Why China still can’t make ballpoint pens. “Made in China has two weaknesses: one is ‘can’t be made,’ which is a real technical chokehold; the other is ‘can’t be used,’ which is often stuck on non-technical barriers.” The piece uses […]
Ideas don’t combine themselves. How long do we wait for new inventions? A nice reminder that “time to invention” is often about social distance between fields, not missing parts. ~ learn more
Texan levels of energy wealth. Sharing another profile of one of my favorite companies that’s making renewable natural gas out of thin air. “The question was not whether we’ll still need hydrocarbons in 2050, but where they would come from.” […]
Biology is a burrito. “But a cell looks more like a burrito,” says Michael Elowitz. The essay makes the case that cells are absurdly crowded, and you need numbers to really feel how fast and chaotic the inside of a […]
Seed hundreds of creators. “I have personally seen 6 brands restart growth using the “hudson method” we are talking 5m a year brands to 100m a year brands.” Sean Frank’s playbook: flood TikTok with lots of small creators, then reuse […]
Why jumping into cold water can stop your heart. “Most cold-water deaths happen in the first three minutes.” It is not hypothermia, it is the nervous system chaos right after immersion, when cold shock and the diving response can collide. […]
Tough calls in transplant surgery. “You’ve got a patient who does not have an infinite window. There are a lot of really tough calls, and it’s always a multidisciplinary team decision.” The piece shows how matching and transporting organs can […]
A team appeared out of nowhere. “All of a sudden, in the slides there were a new team.” Aleix Morgadas on what happens when leadership adds a CX team based on an old operating model, then wonders why the dashboard […]
Outsource thinking, not understanding. “You can outsource your thinking, but you can’t outsource your understanding.” Karpathy’s update on “vibe coding” is that the real skill is developing taste for directing messy, agent-like LLM systems without fooling yourself about what you […]
Where OpenAI’s goblins came from. “Starting with GPT‑5.1, our models began developing a strange habit: they increasingly mentioned goblins, gremlins, and other creatures in their metaphors. Unlike model bugs that show up through a tanking eval or a spiking training […]