My regret minimization framework for venture essentially is: back exceptional people trying to do something extraordinary. You never have to regret that investment decision, even though many of them don't work out financially. They were exceptional people. They tried to do something amazing. I don't ever regret giving my life to this. And enough of them work out so I get to keep doing it.

Long leisure travel, esp. outdoor, esp. experiences

Quick thoughts on Europe 2031, from the flight home. Fun, bleak, and the most effective piece of European AI advocacy I've read. Prescient, too: DC just cut foreign-national frontier-model access on security grounds. Where I agree: make Europe the fastest place in the West to energize a gigawatt and the machine comes to you. That needs permissionless construction and cheap energy, which Europe is bad at. Special Compute Zones point to a way through. Where I differ: I doubt compute scarcity, a US-EU trade breakdown, and the labs abandoning major markets all land at once. Nor would I give up on a European frontier model: China is less behind than the authors assume, and small, distilled, well-scaffolded models punch above their weight. At Heartcore we've backed vertical models that beat OpenAI and Anthropic in healthcare and real estate: refined jet fuel for Concorde, not a gas station. Finally, the US security guarantee remains existential. The time is right to take ~€500B to DC for an Atlantic Compact: Europe trades weapons purchases, regulatory peace, supply-chain alignment, and open capital markets for guaranteed frontier-AI access and American-operated compute on European soil. The successor to NATO. What Vance called for in Munich still needs a resolution in Europe: what the West is for and why its cohesion is worth paying for. The human person, made in the image of you-know-who, possessing a dignity that no state confers or may revoke. This is the actual foundation beneath Western democracy. It is the precise point of contradiction with the systems now consolidating on the other side of the strategic divide with China: the surveillance state, the social-credit apparatus, the subordination of the person to the CCP’s continuity. The contest of this century is over whether the most powerful technology ever built will be governed by polities that regard the person as an end or as an instrument. On that question Europe and America are on the same side, whatever their quarrels. And a European strategy that treats DC and Beijing as interchangeable bullies commits a category error with civilizational stakes. Two sins currently bracket Europe's temptations. One is idolatry: sovereignty pursued as an end in itself, the nation or the Union worshipped as absolute ("Buy European"). The other is acedia, the sloth that cannot rouse itself to act on what it knows, which is the exact sin of the scenario's epilogue: "we told ourselves it could not be done and we went home and we let it happen." Against both stands the virtue the moment requires: hope, which is neither optimism nor a mood, but the settled conviction that the good is worth working for. History is not closed. Europe's patrimony built the cathedral and the university, the hospital and the common law of nations. A civilization that built those can build fifty gigawatts and look its allies in the eye. I think it’s time for a great new compact between Europe and the US.

Artifical intelligence, healthcare, robotics, defense, synthetic biology - one room in Copenhagen. Attendees from London, Paris, New York, Tokyo, and all over the European continent. Heartcore Day 2026 packed a lot more range into 24 hours than most things I have attended this year :) Three things stood out: The incredible breadth. Andreas (Corti) with a healthcare foundational model, Evan (Fora Travel) with a platform empowering travel entrepreneurs, Eden (Nūmi) making breastmilk for moms who can't breastfeed, Fabian (Qualia) making the layer that robotics companies building on VLAs need, Simon (EGIDE) explaining Shahads with Pixar, Mehdi (Davis) outlining their incredible recursive language model in real estate, and a few founders still in stealth, together in one room. The range of incredible businesses Heartcore founders are building has never been wider. The truth. The best moments were the ones where founders and LPs talked openly. We asked for honest conversations and got them! The love. An evening of jazz (way too loud! but apparently world-class), then a day of debate on what we have learned, what we are seeing and where we go next. Thank you to every founder and LP who gave us those 24 hours. The trust, and the honest conversations in this three-sided marketplace business, are what makes us work. $500M returned so far! Already planning Heartcore Day 2027! If you were in the room: say hi and lmk what I missed. And if you want to be in the room: comment and I'll put you on the waitlist for next year!