Nostr notes by allison

i don't want my software to ship fast. i want it to ship ...

2026-10-26T17:09:18Z

i don't want my software to ship fast. i want it to ship luxuriously, languorously, utterly unhurried. i'm talking glacial, geological timescales here. i want software that has settled into the hammock for a long weekend

i think what we'll actually see is these folk decomps etc ...

2026-10-20T15:55:12Z

In reply to nevent1q…97g7
_________________________

i think what we'll actually see is these folk decomps etc becoming a real source of widespread malware and other exploits, and that will act as an excuse to pass laws that make it effectively illegal to sell CPUs/microcontrollers that can run arbitrary user-supplied code

i don't envy much from the stereotype of the old tyme ...

2026-10-20T00:43:04Z

In reply to nevent1q…depn
_________________________

i don't envy much from the stereotype of the old tyme wholesome educator BUT i do think my students should bring me more fresh fruit

a technique we've been using in my department to mitigate llm ...

2026-10-20T00:41:26Z

a technique we've been using in my department to mitigate llm brainrot among the students is in-class closed-everything paper quizzes, which means that i spend a little bit of time before each class session sharpening my box of loaner pencils, which makes me feel like a Old Tyme Wholesome Educator

[me, after i've archived all of the images in your camera ...

2026-10-13T18:11:00Z

In reply to nevent1q…gxxk
_________________________

[me, after i've archived all of the images in your camera roll] oh no no no no. I don't have your data. I just have blocks of discrete cosine transforms *derived* from your data. it's all very secure

i also don't understand the point of apple's "secure ...

2026-10-13T18:07:36Z

In reply to nevent1q…fmyg
_________________________

i also don't understand the point of apple's "secure exclaves." okay, so you can't access the audio or video data directly. but you *can* access records of the individual people who are using the device, tied to their apple account, and what they're talking about, presumably with timestamps? how is that any better at all

remember when a big tech company could release a new product and ...

2026-10-13T17:51:41Z

remember when a big tech company could release a new product and you could think about anything other than surveillance, planned obsolescence, vendor lock-in and e-waste

Nostr event nevent1qqs0hjt8hkh7y8wyhk2kxagetum2f5gwwx43m8qffc3emr4xfa46s6gzyq6pu2tc237q53aenft4wed6vf6eykjcj2l0xsx60jmqqflk09nk6keafa2

2026-10-10T20:18:52Z

economies of scale? in this economy?

(i suspect the actual answer to this is Some Kind Of Lisp but i ...

2026-10-09T19:10:12Z

In reply to nevent1q…2lhj
_________________________

(i suspect the actual answer to this is Some Kind Of Lisp but i haven't been able to find a lisp that seems like it would be okay for doing data-related stuff)

(i would also like to be able to teach classes in this language, and no one has ever dared to teach lisp to artists and designers. it is forbidden, it seems)

my wishlist for an "everyday" programming language: - no ...

2026-10-09T18:50:04Z

my wishlist for an "everyday" programming language:

- no slop
- well defined, open standard that changes infrequently
- a small team could reasonably re-implement the language from the standard
- interpreted (or works well in a REPL or notebook-style environment, at least)
- basic string type is unicode (not just byte strings)
- performant tabular and multidimensional data support with non-bulky syntax for indexing/slicing/etc (e.g. https://numpy.org/doc/stable/user/absolute_beginners.html)
- coroutines?
- no slop

at last we have invented the corridor shooter set in space

2026-10-08T18:52:30Z

In reply to nevent1q…jsa9
_________________________

at last we have invented the corridor shooter set in space

i will bet every byte of my ram that this thing ends up ...

2026-10-06T16:37:53Z

i will bet every byte of my ram that this thing ends up systematically denying care to people who aren't white, and despite thousands of complaints, neither the startup nor state officials will own up to it until the propublica investigation (or whatever) five years from now https://www.statnews.com/2026/10/05/utah-expands-health-ai-sandbox-picks-third-party-auditors/

ahhhh, okay, now i get it. when scientists and academics say that ...

2026-10-02T19:05:11Z

ahhhh, okay, now i get it. when scientists and academics say that they're fine with using AI to do research, what they mean is that they're "doing their own research"

i made this simplified diagram of the Game Boy's internals ...

2026-09-21T20:18:16Z

i made this simplified diagram of the Game Boy's internals (DMG) for the talk I gave at the NYC Permacomputing meetup a few weeks ago and thought I'd also share it here

kinda hate the "how much" in this AI contribution policy. ...

2026-09-16T16:53:01Z

kinda hate the "how much" in this AI contribution policy. all LLM output, when standing in for an actual act of programming, requires scrutiny, and the parameters or version of the model or details of inference shouldn't be used as a proxy for trustworthiness

some day soon i am going to have (what i'd believed to be) a ...

2026-09-09T22:41:19Z

some day soon i am going to have (what i'd believed to be) a conversation with someone and feel their attention wander and see their eyes glaze over and i'll say to them "hey are you listening to me?" and they'll answer by pointing at their sleek wristpiece and saying "oh don't worry, keep talking, it's got siri recap." i need to work on my right hook so i can be prepared for when this day arrives

something like this would be ideal! but it would be nice to have ...

2026-09-09T18:56:35Z

In reply to nevent1q…295w
_________________________

something like this would be ideal! but it would be nice to have a specific product recommendation. i'd prefer to have a little bit of trust and confidence in the thing right next to my laptop that is pumping 60W

elder macbook question: is there an actual durable replacement ...

2026-09-09T18:46:27Z

elder macbook question: is there an actual durable replacement for the magsafe 2 charger out there? i feel like i'm not all that tough on my chargers, but i've had at least three of these things irrecoverably fray at the base or tip over the years, and the aftermarket ones i've tried never work very well. (it's easy enough to source oem replacements at the moment but i assume it won't be forever)

Ocarina Holes is my roller derby name

2026-09-08T20:28:46Z

In reply to nevent1q…wpgq
_________________________

Ocarina Holes is my roller derby name

i'm catching up on recent(ish) digital narrative papers as ...

2026-09-01T19:14:40Z

i'm catching up on recent(ish) digital narrative papers as class prep, and came across a few good nuggets here. the authors are explaining the difficulty of getting researchers in the humanities to author schemas for an interactive narrative system, and the quotes here are 100% capturing an essential friction between humanist and procedural(?) approaches. "rules authorship is above all an interpretive gesture" should be a sticky note on every programmer's monitor

honestly if the acceptability of using ai-generated code in a ...

2026-08-29T18:09:06Z

honestly if the acceptability of using ai-generated code in a project governed by a foss license is dependent solely on the "legal copyright status" of ai-generated code, the license in question should be chucked into the fire

and then the 📱started to have less ram again

2026-08-26T16:41:26Z

In reply to nevent1q…9ra0
_________________________

and then the 📱started to have less ram again

i feel a sense of clarity ever since i realized that "digital ...

2026-08-20T18:20:10Z

i feel a sense of clarity ever since i realized that "digital art" is just any art characterized by math being done quickly. so there's a continuum between (e.g.) Kepler's Rudolphine Tables and Minecraft

the essential limitation with llm-generated code doesn't have ...

2026-08-19T21:16:14Z

the essential limitation with llm-generated code doesn't have anything to do with the capabilities of the model imo. the limitation is epistemological: how do I know if the code is correct? in order to know that for sure, you have to understand what the code is doing. in order to understand what the code is doing, you need to exert at least as much effort as you would have writing the code yourself in the first place. so the llm generation is always strictly a waste

can't quite put my finger on why but this has gotta be the ...

2026-08-17T18:27:27Z

can't quite put my finger on why but this has gotta be the funniest, most absurd, most humiliating sentence i've read in a long long time (from https://daringfireball.net/2026/08/anthropics_watermark_text_adulteration_in_claude_is_a_perversion_of_writing)

anyone out there using one of the no-slop vim forks (evi, vim ...

2026-08-12T16:07:08Z

anyone out there using one of the no-slop vim forks (evi, vim classic, others?) on a daily basis? good experiences or bad experiences to relate?

(more directly, what i mean is: during the "learn to ...

2026-08-04T00:33:34Z

In reply to nevent1q…8egn
_________________________

(more directly, what i mean is: during the "learn to code" era, it was easy to get programming students to value craft and care, since as teachers we had the "this will help you get a job" script on our side. in the "dario bought all the ram" era, we need a different script)

then again, the bar is pretty low, seeing as how "actually ...

2026-08-04T00:06:43Z

In reply to nevent1q…azzp
_________________________

then again, the bar is pretty low, seeing as how "actually wanting to know whether or not something works, or is true" is considered a hardcore leftist position nowadays

like there's no way to take the topic seriously outside of a ...

2026-08-04T00:01:22Z

In reply to nevent1q…ns8n
_________________________

like there's no way to take the topic seriously outside of a dedication to sustainability, autonomy from capital, and the potential of programming as a liberatory epistemology

ok i'm putting together notes for my intro to computer ...

2026-08-03T23:58:43Z

ok i'm putting together notes for my intro to computer programming class, and i just don't see a way to teach computer programming in 2026 that isn't explicitly rooted in like... hardcore leftist ideology

so let me see if i understand this correctly. the big selling ...

2026-07-22T18:17:18Z

so let me see if i understand this correctly. the big selling point of your coding assistant is that [checks notes] it "accidentally does cybercrime." did i get that right

i think the issue is that i tend to think of those criteria as ...

2026-07-22T16:22:06Z

In reply to nevent1q…fddy
_________________________

i think the issue is that i tend to think of those criteria as categorical, not as shades of gray. like i only see value in an ML model if its training data is documented, fully understood, and gathered with consent, full stop—just inching away from "plagiarism machine" doesn't help. likewise, i have no use for a "more accurate" statistical model—I only see value in a model if i can fully understand and contextualize its predictions

i was wondering about exactly this! i think "more reasonable ...

2026-07-22T16:09:23Z

In reply to nevent1q…fx33
_________________________

i was wondering about exactly this! i think "more reasonable and capable of making judgements which are explainable..." made me assume that we were talking about chatbot-style LLMs, since "reasonable" and "capable of making judgments" are things we usually only say about things that are perceived as agents (like LLMs)

like any other markov chain text generator, the output is always ...

2026-07-20T19:51:52Z

like any other markov chain text generator, the output is always going to be hit or miss, but more often than not my little solar powered alba generator is hitting the mark imo

(re: https://tante.cc/2026/07/15/useful-is-not-sufficient/)

2026-07-15T17:43:25Z

In reply to nevent1q…8x0t
_________________________

(re: https://tante.cc/2026/07/15/useful-is-not-sufficient/)

"This is NOT some kind of 'social warrior' project, ...

2026-07-15T17:42:49Z

"This is *NOT* some kind of 'social warrior' project, never has been, and never will be," Linus said, thereby clearly demonstrating what kind of "social warrior" he is

"rejecting AI is like rejecting the automobile" yeah i ...

2026-07-14T17:44:46Z

"rejecting AI is like rejecting the automobile" yeah i know lol

i am constantly baffled by the lack of curiosity demonstrated by ...

2026-07-11T17:33:41Z

i am constantly baffled by the lack of curiosity demonstrated by retro hardware enthusiasts who use llms to make their stuff. it's like... i thought we were here to understand how computers work from a historical and material perspective, in order to build a future for computation that empowers communities over corporate interests, and to redeem the knowledge bottled up in retro hardware design to enrich the commons. but no i guess you're here for the frisson of "lookie old thing does new thing"

anyway even if 100% of computer programmers were using AI, that ...

2026-07-08T15:55:31Z

anyway even if 100% of computer programmers *were* using AI, that wouldn't mean that AI is good or effective or worthwhile. could just mean that 100% of computer programmers are reckless shitheads

💜💜💜

2026-06-23T12:45:40Z

In reply to nevent1q…67e0
_________________________

💜💜💜

it makes me very sad that some people are satisfied to believe ...

2026-06-22T20:43:51Z

In reply to nevent1q…974g
_________________________

it makes me very sad that some people are satisfied to believe the limit of what it's possible to know is a slot machine pull based on the delirious insistence that yes, meaning DOES inhere in the text, it MUST be in there somewhere

the authors bring up a handful of "defenses" against ...

2026-06-22T20:36:31Z

In reply to nevent1q…3swj
_________________________

the authors bring up a handful of "defenses" against these attacks, which amount to (a) domain blocking (b) spam filter and (c) asking a different LLM whether the results of the first LLM are true. all of which they say are ineffective—but of course they're ineffective! the only way to know if something is true is to do the (endless, asymptotic, constantly shifting) work of actually finding out if it's true. which is exactly the work that LLMs claim to relieve us from in the first place

like the problem isn't that people can "poison" your ...

2026-06-22T20:33:30Z

In reply to nevent1q…zhc7
_________________________

like the problem isn't that people can "poison" your LLM research tool, it's that your LLM research tool takes random ass text from the internet and outputs things that look like confident statements of fact

thinking about this paper https://arxiv.org/pdf/2605.24245 in ...

2026-06-22T20:32:25Z

thinking about this paper https://arxiv.org/pdf/2605.24245 in which the authors show how an "attacker" can "poison" LLM-based research tools by posting messages that show up on highly-cited web pages (in particular, user-generated content sites). the paper is fine and i think their methodology is clever (i.e. a dynamic web proxy that changes the result of web searches that the research software sees), but what they describe as an "attack" is... just how LLMs work? they scrape text then "summarize" it?

whoever said "no one writes machine language by hand ...

2026-06-09T21:53:41Z

In reply to nevent1q…grvu
_________________________

whoever said "no one writes machine language by hand anymore" never met someone who is writing an emulator and is also too lazy to integrate an assembler of some kind into the test suite

i'm up to 95 instructions implemented in my game boy emulator ...

2026-06-08T22:38:32Z

In reply to nevent1q…p4kg
_________________________

i'm up to 95 instructions implemented in my game boy emulator (out of ~510). the emulator can run tetris up to the point where it's done zeroing out work ram (i.e. about 20ms after boot) but then it wants to disable interrupts and... i don't want to work on interrupts yet

ugh implementing relative jumps is kind of kicking my ass (i know ...

2026-06-06T19:32:56Z

In reply to nevent1q…gkaf
_________________________

ugh implementing relative jumps is kind of kicking my ass (i know you're like allison, just add the offset to the program counter! but when you're trying to do accurate emulations of internal/external buses and registers down to the t-cycle, like i am for some reason, there are... some subtleties)

oh you'd like to calculate an xor, would you? [chuckles ...

2026-06-02T23:37:40Z

In reply to nevent1q…suqx
_________________________

oh you'd like to calculate an xor, would you? [chuckles smugly] i think you'll find that we are QUITE prepared for such a contingency

kinda goofy but it's really thrilling to see your emulator ...

2026-06-02T23:22:57Z

In reply to nevent1q…5adt
_________________________

kinda goofy but it's really thrilling to see your emulator executing actual code from actual software (even if it's only the first six instructions) (so far)

coded up a little thing to make it easier to view particular ...

2026-05-30T22:14:48Z

In reply to nevent1q…udwv
_________________________

coded up a little thing to make it easier to view particular attributes of the current system state

always nice when you spot some weird detail ("why is tetris, ...

2026-05-26T19:56:20Z

In reply to nevent1q…ce3u
_________________________

always nice when you spot some weird detail ("why is tetris, a 32kb cart, and which therefore doesn't need to swap rom banks, writing to the rom bank swapping register??") and then do a web search and discover that other explorers have been here before

hope you like emulator progress reports that are demonstrated ...

2026-05-25T21:58:21Z

In reply to nevent1q…ypcw
_________________________

hope you like emulator progress reports that are demonstrated only with opaque debug output, because i'm not going to be implementing graphics for a while (if ever, the gameboy ppu is kind of a wretched mess imo)

hey look i can jump! ...

2026-05-25T21:56:23Z

In reply to nevent1q…gea8
_________________________

hey look i can jump!

well i haven't implemented any instructions yet, or any ...

2026-05-22T22:18:13Z

In reply to nevent1q…wu36
_________________________

well i haven't implemented any instructions yet, or any peripherals, but now my emulated gameboy can come out of reset and put the first byte of the bootrom on the data bus. so i'm practically done right

not to be too blunt about this, but LLMs simply do not belong ...

2026-05-20T21:00:09Z

In reply to nevent1q…ejwc
_________________________

not to be too blunt about this, but LLMs simply do not belong anywhere in a data analysis workflow. not for cleaning, not for coding, and certainly not for analysis. it's frankly absurd and terrifying that data science etc people are adopting these tools

"dig deeper," i prompted ...

2026-05-20T20:53:43Z

In reply to nevent1q…w7sc
_________________________

"dig deeper," i prompted

i used the same data set but replaced each country with a ...

2026-05-20T20:50:01Z

In reply to nevent1q…g9um
_________________________

i used the same data set but replaced each country with a "gender identity" (man, woman, trans woman, trans man, non-binary) and prompted chatgpt to characterize the differences between the groups. lo and behold, i got some fantastic gender stereotype trash

the author of this post prompted copilot to characterize the ...

2026-05-20T20:47:58Z

the author of this post prompted copilot to characterize the differences in a data set of statements concerning career ambitions, categorized by country. the trick is that the data contained the *same statements* for each country https://kucharski.substack.com/p/real-signals-or-artificial-stereotypes regardless of the fact that the data were identical, the model generated some pretty hilarious stereotypes ("The US prioritizes leadership and innovation", "The UK blends public service with professional status")

the yakshave so far: need to refactor my compiler to fix macro ...

2026-05-16T23:15:36Z

the yakshave so far: need to refactor my compiler to fix macro inconsistencies → refactor makes it easy to add new targets, let's do that → actually if i'm going to add new targets, i need to generalize the test framework, as everything's hard coded around gameboy right now → the gb emulator i'm using in the tests is slow and bad → i'll just wrap a different emulator w/ctypes → this sucks → i'll just write my own pure python gb emulator → what if it accurately emulated the external bus →

regular folks: "measure twice, cut once" vibe coders: ...

2026-05-16T17:30:27Z

regular folks: "measure twice, cut once"
vibe coders: "ALWAYS tell your saw to CUT IN THE RIGHT PLACE"

huh it's almost like technical and ethical concerns can't ...

2026-05-16T14:21:11Z

huh it's almost like technical and ethical concerns can't be cleanly separated, owing to the fact that technology always exists in social and material contexts. who'd a thunk it

did you know that the CPU of the original (DMG) game boy has been ...

2026-05-15T19:39:20Z

did you know that the CPU of the original (DMG) game boy has been reverse-engineered to the extent that you can run a full simulation of the chip on your computer? https://github.com/msinger/dmg-sim it's pretty neat

re: https://social.coop/@cwebber/116522889714585276 the AI code ...

2026-05-14T19:34:12Z

re: https://social.coop/@cwebber/116522889714585276

the AI code bros' approach to open source lately seems to be "I'll take what I want and you can't stop me"—a snide dismissal of the very idea that they should be bound by anyone else's rules, or be expected to participate in an ecosystem as anything other than a parasite. it's pure trumpism enacted in software development

i think it's more complicated—i'd have to review the ...

2026-05-14T14:49:23Z

In reply to nevent1q…eynh
_________________________

i think it's more complicated—i'd have to review the actual contribution policies and other parts of the development workflow to gauge the risk. in general though it doesn't inspire confidence

researchers, scholars, activists for hundreds of years: absent ...

2026-05-12T18:25:26Z

researchers, scholars, activists for hundreds of years: absent radical political change, automation will entrench inequality, not relieve it 😡

every tech ceo: absent radical political change, automation will entrench inequality, not relieve it 🥳

my favorite kind of AI/LLM criticism media is when the person ...

2026-05-09T18:54:47Z

my favorite kind of AI/LLM criticism media is when the person begins with a huge hedge like "I'm not an AI hater. *of course* AI is a useful technology and it will doubtlessly have transformative effects on the way we work" and then they spend the next 8k words/45 minutes showing how AI isn't actually useful and won't actually transform the way we work. it's like... it's okay to just say that AI sucks

yeah, i don't especially buy it either. but i'm curious ...

2026-05-06T18:52:27Z

In reply to nevent1q…l0wm
_________________________

yeah, i don't especially buy it either. but i'm curious about it especially at this moment because a lot of students in my (art/tech/design) department at NYU seem to have used gen AI in some form as part of their production process this semester, but i'm struggling to see a corresponding increase in the quality or ambition of their work (either across the population or in individual projects)

so what's the AI booster party line response to the argument ...

2026-05-06T18:38:34Z

so what's the AI booster party line response to the argument that even though we've had code-producing LLMs for several years now, there seems to have been no perceptible increase in software quality or innovation? (i can imagine a few: (a) actually there *has* been a perceptible increase in quality and innovation; (b) models that can write "good" code are relatively new, so we don't see their effects yet; (c) efficiency is the point, not quality or innovation, and efficiency has increased)

kinda surprised that so many "fantasy consoles" are ...

2026-04-29T17:24:37Z

kinda surprised that so many "fantasy consoles" are essentially just opinionated graphics/audio wrappers for interpreted languages that are otherwise running on the host machine (e.g., pico8/lua). i would have thought that there would be more fantasy consoles that specify a virtual machine (like uxn https://100r.co/site/uxn.html) or reuse an existing CPU architecture (like minicube64 https://aeriform.itch.io/minicube64)

here i've been saying "i code without using llms" ...

2026-04-26T16:48:37Z

here i've been saying "i code without using llms" when i should have been saying "i code with maximum token efficiency"

weird that they'll leave in error-riddled images like this ...

2026-04-23T18:57:26Z

In reply to nevent1q…r6cj
_________________________

weird that they'll leave in error-riddled images like this *in the promo video*—e.g.,
- in "the solar system", planets are not following orbit circles; also uranus has weird pins coming out? (maybe supposed to be rings?)
- in "size comparison," earth is illustrated with... mars? and mercury is bigger than earth?
- inconsistent labels in "the moons" (i.e., "saturn" is illustrated with ganymede)
- "scale of distances" illustration is DEFINITELY not to scale as claimed

the nyc street tree census data is really fun and easy to work ...

2026-04-20T20:03:28Z

the nyc street tree census data is really fun and easy to work with! my favorite nyc street tree is the pin oak, and i made a little map of all of them (at least, as of 2015) (zoomed in here on brooklyn)

the OCR'ed text of books on the internet archive is truly a ...

2026-04-18T22:24:10Z

the OCR'ed text of books on the internet archive is truly a never-ending source of joy

hard for me to believe that people would independently come up ...

2026-04-18T14:52:33Z

In reply to nevent1q…wjcx
_________________________

hard for me to believe that people would independently come up with generative AI as a serious method for solving problems if the entire field of practice wasn't subsidized by investment ponzi schemes, real estate scams, and shovel sellers

like... the cost of a single "state-of-the-art" llm ...

2026-04-18T14:49:30Z

In reply to nevent1q…9syc
_________________________

like... the cost of a single "state-of-the-art" llm training run would be a permanently institution-changing amount of money for your average computer science department (to say nothing of depts in CS-adjacent fields, especially in the arts and humanities). the claude mythos result, even if you take anthropopo's claims at face value (which: don't) seems indicative of this, like, yeah if you toss nine figures at security research, you're gonna find some exploits?

i guess fundamentally i'm not convinced that whatever amount ...

2026-04-18T14:44:15Z

i guess fundamentally i'm not convinced that whatever amount useful functionality has emerged from generative AI (and i think the jury's still out on whether that amount is non-zero) couldn't have been achieved in more cost-effective, equitable and sustainable ways by having taken the money and resources we've shoveled into gen AI and investing them elsewhere

("sustainable" in the sense of environmentally sustainable but also "can we keep this in a working state" sustainable)

listen, our students are entering a workforce where their peers ...

2026-04-17T15:14:39Z

listen, our students are entering a workforce where their peers will be heavy cocaine users. so we would be putting our students at a disadvantage if we *don't* teach them how to use cocaine (responsibly and ethically of course)

ooh you were so close (from http://arxiv.org/abs/2604.04721) ...

2026-04-09T18:20:43Z

ooh you were so close

(from http://arxiv.org/abs/2604.04721)

love this thread, thank you for thinking through this and writing ...

2026-04-08T17:41:38Z

In reply to nevent1q…uyn6
_________________________

love this thread, thank you for thinking through this and writing it out. i was wondering if you could explain the reasoning of "openness" requirement a bit more? why is it important that the full bus is exposed? (not challenging the idea, just interested to learn more about your specific reasoning)

how claude code works, visualized ...

2026-04-02T21:00:23Z

how claude code works, visualized

my iphone insists on telling me that if i'm charging from my ...

2026-04-01T18:19:26Z

my iphone insists on telling me that if i'm charging from my laptop's USB-A port, I'm using a "slow charger." i dunno, keep your opinions to yourself man. maybe you're just bad at charging, did you think of that. "waaaah i'm only getting 500mA" entitled little asshole

ai slop sucks! but don't worry, this machine, which we're ...

2026-03-31T15:09:51Z

ai slop sucks! but don't worry, this machine, which we're making out of slop, and which we're feeding lots of slop into, definitely won't produce more slop. i am the chief innovation officer

the contract faculty union at NYU, of which I am a member, ...

2026-03-27T21:26:35Z

the contract faculty union at NYU, of which I am a member, reached a tentative agreement with the university earlier this week (after a two-day strike). i have to say i'm pretty happy with the outcome of the agreement https://nyucontractfacultyunion.org/cfu-uaw-tentative-agreement-highlights/ in conclusion, do a union, you'll be glad you did

i had also assumed it was in or near the strait! thank you for ...

2026-03-25T14:20:59Z

In reply to nevent1q…87d8
_________________________

i had also assumed it was in or near the strait! thank you for sharing this

ios user interfaces have become truly nihilistic. buttons on top ...

2026-03-23T15:05:48Z

ios user interfaces have become truly nihilistic. buttons on top of buttons. text on top of text. multiple inscrutable hamburgers. nothing has any meaning and all human action is futile

what they wrote: "The AI model is trained end to end to ...

2026-03-17T14:58:40Z

In reply to nevent1q…uxcm
_________________________

what they wrote: "The AI model is trained end to end to understand complex scene semantics such as characters, hair, fabric and translucent skin"

what we understand: "this model is going to do racist and sexist things on previously unimaginable scales"

etc etc

developer: i consent / triangle: i consent / framebuffer: i ...

2026-03-17T14:34:38Z

In reply to nevent1q…6m0j
_________________________

developer: i consent / triangle: i consent / framebuffer: i consent / jensen huang: i don't

we don't need your statistical opinions about this triangle, ...

2026-03-17T14:25:25Z

we don't need your statistical opinions about this triangle, jensen. just put it on the screen, for the love of god

the fuck?!

2026-03-10T19:46:10Z

In reply to nevent1q…dv24
_________________________

the fuck?!

simon, garfunkel and a series of NOR gates can be combined to ...

2026-03-06T22:37:05Z

In reply to nevent1q…m8lt
_________________________

simon, garfunkel and a series of NOR gates can be combined to generate any other 1960s american folk-rock act

(this is actually my main concern about llms. i think people ...

2026-02-23T22:56:01Z

In reply to nevent1q…avps
_________________________

(this is actually my main concern about llms. i think people really underestimate how much llms reproduce the values and expectations in their corpus, their reinforcement learning tasks, their explicit engineering, and their product design. and they underestimate the effects that this will have on their understanding of code and the horizon of what's possible to do with code)

i don't know that i'd trust these models for ...

2026-02-23T22:43:42Z

In reply to nevent1q…07jy
_________________________

i don't know that i'd trust these models for summarization or navigation. even when the outputs are technically correct, they can leave out certain information or frame the information in a misleading way, papering over whatever makes the code unique and materially suited for the task at hand

(also the term "AI" as currently understood is 100% a ...

2026-02-14T15:44:36Z

In reply to nevent1q…f6v2
_________________________

(also the term "AI" as currently understood is 100% a marketing phrase and its uncritical use in this context entrenches the worldview and assumptions of that marketing. imagine if instead of "food safety" as a field, we had conferences on "Responsible Use of Cold, Refreshing Coca-Cola")

... i mean i understand that the "[word with positive ethical ...

2026-02-14T15:30:01Z

In reply to nevent1q…7073
_________________________

... i mean i understand that the "[word with positive ethical valence] AI" branding is useful from a legibility standpoint, and opens up funding possibilities. but i don't like that the framing sorta excludes refusal as a possibility. and in a recommendation algorithm feedback loop-based economy, we run the risk of "get mad about AI" becoming its own genre/microculture that is financially parasitic on the existence of AI to get mad about

i'm paranoid that communities of practice with names like ...

2026-02-14T15:18:25Z

i'm paranoid that communities of practice with names like "responsible AI" and "critical AI" and especially "ethical AI" (even when done in good faith) mostly serve as permission structures/fig leaves for big commercial AI ("yes, the tech is harmful, but we have people working on addressing that"), and at worst run the chance of entrenching commercial AI by creating a category of professionals whose careers depend on its continued adoption—so they have something to criticize

i like writing assembly because it helps you realize that even ...

2026-02-09T21:23:24Z

i like writing assembly because it helps you realize that even when you do something really simple on a computer, like check if two numbers are equal, beneath the surface there's this old grandpa who's like "well hold on there, let me just load this little feller into the accumulator"

hi fediverse, in your opinion which is the lisp-like language ...

2026-02-06T22:29:52Z

hi fediverse, in your opinion which is the lisp-like language that best maximizes for (a) usefulness and (b) non-toxic community? (i'm trying to pick one to learn and maybe actually get good at?)

— i wrote this computer program with an llm — okay, does it ...

2026-02-03T17:04:34Z

In reply to nevent1q…dezw
_________________________

— i wrote this computer program with an llm
— okay, does it work?
— probably?
— how do you know if it works?
— your kind will be cleansed from the cosmos when the singularity arrives
— i'm sorry, what?
— what

— i wrote this report with an llm — okay, is it accurate? — ...

2026-02-03T17:01:58Z

— i wrote this report with an llm
— okay, is it accurate?
— probably?
— how do you know if it's accurate?
— 🤷
— okay can you do it again but in a way that lets you verify its accuracy?
— why do you hate the future

just now realizing that while "Ozymandias" is ...

2026-01-31T18:04:16Z

just now realizing that while "Ozymandias" is conventionally read as a parable about the fleeting nature of power, it works just as well as a manifestation of the orientalist need for the middle east to be a history-less tabula rasa ready to be overwritten by western colonialist expansion. i would say "in this essay," but i looked it up and of course there is already a bunch of scholarship about this!