@djl@mastodon.mit.edu

David, a Bostonian in Tokyo. @djl@mastodon.mit.edu

· 4mo ago

Replying to

@QasimRashid@mastodon.social "This is the scary socialism they tried to warn you about" Yep. Tax the rich. And tax their capital gains if they try to leave (and sell their local property). It's like, just common sense. One should tax the people and corporations that actually have money, not cut services for people who don't.

54

4

15

0

David, a Bostonian in Tokyo. @djl@mastodon.mit.edu

· 2mo ago

Replying to

@FlockOfCats@famichiki.jp It ain't the parenting that's the problem. It's the tax code that doesn't tax the filthy rich anywhere near enough. That $250,000 a year could provide day care for a lot of kids. djl@democratic.socialists.R.US.com

10

0

1

0

David, a Bostonian in Tokyo. @djl@mastodon.mit.edu

· 2mo ago

Replying to

@garrison@garrison.su It ain't just the Brits. I remember checking out the Egyptian stuff collection at the Boston Museum of Fine Arts and thinking "Sheesh, the Egyptians must be really p!ssed."

4

1

0

David, a Bostonian in Tokyo. @djl@mastodon.mit.edu

· 2mo ago

Replying to

@emilymbender@dair-community.social I disagree with you on some fine print (see my previous note on Turing), but you are a friggin' saint. And not just an ordinary friggin' saint. Anyone who can read through the current AI BS and deal with it is in the highest eschelon of friggin' saints. A friend sent me a video of Dario and someone else talking about AI, and it was the stupidest most inane sci-fi addled insanity I'd ever heard. And that was clear from the first 10 seconds. Whatever. Hang in there. And thanks.

5

0

1

0

David, a Bostonian in Tokyo. @djl@mastodon.mit.edu

· 1mo ago

Replying to

@VogueLaGalere@mastodon.online @gumnos@mastodon.bsd.cafe It's directions for removing AI slop. It's not dangerous, it's anti-dangerous. It's good advice, and correct safety behavior. If you've identified a code base (or anything else) as AI generated, then you should delete it immediately. For the simple reason that you don't have the expertise and/or time to verify that it's not wrong or dangerous. If you do have that expertise _AND_ time, and have invested said expertise _AND_ time, it's not AI slop any more.

1

0

David, a Bostonian in Tokyo. @djl@mastodon.mit.edu

· 5mo ago

Replying to

@ProPublica Paul Krugman is on the more general issue today: "MAGA Will Kill Many Americans" https://paulkrugman.substack.com/p/maga-will-kill-many-americans

paulkrugman.substack.com

MAGA Will Kill Many Americans

Greed, willful ignorance and mortality

17

0

9

0

David, a Bostonian in Tokyo. @djl@mastodon.mit.edu

· 5mo ago

Replying to

@screwlisp@gamerplus.org @tao@mathstodon.xyz @Ardubal@mastodon.xyz No. The automatic theorem proving blokes were writing programs that actually tried to do mathematics. LLMs are exactly and only random text generators. Since they have a lot of math proofs in their training data, they randomly mimic math proofs quite nicely. LLMs do not have a mechanism for checking their output. That's your job. Lean, on the other hand, is about computer representation of proofs; building libraries of valid proof steps. Serious work. Not a joke.

15

1

9

0

David, a Bostonian in Tokyo. @djl@mastodon.mit.edu

· 3mo ago

Replying to

@tante@tldr.nettime.org I'm expecting it all to go negative quite quickly. People are begining to realize how insanely expensive gen AI is. The AI twats figured out that each new model wasn't going to be enough better than the last if they didn't do something radical, so they invented looping the results back into the model (agents, math proofs, and the newer code generation stuff), which explodes the cost. And since it's all still based on random text generation, it still goes crazy 10% of the time.

6

0

1

0

David, a Bostonian in Tokyo. @djl@mastodon.mit.edu

· 2mo ago

I'd like to do some "graphics" programming, but a specific type. I want to display white dots on a black backround. Calculate a new set of dots, and spit them to the screen. Don't mind if I have to turn off old dots myself. SBCL would be optimal, Python OK, C/C++ no thanks city. Do any GUI toolkits support such a widget??? Any recommendations/experience doing this sort of thing??? #graphics #programming #lisp #python #gui

4

1

4

0

David, a Bostonian in Tokyo. @djl@mastodon.mit.edu

· 4mo ago

Replying to

@screwlisp@gamerplus.org @kentpitman@climatejustice.social @yantar92@fosstodon.org "fancy report ... versus having an ai ..." Two problems here: the cost of the ai is currently being massively subsidised and the LLMs can't not halucinate, so verifying that the ai report is accurate has costs. The AI stuff in a specific domain where it's possible to have a domain model works (Go, protein folding), but has to be redone for each domain. As someone who applied to college to study AI in 1971, neither LLM nor domain model games are AI.

6

2

4

0

David, a Bostonian in Tokyo. @djl@mastodon.mit.edu

· 2mo ago

Replying to

@FlockOfCats@famichiki.jp You just demonstrated that proofreading is somewhere between hard and impossible.

2

1

0

David, a Bostonian in Tokyo. @djl@mastodon.mit.edu

· 5mo ago

Replying to

@Ardubal@mastodon.xyz @tao@mathstodon.xyz Svante wrote: " I wouldn't say »proof abundance«, rather »claim abundance«." This is exactly correct. The underlying LLM technology can't do multiplication, let alone any serious math. But it is great at rehashing existing text into beautiful looking text, some of which is dead wrong. So the problem of erronious but beautiful looking text appears in every single "LLM application". At some point, we will all realize that this was a really stupid idea.

6

1

0

David, a Bostonian in Tokyo. @djl@mastodon.mit.edu

· 5mo ago

Replying to

@Natasha_Jay The problem with unboxing cats is that they automatically rebox themselves.

6

0

David, a Bostonian in Tokyo. @djl@mastodon.mit.edu

· 5mo ago

Japanese trainfans (train geeks/otaku) can be something else. (As in Kurosawa's first color film, Dodesukaden. (If you haven't, you need to see it.))

But we Americans have our entry in said competition. The Tanner YouTube channel: it's great fun if you are even a minor train geek.

Today's vid, "This is Japan's UNIQUE Monorail Built By a Hospital!" is hilarious. Especially the reason the monorail got built in the first place.

https://www.youtube.com/watch?v=y3xdRwJUDmc

#Japan #trains

3

0

David, a Bostonian in Tokyo. @djl@mastodon.mit.edu

· 5mo ago

TIL:

Wiki: "Davis earned a Ph.D. in clinical psychology from New York University in 1982. "

Art Davis was a seriously great bebop bass player, who worked with Thelonious Monk, John Coltrane, Dizzy Gillespie, McCoy Tyner and Max Roach.

Also with Art Blakey and Sonny Stitt.

https://www.youtube.com/watch?v=I3f6L0KM4hQ&list=PL-5ntqb8ArlfW8UbP_mXPdJauzGUUHRe-

#jazz #bebop #bass

Cafe

3

0

1

0

David, a Bostonian in Tokyo. @djl@mastodon.mit.edu

· 4mo ago

Replying to

@sue@glasgow.social I love all the articles that say "It only took Claude 14 minutes to do X and 20 to do Y", where there is absolutely no concept of the vast amount of money 14 or 20 minutes of AI datacenter compute would actually cost if you actually had to pay for it. LLM-based AI is stupid on a plethora of levels. I need to copyright this: "The LLM idea is the stupidest idea in the history of computer science."

2

0

David, a Bostonian in Tokyo. @djl@mastodon.mit.edu

· 4mo ago

Replying to

@tomatensson@nrw.social @emilymbender@dair-community.social "heavily "stochastically parroted" student papers. " I know it's radical and rude, but i really believe that the only way to evaluate students these days is to make all grading be based on in-class closed-book closed-device tests or writing. Grading copied homework and computer generated text ain't what you are being paid for. And isn't valid for the students, either. If the school wants to give courses on "Using AI", then that's a different question.

2

0

1

0

David, a Bostonian in Tokyo. @djl@mastodon.mit.edu

· 3mo ago

Replying to

@grwster@mastodon.social @gregeganSF@mathstodon.xyz "didn’t they also just lay off the editorial staff at Scientific American?" It appears not. Wiki reports that Springer Nature sold Sci Am to LabX Media Group earlier this month (June 2026), and it appears the editor (David M. Ewalt) remains in place. There may be more recent news, though, but it wouldn't be about Springer. I've not had much use for Sci Am since Martin Gardner retired, so haven't been following its travails personally.

1

0

David, a Bostonian in Tokyo. @djl@mastodon.mit.edu

· 5mo ago

Replying to

@adamshostack@infosec.exchange @screwlisp@gamerplus.org @tao@mathstodon.xyz @Ardubal@mastodon.xyz I'm not arguing that LLMs cough up _bad_ proofs, I'm pointing out that they cough up _unchecked_ proofs. Whether those unchecked proofs are useful food for thought, or a cesspool of random garbage, is in the eye of the beholder. But the bottom line is that they are randomly generated proofs. Thinking that they are more than that is a category error.

2

0

David, a Bostonian in Tokyo. @djl@mastodon.mit.edu

· 5mo ago

Replying to

@screwlisp@gamerplus.org @adamshostack@infosec.exchange @tao@mathstodon.xyz @Ardubal@mastodon.xyz I'm not a mathematician, just an AI nerd from the 1970s/80s. But I don't have a problem with using computers to enumerate cases. It should be possible to _prove_ that the computer enumeration is correct without actually looking at each case. I really think that that angst is way overblown.

2

0

David, a Bostonian in Tokyo. @djl@mastodon.mit.edu

· 4mo ago

Replying to

@ewen@social.ewenbell.com I resemble that. I had a gismo that holds film for taking pictures of it and screws into a macro lens. But finding a macro lens that works was a bear and I kept buying stepping rings and extension rings and more stepping rings. I finally broke down and bought a copy stand and a film holder with a backlight. So I've got more stepping rings than you do.

1

0

David, a Bostonian in Tokyo. @djl@mastodon.mit.edu

· 4mo ago

Replying to

@david_chisnall@infosec.exchange FWIW, Window NT was (and it's descendents remain) a serious OS. It was a reimplementation (with improvements) of the widely respected VMS. The Mac folks didn't have a protected mode multitasking OS for another ten years, when they finally ported a version of Unix. Prior to that port, MacOS was as bletcherous a disaster as all of the pre-NT Windowses. Nowadays, all the OSes are protected mode multitasking. so which one you use makes no difference.

1

0

David, a Bostonian in Tokyo. @djl@mastodon.mit.edu

· 4mo ago

Replying to

@screwlisp@gamerplus.org @kentpitman@climatejustice.social @yantar92@fosstodon.org My guess here is that the really large LLM stupidity is going to fizzle, and the game is going to be small models that either run locally, or run really cheaply in the cloud. Several monthes ago my friend Larry (hi Larry!) showed me a small model running on a friggin' iphone, and although I didn't admit it to him, I was blown away. Later a music YouTuber talked about this in detail, with small models specialized for what you are doing. This seems sensible.

1

0

David, a Bostonian in Tokyo. @djl@mastodon.mit.edu

· 4mo ago

Replying to

@Gargron@mastodon.social That's a wonderful cat. How could I not watch it??? (Oops. I've got to go off to get my shingles looked at. I'll try to watch it through the pain later today. As a medium term user, I don't know squat about the details of Mastodoning, so thanks!!!)

1

0

David, a Bostonian in Tokyo. @djl@mastodon.mit.edu

· 4mo ago

Replying to

@ewen@social.ewenbell.com Sorry, that's my _college_ (a small engineering school on a badly polluted river (love that dirty water)) reunion. Agreed on the high school reunion bit (and on the barely survived bit; were it not for IBM's generosity, I wouldn't have survived), but there are exceptions. My high school was given an IBM 1130 in 1970, so those of us who figured out programming were an interesting cohort (all three of us). The rest of my class, not so much.

1

0

David, a Bostonian in Tokyo. @djl@mastodon.mit.edu

· 4mo ago

Replying to

@ewen@social.ewenbell.com How about "Growing up in the Post-Hippy Years" I'm class of '76. A friend failed to graduate and went back in the mid 1980s. He reports that the 70s were laid back and the 80s were frantic competition for med and business school. No more learning for the fun of it. Hilariously, though, I'm finding my 50th reunion depressing. I spent the last 30 years hiding under a rock translating Japanese tech stuff, all my classmates started companies, ran research groups, were hot-stuff doctors.

1

0

David, a Bostonian in Tokyo. @djl@mastodon.mit.edu

· 4mo ago

Replying to

@ElenLeFoll@fediscience.org This is a rerun of the Sokal Hoax. https://en.wikipedia.org/wiki/Sokal_affair Like some here, some folks thought the Sokal hoax was ethically problematic, but the postmodernists needed a wake up call, as do the LLM fans. Really: the LLM idea is the stupidest* thing to come out of Computer Science ever. We need to be embarrassed. *: Unnecessary explanation: the idea that random text generation has something to do with intelligence is really really stupid.

en.wikipedia.org

1

0

David, a Bostonian in Tokyo. @djl@mastodon.mit.edu

· 4mo ago

Replying to

@ewen@social.ewenbell.com Yes. But. Like Guy L. Steele used to say "Going to Harvard is like banging your head against a wall: it feels so good when you stop." Here, I'm (trying to) read Jerry Fodor. My head hurts. I need to go out and juggle some clubs. Relatedly, if you ask an LLM for an answer, you may get that answer. But you aren't getting any closer to being able to answer the next question yourself. (Whereas if you read a Wiki page, you are getting closer.) But I still need to throw some clubs.

1

0

David, a Bostonian in Tokyo. @djl@mastodon.mit.edu

· 5mo ago

Replying to

@ewen@social.ewenbell.com FWIW, my problem with most camera eye-level viewfinders is that I can't see the whole screen comfortably/easily to compose through said finder. I have to scrunch to see the corners, even with the best of them. They're designed for folks who don't need glasses. My medium format film cameras ('53 'flex and '85 'blad (both cheaper than my dcam)) allow me to see what I'm getting without problem. Snap up the magnifier to focus, push it down to compose.

1

0

David, a Bostonian in Tokyo. @djl@mastodon.mit.edu

· 5mo ago

Replying to

@CatDragon@mastodon.world @zeldman@front-end.social " I am old enough to remember." Me too. https://pbase.com/davidjl/image/119895297

PBase

PeaceRally, Boston Common, April 1970 by David, a Bostonian in Tokyo

1

0

David, a Bostonian in Tokyo. @djl@mastodon.mit.edu

· 5mo ago

Replying to

@ewen@social.ewenbell.com "Folks don't realize just how good their cameras really are." This is exactly right. I have 12x16" prints taken with an IR-converted 6MP dSLR. That you can walk up to. 24MP is seriously wonderful. My current camera (60MP mirrorless) is insane overkill except for the bit that the viewfinder is the best viewfinder my horrifically nearsighted eyes have ever met. (Rude obnoxious rant about the insanity of Leica prices and the people who pay those prices goes here.)

1

0

David, a Bostonian in Tokyo. @djl@mastodon.mit.edu

· 5mo ago

Replying to

@kristinHenry

My two standard recommendations are:

The State House bulding on Beacon Hill. Seriously lovely, has guided tours.
Massachusetts Historical Society. Lovely Back Bay building, and they sometimes have fun things, like bag lunch talks by researchers. Check their web site.

Also, check Berklee for student or other concerts.

1

0

David, a Bostonian in Tokyo. @djl@mastodon.mit.edu

· 5mo ago

Replying to

@interlisp By 1972 the price of magentic core memory had come down to (drum roll) one million dollars per megabyte. And a megabyte took up the space of a large refrigerator. And it sometimes had parity errors. https://pbase.com/davidjl/image/174368482

pbase.com

1

0

David, a Bostonian in Tokyo. @djl@mastodon.mit.edu

· 5mo ago

Replying to

@FlockOfCats@famichiki.jp You mean like from the roof of a dorm some place???

1

0

David, a Bostonian in Tokyo. @djl@mastodon.mit.edu

· 5mo ago

Replying to

@elduvelle Back in the day (late 1960s, 1970s), I didn't like the computer-generated music/art of the day, but I thought that using computers as a tool in art/music was a valid thing to do. I think I was correct then. But that's dead wrong today. Sigh.

1

0

David, a Bostonian in Tokyo. @djl@mastodon.mit.edu

· 5mo ago

Replying to

@smeg Elephants are seriously kewl. I know Ringling Bros. was obnoxious, but when I saw the show (mid 1980s), the elephant trainer would stand on a teeter board, the elephant would run up to the other end, stomp on it with one foot, and the trainer would land on the elephant's shoulders. The elephant was grinning truly evilly as s/he ran up to and stomped the board. My mom, who grew up on an old New England farm, insisted that animals taught their trainers tricks. I think she was right.

1

0

David, a Bostonian in Tokyo. @djl@mastodon.mit.edu

· 6mo ago

Replying to

@quinn Yep. I have some horror stories from back in the day. The place: An AI conference. The speaker: A woman. Ms. S. The time: The Q&A session. Q: I'm not really happy with this approach. Do you know the paper by Ms. B? Ms. S: Yes, but this work improves on that. Q: No it doesn't. You clearly don't understand that paper. Ms. B. really got it . Ms. S.: I assure you, I'm well aware of that paper. Q: No, you clearly don't understand it. Ms. S: I understand that paper just fine. I wrote it.

1

0

David, a Bostonian in Tokyo. @djl@mastodon.mit.edu

· 4mo ago

Replying to

@heidilifeldman@mastodon.social Finally we're doing something about climate change. YAY!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!

0

1

0

David, a Bostonian in Tokyo. @djl@mastodon.mit.edu

· 5mo ago

Replying to

@lain_7 @mdhughes @screwlisp I wasn't into syntaxy languages in those days (still ain't), but I remember chatting with Ms. Liskov (sometime 1973-1975) about her PhD. work (chess endgames), and we agreed that the less computation an AI program used, the better the AI. She.Was.Ahead.Of.Her.Time.

0

2

1

0

David, a Bostonian in Tokyo. @djl@mastodon.mit.edu

· 5mo ago

Replying to

@chris You probably wouldn't want to try out this core plane... (Click "original" for some detail.) https://pbase.com/davidjl/image/174368482

pbase.com

0

1

0

David, a Bostonian in Tokyo. @djl@mastodon.mit.edu

· 5mo ago

Replying to

@AmenZwa@mathstodon.xyz LLM based AI is bogus. (Since it has no theory of (and thus no mechanism for) how to verify it's randomly generated text output.) And the US is self-sufficient in oil and gas, so the oil and gas companies are selling oil that costs them exactly the same as it did before the Iran war at Iran war prices, so of course their stocks are going up. So there's nothing odd happening here at all.

0

2

0

David, a Bostonian in Tokyo. @djl@mastodon.mit.edu

· 5mo ago

It's not all in-your-face bebop: Emmet also does pretty.

https://www.youtube.com/watch?v=dKcLLYYGg5s

#music #bossa #jazz

0

1

0

David, a Bostonian in Tokyo. @djl@mastodon.mit.edu

· 5mo ago

Replying to

@upmultimedia You called??? (Oops, I'm not in Melbourne. Similar time zone, though.) When Nat. Geo. did their first digital issue, I realized the writing was on the wall, and did my second round (the first was 1970) of MF film photography. Now on my third. I ended up doing 645, mostly using Provia 100. What I did for landscape photography was largely put the sun at my back, place the blue sky at +1 or +2 over medium gray and pray. A few shots from those days. https://pbase.com/davidjl/osakikamijima

PBase

Osaki Kamijima by David, a Bostonian in Tokyo

0

David, a Bostonian in Tokyo. @djl@mastodon.mit.edu

· 5mo ago

Replying to

@ketan It's OK. The whole "New AI Datacenter" thing simply doesn't work. By the time you get the new datacenter physical plant actually built, the GPU chips you bought for it are last generation and hopelessly out of date. And you can't just refurbish an existing AI datacenter with new chips, since the new chips require about 6 times the power and 6 times the cooling of the old chips, and you can't provide either in an existing physical plant. The stupidity of current AI is beyond belief.

0

1

0

David, a Bostonian in Tokyo. @djl@mastodon.mit.edu

· 4mo ago

Replying to

@scy@chaos.social I resemble that. Except for the 60s. Ouch.

0

David, a Bostonian in Tokyo. @djl@mastodon.mit.edu

· 5mo ago

Replying to

@mpe "even if it's told not to, " ... "despite using specific instructions" LLMs have no theory of, or mechanism for, relating text to "meaning". That is, "giving instructions to" an LLM is completely meaningless. They couldn't care, even if they wanted to. It's all random text all the way down, from start to finish. That the random text generation stuff often randomly happens to do the right thing is no guarantee of anything. It's the most unreliable technology in human history.

0

David, a Bostonian in Tokyo. @djl@mastodon.mit.edu

· 5mo ago

Replying to

@dlakelan My laptop is my music server, and drives a tiny Fossi Bluetooth amp that drives some decent speakers (including a subwoofer). Bebop heaven. Inversely, the cheap Sony BT headphones (WH-CH520) do the clarity thing for slightly larger groups. I wasn't expecting much from them but it's quite fun to be able to hear each voice clearly in music I normally listen to as a shmushed-up gestalt.

0

David, a Bostonian in Tokyo. @djl@mastodon.mit.edu

· 2mo ago

Replying to

@dcuthbert@defcon.social @thedarktangent@defcon.social For a plethora of reasons, incluing making our planet unfit for human (or other) life, the private car is a rather bad idea. So I hope that the spying on you bit will be the last straw and people will start figuring out how to do without. Neither my father nor I ever bought a car. It really is possible.

0

1

0

David, a Bostonian in Tokyo. @djl@mastodon.mit.edu

· 5mo ago

Ron Carter friggin' WALKS!!!!

https://www.youtube.com/watch?v=UenXvLuCRjg

#jazz #bebop #monk

0

1

0

David, a Bostonian in Tokyo. @djl@mastodon.mit.edu

· 5mo ago

Replying to

@Cys@ohai.social @mattblaze@federate.social My memory has it that the ISO 50 Polaroid negative film was actually quite good; competitive quality with the films of the time. My memory also has it that they never made it for the SX-70. Since I had a 6x6 enlarger*, I might have been more interested in the SX-70 had they done so. *: Seen here in the background around 1970. https://pbase.com/davidjl/image/170624197

pbase.com

0

2

0