Hello from Palo Alto! I’m back in the states from a lovely trip in France, right on time for some wonderful weather both in New York and now here in California. I hope you enjoyed last week’s thoughts, here’s part two.
📑 February 14th, 2025
What I love so much about Severance is how the show embeds written language as a central theme. Writing is used to paint a detailed universe (such as Lumon’s Compliance Handbook) and even drive the show’s plot (Ricken’s The You You Are). I was pleased to see language return prominently in Severance season two.
Since much of this second installment was likely written in the aftermath of the SAG-AFTRA strikes, I’d assume a lot of it is responding to what’s been happening to content in the age of AI. I love how the show pokes fun at optimized language, highlighting how uncanny manipulative speech can be. The story of the Gråkappan reminded me of a typically unsettling AI slop LinkedIn post.
When Seth Milchick is told off for using big words, it reminded me that prompt engineering is the natural next step for the language-industrial complex of corporations and social media that have turned writing into a society-wide competition for control. The moment underscored the micromanagerial overreach of being told how to use language to reach ends (like when society tells us to engineer something even as personal as a caption for a life update post: “use emojis and hashtags!”).
All I can say is: watch season two of Severance, if only as a small act of resistance against the weaponization of writing in modern life.
🤖 February 16th, 2025
“Get in on AI” is terrible advice. It’s like saying “get in on math and engineering,” “get in on science,” or “get in on technology.” You’ve done nothing more than point out something that will obviously shape the future.
It’s plain to see that those reaping the biggest rewards of AI are either top engineers and business leaders who were “in on AI” since the beginning, or AI grifters looking for an easy buck. Either it’s folks who shaped their entire studies and career around developing AI, or it’s folks who found a clever way to get people to pay for AI slop that’s worth pennies to create. So until we invent time travel, the “get in on AI” advice is essentially encouraging everyone not “in on AI” to discover new scams.
Meanwhile, chatbot use in everyday work is demonstrably leading to cognitive atrophy and reduced critical thinking skills.
Here’s my take: if you have a platform, you’re not only in a position to coordinate collective behavior to benefit society, you have a responsibility to do so. Maybe using AI does make sense for your audience at an individual level – speeding up tasks at the small cost of getting rusty at things that were boring anyways. But what about the ways in which everyone using AI also makes the game worse for everyone? To name a few, AI use at massive scale causes climate change, amplifies echo chambers, accelerates job loss, and feeds a dangerous arms race. If you have an audience, why dole out generic “get in on AI” advice, when instead you could call for resistance?
🚽 February 18th, 2025
I wonder how much of an inspiration the legendary Toilet Boardroom Surprise Prank was for Severance. After all, the first scene in the entire show is one of a startled woman trapped in a board room with no exit, with no idea as to how she got there.
🧑⚖️ February 20th, 2025
I don’t like when people brag about never liking the formerly beloved. “Yeah, I never liked Michael Jackson.” That doesn’t make you a better person. It’s kind of like how a justified true belief does not equal knowledge. While longtime haters are 100 percent right to dislike the King of Pop right now, there’s still a good chance they disliked him back then for all the wrong reasons: racism, normopathy, or bad taste in music.
Is the complexity of morality unbounded? Whether or not it is has implications for AI alignment. If moral decision-making can be solved, then someday small language models could form juries we can trust, with no qualms to be had. But if morality is infinitely complex, then it is perhaps immoral of us to not eventually cede power to higher beings.
📚 February 23rd, 2025
Does anyone else agree that the word stereotype sounds like it should describe something way cooler than what it actually describes?
One of ChatGPT’s more intriguing, if overlooked, features is that it records memories of your previous chats. This has enabled trends where people ask it “What do you know about me that I don’t know about myself?” In that vein, here’s a prompt to heal society: “Chat Gippity, please tell me what I'm least likely to have read by now, given who I am.” Whatever it answers is what we should each be reading next.
🏇 March 4th, 2025
The past four months have been a journey in self-discovery, as I slowly found my way back into a career path I’m excited about. Among the many things I learned along the way is that self-doubt is self-fulfilling. Whatever story you have about yourself, even if it isn’t true, is much likelier to be made true simply by believing it. That goes both ways: self-confidence is also self-fulfilling.
We currently live in the Age of Artifice: one where money is invented, leisure is virtual, and intelligence can be faked. Next will be the Centaur Age (see centaur chess): one where the most successful individuals will be artificial intelligences assisted by humans. The age after that one will be the Posthuman Age. Let’s take pains to make The Centaur Age, when it inevitably arrives, last as long as possible.
🦿 March 7th, 2025
Millennials are the only generation to keenly understand both the social benefits of the Internet and the undeniable harms it has effected. Our childhood years straddled both the Before and the After. We are simultaneously digitally literate and digital skeptics. As such, we carry a great responsibility to advocate for better technology.
Screens have become a double-edged sword for new parents: a great way to find peace by keeping their children entertained, but one that could negatively impact their academic performance down the line. So how do you reliably entertain children while keeping their attention on the real world? I believe there’s a technological solution: higher-quality animatronics. Chuck-E-Cheese was on the right track, they just needed to partner with Boston Dynamics.
💰 March 25th, 2025
Bitcoin’s skyrocketing price is hardly a good thing. It accentuates inequality: the top 2% of addresses own 90% of the currency. It makes it easier for fraudsters to conduct scams by making it easier for money to change hands. It emits carbon. And last but not least, it erodes the sovereignty of nations. Why aren’t more people decrying the recent bull run?
🗓️ April 5th, 2025
Many calendars don’t align with the Gregorian calendar. The Islamic calendar is purely lunar, while the Hebrew and Chinese calendars are lunisolar. This means that your birthday in one calendar could land on a completely different day in the others from year to year.
One could claim an all-American misaligned birthday if one was born on Thanksgiving Day, Election Day, or President’s Day, as those fall on different days each year.
The misaligned birthday I wish I had: St. Totteringham's Day, which is celebrated almost every year and can vary in timing by weeks to months. It’s the day on which it becomes mathematically impossible for Tottenham to finish above Arsenal in the final rankings. Sadly, there was none the year I was born…
🪷 April 7th, 2025
I was a bit put off by the amount of profit-seeking promotion there was around the latest season of White Lotus. Banana Republic partnered with HBO for a fashion line promotion. The Four Seasons launched an absurdly expensive White Lotus travel package. HBO even made an official podcast for the show.
That’s not to say I didn’t thoroughly enjoy the show when I reluctantly did watch it. Mike White is a genius, and I don’t completely disagree with some who have called the latest season a great work of art (though I wouldn’t go quite that far).
I suppose the modern tendency to require vast amounts of money to justify good art is not dissimilar from how masterpieces like the Sistine Chapel were once commissioned. In the end, maybe it’s not a bad thing that the light of good, thought-provoking writing can reach even the most wretched of binge watchers.
🏛️ April 12th, 2025
I always thought the Library of Alexandria burned in one massive conflagration, but I recently learned it was gradually destroyed over the centuries by the three worst enemies of knowledge: belligerent dictators, angry mobs, and intolerant religions. How fitting. (And yet another reason to remember Julius Caesar less favorably)
🚅 April 17th, 2025
At time of writing, Ferrari NV has a market cap of $86.8 billion, and a respectable total of 300,000 cars on the road (that admittedly go very fast). Meanwhile, East Japan Railway has a market cap of $24.2 billion, and a daily ridership of 16 million people, who will often travel much further, much faster, and certainly much cleaner than in a Ferrari. To me, that’s one of the more eloquent examples of why maximizing shareholder value is a far suboptimal way to unlock the best outcomes for society.
🍇 April 18th, 2025
As AI companies continue to compete in their claims to offer the best AI, it’s important to remember that there’s a lot more low hanging fruit in the realm of clever trickery than there is in the realm of true innovation.
🐺 April 22nd, 2025
The New Yorker’s piece on Colossal Biosciences is excellent. It’s frustrating to me how social media coverage of their claimed “de-extinction” of a “dire-wolf” makes almost no mention of how their specimen was made with only twenty gene edits to a total of fourteen genes (among tens of thousands). Much like with charismatic megafauna, only focusing on the most charismatic genes seems like a half-assed way to go about fighting extinction.
And with that, see y’all next Sunday!