It’s the wall on the right – in the wide version (from felesteen.news) it’s all the same blue colour, and a corner, whereas in the Reddit version it’s a white concrete pillar with no corner.
At the very least, someone’s done some infilling on one of them. My most charitable guess was that someone at the news site decided to “punch up” the image for an article header, but the third version and its timing make me think Occam’s razor is the way to go here.
Your version was posted within 3 minutes of this post someone linked in the comments of the Reddit version. So yeah, all credibility lost as far as I’m concerned.
What I don’t get is why, if this is some sort of coordinated campaign, you wouldn’t just generate entirely new images, rather than AI-daydream variants of the exact same one, which makes it obvious they’re manipulated.
Nice find, kind of damning. That version has a totally different wall on the right side to the version posted here, it also seems to be from a slightly different angle. This screams AI manipulation – if not outright fabrication – of at least one of those images. Doesn’t necessarily mean they’re both fake, but a pretty big red flag.
I was gonna say it on the last thread but, before we start appropriating this image as a political symbol as you’ve done here, should someone… maybe… source it slightly better than “Reddit”? There’s plenty of evidence that the school was blown up, but this specific image – conveniently well-framed and poignant – cropped up very quickly on a few random social media accounts, with no photographer attributed, and as far as I know hasn’t actually been verified at all.
TMSJ;BI (Too Much Shitty Javascript; Barely Interacted):
Is there some kind of award for “most irredeemably ‘HN’ HN post”? Just spam a bunch of LLM comments to drive the numbers like everyone else, no need to add the whole extra step of making sure anyone trying to actually engage knows exactly how much contempt you have for their opinion.
Also, the website is so vibecoded that the logo has the wrong number of fingers. Can’t make this shit up.
This definition of social media is new to me as well, thanks for sharing it. This sort of clarifies a term I really dislike, and which you’ve used: “the algorithm”. It’s always seemed a little murky to me which algorithms it refers to. It’s like saying “don’t eat food with chemicals in it”.
Lemmy does have “an algorithm”, it’s just a relatively simple one based on communities one is subscribed to plus some vote/comment data for the various sort orderings.
Lemmy also absolutely implements a social graph – the data about who has interacted with whom is all stored by the system. It’s not explicitly stored as a graph structure, but then we’re arguing database schemas.
As I understand it, however, you’re saying “social media” arises when the “social graph” data structure is used as an input to “the algorithm”. That seems like a pretty robust definition to me.
One bit of pedantry: user blocks on Lemmy are, by a general definition, a form of social graph, and they do affect what content people see. So Lemmy could technically qualify as social media by the definition I’ve written here. I’m not sure what a more precise definition could be that avoids this technicality.
I’m surprised how many people seem offended by this comic. I found it pretty relatable. That doesn’t mean I “don’t understand context”, or think the writers were bad people, or that the shows aren’t worth watching. It just means I find it personally unpleasant to find these jokes in a work I’m otherwise enjoying.
When kids study older literature and media in school, generally when there’s a slur or racist reference or joke, the teacher will stop and explain the context and get the kids to reflect on it and why it’s probably not acceptable today. Even though I understand this and know it going in, I’m still kind of doing an abridged version of that in my head when something like this comes up in a show – I’ve been following along, laughing with the writers, and then suddenly I’m backing up and distancing myself from one joke or idea. It’s jarring, it pulls me out of the show, and it’s just not fun.
In some cases it also comes across as incredibly lazy and unoriginal. So many sitcoms from that era have “the trans episode”, “the gay episode”, “the lecherous character” – and they all make the same unfunny jokes, and it’s a reminder than a lot of these shows, even in their time, were just not that creative. Plenty of modern shows have the same problem, but they don’t draw attention to it by having large classes of “stock jokes” that simply do not land today.
Interesting, thanks for doing the research!
As an extreme non-expert, I would say “deliberate removal of a part of a model in order to study the structure of that model” is a somewhat different concept to “intrinsic and inexorable averaging of language by LLM tools as they currently exist”, but they may well involve similar mechanisms, and that may be what the OP is referencing, I don’t know enough of the technical side to say.
That paper looks pretty interesting in itself; other issues aside, LLMs are really fascinating in the way they build (statistical) representations of language.
This is a good name for one of the main reasons I’ve never really felt a desire to have an LLM rephrase/correct/review something I’ve already written. It’s the reason I’ve never used Grammarly, and turned off those infuriating “phrasing” suggestions in Microsoft Word that serve only to turn a perfectly legible sentence into the verbal equivalent of Corporate Memphis.
I’m not a writer, but lately I often deliberately edit myself less than usual, to stay as far as possible from the semantic “valley floor” along which LLM text tends to flow. It probably makes me sound a bit unhinged at times, but hey at least it’s slightly interesting to read.
I do wish the article made it clear if this is an existing term (or even phenomenon) among academics, something the author is coining as of this article, or somewhere in between.
GPT-4o mini, “Rephrase the below text in a neutral tone”:
This name is appropriate for one key reason: I have not felt the need to use an LLM for rephrasing, correcting, or reviewing my writing. This is also why I have not utilized Grammarly and have disabled the “phrasing” suggestions in Microsoft Word, which often transform a clear sentence into something overly corporate or generic.
Although I wouldn’t categorize myself as a writer, I have been intentionally editing myself less than usual lately to avoid the typical style associated with LLM-generated text. This approach might come across as unconventional at times, but it can also make for more engaging reading.
I also wish the article clarified whether this term is already established in academic circles, if the author is introducing it for the first time, or if it falls somewhere in between.
“avoid the typical style associated with LLM-generated text” – slop!
Yeah, even though I have a bit of background I can’t really make heads or tails of that OpenSearch doc at a glance, it’s dense stuff.
In my experience knowing the keywords to stick in a search engine is often half the battle; there are plenty of resources out there on “vector databases”. “Semantic search” from the lede of the OpenSearch doc might be another good one to have around.
Feel free to ask me any other questions and I can try to answer to the best of my abilities, though again, not an expert and honestly I’ve never actually used these myself beyond toy examples.
I’m not an expert, but it sounds like you want an embedding+vector database. This essentially extracts the part of an LLM that “understands” (loaded term, note the quotes) the text you put in, and then does a lookup directly on that “understanding”, so it’s very good at finding alternate phrasings or slightly differing questions.
There’s no actual text generation involved, and no need to retrain anything when adding new questions.
OpenSearch has an implementation (which I learned about just now while writing this comment and thus cannot vouch for); you could start there.
My assumption for the key icon was something to do with PINs/passkeys, which kind of reinforces OP’s point.
Ngl I really want to know what the tick icon actually does now.
No mention of how it compares to existing spatial indexing methods such as R[*]-trees. That was my first thought reading the article, but they only give a comparison to naïve NxM testing. I assume this method is still an improvement in the presence of sharding, but doubt it’s the 400× quoted.
Is it really not as easy for them as saying “hey btw don’t use this distro if you’re in California” and fully expecting nobody to comply? I’m not sure if Ubuntu is based in Cali in which case I can see it being more difficult.
Also this “age bracket” thing seems to have an obvious flaw in that any application that’s running semi-regularly can just poll the API every day and find out the user’s DOB by checking when they roll into the next bracket. It’s actually leaking more data about children than about adults in that case. Brilliant.