Until now, I’ve been publishing recommendations on a highly scientific schedule known as “whenever I feel like it.” Lately I’ve caught myself forwarding the same articles to a bunch of people, so I figured I might as well centralize the chaos. I’ll (try to) do one of these every month. These are a few things I poked at during February.
My themes are basically whatever I’m currently obsessed with: healthcare, technology, AI (both the shiny upside and the creeping doom), art, and books.
Anyway, let’s see if this works. Here are 5 I genuinely enjoyed this past month:
Wall Street: Stewart vs the suits
So I’ve been going down a rabbit hole with this one and honestly can’t stop thinking about it. Jon Stewart sits down with Oren Cass, and they basically put into words something I’ve felt for years but couldn’t quite explain: the economy stopped working for most people not by accident, but by design. They call it “financialization,” which sounds boring but is actually kind of a bombshell once you get it. It’s the idea that at some point America stopped making things and started just reshuffling who owns things, and Wall Street got filthy rich in the process while the rest of us wondered why everything feels so expensive and precarious. Cass comes from the right, which makes it even more interesting because this isn’t your usual anti-capitalism rant. It’s a really honest, funny, and weirdly bipartisan conversation that left me genuinely rethinking a lot of assumptions. Give it a shot.
Galleri: blood, sweat and false positives
Most people who know me know I’m pretty bullish on molecular biology and what it promises for the future of medicine. So this one hit different. Grail’s Galleri test was supposed to be the thing, a simple blood draw that could detect over 50 types of cancer before you even feel a thing. They ran it on 142,000 people in the UK over three years, the biggest trial of its kind ever, and it just didn’t work the way we hoped. It failed to meaningfully reduce late-stage cancer diagnoses, which was the whole point. The stock dropped 50% overnight, and honestly, the science community’s reaction was just as brutal. One cancer genetics expert basically said “end of story.” I still think the underlying biology is real and the direction is right, and there were some encouraging signals in the data. But this is a reminder that hope is not a clinical endpoint, and that even the most elegant molecular premise has to survive contact with a messy, complicated, 142,000-person reality. Worth reading if you follow this space.
Dario Amodei: the calm voice of existential dread
If you like your tech bros equal parts visionary and vaguely apocalyptic, this Dario Amodei piece is a tidy little panic attack. In one sitting you get “AI might cure cancer,” “AI might wipe out half of white‑collar work in five years,” and “also, we’re not totally sure my chatbot isn’t conscious,” all delivered in the soothing voice of someone insisting the turbulence is normal while casually mentioning the engines might fall off.
AI in medicine: the new junior doctor
AI is basically crashing the hospital party like an overachieving intern: it wants to write the notes, read the scans, clean up the paperwork, and whisper differential diagnoses in your ear, all so doctors can finally look patients in the eye again instead of worshipping the EMR. Assuming, of course, we manage not to turn medicine into one big, beautifully efficient, algorithmic bureaucracy with great throughput and terrible judgment.
Spotify: no-code, all hubris
Somewhere between a productivity fairytale and every junior dev’s worst nightmare, Spotify says its top engineers haven’t so much as poked a semicolon since December, instead barking orders into an internal AI called Honk while Claude Code dutifully ships features like a tireless intern who never asks for stock options or a standing desk.