Sascha Fast of the Zettelkasten blog writes, in a post titled The Scam Called “You Don’t Have to Remember Anything”:

Rowlands et al. wrote about the so called “digital natives” that they lack the critical and analytical thinking skills to evaluate the information they find on the internet. We need a fully developed mental map of the subject in order to derive value from the results of an internet search.

In short: You need a trained brain to actually benefit from the internet.

But not just from the internet, as the post elaborates. This applies equally or even more to LLM outputs. A great example comes from a recent post on Andrew Gelman’s blog, though not from the man himself, where a human and ChatGPT 5 both try to improve upon a statistical model in a new-to-me language called Stan. Now if you don’t know Bayesian statistics or Stan this will all look like gobbledygook and ChatGPT won’t help you understand.

LLMs are also seeping into the everything-bucket software, the one whose primary purpose is to black-hole every article and textbook you will never read or video you will ever watch. Well now it can also give you the illusion of knowledge and control because you can ask questions about the contents. This is something Casey Newton learned this year:

I can give Notion a sprawling question like “how did the Cambridge Analytica case resolve” and get a good summary of regulatory actions across several years and countries. And by default, web search is off, meaning I know that its AI systems are drawing only on the vetted journalism that I have saved into my database.

This is a dream come true. I finally have a meaningful way to sift through millions of words of article text, ask follow-up questions, and get citations that I can use in my work. Notion may yet prove to be the AI librarian that Readwise never became.

One more thing I’m trying: I mentioned above that I continue to experiment with different ways to save material that might be useful later. Recently a Reddit post turned me on to Recall, which positions itself as a “self-organizing knowledge base.” Currently available as a web and mobile app, Recall lets you save web pages, YouTube videos, PDFs, podcasts, Google Docs, and other materials into a single database that it then organizes on your behalf.

To be clear, I myself have asked for something like this from my everything-bucket software of choice, DEVONthink. And it delivered! But now I am realizing, and Sacha’s post was a good reminder, that these are becoming command line-level power tools — Hole Hawgs of the mind if you will — which can and will do great damage if not used carefully. And unlike the Hole Hawg they are freely available and come with no instruction manual. Caveat utilitor.

I am writing this from an Apple Vision Pro, VisionOS 26 Beta 8, no additional hardware, just pecking away at the virtual keyboard. A few thoughts:

• I am convinced that no person younger than 40 was involved in the decision-making during its development.
• The basis for this is that the pass-through looks better each time I use it, not because the camera, screens or software have improved but because I am developing presbyiopia (near-sightedness of old age), and rapidly.
• This is also why I suspect there are still no good VR-first games or even a proper controller.
• Why can’t I stream a PS5 game onto a huge screen? I thought Apple and Sony were friendy.
• The carrying case is too bulky for short work trips, which would have been my primary use.
• They had 18 months to work on the virtual typing experience and it is still a nightmare by default.
• For a tolerable experience I suggest enlarging it to maximum size, posiyioning it so that you can use both index fingers with elbows supported, then imagine you are using an old typewritter with sticky keys that are prone to jamming.
• Still, no regrets.

• Bart’s Watch Stories on YouTube: How Casio Made the Most Sold Watch Ever. It is about Casio F-91W, my default watch, and I can attest to its — excuse the pun — timelessness.
• @StatisticUrban on X: What states do Americans most approve of? DC is the lowest at +1 net approval.
• Ruxandra Teslo: Your newborn is not Hepatitis B vaccinated because of wokeness. See also her exchange on X with Curtis Yarvin, aka Mencius Moldbug, aka Mr Head-Up-Ass who loves the sound of his one voice even as he recommends disastrous remedies for obvious problems.
• Marioun Fourcade and Kieran Healy: Authenticate thyself. The subheading is “The sovereign individual and the paradox of the digital age”. It is more highfalutin than what I usually read — “The combination of epistemological self-centredness and hyperconnectivity makes people susceptible to diffuse forms of ‘supersense’-making (to borrow a term from Hannah Arendt).”, reads one sentence — but the essay does some important sense-making of its own:

As the deployment of digital technologies continues to generate ever-more stratospheric concentrations of wealth, the masses sink deeper into the void left by the evisceration of social solidarity and the rise of automation. The often-missed point about sovereign individuals is that not everyone gets to be one. But everyone should aspire to be one, and in the meantime follow one, as they walk down the road to selfdom.

Worth reading for that last sentence alone.

This week was a good one for learning about new apps:

• Zen browser puts an Arc skin over a Firefox core. It has already become my default browser. (ᔥBen Werdmuller)
• Macrowave enables AV streaming to anyone. I do plan on using it to stream but for now I am just marveling at the WinAmp esthetics. (ᔥGus Mueller)
• Inksightful digitizes paper notes. I am on the waitlist for the beta. (ᔥThe Iconfactory)

On the topic of apps, here are some fairly new ones (i.e. I have been using them for less than 5 years) that are passing the test of time, with honors:

• Tot, which is still my default for digital note-taking on the both the Mac and on iOS
• Flighty, which is invaluable for those flying in the US. Alas, it is not as useful, I have been finding out this summer, for travel in Europe. Those Europeans are not very good at updating their systems.
• The Cardhop/Fantastical combo, without which I could not imagine juggling between four different contact lists and a half-dozen calendars.
• MarsEdit, which I missed out on listing initially even as I was using it to write this very post. It is that essential that I consider it a part of the OS at this point.

And the apps that make me stick to the Mac even though I entertain from time to time the possibility of switching over to Linux:

• MailMate (since 2013)
• DEVONthink (since 2015)
• OmniFocus (since 2016)
• Tinderbox (since 2017)
• Audio Hijack and Sound Source (since 2020)

As much as I like the Mac hardware and the sleek aluminum esthetics, it is the software listed above that keeps me in the ecosystem. None of it is made by Apple.

The most recent issue of the FT Weekend Magazine is about games of all kinds, but the highlight is a massive article about the tragedy of Disco Elysium. It is depressing throughout, with a glimmer of hope buried near the end:

Kurvitz is making his next game at a new studio, Red Info, with Aleksander Rostov, Helen Hindpere and Chris Avellone, lead writer of the 1999 video game Planescape: Torment, a huge influence on Disco. “[Kurvitz] felt that Disco was the project in his head, and once he was cut off from the franchise, he was worried he didn’t have any other ideas in him,” Avellone told me. “I felt that was bullshit . . . Robert’s too creative to simply ‘not’ create something or rely on a single world idea in his head.”

Creators of Planescape: Torment and Disco Elysium working together on a new game? Be still, my heart.

• 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.

For a contrarian take on LLMs as intelligent machines, here is Alexey Guzey saying that:

• ChatGPT understands what it reads
• LLMs are creative
• Compression is intelligence and ChatGPT compresses really well
• The idea of AGI is stupid
• It doesn’t matter if AGI is real or not

I remain dubious.

An important note from Dave Winer:

I say ChatGPT instead of “AI” because I’m not comfortable characterizing it as intelligence. Deeper you get into it you learn that these beings whatever they are have serious character flaws that are counter-intelligent.

Exactly. LLMs are closer in intelligence to a screwdriver than to a human.

Flighty does not seem to be as up-to-date traveling internationally as it is on domestic flights. The IST airport departures board had our flight listed as delayed as soon as we got there, yet the app thought everything was fine. Trust no one.

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.)