish AI written texts without disclosing that it was fully or partily AI written. It’s like an athlete who is secretly using performance-enhancing drugs while competing against athletes who are not cheating, and then pretending it was all natural hard work.

The beautiful thing about being authentic and direct is that you filter out people you don't want to have in your life anyways. Being a people pleaser will make more people like you; people that you don't even enjoy talking to — and perhaps don't even want to be seen with. If you are authentic and say out loud what you think, you will naturally have a circle of friends and fans that you enjoy being with.

Test. This is posted on Nostr first, then mirrored to X.

This is a side project we've been working on: a PoC prototype for accurate weight measurements of car seat passengers. This is a 98.55kg static load test where we achieve ±97% accuracy for measurements across relevant weight classes (46kg to 98kg). This is not a final product, it still is a proof of concept prototype. We already found multiple vectors where we will improve it; but for that we are looking for an OEM or Tier 1 partner.

If you seriously think about what current AI is capable of, it should blow your mind. If you imagine what it will be capable of 3 years from today, you will reconsider everything you're currently doing.

Today I saw a nice Mac app, but it was sold at 30 USD. I gave the link of the landingpage to Codex, told it to define and plan the necessary goal in order to build the same features for me. It took 3 hours 28 minutes, without a single intervention on my side, using 1,554,809 tokens (20% of my weekly $20USD usage) and I had a FULLY functional Mac app, that works perfectly!

The most valuable decisions happen beyond logic. Beyond logic means: the spreadsheet is accurate, but incomplete. This is what Tim Cook described in a conversation at Duke University in 2013. It was Intuition that led him to work for Apple, not logic. For him, at that time, Compaq was the superior, most rational choice. But that choice was based on a rational, static model of the present; which is a snapshot of today plus the accumulation of historical data points. The superior choice was what logic did not agree with: joining Apple. This is what he felt as intuitive truth, he was knowing what will become possible. AI is today the plus and minus list Tim Cook worked on. Data backed analysis and logic work when variables are known and the future will resemble the past. This is the strength of AI. The truly important and world changing decisions are not that. These decisions happen BEYOND LOGIC; usually in situations that are genuinely novel with a future that is nonlinear and unpredictable. Tim Cook repeatedly tried to make the analytical pro and contra list work in his favor, to tell him what he already KNEW. But he could not. The individuals who will OVERGO unquestionable data-backed logical recommendations of AI systems in favor of what seemingly makes no sense at that time, will win. Only humans can sense the emerging reality, AI systems will never be able to.

One in three top LinkedIn posts is now AI-generated. Which means two things: first, that sounding like an expert now proves nothing, and second, that your LinkedIn feed is now drowning in fake authority. Pangram – the leading AI-detector – did an intersting analysis with unsurprising results: 1) 40 % of long LinkedIn articles flagged as fully AI-generated 2) LinkedIn is accountable for 62 % of all flagged AI content, even though it representet only 1/3 of scans. What it already shows is that long writing is no longer scarce on LinkedIn, they became abundant – but what is missing for the reader is the evidence how real the mind / human behind an article actually is. In other words: The risk is not the fake writing (which tools like Pangram can detect), but the fake authority behind it. Zero marginal cost -> infinite polished posts -> synthetic authority inflation For example, a high alpha strategic view on a specific subject previously required and thereby signaled competence. Now exactly these type of posts become increasingly meaningless — which discourages those with real competence to publish anything at all. It is clear, that content as a category becomes meaningless and competence signalling will shift entirely to costly signalling: deals closed, decisions made, documented mistakes with consequences, forecasts that become timestamped in a blockchain before outcomes, etc. What most on LinkedIn overlook: generic AI content without costly signalling will not only underperform, but because of detection tools (like Pangram) and with increasing audience skepticism (like myself), it becomes a negative reputational signal: this person wants authority without the cost of thought. Some – myself or Paul Graham included – started blocking anyone commenting with AI slop below posts and articles. Not only slop, but anything generic is becoming reputationally dangerous. Even senior executives, who built their career over decades, publishing interchangable insights will now signal less competence than an imperfect writer publishing a specific accountable observation. Ergo? Treat LinkedIn engagement as a weaker evidence of expertise than ever before. The important question for Nostr: How can we avoid becoming a protocol of AI generated slop? Source:

AI Content Is Everywhere on Social Media, Especially LinkedIn

We scanned over 1 million social media posts for AI content. It turned up on every platform we checked, and 1 in 3 top LinkedIn posts flagged as AI...

Nobody who is not using X.com actually has an idea of what is going on in the world.

what is the most underrated PDF online?