Timely
Mistaking Politics For Religion - I’ve been a huge Martin Gurri fan since I read the Revolt of the Public (see below). Gurri outlines how much zealotry is in today’s politics, reminds us that we don’t live in a democracy, and defines the aptly named Thunberg Syndrome for those who find meaning in shouting and feeling superior.
How To Succeed In Mr. Beast Production - Mr. Beast is a perfectionist and you get to see just how much this is true in the onboarding document for his production company. He defines the YouTube metrics he cares about (and why) in excruciating detail and explains what responsibility really means. There’s a lot to learn about teams, psychology, and what it takes to be the best. (And here’s the doc itself.)
What The AI Debate Is Really About - Matt Yglesias is one of my favorite left-leaning writers and here it seems like he’s been AGI-pilled by none other than Leopold Aschenbrenner? A good mainstream take that moves beyond the doomer cult.
Terry Tao uses O1 - This is the most impressive feedback I’ve seen yet for pretty much any model ever. Tao is one of the world’s leading mathematicians. His experiences with O1 suggest that it’s equivalent to a “a mediocre, but not completely incompetent […] graduate student” (vs the previous experience which was worthless).
In addition to this appearing to be a step change in capability, it’s an interesting commentary on how we need to evaluate the models at this point. As Ethan Mollick points out: “Note o1 is showing how unprepared we are for testing high level AIs to figure out what they are good or bad at. Instead we are turning to experts, like one of the greatest living mathematicians, to give it a vibe check”
Timeless
Sorry Ted, Humans Aren’t Very Original Either - Many humans are currently looking at LLMs and AI output and studiously ignoring the mirror-like quality they have as a tool for understanding ourselves. The term “stochastic parrots” was coined to think about LLM outputs but it describes us humans just as much. Erik Hoel makes me feel much better about that.
Burkean Longtermism - Burke keeps coming up as something I need to go read. I am totally on board with this optimistic, conservative, nuanced view of long term thinking with room for change and evolution.
We Don’t Sell Saddles Here - A properly famous quote of Antoine de Saint-Exupery is “If you want to build a ship, don’t drum up the men to gather wood, divide the work, and give orders. Instead, teach them to yearn for the vast and endless sea.” This is that quote wrapped in the hype and economics of 21st century tech.
What You’ll Wish You’d Known - In honor of all the hype around Founder Mode (aww come on, you missed that? Don’t worry you didn’t miss much besides a false binary distinction), here’s one of my favorite old Paul Graham essays. Actually, a talk he was supposed to give to high school students. It is not just profoundly good advice for smart, ambition 17 year olds but for practically any stage of life.
Books
Revolt of the Public by Martin Gurri - First let’s just say that the cover and typography in this book are beautiful. Stripe Press is high quality. And the arguments are high quality. Gurri explains how the information firehose we now sit constantly in has eroded the nature of authority and our systems. He talks about the Arab Spring and the rise of populism and it’s all still just as relevant today as it was 10 years ago.
How To Win An Election by Cicero - Think it’s changed much since Rome? Not really! The basics are still the same and this is a succinct outline for the necessary tactics.
The Mom Test by Rob Fitzpatrick - Happy birthday Mom! This book is why I never ask you business type questions, because you like me too much to answer in the negative. This book gives great advice on how to craft questions so that you get actionable responses rather than just bland, feel good lip service.
Tweets
Watch this young PhD get his mind blown when ChatGPT o1 is able to reproduce the code based on the Methods section of his PhD thesis.
Cheers!