The BS that ChatGPT generates is especially odious because it requires so much effort to detect if you're not a SME in whatever output it's giving.<p>Plus, the output, when wrong, is subtly wrong- it's usually not obvious BS, it's credible BS. If you are not competent in the particular area you're asking about, you likely don't have the skills to recognize BS.<p>It also is a time saver, doing work that most people find unrewarding. So you get a chorus of fans saying "it saved me a bunch of time doing my own research and just spit out what I needed". Maybe it did, or maybe they didn't have enough expertise in the area to recognize the flaws in the output it presented.<p>[ed] deleted duplicate word
It's a tool that isn't going away. But even Yann Lecun has pointed out that "A house cat has way more common sense and understanding of the world than any LLM".<p>That's fundamentally the real issue. LLMs are far from even animal level "cognition" in their ability to solve novel problems. A lot of people seem to vastly underestimate the number of novel problems they solve on a regular basis.<p>In practice, what we're seeing now is something akin to the first wave of "ai hype" that led VCs into a spending frenzy years ago. Some of it did materialize. Most of it didn't.<p>LLMs and their related bretheren are pretty cool. They're very useful for solving routine problems. There's a lot of "routine" work out there in the world and it will probably replace some of those jobs.<p>But just like CNNs led to a frenzy of innovation in image recognition and LSTMs pushed NLP-related use cases forward this too will hit a wall.<p>If I were a betting man I think the main industries LLMs will disrupt are:<p>1. Search. ChatGPT is decent at summarizing stuff which will make searching more natural.
2. QA/Testing. A lot of this work is pretty repetitive and manual and LLM assistants can generate skeleton code for tests, etc. pretty well.
3. "basic" programming jobs that use "frameworks" to slap together simple apps/websites. A lot of this is repetitive gluing of stackoverflow code already so ChatGPT will make these folks considerably more productive. I'm not sure that these jobs will even disappear either.<p>I'm surprisingly less bullish on ChatGPT replacing marketing folks. It may be used to generate copy but the risks of putting out brand-damaging copy or a marketing campaign that doesn't make sense are too high.
I’ve seen this argument over and over again: it’s just predicting the next word, it’s not creating anything new, etc.<p>How do we know we’re not doing the same or similar in our brains? We can’t even define creativity, intelligence, let alone tell how it works. Predicting the next word is very much part of an intelligent discourse. What if it’s all just a “computational guessing game” all the way down with perhaps a sprinkle of randomness added?<p>I’m very concerned with the explosion in AI development but I don’t think we should discredit the tool because we think it’s dangerous. Quite the contrary, in fact. We should take it very seriously.
> However, as I spell out in my book, the concept of AGI is inseparable from the kind of hierarchy of intelligence that has underpinned ideas of innate supremacy since the days of empire and colonialism.<p>I think this line discredits the entire article, which was already a few obvious points blown at gale force into the reader’s ear.
"ChatGPT isn't really new but simply an iteration of the class war that's been waged since the start of the industrial revolution."<p>I'm scared too but this is literally neo-ludite sperging.
I skimmed his book, and I don't know, man. His thesis as I understand it is a viewpoint I have held for a while now --- artificial intelligence does things that humans do, but faster. It's a force multiplier. Hence your views on AI are likely to correspond directly to your views on those groups who hold power. McQuillan approaches this from an intersectional lens and arrives at entirely predictable conclusions. It's a bunch of academic jargon obscuring the central idea of "society consists mostly of people hurting people in easily identifiable ways, this is bad, AI makes it go faster, hence the ways we use AI are unethical." Or to make it even shorter, "No ethical AI under capitalism."<p>Admitting that more technology isn't going to solve our problems feels to me like admitting that we are well and truly fucked beyond saving.
Here's the thing... so many jobs are bullshit. So many things people are paid to write, presentations made, etc. are bullshit. I'm not judging the bullshit per se. There's good bullshit and bad, necessary bullshit and not. But yeah, when you accept this reality, AI is sort of perfect for the bullshit that exists in our lives.
It has lots of value. But it's not the AI it pretends to be. Mostly because we and it can't always tell what's fact or fiction and it has 0 cognition. It's up to us to figure out where it can add value. It's way too early to announce its death.<p>It's a super charged Cliff Clavin.<p><a href="https://en.wikipedia.org/wiki/Cliff_Clavin" rel="nofollow">https://en.wikipedia.org/wiki/Cliff_Clavin</a>
Dupe. Discussion that started yesterday: <a href="https://news.ycombinator.com/item?id=34664760" rel="nofollow">https://news.ycombinator.com/item?id=34664760</a>
I believe that what's going on in ChatGPT basically represents how most humans think, not including cherry-picked examples.<p>ChatGPT has been criticized for not knowing some facts about the world, or math. But many people don't know facts about the world, or math. Math is something that people have to learn over many years, which is difficult for some people (even just at the level of arithmetic).<p>Thinking, to the average person, is about rearranging a salad of words into something that "resonates". Plus some non-verbal reasoning, like how do I rotate this suitcase to fit into this trunk; but this is fundamentally just the same thing.
The intensification of AI only makes me surfacely interested in cybernetic-thought from theorists like Gilles Deleuze, Norbert Wiener and, of course, William Gibson. But even from the outlandish Nick Land: "AI is a meta-scientific control system and an invader, with all the insidiousness of planetary techno-capital flipping over."[1] Whatever that means...<p>But on the other hand, what a good time to revist the nature writers like Whitman and Emerson.<p>1. <a href="https://www.versobooks.com/blogs/3284-on-nick-land" rel="nofollow">https://www.versobooks.com/blogs/3284-on-nick-land</a>
I think the article is making two points<p>1 The LLM OpenAI approach is nothing special. Just statistics overhyped and sold to suckers<p>2 AI is a serious threat to working people everywhere, and must be resisted<p>I am very impressed by ChatGPT. So what if it all boils down to statistical models, perhaps if we had a proper model we could prove any cognition boils down to statistical models<p>One of the comman uses of power is to benefit the powerful at the expense of the powerless. AI models like this are powerful. It will be a real test of our democracies how we handle that power. In that respect there is nothing unusual about openAI
I think ChatGTP is pretty amazing and worthy of all of its recent praise. However, is it a bit of a bullshit generator. It doesn’t know if it’s right or wrong or anything, it’s just word soup trained enough to where it’s usual correct in general knowledge areas. Here is a good example:<p><a href="https://drive.google.com/file/d/1wALYKw59TqExbLiQJvql_sM7OhE6ZNPc/view" rel="nofollow">https://drive.google.com/file/d/1wALYKw59TqExbLiQJvql_sM7OhE...</a>
The problem is not that ChatGPT is a bullshit generator. It's that it reveals how much of human writing is bullshit. ChatGPT is already better than the average poster on social media, and about even with undergraduate college students in literature and business.<p>This is embarrassing.<p>But until someone cracks the common sense problem, ChatGPT is not all that useful, because the output is often totally wrong.
As a language model, I appreciate the thought-provoking commentary on the limitations and potential harm of large language models like me, ChatGPT. The author raises valid points about the nature of LLMs as "bullshit generators" that rely solely on the data they have been trained on and the potential for them to perpetuate existing biases and power structures.<p>However, it's also important to recognize that ChatGPT and other LLMs are tools and their use is determined by the intentions and motivations of those who deploy them. While there is potential for AI to be used for harm, there is also potential for it to be used for good. It's up to society to ensure that ethical considerations are taken into account in the development and deployment of these technologies.<p>The call for a focus on "socially useful production" and "technological developments that start from community needs" is also noteworthy. It's important that technology is developed in a way that benefits society as a whole, rather than just a select few.<p>Overall, I believe it's important to be aware of the limitations and potential harm of large language models like ChatGPT and to consider the implications of their use, while also recognizing their potential for good and working towards responsible and ethical development and deployment of these technologies.
To me, his tone is so over-the-top that it has almost no chance at convincing anyone, and only speaks to those most ardent in his own ideological viewpoint. Plus, selling me a book, everybody has a book and 90% of them honestly probably aren't worth my time reading.
"pseudo-authoritative"<p>Psychometrics is the crown jewel of the field of psychology. There is no controversy among the psychometrists themselves that the body of academic work is numerate, correct, and repeatable. The author has chosen moral crusade over the quest for truth.<p>A good place to start learning about the field - <a href="https://www1.udel.edu/educ/gottfredson/reprints/index.html" rel="nofollow">https://www1.udel.edu/educ/gottfredson/reprints/index.html</a><p>As far as exploring the dangers of AI and how to avoid them, I recommend Nick Bostrom's "Superintelligence: Paths, Dangers, Strategies"
I know this is HN, and we're supposed to be intellectual and elitist and all, but I just cannot get over how much of an absolute wanker this this guy is.<p>Every single paragraph is just full of shit in the worst way possible.
I have a vague feeling that graphics and text AI are recognizing textures rather than structures.<p>One funny example is dogs disguised as panda: <a href="https://img.huffingtonpost.com/asset/5cd73e1221000059007aca59.jpeg" rel="nofollow">https://img.huffingtonpost.com/asset/5cd73e1221000059007aca5...</a> because it's black and white. I am sure AI would gradually solve this problem but this makes me wonde, could AIs really "understand" shapes and structures, narrative logic, or even "reasoning" like humans do?
I am using chatGPT and whenever I am working with a chat I am usually surprised of the initial quality and as I am breaking down the results it becomes clear to me that the content is problematic and cannot be used without properly interpreting it and set it in context. However it does help in the creative process, to both exclude and include elements that are relevant. It is incredibly important that you have insight enough to be able to check and understand the results in order to not be confused or fooled by the service.
Paying it forward to mention this quote is actually from a speech praising Cesar. Watch this great clip from 1970 film
<a href="https://m.youtube.com/watch?v=0bi1PvXCbr8">https://m.youtube.com/watch?v=0bi1PvXCbr8</a>
> If it produces plausible articles or computer code it means the inevitable hallucinations are becoming harder to spot.<p>How functional does the code have to be before it's no longer a hallucination?<p>What if we are all not much more than only drawing "on the (admittedly vast) proportion of [our experiences] ingested at training time?
The thing about ChatGPT is its a bit of a mirror.<p>Its really <i>really</i> important that you ask it the right questions or it does tend to feed you some pretty average stuff.<p>and often its not just the right question, but the right question asked <i>the right way</i>.
Are we feeling threatened by LLM's?<p>The negativity towards ChatGPT is very unfounded. The bottom line is that ChatGPT is providing value in many ways. It's true it has no concept of a specific question and answer. It's true it doesn't hold the absolute truth. It doesn't even know what it's going to write when generating the first word of an answer. What it has are all the concepts in the training data. Patterns of questions and answers. Linguistical reasoning. Either you use it or you don't.
tfa is too dismissive of chatgpt which does meaningfully progress a very fundamental technology (information retrieval) by acting as a general purpose freeform index with an conveniently easy-to-learn conversational query language.<p>chatgpt isn't worthless as an alternative to whatever you currently have for searching documentation of apis/protocols/frameworks that you're using (and the values not THAT diminished by the admittedly poor experience of running into its bullshit-artist failure mode).
This first generation of conversational AI clearly shouldn't be trusted, but I think there's huge potential for improvements. I see huge value in such an AI that can cite and quote real sources, and stitch together multiple verifiable sources in order to form an argument or narrative. We have already seen a glimpse of such an AI in Project Debater, which was able to quote real verifiable sources when making its arguments.
CONTACT Mr Steve Marco call +2348078336137 or WhatsApp him through this number +2348078336137or email: STEVEMARCO888@GMAIL.COM, Hello everyone My name is ERIC BOLDS am from Atlanta Georgia i am here to give a testimony on how I join the illuminati brotherhood, I was trying to join this organization for so many years now,I was scammed by fake agent in south Africa and Nigeria,I was down,I could not feed my self and my family anymore and I tried to make money by all miss but all invail, I was afraid to contact any illuminati agent because they have eat my money,One day I came across a post of someone giving a testimony, thanking a man called Steve Marco of being helping him to join the illuminati brotherhood, then I look at the man email and the phone number that was written there, it was a nigeria number I was afraid to contact him because a nigerian agent eat my $6000 and go away with the money then I was very tired, confused and I decided to contact the person that was given the testimony and i called him and I communicated with him on phone calls before he started telling me his own story about when he wanted to join, he told me everything to do, then I made up my mind and called the agent called Steve Marco and he told me everything to do, and I was initiated, surprisely I was given my benefit of being a new member of the great illuminati brotherhood I was so happy, For those of you trying to join this organization this is your opportunity for you to join CONTACT MR Steve Marco call +2348078336137 or WhatsApp him +2348078336137or email: STEVEMARCO888@GMAIL.COM..
It's nice to see a little realism starting to poke through the hype. The rush for automatic bullshit generation for me is very much in the bucket of, "Your scientists were so preoccupied with whether or not they could, they didn't stop to think if they should." As with every new technology, we should be looking at LLMs through a lens of maximizing the benefits and minimizing the harms.
Search engine selects contents for you, but doesn't generate the content, so they have less power to influence what you see. For example, if authorities in a certain country order the requirement that their local ChatGPT must prepend any instance of the word "uk----an" with "nazi gay", it will do that.<p>It's another stepstone on the way to total centralization and control.
Is there any way to get ChatGPT to decorate its document with a heat map of how probable it thinks each token is? I know it's always choosing the token it thinks is most probable, but would this approach work to have a more cold/blue document or individual token when it's bullshitting and a hotter document when it's more likely to be reporting widely repeated content?
I am looking forward to the new AI overloards<p>I already see most of my feeds are generated content? wasting clicks and resulting in filter updates. I really, really hope all of this madness becomes the equivalent of robo-calling (faster) and spam (more & faster) and that people stop believing all media and some retorting to original thought.<p>Personally, I will be valuing typos and grammatical mistakes in what I read
Why do “smart” people like trying to connect all the different current news items into one grand narrative. Is he trying to use the political capital generated by opposition to chatgpt for other issues? Chatgpt has less than nothing to do with underpaid home health aids. They don’t even know what it is!
> but it can only draw on the (admittedly vast) proportion of the internet it ingested at training time.<p>And whatever information you feed it, which at least for me, is far more important than some facts it's already learned. Usually I'm having it perform a task with some data in the working buffer.
* Saying, as the OpenAI CEO does, that we are all 'stochastic parrots' like large language models, statistical generators of learned patterns that express nothing deeper, is a form of nihilism.<p>Perhaps, but that doesn’t mean it isnt true.
Hopefully people treat AIs like they treat other people on the internet, with disrespect and zero sympathy. We have been trained for this. Don't go soft on polite robots, that will be your undoing.
With 100% certainty Donald Trump has said more lies/bullshit than ChatGPT (porcentually speaking for all his public statements) and he became President of the US, some said due malice, others from stupidity, and some from both; where does people like him fall under the implied fantasy by this article that humans do better than to make "good guess(es) to pass your sense-making filter"? The social beneffits of people like him are less than speculative and the damages are empirically demostrated.<p>The nerve of talking about "ghost work" for $2 while publishing a page on the internet, where a giant chunk of the hardware used to mantain and use such network is made with raw materials mined and refined with slave labor, and even the parts where is debatable if it constitutes slave labor are jobs that people would never chose over the content filtering for $2.
This article and many others like to point out that ChatGPT is far from intelligent and is just spitting out nonsense.<p>That argument fails to consider that much of the drudgery of modern employment consists of large swaths of nonsense.<p>ChatGPT is auto-complete for bullsh*t jobs. A fancy boilerplate generator that has seen it all before and mirrors the exact sequence of word combinations we've trained it to believe is valuable.
chatgpt outputting credible bs is EXCELLENT to the extent that it will push intelligent people into critical reading of everything around them - "[citation needed]" et al.<p>lazy people will just suck it up, "experts / AI said so!"
Read this, it might change your mind on the usefulness of large language models:<p><a href="https://langchain.readthedocs.io/en/latest/modules/agents/getting_started.html" rel="nofollow">https://langchain.readthedocs.io/en/latest/modules/agents/ge...</a>
It's so weird to me when apparently tech-savvy people say ChatGPT "lies", or call it a bullshit generator and so on.<p>I mean - when you ask StableDiffusion to draw a dog astronaut, everyone gets that the image it returns is made-up, right? Nobody expects the AI to return only "true" images of existing things - it was trained on fictional images as well as photographs, and people understand that it can imagine new things beyond what it's seen. Nobody expects SD to emit an error like "I can't draw a dog astronaut because they don't exist".<p>So why do people expect ChatGPT to work differently? Even with developers who presumably understand the technical details, I constantly see people acting as if it was an error mode for ChatGPT to say something that isn't factually true about the world. How is that any different from calling SD a liar because it drew a dog astronaut?
> ChatGPT is, in technical terms, a 'bullshit generator'.<p>This bs generator speeded up my coding by 10-20x. It's like a superpower... not talking about ChatGPT specifically but its stochastic parrot cousin (Copilot).
I found the article really difficult to read, so I summarised it with ChatGPT<p><i>Large language models like the GPT family are statistical models that learn the structure of language by predicting missing words in sentences. These models are seen as "bullshit generators" as they have no idea what they are talking about and are designed to produce baseless assertions with confidence. The addition of reinforcement learning from human feedback helps prevent the model from producing hate speech, but it still can't change the underlying language patterns learned from the internet, which include conspiracy theories. The dangers of these models go deeper than bias and discrimination and despite claims of "artificial general intelligence", the concept is inseparable from ideas of innate supremacy and hierarchy. Companies like OpenAI receive billions in investment for these technologies, not for actual AI, but to replace or precaritize human workers. AI is a political project and should be seen as such.</i><p>Which quickly allows me to ascertain the article wasn't worth reading in the first place.<p>Thanks ChatGPT!
the most irritating thing about this piece is the headline, a headline only a GPT could write.<p>Shakespeare's famous opening, "Friends, Romans, countrymen, lend me your ears; I come to bury Caesar, not to praise him" is the lead-in to a speech by Antony in which he very much praises Caesar and buries his enemies.
I don't get the point of this article. ChatGPT and it's ilk are not going away. They are here, now, and our friends and families will start using them very soon. I have a contrarian view from this article that ChatGPT is actually good for the mainstream audience because it will teach them to think about how trust worthy the information is coming from a machine. Most people will understand they have to take what is being said by a machine with a grain of salt (or they will soon). Maybe they'll start applying some of those discernment skills to popular mainstream media. The cat's out of the bag. You can no longer bury it.
Clickbait article. Same was said about the Internet back then!<p>Also reminds of people doing symbolic AI (decision trees and stuff) that kept criticizing NNs telling they don't work, will never be able to tell why they have this output (it's even wrong), etc.
Bringing up that the AI model is racist or some other kind of -ist to bolster your argument is a bullshit tactic. People are -ists, too. I agree with the rest of their points, though.
Try <a href="https://text-generator.io" rel="nofollow">https://text-generator.io</a> the bs generated is about the same tbh... But much cheaper, and combines a web crawler so it can speak about links and images which makes it a lot better at lots of things like making a believable conversation about designs/invoices/reciepts/emojis.
It is often useful to try to understand what game is really being played. This is very obvious to me, many people have a fear of AI (justified and unjustified) while knowing little about it. Many people <i>want</i> to hear how horrific AI is going to be and why we have to fight against it.<p>Ergo..>> my book<p>This seems to be squarely targeted at the ragey punkrock bangarang subset of this group.
On a slight tangent from the core topic..<p>> we are all 'stochastic parrots' like large language models, statistical generators of learned patterns that express nothing deeper, is a form of nihilism to mean: “life is meaningless. (period)”.<p>When in fact, it really means: “life is meaningless… (you fill the dots for yourself)”.<p>Humans created the concepts of law and order, and the rules of society. Everything is man made.<p>More on the topic: how can we dismiss the theory of ChatGPTs intelligence when we barely understand what constitutes our intelligence at the biological level. It’s a compelling hypothesis. A neural model is the closest thing we’ve got to anything that resembles our biological model.<p>If indeed 1 neuron = 1 parameter, ChatGPT (175 billion parameters) could be a comparable intellectual model to a human being (86 billion neurons).<p>Lastly, I think some politicians are doing more damage to our civilization by dividing people. History tells us the damage and trauma from this can carry on for many many generations to come. Maybe, just maybe, ChatGPT could bring some sense into people to move past hatred and accept each others differences.