• Nick Maggiulli: Optimizing Ourselves to Death. This is nothing new, of course. Modus omnibus in rebus was first written down some 2600 years ago and has withstood the test of time. It is good to have an occasional reminder.
• Joe Stone: A moment that changed me: I resolved to reduce my screen time – and it was a big mistake. Case in point to the above. Smartphones are magical, when you remember to tone them down.
• Dwarkesh Patel: Lessons from The Years of Lyndon Johnson by Robert Caro. Moderation in everything, even in moderation. Here are two examples of people, Johnson and Caro, who did not hold back on their own thing, with great success.

Robert Caro’s books are about formidable, single-mindedly devoted characters with storybook life arcs. It may be the case, then, that the only person who could write the biography of Robert Caro is the man himself.

• John Gruber: Gold, Frankincense, and Silicon. The amount of money and power one obtains in life are proportional to the size of frogs — or, if you are feeling less charitable, turds — one swallows throughout life, and at some point you either acquire a taste for frogs (maybe you’ve always liked them!), spend a lot of money on therapy and/or drugs, or drop out. And Tim Apple hasn’t dropped out just yet.
• John Gruber again: OpenAI Brings Back Legacy ChatGPT 4o Model in Response to Outcry From Users Who Find GPT-5 Emotionally Unsatisfying. To quote Gruber, “These people need help, and that help isn’t going to come from a chatbot.”
• Duncan McClements: The Sun Never Leaves. The subtitle is “How emigration ended the British Empire”, and it could not have happened to a nicer bunch of overindulgent cut-throats.

Happy reading.

• Why I Hate Your Podcast (Because it is unnatural, overproduced, and tries too hard to imitate Joe Rogan, for the most part.)
• Everything I Know about Self-Publishing (From Kevin Kelly. I am getting more serious about putting words to paper, and this is an invaluable resource.)
• Oreos Combined With Reese’s? Inside the Manhattan Project of Snacks. (Capitalism at its best, and most hilarious. Jerry Seinfeld should have made a movie about this.)
• A Good and Faithful Servant (A wonderful reminiscence on a recently deceased person the mere reading of which makes me want to be a better person.)

Enjoy.

• From the ever-excellent Kyla Scanlon: How AI, Healthcare, and Labubu Became the American Economy
• From a good friend, Timothée Olivier: Can I confidently say to patients that a 3-year structured exercise program may save their lives?
• From the recently-returned-to-blogging Thomas Basbøl: A simple test
• From the never-stopped-blogging Derek Lowe: In Praise of Weirdness
• From Dorothy L. Sayers, written in 1947, and touching upon everything above: The lost tools of learning

Happy Friday, etc.

• Razor, Gun, Fence and American, both from Kieran Healy
• The Paradox of India from Samir Varma (ᔥTyler Cowen (as a good example see Strategic Ambiguity as Wisdom: “Not everything needs to be defined, categorized, and resolved. Sometimes the wisest response is a head waggle that means yes, no, and maybe all at once.”. Indeed.)
• The growth and collapse of autonomy at work (after age forty, feelings of career autonomy apparently tend to decline; perhaps for some)
• Fortune Dark, a TTRPG system with an update here

• How to travel from Janan Ganesh (I agree with most, and the discerning reader should have no trouble identifying the minor point of disagreement)
• Journal articles are trying to do six things at once — no wonder they’re unreadable (yes!)
• Is atheism like a point null hypothesis? and other thoughts on religion (Tyler Cowen was correct)
• The Sacrifices We Choose to Make from Michael Nielsen (in the same vein as the above)
• The Herd of Independent Minds: Has the Avant-Garde Its Own Mass Culture? (written in 1948, but more and more herds of ever-more independent minds are still roaming)

• From the FT Editorial Board: The risk of letting AI do your thinking (the only nit I have to pick here is that by “AI” the esteemed Board means “LLMs” or, if you want to be kind and stretch the definition of intelligence, “generative AI”)
• From Dave Winer: AI should behave like a computer (see previous note)
• From a person on the Internet: Cognitive Hygiene: Why You Need to Make Thinking Hard Again (file under YouTube videos that should have been blog posts)
• From Michael Lopp: Every Single Human. Like. Always. (subtitle: The robots… They did the thing.)

And with these four links I hereby declare a moratorium on LLM-related matters on this blog, until further notice.

• Vinay Prasad Is a Bernie Sanders Acolyte in MAHA Drag (Apologies for linking to the WSJ. Please file it under “textbook hit piece”, that worked.)
• Zero-sum Thinking and the Labor Market (From Kyla Scanlon, whose substack blog is a jewel.)
• Design Your Own Rug! (We just came back from Istanbul, where rugs are no longer made by hand and the ones who are sold as such were made 40+ years ago and “refreshed. These are from Afghanistan and are “only” around $1K.)
• Belgrad Forest on Wikipedia (Speaking of Istanbul, the forest closest to it was named after the village next to it, “settled by thousands of Serbs who were deported to the capital Constantinople from the city of Belgrade in 1521” — to maintain a millennium-old Byzantine aqueduct. Please file it under “isn’t history fascinating”.)

• Breakneck: China’s Quest to Engineer the Future, an intro to this book that will make its way onto the pile as soon as it comes out
• Inclusivity In Healthcare Should Not Be Valued Above Our Paramount Mandate: First, Do No Harm (controversial?)
• New capitalism III: Capital from Branko Milanović, who should be on the short list for the “Nobel” in economics for this series of posts alone
• Things: The Surprising Power of Stuff That Exists is a brilliant description of every book Malcolm Gladwell and his lookalikes ever published

Note: While these link posts are usually untitled, this one is in reference to recent troubles at the Marginal Revolution blog. Isn’t HTML great?

Casey Handmer on LLMs:

Every time one of the labs releases an updated model I give it a thorough shakedown on physics, in the style of the oral examination that is still used in Europe and a few other places. Claude, Grok, Gemini, and GPT are all advancing by leaps and bounds on a wide variety of evals, some of which include rather advanced or technical questions in both math and science, including Physics Olympiad-style problems, or grad school qualifying exams.

And yet, none of these models would be able to pass the physicist Turing test. It’s not even a matter of knowledge, I know of reasonably talented middle schoolers with no specialized physics training who could reason and infer on some of these basic questions in a much more fluent and intuitive way.

Alexander the Great had Aristotle, some poor kid will have a brain-dead version of Wheatley.

(Casey’s post is deeper than simple LLM-trashing for he gives the actual 8-step process of reasoning through physics problems, so please do read the whole thing.)

In what feels like a troll but is in all likelihood completely serious, some parents are deciding to have their children fully immersed in AI LLMs:

We’re declaring bankruptcy on humans. Bring on the AI. In addition to integrating AI into as many facets of our lives as possible (our health, our work, our entertainment, and our personal lives), we’re designing an AI-integrated childhood for our kids—all while feeling like we’re helping them dodge a major bullet.

Did CS Lewis suspect, when he wrote The Abolition of Man, that the anti-human sentiment would be expressed as freely and overtly as the first sentence of this intellectually bankrupt paragraph? A paragraph that would be horrifying even if the AI it touts was actual intelligence, an AGI, but what these families are actually immersing themselves in is industrial-grade bullshit. As useful as bullshit can be — I hear it makes for great fertilizer! — one should not drink it as one would do Kool-Aid.