Mid-February Reading Thread
Links
Have You Ever Thought About Ordinals Closed Under Addition?
Comment: Joel David Hamkins has, and it’s interesting.
Comment: I disagree with this more than I agree. Yes, people can get carried away “playing the game.” But lots of weird, interesting stuff gets published and there are significant rewards for being the kind of person who can get weird, interesting stuff into well-regarded journals. Most importantly: it’s a super old field, and if your “weird, interesting” idea about the meaning of life isn’t making contact with the ongoing conversation about the meaning of life, that’s a red flag in itself. These complains often read to me like “why isn’t Mind publishing more stuff in my AOS?”
Comment: this is very clear and well-argued. My primary takeaway is that the term “AGI” is pretty vague, and not all reasonable precisifications are going to correlate to certain things (like massive economic benefits or an intelligence explosion) that are often associated with the idea that someone has made AGI. To use a move the paper continually makes: humans have had general intelligence for many millennia without triggering massive economic growth (on short human time scales) or an intelligence explosion!
Comment: chalk up another for the Leopard/Face Award.
Federal Prosecutor’s Courtroom Breakdown
Comment: this is extraordinary.
Comment: a good complement to the Nature piece arguing that AI is AGI already.
Comment: some politicians just don’t want productive research centers.
From Dating App to Networking App?
Comment: “both the job and dating markets are terrible” in one headline.
The Unsung Potential of Rare Earth Recycling
Comment: worth trying.
A Lament for the Gutting of the Washington Post
Comment: our rich people are the worst.
DHS Buys Up Warehouses for Humans
Comment: mass deportation means camps where you concentrate the deportees.
Comment: it does not sound like these agents are backup drivers to me, but it does throw into sharp relief the need to clearly define terms like ‘autonomous.’
The Value of Being Vague and Right About AI
Comment: Stefan Schubert argues for less precise measures of what AI can do to get a better sense of progress.
What Can We Learn from Educational Improvements in the Deep South?
Comment: standards are underrated.
The NYT Makes a Case for Pot Taxes
Comment: I am in favor of vice taxes on many things, including pot.
Comment: it would be really complicated if he opposed it, and if he supported it he would just resign.
Comment: I usually find Nate Silver takes sensible. In particular, the arrogance of Silicon Valley has Ancien Regime vibes.
Are Abundance Political Interest Groups Hypocritical?
Comment: Jerusalem Demsas weighs in.
The FDA Won’t Even Look at mRNA Flu Vaccine
Comment: Moderna shoulda bribed Trump with crypto. My disgust with this administration and the people who support it grows ever deeper.
Kevin Dorst on Bayesian Gambler’s Fallacies
Comment: there are more of these “ackshually this mistake is kinda just good Bayesianism” than we’d like to admit out there. Exhibit A: Affirming the Consequent. You can also Dutch Book some conditionalizers.
Chapel Hill Asserts Right to Secretly Record Classes
Comment: it’s a bad look.
The Multiple Evolutions of Intelligence
Comment: fascinating how much more efficient the bird brains are.
The Role of Prop Bets in Problem Gambling
Comment: I like prop bets, but you need discipline that many people lack to know when to hold and when to fold.
Comment: enjoying this with a strong brewed tea.
Deutsch Bank Sending Moon Stocks… to the Moon?
Comment: someone is getting rich on moon stuff in the medium term. The trick is finding them.
The Weird World of Rapture Derivatives
Comment: God could do the funniest thing ever…
Remember when Grok Became a Sexual Harassment Bot? The EU Does…
Comment: xAI deserves a whale of a fine for that.
I find these threads of yours super helpful. Thanks for them!