What I Am Reading:
The 5AM Club: Own Your Morning. Elevate Your Life.
Robin Sharma wrote this as a fable rather than a manual, which is either the best or worst decision depending on your tolerance for storytelling in a business book. The core of it is the 20/20/20 formula for the first hour after waking: twenty minutes moving, twenty minutes reflecting, twenty minutes learning. What I find useful is not the specific hour but the underlying claim that the first sixty minutes set the tone for everything that follows, and most people surrender that window to a phone screen before their feet hit the floor. I do not wake at five and I am not going to pretend otherwise. But the argument that you should own the front of your day rather than react to it has held up every time I have actually applied it.
We're All One Crisis Away From Taking Unlicensed Research Peptides
Elizabeth Van Nostrand wrote this for Asterisk, which is one of the few publications where the writers actually chase every claim back to its source before printing it. The premise sits uncomfortably close to home for anyone who follows longevity seriously: the line between responsible self experimentation and buying unregulated compounds off the internet is thinner than most people admit, and it usually gets crossed when something goes wrong with your health and conventional medicine runs out of answers. Peptides are the current frontier of that gray zone. This is the most honest treatment of the topic I have come across, and it is neither a hit piece nor a promotion.
Why American data centers can't plug in
Chris Gillett makes an argument that reframes the entire AI energy conversation. America is not short on electricity. The problem is moving it from where it is generated to where the data centers need it, and the interconnection queue and transmission buildout are the actual constraints. That distinction matters enormously if you are trying to figure out where capital should go. Generation is not the bottleneck. The wires, the substations, and the permitting process are, and those are much harder problems to solve with money alone.
The world's most complex machine
Neil Hacker tells the story of how ASML bet on extreme ultraviolet lithography for decades before anyone knew whether it would work, and in doing so became the single chokepoint for every advanced chip on the planet. What strikes me is the timeline. This was not a clever pivot or a fast follow. It was a multi decade commitment to a technology that repeatedly looked like it might never function. Every conversation about chip sovereignty and export controls eventually runs into one Dutch company in Veldhoven, and this explains exactly how that happened.
The algorithm will see you now
Radiology was supposed to be the first medical specialty that AI fully replaced. It has digital images, clear benchmarks, and repeatable tasks, which is close to an ideal setup for a model. Deena Mousa walks through why it has turned out to be far harder than anyone predicted, and the reasons are instructive well beyond medicine. The gap between benchmark performance and real deployment keeps showing up in every field I look at, and the people who understand why that gap exists tend to make much better decisions about where automation actually lands.
Aria Schrecker makes the case that the answers to human longevity may already exist in animals that routinely outlive us by an order of magnitude. Naked mole rats live thirty years and rarely develop cancer. Bowhead whales pass two hundred. Some of these species have evolved cellular repair mechanisms that we simply do not have. Rather than inventing longevity from scratch, the more promising path may be reverse engineering what evolution already solved. That framing changed how I think about which biotech bets are actually worth taking seriously.
In 1958, a Ukrainian virologist serving as deputy health minister of the Soviet Union stood up at the World Health Assembly and argued that smallpox could be eradicated from the face of the earth. A nearly identical proposal had been rejected three years earlier as unrealistic. His argument was convincing enough that the vote came back unanimous, and smallpox was declared eliminated worldwide in 1980. Some people have argued Zhdanov did more good for humanity than any individual in history, and almost nobody knows his name. Saloni Dattani tells it well, and it is a useful reminder that the highest leverage moments often look like one person making one argument in one room.
How Long Until AI Doesn't Need Humans?
Asterisk put Ajeya Cotra and Timothy B. Lee in conversation on the question everyone is arguing about badly, and the format is what makes it worth reading. Cotra comes from the forecasting world and Lee is a technology journalist who has spent years pushing back on inflated timelines. Neither is performing certainty. What you get instead is two well informed people disagreeing carefully about where the automation of knowledge work actually breaks down, which is far more useful than another confident prediction from someone with a position to talk up.
What I Am Listening to:
BI 241 Johannes Jaeger: Agency and the Cyborg Myth
Paul Middlebrooks continues to run the most consistently interesting show at the intersection of neuroscience and AI, and this episode makes an argument I have not heard anywhere in the mainstream conversation. Johannes Jaeger contends that from single cells to whole organisms, living things have to continuously regenerate themselves and judge what matters in order to keep living, and that this is the actual source of agency. AI does not do this and, in his view, cannot. Whether or not you land where he lands, it is a far more rigorous version of the consciousness debate than what usually circulates, and it comes from someone doing the actual biology.
The Dark Side of the Longevity Industry, with Matt Kaeberlein
Matt Kaeberlein is a career geroscientist who ran the Dog Aging Project, and this episode is him calling out his own industry with more candor than anyone in his position usually risks. He had just come from two conferences in the same stretch: a high end longevity investors event in Gstaad and an academic aging conference in Jena, and the contrast between them is the whole story. He gets into why the term longevity medicine is already acquiring a bad reputation, the ethical problem with selling biological age clocks in clinical practice today, and the widening gap between real science and expensive snake oil. If you spend money in this category, listen to this before you spend more.
Jordan Schneider brings on the SemiAnalysis robotics team to make the case that robots are the real general purpose technology, because for the first time in history we can decouple capital from human labor. The part that stayed with me is their argument on Unitree, which most Western observers keep dismissing as a maker of dancing robot dogs while it quietly runs the same playbook DJI and BYD ran. Vertical integration, in house actuator manufacturing, and a supply chain that makes Chinese humanoids dramatically cheaper than American ones. Their closing line is one every investor in this space should sit with: you cannot AI your way out of a hardware problem.
What I Am Watching:
The Art Of Winning Requires One Thing Most People Won't Give
Arnold Schwarzenegger has told versions of this story for forty years and it still works, mostly because he is not selling anything abstract. The thing most people will not give is the unglamorous volume of reps, done when nobody is watching and nothing is happening. He is credible on it because he did it in three separate careers that had nothing to do with each other. I am generally allergic to motivational content, and this one still got me. Worth watching before a stretch where you know the work is going to be boring for a while.
How To Explain Anything To Anyone
Sharran Srivatsaa packs a genuinely useful communication framework into about sixty seconds. Clarity wins when you give people four things: context, understanding, practical guidance, and one obvious next step. Most explanations fail because they deliver the middle two and skip the bookends, so the listener follows the logic and then has no idea what to do with it. I started thinking about my own investor updates and board conversations differently after watching this. The one obvious next step is the part almost everyone leaves out.
Synthetic Brains and Minds: What Is Possible?
Ricard Solé runs the Complex Systems Lab in Barcelona and holds external faculty status at the Santa Fe Institute, and this is the second lecture in his Ulam Memorial series on evolving brains. His question is whether we can build genuinely new kinds of minds rather than only studying the ones evolution already produced, and whether synthetic cognition would be recognizably different from ours. He pulls from synthetic biology, artificial life, physics, and AI in the same talk, which almost nobody does well. SFI puts these lectures on YouTube for free and they remain one of the most underused resources on the internet.
