lel's recent activity
-
Comment on Report: Potential New York Times lawsuit could force OpenAI to wipe ChatGPT and start over in ~tech
-
Comment on This is how we finally kill TurboTax in ~finance
lel Link ParentYeah, it's a crime to give them incorrect information intentionally.Yeah, it's a crime to give them incorrect information intentionally.
-
Comment on Baldur’s Gate 3 is causing some developers to panic in ~games
lel Link ParentIf anything, it feels to me like the title has some serious risk of hurting it, honestly. I know, as someone who has never played Baldur's Gate but knew it was a '90s RPG game series, that I saw...And lastly, this reoccurring point I see about it being successful because it has the name "Baldur's Gate" on it is kind of silly. BG2 came out almost 24 years ago. It is not the powerhouse name that older gamers think it is anymore.
If anything, it feels to me like the title has some serious risk of hurting it, honestly. I know, as someone who has never played Baldur's Gate but knew it was a '90s RPG game series, that I saw the 3 in the title and let air out through my nose and then never bought it because it can be intimidating getting into any franchise with a continuing story in any medium 25 years late, especially one that has a metatextual story spanning several videogames and dnd game books.
Obviously the popularity of it means they must have made it very easy to get into 25 years late, and I'm more tempted now, but I have to imagine there are people out there who saw that it was the newest entry in a 25 year old franchise (or for a lot of people that it was the direct sequel to a game that looks like this) and figured it was inaccessible to them.
-
Comment on Backward compatibility, Go 1.21, and Go 2 in ~comp
lel Link ParentJesus, I used to use lobste.rs and I kind of forgot how much of a colosseum it is in there. Now that I see all the people yelling at each other in that thread, I do remember a period of a few...Jesus, I used to use lobste.rs and I kind of forgot how much of a colosseum it is in there. Now that I see all the people yelling at each other in that thread, I do remember a period of a few years where any mention of Go anywhere in any thread on the entire website would lead to whatever thread it was in becoming a massive Go vs. Rust brawl for no reason. Just goes to show that making your site invite only doesn't necessarily create what Tildes has absent moderation and a community that is interested in keeping that kind of stuff in check.
-
Comment on Batman is a jerk in ~comics
lel Link ParentWhich is funny because this is the exact opposite of the argument made in Bill's monologue at the end of Kill Bill. I've never read comics but one of my old roommates was into them, and when he...Which is funny because this is the exact opposite of the argument made in Bill's monologue at the end of Kill Bill. I've never read comics but one of my old roommates was into them, and when he first saw that scene when I showed him Kill Bill he started yelling that it was wrong. I wouldn't know. My understanding now is that apparently Clark Kent is best understood as Superman's alter ego in like golden and silver age comics, but that this characterization seems to have flipped since then?
-
Comment on Unicode thanks Blue Blocker, our newest Silver Sponsor! in ~tech
lel Link ParentEverything he has done has been so funny. He bought an asset that essentially got its value from three places: (1) Being advertiser-friendly and having one of the biggest userbases of any website...Everything he has done has been so funny. He bought an asset that essentially got its value from three places:
(1) Being advertiser-friendly and having one of the biggest userbases of any website in the world to draw advertising revenue,
(2) Having an existing code base that was solid and worked, and
(3) Being one of the most famous and recognizable brands to literally ever exist, with multiple dictionary entries, etc.And he (1) drove away oodles of users while also pissing off advertisers by getting rid of verification for companies and turning the website into a nazi / child porn-friendly nightmare realm, so now all the ads are for dick pills, pyramid schemes, and fake colleges, (2) fired every employee which then broke the code base so the website crashes every day and everyone can only view a few posts a day, and (3) just replaced, again, one of the most recognizable brands to ever exist with something completely unmarketable.
It's a crash course in how to bring an asset's value to zero.
-
Comment on ChatGPT's odds of getting code questions correct are worse than a coin flip in ~tech
lel Link ParentAccording to the paper they used 3.5. That being said, in my experience with GPT-4, it is overhyped and hallucinates just as often, though I would really like to see this repeated with 4.According to the paper they used 3.5. That being said, in my experience with GPT-4, it is overhyped and hallucinates just as often, though I would really like to see this repeated with 4.
-
Comment on Clarence Thomas’ thirty-eight vacations: The other billionaires who have treated the US Supreme Court Justice to luxury travel in ~misc
lel LinkJust took the bar ethics exam yesterday. There were two questions about the rules relating to reporting gifts -- naturally those don't apply at the Supreme Court level.Just took the bar ethics exam yesterday. There were two questions about the rules relating to reporting gifts -- naturally those don't apply at the Supreme Court level.
-
Comment on The Reddit protest is finally over. Reddit won. in ~tech
lel Link Parent99.99% of online debates (actually debates in general) are fundamentally performative. The fact that online there is always an audience means that every discussion between two people who disagree...99.99% of online debates (actually debates in general) are fundamentally performative. The fact that online there is always an audience means that every discussion between two people who disagree about something turns into that because both sides just want to be perceived as having won.
Hypothetically the goal of this is to change the onlookers' minds or whatever by appearing to have the victorious idea, under the assumption that someone seeing your idea prevail over another in a public space will convince neutral parties, but the reality is that like the rest of the internet, it's all just dopamine buttons. So if I can get a TKO on you by doing wordplay that works just as well.
I like the way tildes reply chains become unusable after a certain number of replies because they stop nesting and turn into that "(reply to the previous post)" thing because it (along with the culture of course) has resulted in me very, very rarely having seen a massive flamewar shitstorm like that on here.
-
Comment on Language is a poor heuristic for intelligence in ~comp
lel (edited )Link ParentYeah I'm gonna bow out after this one before this truly just gets stuck in the mud as a definition debate, but: Yes, an LLM can often give an answer to a question correctly. Is being able to...Yeah I'm gonna bow out after this one before this truly just gets stuck in the mud as a definition debate, but:
Yes, an LLM can often give an answer to a question correctly. Is being able to produce text that encodes some information the same thing as knowledge? I mean, maybe! That's not really relevant though. What is relevant is that an LLM does not conceive of that as information, per se, it conceives of it purely in terms of language. What I'm getting at is that you cannot just give an LLM new information, you can only give it new language, and though that language may encode useful information, the LLM is only ever capable of using that information as language.
I brought this up because it means you can't just teach an LLM to start introspecting, or fact-checking, because those are new processes, and you cannot teach an LLM a new process, you can only give it new language. You can give it language that tends to indicate introspection, but that doesn't actually result in anything resembling an introspective process or a fact-checking process occurring under the hood, because there's no code behind that. What it results in is exactly what LLMs do when you ask them to introspect: they say "Sorry! On second thought I got that wrong, didn't I?", but no introspective process has occurred. In fact you can easily get them to say sorry even if they were originally right! Saying sorry is the simple result of mimicking the language that people tend to have after introspecting.
You can teach an LLM how to introspect, or how to fact-check, by adding language associated with these tasks to its training data, and this will result in it being able to explain to you how to introspect and fact-check, and it will result in it confidently speaking in the language of introspection and the language of fact-checking when prompted, but it will not prompt it to conduct introspection or fact-checking. The LLM does not even understand these processes as things that it has the ability to do, or things that anyone has the ability to do. This is relevant because I was responding to someone talking about introspection, fact-checking, etc as new processes that they believe you can teach LLMs how to do.
Whether a mimic of language which is sufficiently advanced that it often correctly responds to queries can or cannot be said to "know" something is not the question I was raising and is ultimately not very interesting to me because it's just a definition debate. What I was saying was that an LLM by definition does one single process (transforming text) very well and that single process also happens to give it the ability to mimic a lot of other processes (finding information, etc etc etc) but it cannot actually perform any other processes, nor can it be made to perform any other processes because it does not have a model for anything except for text and the transformation thereof.
-
Comment on Language is a poor heuristic for intelligence in ~comp
lel (edited )Link ParentIntrospecting with oneself is not something LLMs have the capacity to do by their nature. You can very easily imagine apolz's point that they might be able to construct SQL queries. Maybe you have...Introspecting with oneself is not something LLMs have the capacity to do by their nature. You can very easily imagine apolz's point that they might be able to construct SQL queries. Maybe you have a pipeline of LLM inputs and outputs where one run of the LLM generates an SQL query based on the user's query, then the program polls a database using that query, then the LLM is again asked to convert that into human-readable content with something like "answer this question assuming the following to be true" (however I definitely have questions about the utility of this because you still end up in the same position of unverifiability you're in right now, though that's not a conversation I really care about having). This is so easy to imagine because converting text from one form to another is one of the things LLMs are actually pretty good at, and you can just string some dumb code or more traditional AI in the middle between the two text transform steps to make the more mechanical and information-oriented parts of it work and you're good.
But LLMs cannot query external sources for information, nor cross-reference sources, nor fact-check, nor introspect, because all of these are tasks that require some concept of information, and information is fundamentally incompatible with how LLMs work under the hood. I respectfully think you're not grasping the point of the article which is that an LLM does not and cannot possess information, it mimics language, and it turns out this gets you a shockingly close simulacrum of possessing information, which is good enough for a lot of uses! But that doesn't get you any closer to knowing how to create an AI that does the things you're describing because to do those things you'd need to also invent a completely different type of AI that works via a completely different method. There is no iterative development process that can be performed on an LLM that gets you a machine that knows things.
Maybe if AGI is possible then LLMs will be part of the pipeline for how they process input into and generate output from whatever currently future-tech AI actually does the "knowing" and the "thinking", and that currently future-tech AI could do the things you're describing, but that's not how LLMs work or are able to work, and our (current?) inability to create an AI that could do those things is the entire reason why we instead settled for making an AI that subjectively feels like it's doing those things when it actually isn't.
-
Comment on William Friedkin, acclaimed director of ‘The French Connection’ and ‘The Exorcist,’ dies at 87 in ~movies
lel LinkI only just learned yesterday that, when it was discovered that the counterfeit money from To Live and Die in L.A. had entered circulation, the feds interviewed Friedkin to find out what had...I only just learned yesterday that, when it was discovered that the counterfeit money from To Live and Die in L.A. had entered circulation, the feds interviewed Friedkin to find out what had happened to it after filming wrapped. He didn't give them a thing and told them to fuck off. When the statute of limitations expired he admitted it was him -- he had been spending it as real money all those years. First thing he had done after the cops were done grilling him was he went and spent some more. One of the realest filmmakers to ever do it. He'll be missed.
-
Comment on Language is a poor heuristic for intelligence in ~comp
lel (edited )Link ParentA point the author spends good portion of this article proving is that the creators of these models themselves have put a lot of time and effort into deliberately blurring the line between AI and...A point the author spends good portion of this article proving is that the creators of these models themselves have put a lot of time and effort into deliberately blurring the line between AI and AGI, and trying to market these products as AGI, or at least consciously trying to evoke the aesthetics of it. She talks about a number of choices that these companies make to try to suggest to end users that the AI is more than it is. I suspect a lot of this is because of how much we used to hear about the Turing test. AI research used to be very obsessed with creating an AI that could appear and act human to the end user to pass the Turing test, and since we didn't have anything as advanced as LLMs yet, the way you did that was primarily through presentation and smoke and mirrors and language, which is how we've gotten where we've gotten.
Regardless, the effect is that most layman end users come out of an interaction with ChatGPT et al thinking that this is something qualitatively similar to AGI that just has to be iterated on to get the Star Trek computer. The Snapchat AI was a particularly good test of this for me because I got to watch dozens of friends have their first experience with an LLM and come away from it with their own impressions. And I don't know anyone who came away from that experience thinking correctly that the way it works is that it can't do anything but mimic language patterns, and a side effect of mimicking language patterns is that sometimes the language it's mimicking happens to have correct information in it. Instead I saw a lot of people come away thinking they can ask an LLM a question and get a correct answer. On the other hand, I'm in law school, and my law school friends by and large just decided it was broken and useless and not-quite-there-yet because basically any question you ask any of the existing LLMs about anything related to the law causes hallucinations. And yet even they were thinking like "wow, well once they add information about the law to the databases and teach it about the law this will be really cool," not understanding that teaching an LLM information or adding information to a database is fundamentally incompatible with how LLMs work.
And ultimately this feels like it's just a stupid semantic argument about whether AI is really AI or not really AI and a different word means what some people mean when they say AI bla bla bla bla bla. But I don't think that's really the crux of the argument the author is making, which is important in a world where people bullish on AI regularly write op-eds or get hundreds of thousands of Twitter likes discussing the integration of LLMs into education. Laypeople en masse have a fundamental misunderstanding of how LLMs work that has been deliberately promulgated primarily by people that are financially invested in the proliferation of LLMs, and that fundamental misunderstanding is driving public opinion on policy related to them. That seems bad!
-
Comment on Language is a poor heuristic for intelligence in ~comp
lel Link ParentI don't think norb is necessarily saying that consciousness doesn't come from the brain, I read that instead as a reference to what you call the "inaccessibly complicated" nature of it. We have a...I don't think norb is necessarily saying that consciousness doesn't come from the brain, I read that instead as a reference to what you call the "inaccessibly complicated" nature of it. We have a very abstract and high level descriptive model that tells us a strong relationship between the brain and consciousness exists, and we can use it to make very broad predictions, but that only gets you so far. Anyone can gather patient reports about different types of brain damage and subjective experiences, and use those reports to predict that the next guy with a substantially similar kind of brain damage will have a substantially similar subjective experience, but that doesn't really get you anywhere near an understanding of the connection between those two things.
-
Comment on Red Dead Redemption and Undead Nightmare coming to Switch and PS4 in ~games
lel Link ParentI'm hopeful it will emulate it well. I've seen a lot of doom and gloom about that because of how the GTA ports went, but I think it's generally accepted that the problems with those came from...I'm hopeful it will emulate it well. I've seen a lot of doom and gloom about that because of how the GTA ports went, but I think it's generally accepted that the problems with those came from Grove St, who apparently isn't involved in this, trying and failing to remaster it, whereas this is just a straight port. The Switch has far more than enough power to run a 360 game, so unless they really butcher it I have to imagine it'll run okay.
-
Comment on ‘Barbie’ reaches US$1 billion box office sales globally; 'Oppenheimer' is the biggest World War II pic of all time at US$552.9 million. in ~movies
lel (edited )Link ParentOh yeah, for sure, I guess the real thing here is that WWII movies basically just vanished somewhere in the 2000s/2010s, and almost all of the highest grossing movies are from the last ten years,...Oh yeah, for sure, I guess the real thing here is that WWII movies basically just vanished somewhere in the 2000s/2010s, and almost all of the highest grossing movies are from the last ten years, because that's when studios shifted to the billion dollar return model you mention. Like I guess my point is just that if you just peek the Wikipedia highest grossing movies list, it's pretty shocking that there isn't a single (as far as I can tell) WWII movie on there (or even stranger, a single historical war movie, I'm pretty sure???) given just how reliable a cash cow WWII used to be for Hollywood. A lot of that is obviously inflation and a lot of that is that Hollywood has changed its model but the premise of each of those arguments is that studios haven't been making WWII tentpoles either since before the entire film market changed in the 2010s or since long enough ago that inflation is relevant, which probably shouldn't be surprising to me but somehow still is.
-
Comment on ‘Barbie’ reaches US$1 billion box office sales globally; 'Oppenheimer' is the biggest World War II pic of all time at US$552.9 million. in ~movies
lel Link ParentIt certainly says something about the way the market has shifted in the modern era that WWII films, which used to be really common tentpole movies, aren't anywhere in the list of highest grossing...It certainly says something about the way the market has shifted in the modern era that WWII films, which used to be really common tentpole movies, aren't anywhere in the list of highest grossing movies of all time. Like, the fact that the highest grossing WWII movie of all time wasn't even the highest grossing movie released that day is pretty shocking to me.
-
Comment on Among US whistleblower claims: Nonhuman biologics have been recovered in ~news
lel (edited )Link ParentThat's what the guy says when he isn't under oath. He has explicitly made these claims before. When people say that he refused to give his most extreme claims under oath, this is what they're...That's what the guy says when he isn't under oath. He has explicitly made these claims before. When people say that he refused to give his most extreme claims under oath, this is what they're referring to. He's also said the aliens are from another dimension and has spun some wild stories about the secret battle between Pope Pius XII and Mussolini over alien remains that Mussolini got from a crash site. Buddy's a crank.
-
Comment on Among US whistleblower claims: Nonhuman biologics have been recovered in ~news
lel Link ParentI don't think that's the alternative. There's a lot of wiggle room between "there are no aliens because civilizations will inevitably destroy themselves" and what the whistleblower claims, which...I don't think that's the alternative. There's a lot of wiggle room between "there are no aliens because civilizations will inevitably destroy themselves" and what the whistleblower claims, which is that the Vatican and every world government have teamed up to assassinate people who want to share the truth about aliens, and even if we accept that those are the only two positions I'm really not sure which one is more depressing.
-
Comment on Among US whistleblower claims: Nonhuman biologics have been recovered in ~news
lel Link ParentWorth noting that they're getting a ton of play on Fox News the last 24 hours too.Worth noting that they're getting a ton of play on Fox News the last 24 hours too.
It's so interesting how AI written text always feels like AI written text, no matter who it's told to copy. I think part of it is its unwavering but utterly neutered positivity, like an HR email or something, but that doesn't capture the whole thing. Part of it is the way it structures every output like a middle school hamburger essay -- it goes so far as to say "In conclusion" to start the last paragraph of the "New York Times" article. And when it's told to try to be someone funny, the way it does that is by tactically inserting what it thinks are jokes at regular, precise intervals, like someone writing their first ever standup set after doing a full scansion of a Bill Hicks appearance or something. But even when someone does that, it feels clumsy, not deeply unnatural in the way AI's attempts are. And usually its attempts at being funny are just saying things in a weird way or using a goofy word like a middle schooler.
I dunno. The most intriguing part about AI text to me is that I don't think I've ever seen an AI text output that doesn't ooze "this is an AI text output" from every line, in the same way I don't think I've ever seen an AI image output that doesn't ooze "this is an AI image output" from every pixel. But the human uncanny valley spidey-sense is really only geared toward images, so the uncanny valley presents itself far differently in text form with AI text than it does with AI generated images.