Fit For Science

Rob and Stephan dissect the rise of minimalist, screen-free health trackers, reviewing the popular Google Fitbit Air against established giants like the Oura Ring and WHOOP.

📝Summary
In this episode, biological data scientists Rob and Stephan explore the growing market of screenless wearables, boosted by the recent launch of the Google Fitbit Air and massive investments in companies like Oura and WHOOP. They define the specific characteristics of these minimalist trackers, emphasizing their passive 24/7 tracking, long battery life, and deliberate lack of attention-grabbing notifications. Stephan shares his initial hands-on review of the Fitbit Air after being an Oura user for eight years, highlighting its impressive AI-powered food logging, integrated workout builder, and smart alarms, while critiquing its lack of custom journaling and granular daytime heart rate variability data. The hosts also discuss the inherent form-factor limitations of smart rings during exercise, analyze the dubious claims surrounding the drop-shipped Hume band, and examine how tech giants like Google are using these affordable devices to become ecosystem leaders. Finally, they offer practical tips on syncing multiple wearables via Apple Health without data conflicts. 

⏳Chapters
00:00:00 Screenless Wearables: The market shift towards minimalist trackers and the new Fitbit Air
00:01:21 Early Adopter: Why Stephan bought the Fitbit Air after 8 years with Oura 
00:11:03 Defining the Category: What makes a wearable truly "screenless"
00:18:57 The Wearable Landscape: Amazfit Helio strap, Polar Loop, and smart rings 
00:22:03 The Hume Band Scam: Drop-shipping, AliExpress clones, and fake scientists
00:26:23 Form Factor Limits: Why smart rings struggle with cycling and weightlifting
00:28:50 Divergent Data: Comparing wildly different readiness scores between Oura and Fitbit
00:34:05 Subscription Models: Device costs and the necessity of paid app memberships
00:50:42 Fitbit Air Review (Likes): AI nutrition logging, workout builders, and smart alarms
00:57:48 Fitbit Air Review (Dislikes): Missing custom tags, daytime HRV, and slow activity detection
01:05:30 Data Management: How to safely sync Fitbit Air with Apple Health
01:11:18 Final Verdict: Will Stephan ditch the Oura Ring for the Fitbit Air?

📚Resources
 The New Google Fitbit Air and Other Fitness Bands Are Losing Screens—and Gaining Fans - WSJ
Google Fitbit Air
Sharing upcoming roadmap and improvements - Google Health
Oura Ring 5
WHOOP 
Correction: The first Whoop band officially became available and started shipping to elite athletes and professional sports teams already in 2015. Same year as the first generation Oura ring, which was sold as an end consumer product.
It looks like Garmin is finally preparing its Google Fitbit Air-style screenless band, the Cirqa, for launch — but I hope it doesn't copy Google's approach to revamping its fitness app | TechRadar 
Helio Strap 
Zepp App 
Amazfit Helio Strap does not require a subscription, but there is an optional premium add-on called Zepp Aura, a personalised Rest and Wellness Service.
Ultrahuman
Polar Loop
Hume Band2.0 
Explore Galaxy Ring
Correction: Google Health loads for most, but not all, relevant Apple Health entries the last three months.
…There is more: complete show notes here

🎙️About
Fit For Science is a deep-dive podcast hosted by two biological data scientists, Rob and Stephan, exploring the intersection of research, health tech, and data-driven lifestyle design. The hosts provide evidence-based systems, layered with practical "N=2" personal experimentation, to cut through the noise and enable everyone to become their best N-of-1.

Learn more and subscribe on your favorite platforms:
YouTube
Spotify
Apple Podcasts
Amazon Music
Collection of all show notes

⚠️Disclaimer: This podcast represents our own opinions and is for informational purposes only. It does not constitute medical or financial advice or a professional relationship.

Rob and Stephan dissect Google’s groundbreaking "Sensor FM" paper, exploring how a foundation model, trained on trillions of wearable data minutes, could revolutionize preventive healthcare and disease prediction.

📝Summary
In this episode, biological data scientists Rob and Stephan break down Google's latest research introducing Sensor FM, a massive foundation model trained on over one trillion minutes of multimodal wearable data across 5 million users. They explore the technical mechanics under the hood, including its encoder-decoder architecture, its alignment with deep learning scaling laws, and a unique "AI classroom" setup where collaborating virtual agents optimize 35 distinct clinical health prediction tasks. The hosts discuss the profound implications of using generative AI like Gemini to transform these raw sensor representations into clinician-approved lifestyle recommendations, while critically evaluating the delicate trade-offs between continuous behavioral nudging, hidden human blind spots, and systemic data privacy risks. Finally, Stephan shares his upcoming personal testing protocol for the new Fitbit Air as a potential alternative to his long-term Oura ring setup.

⏳Chapters
00:00:00 Wearable Data Predictions: Google’s motivation for tracking disease and lifestyle indicators 
00:01:48 Google Health Platform: The strategic endgame of consolidating consumer health tracking 
00:04:00 The Interface Layer: Transforming raw hardware signals into actionable health insights 
00:07:13 Introducing Sensor FM: A deep learning foundation model for wearable health data 
00:10:15 Overcoming Data Labels: High phenotypic diversity and few-shot learning 
00:12:37 Data Privacy vs. Preventive Care: Evaluating the societal trade-offs of deep tracking 
00:17:18 Encoder-Decoder Architecture: The science of signal compression and reconstruction 
00:18:43 Trillion-Minute Dataset: Mapping 5 million global wearable users
00:22:27 Empirical Scaling Laws: Maximizing compute and parameters for improved performance
00:27:19 The AI Classroom Experiment: Peer collaboration mechanisms among virtual agents 
00:39:07 Gemini Judged by Clinicians: Blind evaluation of AI-generated health recommendations 
00:44:57 Behavioral Nudging and Alignment: Addressing the blind spots of metric-driven optimization 
00:50:42 Insurance Risk Paradigms: The dangers of continuous data history in for-profit healthcare 
00:53:26 Fitbit Air vs. Oura Ring: Stephan’s upcoming personal logging experiment

📚Resources
SensorFM from Google
SenseFM: Towards a General Intelligence and Interface for Wearable Health Data 
Google Fitbit Air, das Fitness - Tracker-Armband 
Google Health 
Foundation model
Embedding (machine learning) 
Generative AI 
Supervised learning
Few-shot learning
Wearables with hypertension features: Huawei D2, Apple Watch, Whoop, Oura
Nudging
Dimensionality reduction
Prince Charles’ “middle finger” dimensionality reduction example
Sam Altman tweet: “there is no wall”
Correction: The Netherlands is divided into 12 provinces and 3 special overseas municipalities
"Life can only be understood backwards; but it must be lived forwards." - Søren Kierkegaard
…There is more: complete show notes here

🎙️About
Fit For Science is a deep-dive podcast hosted by two biological data scientists, Rob and Stephan, exploring the intersection of research, health tech, and data-driven lifestyle design. The hosts provide evidence-based systems, layered with practical "N=2" personal experimentation, to cut through the noise and enable everyone to become their best N-of-1.
Learn more and subscribe on your favorite platforms:
YouTube
Spotify
Apple Podcasts
Amazon Music
Collection of all show notes

⚠️Disclaimer: This podcast represents our own opinions and is for informational purposes only. It does not constitute medical or financial advice or a professional relationship.

In this episode, biological data scientists Rob and Stephan examine Oura's cardiovascular age validation study, discuss WHOOP's new medical concierge service, critically analyze the overhyped commercialization of the longevity bubble, and explore the citizen science potential of the Google Fitbit Air.

📝Summary
In this episode, biological data scientists Rob and Stephan break down recent structural shifts in the consumer health-tech industry, critically analyzing the transparency and financial conflicts of interest within a newly published Oura validation study on cardiovascular age tracking. They critique WHOOP’s new opt-in medical concierge service, highlighting that well-informed users can often interpret their own baseline metrics more effectively. Transitioning into the commercial "longevity bubble," the hosts dismantle overhyped, unproven, and highly expensive products, such as bespoke blood panels and premium health clubs, by explaining that 80% of longevity gains stem from free or cheap, evidence-based fundamentals like consistent sleep, regular activity, a balanced diet, and resistance training. Finally, they evaluate the hardware and subscription-free utility of the new screen-free Google Fitbit Air, exploring its practical applications as a robust tool to capture population-level sleep and activity patterns across diverse demographics within the large-scale Vienna Prevention Project (ViPP).

⏳Chapters
00:00:00 Wearable Updates: Oura, Whoop, and the Google Fitbit Air
00:01:03 Cardiovascular Age: Oura's validation study and conflict of interest concerns
00:06:57 Medical Consultations: Evaluating Whoop's physician integration
00:13:01 The Longevity Bubble: Why extreme spending yields diminishing returns compared to basic health habits
00:19:34 Bryan Johnson's Blueprint: Analyzing the $1 million protocols versus free lifestyle fundamentals
00:27:01 Communicating Science: Why Dr. Mike effectively cuts through health influencer noise
00:30:06 Google Fitbit Air: Discussing the new screen-free, subscription-free tracker
00:35:55 Sleep Inertia: The psychological and physiological benefits of timed wake-ups
00:40:43 The Vienna Prevention Project (ViPP): Deploying wearables to 20,000 citizens for public health
00:48:42 Wearable Accuracy vs App Experience: Finding the Goldilocks zone for tracking devices

📚Resources
Pulse wave velocity (PWV)
New NUS Research Validates Oura’s Vascular Age Estimation, a Key Indicator of Cardiovascular Health - The Pulse Blog 
Vascular age estimation using a consumer wearable sleep tracker | PLOS Digital Health
WHOOP just hired doctors. - LinkedIn post 
VIP medicine
Bryan Johnson 
Methylation clocks and epigenetic aging
[...] the variation from test to test is so high that any given result is essentially statistically meaningless.[...] - Matt Kaeberlein on LinkedIn
The Truth About Biological Age Tests 
Dunedin Pace | Rejuvenation Olympics 
Bryan Johnson's $1M 42 point longevity protocol and Stephan's comment  
Doctor Mike: Evidence-Based Medical Communicator on YouTube 
Google Fitbit Air, Fitness Activity Tracker Band 
Rob's videos on Google's FitBit Air (so far)
Fitbit Air: The $99 Future of Fitbit (WHOOP alternative) 
The Fitbit Air Found WHOOP’s Weak Spot! 
Fitbit: Scientific Sleep Test! 

…There is more: complete show notes here

🎙️About
Fit For Science is a deep-dive podcast hosted by two biological data scientists, Rob and Stephan, exploring the intersection of research, health tech, and data-driven lifestyle design. The hosts provide evidence-based systems, layered with practical "N=2" personal experimentation, to cut through the noise and enable everyone to become their best N-of-1.
Learn more and subscribe on your favorite platforms:
YouTube
Spotify
Apple Podcasts
Amazon Music
Collection of all show notes
⚠️Disclaimer: This podcast represents our own opinions and is for informational purposes only. It does not constitute medical or financial advice or a professional relationship.

Stephan and Rob explore the ambitious claims made by AI industry leaders about rapid scientific advancement due to AI. They provide an introduction to the "AI for Science" field and analyze the fundamental physical limitations that govern the acceleration of biomedical discovery through AI.

📝Summary
Biological data scientists Stephan and Rob evaluate the grand claims made by major tech executives regarding artificial (general) intelligence compressing a century of scientific breakthroughs into a single decade. By analyzing a recent paper co-authored by Stephan, the hosts break down the theoretical limits of general-purpose AI systems when faced with physical world restrictions. They emphasize that while cognitive tasks like literature synthesis, data analysis, and manuscript preparation can be massively accelerated, the time constants of the physical world remain irreducible bottlenecks. The conversation balances the promise of AI for science, including the hosts contributions to and beliefs in the field, with realistic infrastructure and policy demands, and the psychological and technical risks of relying on systems we do not fully comprehend.

⏳Chapters
00:00:00 Machines of Loving Grace: Dario Amodei and compressing a century of progress into a decade
00:04:10 Theoretical Scaffolding: Defining general purpose AI versus narrow machine learning systems
00:06:20 Cognitive vs Physical Domains: Splitting the lifecycle of a scientific research project & irreducible bottlenecks
00:15:35 Human Creativity and Technical Debt: The risk of losing comprehension via vibe engineering
00:19:01 Strategic Proxies: Using predictive biomarkers to capture outcomes early and bypass constraints
00:25:48 Emergence of “AI Co-Scientists”: Discovery of digital biomarkers from wearable datasets
00:38:15 Discovery Deficits: Why modern molecular biology is data-rich but discovery-poor
00:39:23 AI for Science: FutureHouse, Marinka Zitnik's ToolUniverse and James Zou's virtual lab
00:43:54 Simulating biomedicine with AI: What we did with early access to GPT-4 in 2023
00:49:57 Matthias Samwald, the EU General-Purpose AI Code of Practice and Accelerate Europe: Balancing trustworthiness and acceleration
00:54:55 MrBiomics: Automating biomedical data analysis using workflows and AI
00:58:44 Summary: Accelerating scientific discovery is possible, but not easy

📚Resources
⁠Stephan's recent paper: What are the limits to biomedical research acceleration through general-purpose AI?⁠
Social meda: ⁠LinkedIn⁠, ⁠X⁠, ⁠Press release: Potential and limitations of AI in biomedical research⁠ 
⁠A multimodal sleep foundation model for disease prediction⁠ -> we discussed this paper before in episode 8: ⁠AI is Changing Wearables in 2026(?) and Predicts 130 Diseases from Sleep! (Episode 8)⁠ 
⁠Rob's and Stephan's 2023 AI paper: GPT-4 as a biomedical simulator⁠
⁠Press release: "ChatGPT" for biomedical simulations⁠
Correction: GPT-4 predates o1-preview by 1 year and 6 months, not 6 months
⁠Matthias Samwald⁠ 
Previously: Co-chair of the Safety & Security chapter of the ⁠EU's General-Purpose AI Code of Practice⁠ 
Now: ⁠Accelerate Europe⁠ coordinator
Stephan's passion project: ⁠MrBiomics: Composable modules and recipes automate bioinformatics for multi-omics analyses⁠

…There is MUCH more: complete show notes here

🎙️About
Fit For Science is a deep-dive podcast hosted by two biological data scientists, Rob and Stephan, exploring the intersection of research, health tech, and data-driven lifestyle design. The hosts provide evidence-based systems, layered with practical "N=2" personal experimentation, to cut through the noise and enable everyone to become their best N-of-1.
Learn more and subscribe on your favorite platforms:
YouTube
Spotify
Apple Podcasts
Amazon Music
Collection of all show notes

⚠️Disclaimer: This podcast represents our own opinions and is for informational purposes only. It does not constitute medical or financial advice or a professional relationship.

Rob and Stephan break down the three critical dimensions of wearables—hardware, algorithms, and UI/UX—to explain what truly drives accurate health and sports tracking.

📝Summary
Biological data scientists Rob and Stephan explore the three foundational pillars that determine the quality of health and sports tracking wearables: hardware, algorithms, and Apps (UI/UX). They begin by evaluating the maturity of physical sensors like PPG and accelerometers, noting that while hardware capabilities have largely plateaued in high-end devices, energy density and battery technology continue to improve. The conversation then shifts to the critical differentiating factor of algorithms, breaking them down into three levels of complexity: direct on-device processing of heart rate, second-order computations for metrics like sleep staging, and highly advanced long-term disease risk predictions. Finally, the hosts discuss how the user interface and user experience tie these elements together, highlighting the importance of data presentation and the emergence of pure data aggregators in the wearable market.

⏳Chapters
00:00:00 The Three Dimensions of Wearable Performance
00:02:26 Hardware: The Foundation of Wearable Sensors
00:06:15 Understanding Raw Signals and Sensor Interference
00:09:46 Battery Technology and Hardware Durability
00:15:41 Level 1 Algorithms: Direct On-Device Processing (e.g., Heart rate)
00:23:51 Level 2 Algorithms: Derived Metrics (e.g., Sleep Stages)
00:55:50 Level 3 Algorithms: High-Level Aggregations (e.g., Long-Term Disease Risk)
00:56:20 Apps (UI & UX): The Final Wearable App Experience

📚Resources
Photoplethysmogram (PPG)
Accelerometer
Global Positioning System (GPS)
Pulse oximetry (SpO2 Sensor)
Holter monitor (ECG)
Polysomnography (Sleep Study)
Heart rate variability (HRV)
Dual carbon battery 
Edge computing 
Embedded system
Pulse wave velocity (PWV)
Foundation model (AI)
User experience (UI/UX)
Garmin
Oura Health
Apple Watch
The accuracy of Apple Watch measurements: a living systematic review and meta-analysis
Whoop
Bevel 
Athlytic 
Garbage in, garbage out (GIGO)
Introducing the new Google Fitbit Air 
A Systematic Review of Chest-Worn Sensors in Cardiac Assessment: Technologies, Advantages, and Limitations 
…There is more: complete show notes here

🎙️About
Fit For Science is a deep-dive podcast hosted by two biological data scientists, Rob and Stephan, exploring the intersection of research, health tech, and data-driven lifestyle design. The hosts provide evidence-based systems, layered with practical "N=2" personal experimentation, to cut through the noise and enable everyone to become their best N-of-1.
Learn more and subscribe on your favorite platforms:
YouTube
Spotify
Apple Podcasts
Amazon Music
Collection of all show notes

⚠️Disclaimer: This podcast represents our own opinions and is for informational purposes only. It does not constitute medical or financial advice or a professional relationship.