Bart Collet

None of the AI app-builders sign a BAA. I checked 18 platforms against their own terms: the layer underneath (hosting, database, model APIs) is what signs, not the code generator. cliniciansthatcode.com/guides/whic...

Before buying a partner-scouting tool: can you describe the ideal partner in one sentence? If not, filters won't help. A better B2B partnership stack: AI to widen the field, a database to qualify, an intro to warm up. da.gd/2LIsx8

A Paris intensivist vibe-coded a ventilator titration tool for ARDS in 4 min 25 s, then verified it with 1,400 automated checks. Zero formula errors. The speed is the headline. The clinical spec and the tests are the medicine. da.gd/Vj2e9

AI agents are finding software flaws faster than humans, and attackers are using both open and closed models. Cybersecurity likely gets worse before it gets better, and the fix may be a safer but less open internet. Acceptable trade?

50+ longevity events for 2027 are already on our calendar. Europe or the US? Science and policy: Europe. Founders, investors, consumer wellness: the US. Most events per trip: Europe, with denser clusters. da.gd/2gJdp

25,000 people wait for a $499 Neko Health scan in New York. England's NHS list is 7.33 million treatments, 111,000 over a year. Same signal, opposite verdict: the difference is whether you chose to stand in the queue. da.gd/ak7A

$45,000 a year is the top of the hospital longevity market. Weill Cornell's FAQ says the goal is proving the model so insurers eventually reimburse it: a trial where the participant pays the hospital, not the other way round. da.gd/cDSVe

We track 500+ longevity events; the market behind them has focused overwhelmingly on men. Women's biology needs a different approach, not a smaller dose. In Q1, 115 consumer longevity startups were mapped and Fonda and Hyman were named to headline a women's health summit. da.gd/TlRTq

Enron booked $110 million of 'profit' from a film-streaming deal with Blockbuster before one movie reached one customer. Dead in eight months. None of the six unusual pairings here died of being weird; they died of skipped diligence. da.gd/hdKxNT

Tesla's Roadster funded the Model S and Model 3: premium first, cheap second, on purpose. Preventive healthcare keeps reversing the order, or sells the premium tier and keeps the margin. In 2014 Tesla open-sourced its patents. da.gd/Fu1NKi

118 breakout attempts across nine frontier models, zero escapes, says Perplexity of its open-sourced SPACE agent sandbox. Open code lets others test that claim, which matters. But 118 is a small sample for AI agent security.

One engineer built Stripe's Harbor v1; Ramp needed two for Glass. Internal vibe-coding tools look cheap next to Lovable or Replit once internal data and design systems matter. The unknown is who maintains the apps later. da.gd/qzF3C

55% of free-trial cancellations happen on Day 0 (RevenueCat, 115,000 apps). About half of paid conversions happen then too. For product onboarding and retention, the first session is most of the game. Does that punish tools that need setup? da.gd/R1LdCP

The Handover #11: a Centre for AI and Digital Health. The MHRA's 2030 Strategy commits to a Centre for AI and Digital Health and an adaptive framework for AI as a medical device. cliniciansthatcode.com/briefing/is...

We archive longevity events instead of deleting them once they end. With 550+ events and 4,700+ speakers on record, the history shows which topics gain programme time and which cities become hubs. da.gd/XbJH

Product-led growth is love at first sight. 55% of trial cancellations happen on Day 0. So do about half of all paid conversions. But instant value only gets you the first date. Month three is the marriage. vcunfiltered.beehiiv.com/p/your-user...

77% accuracy reading thoughts via wearable cap, no surgery. Every serious BCI so far required a cranial procedure. Non-invasive changes the access equation entirely. Brain computer interface without implants is the real unlock to watch.

Oura x Dexcom, WHOOP x Quest, Garmin's €40m study with King's College London (40,000 participants). Wearables are becoming a biological data platform. Who owns the data? da.gd/KgAD6

Nvidia agreed to buy a minority stake in Doctolib. My question: what happens when Doctolib merges with Alan? live.euronext.com/en/financia...

$250M, six years stealth, and a factory that adapts to code changes without retooling. Atomic Machines targets microscale components chip fabs cannot make. Prompt-to-product manufacturing is the pitch; a power switch is the proof point.

A 3T MRI: about €1.2m and three staff. A low-field MRI: €1,450 a month over five years. The hard part of subscription diagnostics is not the hardware, it is keeping idle devices busy. da.gd/k1DudL

Abu Dhabi and Ultrahuman are building a biological age model from biomarkers plus wearable data. The real test: does it predict illness better than existing risk scores? Two years to find out. da.gd/dlXcB

GLP-1 drugs outpacing generative AI on revenue is the kind of signal that reframes a whole decade. Prices falling, oral formats coming, adherence still unsolved. The science is lagging the commercial curve.

Venture clienting in three comparisons: accelerators back many startups, acquisitions absorb the whole company, venture clienting buys one product for one problem and stays reversible. Test in months, not years. da.gd/e8uqE

4,700+ speakers across 550+ longevity events: corporate and industry voices now outnumber academics on stage. A typical summit has ~15 speakers over 2-3 days. Know who is speaking before you book. da.gd/qCJo

40% of US adults see a dentist yearly despite 80% wanting more care. Wally's $249 membership model strips out insurance friction entirely; 50K members across 15 clinics. The oral health as preventive wellness reframe is the interesting part, not the funding.

Longevity got a hearing at the European Parliament on 1 October. Geroscience and prevention pitched as an EU health priority. Still advocacy, not policy. The test: a line in FP10. longevity.technology/news/eu-lon...

73% of committed wellness travellers plan to integrate more health into trips. Minor Hotels is building portable clinical records across resort and urban locations; longitudinal wellness data following guests between properties is the actual bet here. da.gd/3Ajl4z

108 things clinicians shipped in nine weeks, classified by hand. 59 contain no AI at all. 20 hold real patient data; 19 say nothing about assurance. Clinician builders are making tools for their own shift, not an AI doctor. da.gd/iJgnF

The Handover #10: the 44 recommendations, accepted. The UK government accepted every recommendation of its doctors-led AI Commission and opened AI Airlock phase 3. cliniciansthatcode.com/briefing/is...

Wearables went from fitness gadget to front door of the lab in one September: a $119, 50+ biomarker panel bookable in Apple Health, Oura's IPO at up to $15.6bn. Yet no more than 6% of doctors in any country have wearable data fully in their workflow. da.gd/u0lX2

Minor Hotels launches Life Longevity: a wellness brand across resorts, city locations and future residences. Health records and care plans follow guests between locations. www.spabusiness.com/index.cfm?p...

Ageing biomarkers validated for populations, not yet for individuals. That one gap is blocking personalised longevity medicine and shorter trials more than any AI limitation. Worth watching closely.

Longevity conferences are no longer just biotechs and biohackers: insurers, wealth managers and hospitality groups now send people. 2,000+ sponsors across 550+ longevity events; the calendar is a proxy for where capital moves. da.gd/2EiOKG

Empathy mapping splits user quotes, thoughts, actions, and feelings into four sections. The gap between what users say and what they do is where most product failures originate. design thinking user research product design www.nngroup.com/articles/em...

Four ways teams find partners: databases, matchmaking platforms, curated events, generative AI. Good results come from stacking two: widen with generative methods, qualify with a database or curated programme. Open innovation in practice. da.gd/l3YXf6

The Q3 Longevity Events Pulse is out: 74 longevity events this quarter, Europe level with North America at 32 each, and September plus October carrying 30% of the year. Free eleven-page PDF; three pages attached. da.gd/xrXIgl

Looking for hackathons and radical healthtech innovation events in Europe? Here you go! hackathonhub.eu/events?them...

Utah's health AI sandbox is reviving independent assurance labs; third-party testing with findings shared back to regulators. The model is sound in theory. The hard part is picking reviewers who are actually independent. Other states are watching.

Longevity vocabulary is a price list: healthspan, hotspan, sexspan ($30-40k a year), peakspan. Each -span starts as a scarcity product; software (Midi, about $50 a visit) climbs after it. Count the -spans in your next longevity pitch. da.gd/J5w7i7

Great read at Fast Company: concise, easy-to-understand basic principles of vibe-coding in healthcare. www.fastcompany.com/91615347/th...

$42m raised for a one-year post-high-school programme where OpenAI and Palantir co-author the curriculum. Proof of work over grades for admissions. The vertical integration of education and hiring in one VC ecosystem is the real story here.

26 regenerative medicine events ran Jan to Aug 2026; 12 are still ahead. Pick by job: data (ISSCR, CSHL), CME credits (AASCP, AMMG) or deals (Cell & Gene Meeting on the Mesa, Phoenix this week). da.gd/1ZG5vF

56% of workers rarely use AI and only 40% can access learning resources. That's not a skills problem; it's an organisational infrastructure problem. AI workforce stratification workforce futures

Over half of AI front-runners have shifted from assigned tasks to outcome ownership; same accountability for the what, full autonomy on the how. That's organisational redesign, not tooling. Most firms aren't ready for that conversation. da.gd/YK2DPo

"De toekomst behoort toe aan probleemoplossers" ... en morgen komen die probleemoplossers samen in Antwerpen voor #careandcode www.uptodatewebdesign.com/2026/09/zor...

Tavus Griffin: a ‘Human Interaction Model’ that renders a lifelike person who can hear, see, talk, and react over live video. This is getting spooky good, very fast. www.tavus.io/griffin

If you keep the dead happy, they are less likely to haunt you. liquiddeath.com/en-be/pages...

AI shifts the founding bottleneck from execution to selection. Steve Blank's point lands hard: AI makes bad ideas go faster. The moat is now judgement, not speed. abundance entrepreneurship customer discovery startup da.gd/ezHV

The Handover #9: the AI search that misses the paper. University of Florida researchers tested five AI evidence-search tools: 12% of the relevant evidence was never retrieved by any of them. cliniciansthatcode.com/briefing/is...