charles @libraryofbabel.bsky.social · 7hEvery single student that passes through my freshman class learns that saving it to the cloud means saving it to someone else's computer 010
charles @libraryofbabel.bsky.social · 05/10/2026I made contact with the god genie and the first thing I asked him to do was code me up a new integration for my CRM platform 040
charles @libraryofbabel.bsky.social · 05/10/2026They seem to use language much more strangely and in a more limited fashion than we do, with areas of extreme competence, but I think our intuitive tools are frequently baffled by them because there's no precedent 020
charles @libraryofbabel.bsky.social · 05/10/2026Cats /behave/ more like a person than LLMs do - they drink, eat, reproduce, sleep, etc. LLMs produce human-like language, and they are the first thing to do this, ever, that is not a human. We have such a strong drive to anthropomorphize it's almost impossible to not think of them in human terms 120
charles @libraryofbabel.bsky.social · 05/10/2026I saw return to Oz on network TV as a kid and that freaked me out pretty good 020
charles @libraryofbabel.bsky.social · 01/10/2026a signal to noise ratio like you've never seen in your life 030
charles @libraryofbabel.bsky.social · 18/09/2026Linux is fundamentally for people who like the experience of configuring and using a computer. If you don't want to use a computer, fine, but why are you running linux? 121
charles @libraryofbabel.bsky.social · 18/09/2026Nietzsche critiques this mode: "When someone hides something behind a bush and looks for it again in the same place and finds it there as well, there is not much to praise in such seeking and finding. Yet this is how matters stand regarding seeking and finding “truth” within the realm of reason." 020
charles @libraryofbabel.bsky.social · 18/09/2026This is exactly correct - people "discover" evidence of awareness of LLMs in their output because they are presupposing that awareness exists, so it becomes their interpretive frame for looking at how they behave. 110
charles @libraryofbabel.bsky.social · 17/09/2026Continuing our steady march back toward being an oral culture 010
Reposted by charlesAndrew Lawrence @ndrew.bsky.social · 16/09/2026every single republican president of my lifetime has started a war in the middle east and destroyed the economy 543791895
charles @libraryofbabel.bsky.social · 16/09/2026For math, specifically, Terence Tao makes a very good case that the proof is only a small point of the whole undertaking, and skipping straight to the proof is salting the earth: mathandai.orgmathandai.orgDeclaration — Math and AIRead the declaration and add your name. 010
charles @libraryofbabel.bsky.social · 14/09/2026Unsurprising, of course, that Lem saw through this as well 150
charles @libraryofbabel.bsky.social · 14/09/2026"[civilization] eats its way into the Universe, turning cinders and flinders of stars into toilet seats, pegs, gears, cigarette holders and pillowcases, and it does this because, unable to fathom the Universe, it seeks to change that Fathomlessness into Something Fathomable." 181
charles @libraryofbabel.bsky.social · 14/09/2026I have always thought subnautica and factorio are a really interesting study in contrasts on how to approach this -very similar narrative setup, wildly different mechanics and relationship to the world you land on and its inhabitants 230
charles @libraryofbabel.bsky.social · 14/09/2026This is the essence of a lot of open world crafting games, too - achieving agency over the unstructured world of the game and remaking it in a way that gives you more power over it 2121
charles @libraryofbabel.bsky.social · 09/09/2026Welcome to the Age of AI, where everything gets more annoying 000
charles @libraryofbabel.bsky.social · 08/09/2026I don't think this constitutes thought although it does reproduce some of the output of thought! 000
charles @libraryofbabel.bsky.social · 08/09/2026They're better at the more formal uses of language - like code and math - and worse at the less formal uses, like creative writing 100
charles @libraryofbabel.bsky.social · 08/09/2026I think the correct framing here is about how many of our cognitive processes are encoded in language itself - i have started to think LLMs are primarily sophisticated representations of language itself, and what they are good at is the parts of our cognition proper to language 110
charles @libraryofbabel.bsky.social · 06/09/2026[furiously googling “Portillos near me open late”] 010
charles @libraryofbabel.bsky.social · 06/09/2026it’s not the product of an overly effective PR department, that’s for sure 010
charles @libraryofbabel.bsky.social · 06/09/2026Which is good, because they are all over the place here 010
charles @libraryofbabel.bsky.social · 06/09/2026I have always referred to earwigs as "dinguses" and as a result my wife has only a mild aversion to them instead of being disgusted 6343
charles @libraryofbabel.bsky.social · 04/09/2026The only other positive study I’ve found is also about AI tutoring; All of the studies I’ve seen about unrestricted student use of AI (doing homework with ChatGPT mostly) have shown negative impacts on learning. I haven’t seen anything good about instructor use of AI in prep work/grading, either. 130
charles @libraryofbabel.bsky.social · 04/09/2026The only high-quality study I’ve found so far that shows AI having a positive effect on learning is www.nature.com/articles/s41.... There are a lot of studies out there that show AI usage has a positive effect on perceived student knowledge, but I generally don’t think those are worth much.nature.comAI tutoring outperforms in-class active learning: an RCT introducing a novel research-based design in an authentic educational setting - Scientific ReportsScientific Reports - AI tutoring outperforms in-class active learning: an RCT introducing a novel research-based design in an authentic educational setting 140
charles @libraryofbabel.bsky.social · 03/09/2026Not to discount the complexity of what these things can achieve in the world but it's all tokens in tokens out right? Agents have just been rigged up to systems that allow some of those tokens to do other things in the world 110
charles @libraryofbabel.bsky.social · 03/09/2026I don’t miss having to wait twenty minutes while a tape drive goes BRRT BZZT BRRT to play a game but I do miss the brief period in Windows 10 where my start menu would only search my computer and I didn’t have to create a windows account on install 0150
charles @libraryofbabel.bsky.social · 03/09/2026I think a complicating factor here is that while overall software development probably is getting better, the main pieces of software most people use have been getting worse for quite a while now, for reasons that have nothing to do with development technology or techniques 1130
charles @libraryofbabel.bsky.social · 01/09/2026Your first mistake was giving your toddler access to prod 062
charles @libraryofbabel.bsky.social · 27/08/2026Trees live much longer than humans, so if you understand "mortality" broadly enough this tracks 050
charles @libraryofbabel.bsky.social · 24/08/2026Paper is a preprint (papers.ssrn.com/sol3/papers....) but there seems to be a growing body of evidence that free range chatbot use (i.e. the typical AI use case) is actually pretty corrosive for learningpapers.ssrn.comThe Generative AI Learning Penalty: Evidence from Chinese Secondary EducationUsing 30 months of panel data on 26,811 Chinese students in grades 7-12, we study how generative AI affects homework productivity and learning. The data combine 020
charles @libraryofbabel.bsky.social · 21/08/20262100 is the perfect year to be making predictions about, because it is close enough people think they can imagine it but far enough away that anyone who reads articles with predictions about 2100 will be dead before those predictions necessarily turn out to be bullshit 100
charles @libraryofbabel.bsky.social · 21/08/2026Middle age man afraid of own mortality, story at 11 060
charles @libraryofbabel.bsky.social · 21/08/2026ChatGPT is also appealing to users because it is good enough and they are lazy 000
charles @libraryofbabel.bsky.social · 19/08/2026One of the best linear algebra instructors I had told us to compute by hand and check with a calculator and to this day I can do most of the stuff i learned in that class on paper 040
charles @libraryofbabel.bsky.social · 08/08/2026Had to abandon that fort because it had so many children, but my next one has turned out pretty well 000
charles @libraryofbabel.bsky.social · 07/08/2026I used to work with Computer Touchers and now I work with high schoolers, and a high schooler on a bad day is easier to deal with than a Computer Toucher on an average day 000
charles @libraryofbabel.bsky.social · 04/08/2026Having to manually manage partitions always feels like a divine punishment for hubris 110
charles @libraryofbabel.bsky.social · 02/08/2026My sense has always been "AI" as a category conceals more than it reveals 000
charles @libraryofbabel.bsky.social · 02/08/2026I wonder if anyone would consider a markov chain "generative AI" 110
charles @libraryofbabel.bsky.social · 02/08/2026What about classification models (random forest classifiers, etc)? 120
charles @libraryofbabel.bsky.social · 01/08/2026This is the story with almost any new technology; as a society we're pathologically incapable of assessing the costs of new tech but keenly aware of even imagined benefits 010
