Gregory's Server
|
25 posts total
12
i enjoy Dr. Becky's videos https://www.youtube.com/@DrBecky in which she (an Oxford astrophysicist) explains recent astrophysics research for non-expert audiences. (usually this means "debunking" pop science journalism about pre-prints, but not always.) are there video-makers out there like this, but for other academic areas? not just "here's science explained in an engaging way" but "here is what's currently happening in this scientific field in which i am a participant"?
Show previous comments
@aparrish Dr. Becky is great. When I remember any others I will post them.  @aparrish i made this custom game boy flash cart by "dead bug" soldering a DIP parallel flash chip to a cartridge edge breakout board that i designed. there's no memory bank controller in here, so you can only flash roms up to 32kb (i have tetris on there now haha) but i like how it turned out!
Show previous comments
the game boy is the perfect game platform imo. there are >100 million hardware units in existence (probably one in your junk drawer); the hardware is low-power, portable, well-documented and user-serviceable; free, near-perfect emulators run on basically any platform. there are several foss dev toolchains, and the limited capabilities of the machine means naturally lower dev costs. and it's not online so it's basically impossible to make exploitative free-to-play shovelware this table of C escape characters is making the same noises as I do when I have to wake up early
Show previous comments
 @aparrish \xhh... needs no explanation, we've all felt that. \? cannot be explained with a mere note. @aparrish reminds me of working on part of our code where for some reason we used strings to do some configurable pattern matching. Anyway, autocomplete picked up on the strings and, well, it's a mood.  @aparrish I don't even understand that coding language and the table still made me laugh so much that I snorted a cup of tea. oh my god "AI" is such a clown show https://www.theverge.com/features/23764584/ai-artificial-intelligence-data-notation-labor-scale-surge-remotasks-openai-chatbots just a big funhouse mirror of useless labor, all because billionaires and computer scientists can't get it into their heads that the task of accurately answering a question is fundamentally incompatible with statistical prediction my department at NYU is now accepting applications for our 2023/2024 Project Fellowship: https://tisch.nyu.edu/itp/itp-ima-project-fellowship basically we'll pay you US$5K to come use our facilities to make a cool project over the course of a semester. pretty good deal if you ask me? I'm re-reading Sueyeun Juliette Lee's insightful essay "Shock and Blah" for class tomorrow https://thevolta.org/ewc41-sjlee-p1.html and realizing that... "stuplimity" is exactly the aesthetic experience that LLMs like ChatGPT evoke in me Lee is discussing what she calls the "traumatic stuplime" in conceptual poetry of the 2010s—works like Kenneth Goldsmith's _Seven American Deaths and Disasters_, which collects "raw" and unadorned transcripts of radio and tv coverage of traumatic events in American history. she quotes a question that Stephen Colbert asked Goldsmith during his appearance on the Colbert Report in 2013 to promote that book, a quote that I think about all the time. imo Colbert's question also applies to LLMs wait did they really retcon <b> as the "bring attention to" element lol https://developer.mozilla.org/en-US/docs/Web/HTML/Element/b
Show previous comments
 every paragraph in this paper is blowing my mind https://www.cell.com/trends/cognitive-sciences/fulltext/S1364-6613(22)00236-4 even as someone who already believed in the infinite variety of language, there are just so many deeply researched counterexamples to basically anything you think is universal (especially if you're using english as your basis for "universal") I set today aside for reading tabs, this paper among them, but I have ended up with a net tab surplus because of all of this papers interesting citations 😔 as a quick reminder, or for new followers: for a while now, I have been writing critically about large language models from the perspective of a poet, linguist, computer programmer and trans person (which I think is an at least somewhat unique perspective?). here are some links to things I've written, if you're interested... "Language models can only write poetry" (on language models, speech act theory, and the definition of poetry): https://posts.decontextualize.com/language-models-poetry/ "Material paratexts" (on "cistextuality" and the surprisingly mundane ways that people relate computer-generated texts to the social world): https://posts.decontextualize.com/material-paratexts/ I have posted a very detailed write-up of my Game Boy Pocket SP mod, including details on design process and fabrication. I also included instructions on how to replicate the build yourself! https://posts.decontextualize.com/pocket-sp/ i have discovered a new musical instrument: cat chasing her tail in a wooden box. here is althea performing a solo   Thinking of it, a cat is an entire band worth of musical instruments. These sounds can be used as percussions, the meows as the main instrument, and the purrs as bass. I have a new book out on Aleator Press! https://www.aleator.press/releases/wendit-tnce-inf it's a collection of asemic prose poems in a small handmade letterpress edition. the book was generated with code from a nanogenmo project of mine from a few years ago (source available here: https://github.com/aparrish/word-gan-book-generator) james did such lovely work with the book design and printing. it was rewarding to work with him and I'm so glad this is finally out in the world! hi fediverse, I just posted transcripts of a few talks and lectures I've given over the past few years, mostly concerning the connections between machine learning, language and poetry: https://posts.decontextualize.com (notes and summaries in individual posts below) the university of milan has released over four hundred meows for non-commercial and research purposes https://zenodo.org/record/4008297 (via https://www.data-is-plural.com/) hey fediverse, one of my (computer-generated) poems is on the cover of BOMB Magazine's Winter 2021 issue! there are also several new poems from the Compasses series inside https://bombmagazine.org/issues/154/ 12
|
it falls prey to every fallacy of AI creativity research (and AI research in general), e.g., that "AI" is a monolithic technology, that "AI" is independent of human intention, that "AI"'s telos is to produce artifacts "indistinguishable" from "humans," that the ability to "replicate" certain genres of art (especially genres positioned as highly "creative," like poetry) are benchmarks along that telos, etc.