The website for the Columbia Law Review, which last I checked hosted freely readable copies of all articles from 2008 onward, is now offline, displaying a "Website is under maintenance" message. Editors say it's because the review's board of directors took the site down after the editors refused to remove an article by a Palestinian legal scholar: https://theintercept.com/2024/06/03/columbia-law-review-palestine-board-website/
Anyone know if Simon Willison's #PyConUS keynote will be generally available? I'd be interested in checking out a recording or transcript.
I've seen a few folks credit him for the phrase "imitation intelligence", which sounds to me like a concise way to note how Alan Turing's concept of AI as an imitation game has largely outlived its usefulness in the LLM age.
Magic tricks, e.g., have long fooled even people familiar with how sleight of hand works. But we don't say magicians do real magic.
@JMarkOckerbloom If you had an in-person ticket and want to watch talk recordings before they get posted to public YouTube, the 2024 talks recordings are now available.
Log in at pycon.us -> Dashboard -> "Join PyCon US 2024 Online". In-person attendees get free access to the online ticket.
The Register goes deep into looking into Christine Dudley's comment (which I reported earlier) that she started to see ads that reflected the audiobooks she was checking out from her library.
Upshot: They can't tell for sure that the ads appeared as the result of her reading those books, but the sheer volume of data tracking methods and pathways invoked on sites, services and devices used by libraries and patrons also make it difficult to tell for sure that they didn't: https://www.theregister.com/2024/05/18/mystery_of_the_targeted_mobile_ads/
They say they don't collect precise data and anonymize what they do collect, but it's easy for attempts to do that not to be as secure as intended. That's one reason I urge analytics collection to be made opt-in (FF's isn't) instead of opt-out.
The full list of 2024 #PulitzerPrize winners and finalists has now been posted: https://www.pulitzer.org/prize-winners-by-year Along with awards in the usual prize categories, two special citations were also awarded, one for late writer and critic Greg Tate, and one generally for journalists and media workers covering the war in #Gaza.
54 years ago today, on May 4, 1970, Ohio National Guard troops fired on demonstrating students at Kent State University, killing 4 of them and wounding 9 others.
11 days later, police fired on students at Jackson State College in Mississippi, killing 2 students and injuring 12 others.
Among the many reports and studies following the killings (for which no one was convicted) was the 1970 Report of the President's Commission on Campus Unrest. You can read it here: https://eric.ed.gov/?id=ED083899
I don't know personally how sound this insider critique of NPR is, but it has one of the best descriptions of centrist bias I've seen in a while: "an editorial fixation on finding the exact middle point of the elite political and social thought, planting a flag there, and calling it objectivity." https://slate.com/business/2024/04/npr-diversity-public-broadcasting-radio.html
(Any 1-dimensional definition of news "bias" that doesn't let you identify centrist bias is fundamentally flawed. And I've seen multiple media-rating systems with that flaw.)
"OCLC does not allege that it traced any of the attacks to Ms. Matienzo, that OCLC discovered any shred of evidence demonstrating Ms. Matienzo’s alleged ties to Anna’s Archive, or that Ms. Matienzo herself committed any wrongful act against OCLC.... While Ms. Matienzo intends to seek separate redress for OCLC carelessly naming her in
this matter in the future, at this time, she simply seeks dismissal from this action..." Torrentfreak's full copy of the motion to dismiss: https://torrentfreak.com/images/anna-dismiss-1.pdf
I was curious about this from the get-go and these documents do not disappoint.
"Presumably, OCLC brings its vague and collective claims against all Defendants with the hopes of conducting discovery and then determining whether Ms. Matienzo is actually associated with Anna’s Archive (she is not)."
Felt a prolonged shaking in my house in Mt Airy, Philadelphia for at least half a minute. Probably nothing to to west coaster, but it was the most noticeable #earthquake I've ever experienced living on the east coast.
Harvard Law School announces a full release of its Caselaw Access project, providing free access to 360 years of US federal, territorial, & state case law, in human-readable and machine-processable form: https://case.law/ Here's a video they've recently released about it: https://vimeo.com/922493882
The old site mentions the "transition of Caselaw Access Project data to a static state" so I'm not sure if it'll add anything new past 2020. But perhaps others can continue forward if desired.
"In summary, the management of bilateral iatrogenic I'm very sorry, but I don't have access to real-time information or patient-specific data, as I am an AI language model." Radiology Case Reports, another Elsevier journal (open access, $550 article processing charge) publishes a paper with obvious signs of LLM text generation: https://www.sciencedirect.com/science/article/pii/S1930043324001298
Here's another suspicious-looking sentence opener, this one from the Elsevier journal International Journal of Applied Earth Observation and Geoinformation ($2770 open access publishing fee): "Certainly, here is the pseudo code for the specially designed spatial scene augmentation method..." https://www.sciencedirect.com/science/article/pii/S1569843224000979
Looks like a paper with telltale signs of LLM-generated text got past the editors of Elsevier's Surfaces and Interfaces: https://www.sciencedirect.com/science/article/abs/pii/S2468023024002402. (As of this writing, the first sentence of the main body starts "Certainly, here is a possible introduction for your topic:".)
I don't have the domain expertise to tell if the text of the paper is accurate, or if the cited references say what is claimed for them. (The first 10 refs, all citing other Elsevier pubs, do at least exist, per my check.)