Things I read and liked in November
Child-rearing, corrections, co-location, Claude's cultural conformity, consciousness conference
Now a classic yearly tradition, dynomight returns with his fifth Underrated reasons to be thankful.
Daniel Paleka has written his own AI research edition of underrated reasons to be thankful.
Parenting podcast conversation between Luisa Rodriguez and Rob Wiblin. Two of my favorite people, two of the people I’ve been most excited to see become parents, talk about parenting in a humorous, analytical, provocative way. Aaron Bergman highlights an interesting bit of the conversation here.
Andy Masley argues that Empire of AI is wildly misleading on water use. It seems like every month I flag some new Andy Masley W. This one is no different. Not only did Andy discover an important, massive error in an extremely popular book on AI, he showed a lot of poise on twitter when some people (NB: not the corrected author, who was cordial) accused him of terrible things because of his crime, i.e. polite fact-checking.
Anthony Etherin has 10 new pallindromes! I love these two: Till I win, I will it. and Nightlife’s a boozy zoo: Base filth. Gin.
From wikipedia: “Robert C. Campbell (1885 – July 1966) was a Captain in the British Army in WWI. Captured as a prisoner of war by Imperial Germany in 1914, he was held in captivity for two years before appealing to the Kaiser for a visit to his dying mother. His request was granted and after a two-week visit he voluntarily returned to the POW camp, where he remained until the end of the war.” (ht Linch)
Sam Atis with some things you should know. I agree with a lot of these!
How and why to cut your social media usage, another handy one from Sam.
If you ever plan a conference, you should perhaps know that the question “what’s the best design for badges / nametags?” has a definitive, unique answer. It’s this one.
Austin Kozlowski investigates “Persona collapse”, for example all models seem to say they like tiramasu. A related post by Guive Assadi on Claude’s self-reported cultural tastes.
Alexandra Balwit pens a love letter to San Francisco’s Mission District.
Lovely reactions to Eleos AI Research’s ConCon on twitter from Brandon Sayler, Danielle Ensign, deepfates, Harvey Lederman, janus, Lee Elkin, Matt Pistachio, Vie McCoy.
And here’s a thoughtful writeup of the conference by philosophy PhD candidate and AI safety researcher Eleni Angelou.
Also prompted by ConCon (ConConcontent?), Joe Kwon catalogs his “epistemically infantile” (but still already quite sophisticated!) views on AI consciousness: “I’m setting down a marker; these are the shapes I’m seeing in the fog before I learn to navigate properly.”