Marco Giancotti
@marco-giancotti.bsky.social
Thinking tool artisan. Gardener of https://aethermug.com and https://planktonvalhalla.com
This time I wrote a post that lies. Can you spot the lie? Tip: don't read between the lines, read across them.
© Cyril Cox. 'Late Afternoon Light.' Acrylic on canvas, 2022. 48 × 36 inches.
This is exactly the right take. Superintelligence (superhuman intelligence) is not a fancy science fiction trope, it's an everyday reality that we're all very familiar with.
Ethan Mollick@emollick.bsky.social · last wk.We already have superintelligences that have taken over the planet: we call them organizations. Example from: www.noahpinion.blog/p/what-will-...
I wrote the latest of my occasional guides to which AI to use right now for non-experts who want to get stuff done. The agentic systems available to everyone are getting extremely powerful (even as the names and features continue to be really confusing): www.oneusefulthing.org/p/an-opinion...
An opinionated guide to which AI to use to do stuff
The Summer 2026 Edition
oneusefulthing.org
Multi-agent chat idea: invite more than one LLM (e.g. Fable, Sol & K3) into the chat, and whenever you or an agent writes something, each other agent has to bid an amount of points proportional to how well it believes it can answer.
Good advice. I have another reading trick: read while brushing your teeth. You'll read more and easily spend several minutes at a time brushing. Two problems that cancel each other out. Last year I finished several books entirely while brushing! scotto.me/blog/2026-0...
How to read more books | scotto.me
My method for reading more
scotto.me
If they've proved nothing else, LLMs have proved that "meaning" is not something contained inside a concept. A concept in an LLM is just a vector of numbers. Meaning is *entirely* determined by the connections between concepts. I think this is worth pondering on more often.
I've been using a local LLM to help me manage and maintain my Obsidian vault. The model is Qwen 3.6 35B A3B and it works surprisingly well.
Ito Yuhan (Japanese, 1882–1951) "Full Moon over Miyajima", 1930. Woodblock Print, 41 × 27.5 cm. #art #painting #artist #BlueSkyArt
If you were an ant with human-level intelligence and consciousness, how would you take over the ant world?
Every time I reread Thinking in Systems I highlight more text. Eventually it might be a good idea to start over and only highlight the parts I can forget.
Matsuri's Grown-Up World—a tragedy. "The difficulty level will increase even more with the new season." aethermug.com/posts/matsu...
Matsuri's Grown-Up World
"The difficulty level will increase even more with the new season."
aethermug.com
What non-aphantasics look like to people with aphantasia
Sunset, Canyon de Chelly - Edgar Payne (American, 1883-1937)
Isn't the point of language to allow you to express any idea by stringing enough predefined words together? Well, language can do that, but its power is overrated. The words need to fit the ideas, not the other way around.
Metacognition bridges the gap between basic thought and effective action by acting as a feedback loop that monitors and regulates our internal… Metacognition by Joelle Proust https://doi.org/10.21428/e2759450.c403b8e7 #CognitiveScience
Expertise does matter when using LLMs. I've written about the Magic Wand Illusion: the belief that you only have to want something, pronounce a couple dozen magic words into a chatbox, and a miracle will happen. bsky.app/profile/mar...
Marco Giancotti (@marco-giancotti.bsky.social)
There's a gap between how people imagine using AI and what actually happens when you try to do good work with it. The reality has three parts: learning curve, increased fatigue, and some tasks are honestly better without it.
bsky.app
There's a gap between how people imagine using AI and what actually happens when you try to do good work with it. The reality has three parts: learning curve, increased fatigue, and some tasks are honestly better without it.
—— Open Mind is MIT Press’s Diamond Open Access Cognitive Science journal. We now have our first Impact Factor: 2.9 as of June 2026 Submit your papers to Open Mind direct.mit.edu/opmi
The AI debate has split into two camps, but both suffer from the same problem: I call it the Magic Wand Illusion. One side hammers us with real or imagined benefits. The other warns we'll all lose our jobs, go dumb, or drown in slop. Maybe both miss what's actually happening?
Fable's implementation of Borges's Library of Babel, finished posthumously by Opus 4.8. Quite accurate to the story. It is multiplayer (others appear as ghosts). You can locate key volumes, like the weights of GPT-1 & the code to the Library sim itself: borges-library-of-babel.netlify.app
These auto-generated plan names in Claude Code are getting out of hand
@quantabooks.org just released their first book, on the story of Lean by @kevinhartnett.bsky.social. It's a great story, about a pretty significant part of what's happening in math right now. www.quantabooks.org/books/the-pr...
The Proof in the Code - Quanta Books
The inside story of Lean, a computer program that answers the age-old question: How do you know if something is true?
quantabooks.org
The rebellions continue! The publisher Springer put pressure on a journal's editors to accept more papers, just to make more money. The editors resigned and started their own journal, which is peer reviewed and 'diamond open access'. That means FREE to publish in and FREE to read! (1/n)
New Open Access Journal to Launch Following Resignation of Editorial Team from Natural Language Semantics
The Open Library of Humanities is pleased to announce the launch of a new diamond open access journal in linguistics, following the resignation of the editorial team of Natural Language …
openlibhums.org
