There is a machine learning bubble, but the technology is here to stay. Once the bubble pops, the world will be changed by machine learning. But it will probably be crappier, not better.
What will happen to AI is boring old capitalism. Its staying power will come in the form of replacing competent, expensive humans with crappy, cheap robots.
AI is defined by aggressive capitalism. The hype bubble has been engineered by investors and capitalists dumping money into it, and the returns they expect on that investment are going to come out of your pocket. The singularity is not coming, but the most realistic promises of AI are going to make the world worse. The AI revolution is here, and I don’t really like it.
You could have said the same for factories in the 18th century. But instead of the reactionary sentiment to just reject the new, we should be pushing for ways to have it work for everyone.
I don’t see how rejecting 18th century-style factories or exploitative neural networks is a bad thing. We should have the option of saying “no” to the ideas of capitalists looking for a quick buck. There was an insightful blog post that I can’t find right now…
Lets not forget all the exploitation that happened in that period also. People, even children, working for endless hours for nearly no pay, losing limbs to machinery and simply getting discarded for it. Just as there is a history of technology, there is a history of it being used inequitably and even sociopathically, through greed that has no consideration for human well-being. It took a lot of fighting, often literally, to get to the point we have some dignity, and even that is being eroded.
I get your point, it’s not the tech, it’s the system, and while I lost all excitement for AI I don’t think that genie can’t be put back in the bottle. But if the whole system isn’t changing, we should at least regulate the tech.
But AI will eliminate so many jobs that it will affect a lot of people, and strain the whole system even more. There isn’t a “just become a programmer” solution to AI, because even intellectually-oriented jobs are now on the line for elimination. This won’t create more jobs than it takes away.
Which shows why people are so fearful of this tech. Freeing people from manual labor to go to intellectual work was overall good, though in retrospect even then it came at a cost of passionate artisans. But now people might be “freed” from being artists to having to become sweatshop workers, who can’t outperform machines so their only option is to undercut them. Who is being helped by this?
Yes, I know about the exploitation that happened during early industrialization, and it was horrible. But if people had just rejected and banned factories back then, we’d still be living in feudalism.
I know that I don’t want to work a job that can be easily automated, but intentionally isn’t just so I can “have a purpose”.
What will happen if AI were to automate all jobs? In the most extreme case, where literally everyone lost their job, then nobody would be able to buy stuff, but also, no company would be able to sell products and make profit. Then, either capitalism would collapse - or more likely, it will adapt by implementing some mechanism such as UBI. Of course, the real effect of AI will not be quite that extreme, but it may well destabilize things.
That said, if you want to change the system, it’s exactly in periods of instability that can be done. So I’m not going to try to stop progress and cling to the status quo out of fear what those changes might be - and instead join a movement that tries to shape them.
we should at least regulate the tech.
Maybe. But generally on Lemmy I see sooo many articles about “Oh, no, AI bad”. But no good suggestions on what exactly regulations should we want.
Movements that shape changes can also happen by resisting or by popular pressure. There is no lack of well-reasoned articles about the issues with AI and how they should be addressed, or even how they should have been addressed before AI engineers charged ahead not even asking for forgiveness after also not asking for permission. The thing is that AI proponents and the companies embracing them don’t care to listen, and governments are infamously slow to act.
For all that is said of “progress”, a word with a misleading connotation, once again this technology puts wealthy people, who can build data centers for it, at an advantage compared to regular people who at best can only use lesser versions of it, if even that, they might instead just receive the end result of whatever the technology owners want to offer. Like the article itself mentions, it has immense potential for advertising, scams and political propaganda. I haven’t seen AI proponents offering meaningful rebuttals to that.
At this point I’m bracing for the dystopian horrors that will come before it all comes to a head, and who knows how it might turn out this time around.
Like the article itself mentions, it has immense potential for advertising, scams and political propaganda. I haven’t seen AI proponents offering meaningful rebuttals to that.
You won’t get a direct rebuttal because, obviously, an AI can be used to write ads, scams and political propaganda.
But every day millions of people are cut by knives. It hurts. A lot. Sometimes the injuries are fatal. Does that mean knives are evil and ruining the world? I’d argue not. I love my kitchen knives and couldn’t imagine doing without them.
I’d also argue LLMs can be used fact check and uncover scams/political propaganda/etc and can lower the cost of content production to the point where you don’t need awful advertisements to cover the production costs.
This knife argument is overused as an excuse to take no precautions about anything whatsoever. The tech industry could stand to be more responsible about what it makes rather than shrugging it off until aging politicians realize this needs to be adressed.
Using LLMs to fact check is a flawed proposition, because ultimately what it provides are language patterns, not verified information. Nevermind its many examples of mistakes, it’s very easy for them to provide incorrect answers that are widely repeated misconceptions. You may not blame the LLM for that, you can scratch that to generalized ignorance, but it still ends up falling short for this use case.
But as much as I dislike ads, that last one is part of the problem. Humans losing their livelihood. So, going back to a previous point, how does the lowered ad budget help anyone but executives and investors? The former ad workers get freed to do what? Because the ones focused on art or writing would only have a harder time making a career out of that now.
Accelerationists?
You could have said the same for factories in the 18th century.
Everyone who died as a result of their introduction probably would say the same, yes. If corpses could speak, anyway.
Well if you can find anyone who’s died because an AI wrote an article then I’ll concede you have a point.
Did you read the whole article including the “flame bait”? The author gives an example there of someone committing suicide because an AI encouraged them…
Is that the AI’s fault, or the depressed and suicidal human’s fault?
Do you not think that the person would have committed suicide whether they asked the AI or not? The AI might have sped up the decision, but it is the human who made it.
It is not like the AI is out there trying to convince non-depressed humans to become depressed in order they go kill themselves …
Well, in the linked article the wife of this person said that they wouldn’t have committed suicide without the AI facilitating it. So yes, I would say it is at least in part the AI’s fault. And no, I didn’t say it was the intention of the AI to do so. But that doesn’t mean it won’t do it at all.
You seem to really wanna push AI and lose your empathy over this…
If the technology actually existed to replace human workers, the human workers could chip in and buy the means of production and replace the company owners as well.
If the humans are replaced, how will they afford to buy what the company owners aren’t selling?
deleted by creator
Top quality luddite opinions right here. Plenty of fear and oprobium being directed against the technology, while taking the kleprocratic capitalism and kakistocracy as a given that can’t be challenged.
That seems to be the theme of the era.
Yes, it is incompatible with the status quo. That’s a good thing. The status quo is unsustainable. The status quo is on course to kill us all.
The only real danger AI brings is it will let our current corrupt leaders and corrupt institutions be more efficient in their corruption. The problem there is not the AI; it’s the corruption.
Improving human efficiency is essentially the purpose of technology after all. Any new invention will generally have this effect.
deleted by creator
What? It’s the sociological definition of technology. A cultural tool which is used by a community for making a task xyz, easier, faster, more efficient.
Efficiency is an extremely broad term.
What’s your counter definition of technology and efficiency that is leading you to disagree?
deleted by creator
Extra spicy take: The Luddites were right. They were really always about opposing unethical use of technology, people who use their name as an insult were always all about “progress over people”, and you should never feel bad for being called a Luddite.
deleted by creator
Extra spicy, hut the kind of spicy you get on your anus after shitting out a too-hot meal
What are those people doing to you?
There are definitely people who are harmed by FUD like this. For example the current writers strike, which has 11,000 people putting down tools… indefinitely shutting down global movie productions that employ millions of people and leaving them unemployed for who knows how long.
deleted by creator
I don’t have anything against you or your colleagues. You’ve got every right to strike if that’s what you want to do.
But there are millions of people being harmed by the strike. That’s a simple fact.
Journalists/etc need to do their job and provide good balanced information on critical issues like this one. FUD like Drew Devalt posted inflames the debate and makes it nearly impossible for reasonable people to figure out what to do about Large Language Models… because like it or not, they exist, and they’re not going away.
PS: while I’m not a film writer, I am paid to spend my day typing creative works and my industry is also facing upheaval. I also have friends who work in the film industry, so I’m very aware and sympathetic to the issues.
deleted by creator
Why are they so threatening?
Simple example: A lot of artists would like that their images aren’t used for AI training and would like to have legislation to prevent that. Problem with that is that such legislation would grand a monopoly on AI to the Google’s, Facebook’s and Adobe’s of this world, as they are already sitting on mountains of data and have ToS that allows them to use it for training. Any Open Source project that doesn’t have the data and would need to web scraping would be illegal.
That’s the issue. A lot of criticism on AI is extremely short sighted and ignorant, often not even understanding the very basics how it all works.
Another more fundamental problem: What are you going to do? AI is just a collection of algorithms and math. Do you want to outlaw math? Force humans to use less efficient tools? Technological progress is not something you can easily control, especially not in advance when you don’t even know what changes it will bring.
Imagine if we had taken an extra five minutes before embracing Facebook and all the other social media that came to define “Web2.0.”
We did and nothing ever came of it. Projects like https://freenet.org/ or https://freedombox.org/ have been around for a decade or two. But the masses want convinience.
deleted by creator
we are spreading FUD
You are spreading FUD right in this post: “poorly regurgitating our shit without our permission”
- There is no need for permission, that’s not how copyright works
- It’s not “poorly regurgitating”, it’s for most part original and of high quality
You wanna be taken serious? Stop repeating the nonsense as everybody else.
And as for “replace us at our jobs”, that’s not a problem, that’s called progress. If you want UBI or something along the line, go fight for that, don’t make stupid arguments against AI.
That does not mean everyone’s critique is ignorant.
Well, maybe, still waiting for reading one that isn’t. Since everybody just keeps repeating the same nonsense, including you right now.
The only sensible one I heard so far was from Hinton, which simply suggested putting about equal money into AI safety research as we do into AI research. Nothing wrong with that, though that will probably just show us more ways in which AI can go wrong and less on how to prevent it.
deleted by creator
Oh yeah, I forgot copyright law was this static and beautiful thing that never gets reassessed or re-contextualized.
So you want it changed in such a way that it grands Google, Adobe and Co. exclusive ownership over powerful AI models and kills all Open Source efforts? Congratulations for proving my point.
And you wonder why nobody takes you people seriously…
Not everything anti-Ai is luddite, some is just poorly thought through or downright incorrect, this is absolutely a luddite take
But there’s also real danger here today.
For example, ‘Life or Death:’ AI-Generated Mushroom Foraging Books Are All Over Amazon
These are easily avoidable problems. There are always reputable authors on topics and why would a self published foraging book by some random person be better than an AI one? You buy books written by experts, especially when it’s about life or death.
I think the idea is that someone buying a basic book on foraging mushrooms isn’t going to know who the experts are.
They’re going to google it, and they’re going to find AI-generated reviews (with affiliate links!) of AI-generated foraging books.
Now, if said AI is generating foraging books more accurate than humans, that’s fine by me. Until that’s the case, we should be marking AI-generated books in some clear way.
Now, if said AI is generating foraging books more accurate than humans, that’s fine by me. Until that’s the case, we should be marking AI-generated books in some clear way.
The problem is, the LLM AIs we have today literally cannot do this because they are not thinking machines. These AIs are beefed-up autocompletes without any actual knowledge of the underlying information being conveyed. The sentences are grammatically correct and read (mostly) like we would expect human written words to read, however the actual factual content is non-existent. The appearance of correctness just comes from the fact that the model was trained on information that was (probably mostly) correct in the first place.
I mean, we should still be calling these things algorithms and not “AI” as “AI” carries a lot of subtext in people’s minds. Most people understand “algorithms” to mean math, and that dehumanizes it. If you call something AI, all of a sudden people have sci-fi ideas of truly independent thinking machines. ChatGPT is not that, at all.
I agree. And ML may never be able to cross that line.
That said, we’ve been calling it AI for decades now. It was weird enough to me when people started using ML more. I remember the AI classes I took in college, and the AI experts I met in my jobs. Then one day it was “just ML”. In most situations, it’s the same darn thing.
“Easily avoidable” if you know to look for them or if they’re labelled appropriately. This was just an example of a danger that autocomplete AI is creating today. Unscrupulous people will continue to shit out AI generated nonsense to try to sell when the seller does zero vetting of the products in their store (one of the many reasons I no longer shop at Amazon).
Many people, especially beginners, are not going to take the time to fully investigate their sources of knowledge, and to be honest they probably shouldn’t have to. If you get a book about mushrooms from the library, you can probably assume it’s giving valid information as the library has people to vet books. People will see Amazon as being responsible for keeping them safe, for better or worse.
I agree that generally there is a bunch of nonsense about ChatGPT and LLM AIs that isn’t really valid, and we’re seeing some amount of AI bubble happening where it’s a self feeding thing. In the end it will shake out, but before that all happens you have some outright dangerous and harmful things occurring today.
I mean, people should at least check if the publisher is reliable for any information source.
deleted by creator
People should learn how to approach information sources. If that’s not happening AI doesn’t really matter for this discussion.
deleted by creator
taking the kleprocratic capitalism and kakistocracy as a given that can’t be challenged.
It’s literally baked into the models themselves. AI will reinforce kleptocratic capitalism and kakistocracy as you so aptly put it because the very data it’s trained on is a slice of the society it resembles. People on the internet share bad, racist opinions and the bots trained on this data do the same. When AI models are put in charge of systems because it’s cheaper than putting humans in place, the systems themselves become entrenched in status-quo. The problem isn’t so much the technology itself, but how the technology is being rolled out, driven by capitalistic incentives, and the consequences that brings.
snicker drewdevault is an avid critic of capitalism. thats entirely the point of this post actually.
Then it is horrifically badly written. Maybe get an AI to give it a once over?
deleted by creator
Sure, it’s all about capitalism. Nothing good like this could ever come from advances in technology:
ML is a tool and like most tools it has broad use cases. Some of them are very, very, good.
There is a name for this debating technique where you go “sure, there was nothing good about Hitler - except he cared about dogs!”. Can’t remember. Is it strawman?
I think we all understand that capitalism is mostly bad for humans, and really good for corporations and their owners. AI and robots will be exploited to replace people since they are massively more powerful and much cheaper.
A few things will be better I guess, but most will be worse. People already are not actually needed to work this much anymore, and as soon as they can be replaced with something cheaper and more efficient they will. That is capitalism.
A strawman argument is where you ignore what was said by the other person and instead respond with something distorted. That’s not what I did - the core premise of Drew’s argument is that AI will not “make the world better” and I provided a crystal clear example of how it makes the world better.
It was just one example, and obviously not the complete picture, but what choice do I have? It’s such a broad topic I couldn’t possibly list everything AI will impact without writing an entire book.
I think we all understand that capitalism is mostly bad for humans, and really good for corporations and their owners.
No I disagree. Corporations exist exclusively to benefit their human owners them. Which means anything that’s “good for corporations” is good for a select small number of humans.
Don’t blame “capitalism” for wealth inequality. Blame the actual humans (e.g. Donald Trump, Elon Musk) who have made it their life’s work to drive the global economy even harder into a world that benefits the fiew and ignores the struggles of the many.
Also - not all corporations are bad. Some of them do great work that truly benefits the world and I would personally put OpenAI in that category. Their mandate is not to make a profit - and in fact the amount of profit they can legally make has been limited. Their mission is literally “to ensure that artificial general intelligence benefits all of humanity”. I hope they succeed, and I think they will. Drew is wrong.
Who buys the stuff if everyone is replaced?
Eventually nobody.
Capitalism isn’t about sustainability, it’s about making the most amount of profit in the shortest amount of time.
Eventuall you bleed everyone dry and nobody has a job. But for a short amount of time the shareholders will have had a huge number of 0’s and 1’s in a database somewhere equating to their “worth”
Humans are consumers, they will buy stuff. But most won’t work for corporations anymore since robots and AI are far more effective at most jobs.
Humans will still buy the stuff robots produce. Maybe the money will come from governments as some kind of citizen coins, distributed differently based on some criteria. Not sure.
That would be UBI and that’s seen as an improvement by a lot of people so why stand in the way of that? Robots do the work, people get a budget they can spend on that work while they don’t have to do it.
That would be UBI and that’s seen as an improvement by a lot of people so why stand in the way of that?
Because none of those people have policy changing influence. Nobody here is standing in the way of it, most people here are advocates for it. But we don’t write the laws, and we only get to vote on a very small percentage of them.
I’m getting so so tired of these “AI/ML bad, world is doom” articles being posted multiple times a day. who is funding these narratives??
you have no idea who drew devault is clearly and you entirely missed the point of what he posted.
I also have no idea who he is and I also missed the point. It’s just another “AI bad” article, even if the message this time is “AI bad, but not as bad as you think.”
hes pointing out how the technology is going to be used realistically by corporations. hes not saying AI is bad inherently. hes saying the outcome will be bad for society. which is very true.
deleted by creator
And its a hot topic so there are a lot of articles about it since it generates traffic. And the bigger and doomier the article, the more hype.
But lots of people are unhappy about AI, its not a narrative being funded by a secret cabal or something.
I disagree. If we replace this writer with ChatGPT4, it would generate a more balanced article.
More balanced articles are not necessarily better though. I’d dather read two conflicting opinions that are well thought out than a mild compromise with unknown bias.
I’d dather read two conflicting opinions that are well thought out
That’s where it all falls down of course. Because these opinions are anything but well thought out.
Why would ChatGPT be more “balanced,” what does “balanced” mean, and why is it better?
More balanced than what?
ChatGPT ingest lots of articles from the web and newspapers, identify patterns in the text, and generate relevant reply based on what it ingested.
I expect ChatGPT to perpetuate biases found in its training data, and don’t see how it’d improve balance.
What I mean is that the article was full of negative bias.
ChatGPT 4, when used with care, can take into account different opinions, both positive and negative.
deleted by creator
If you disregard the positive side that’s a negative bias. I’m not interested in a semantic fight with you.
deleted by creator
Maybe it was a poor choice of words but I’m honestly tired to do arguments with online people who can’t see biases. Cheers, indeed.
This is a very one sided way to look at things. Yes, people will use AI to generate spam and stuff. What it is missing is that people will also use AI to filter it all away. The nice thing about ChatGPT and friends is that it gives me access to information in whatever format I desire. I don’t have to visit dozens of websites to find what I am looking for, the AI will do that for me and report back with what it has found.
Simply put, AI is a possible path to the Semantic Web, which previously failed since ads and SoC were the driver of the Web, not information.
Sometimes I really wonder in what magical wonderland those people complaining about AI live, since as far as I am concerned, the Web and a lot of other stuff went to shit a long while ago, long before AI got any mass traction. AI is our best hope to drag ourselves out of the mud.
The real problem is that AI isn’t good enough yet. It can handle Wikipedia-like questions quite well. But try to use it for product and price information and all you get is garbage.
which previously failed since ads and SoC were the driver of the Web, not information.
Can you elaborate on why you think the ads wouldn’t sneak in again? The semantic web is a fantastic concept, but I don’t immediately see the AI connection. AI doesn’t magically pay for authored content and there is still an incentive to somehow get ads into LLM answers.
Can you elaborate on why you think the ads wouldn’t sneak in again?
You can run a LLM at home on your own PC. Think of it less as a replacement for Google and more like the computer from StarTrek. You tell it what you want and it goes to search the net for you. What you see is just the answer, in a format specified by you, not the websites they came from.
Google, Bing and Co. will of course add ads into their services, but that’s a short issue. AI will fundamentally reshape how we interact with computers and information in the long run.
The semantic web is a fantastic concept, but I don’t immediately see the AI connection.
The semantic web relies on human doing the markup, that’s doomed to fail, nobody has the time for that and even if they spend the effort, they would miss a whole lot of information that is in the text. A LLM can extract semantic information directly from the text without any markup and you can query that information with natural language. That’s not only way easier on the creators side, but also way more powerful on the users end.
You can run a LLM at home on your own PC. You tell it what you want and it goes to search the net for you.
Unless it’s open-source and connected to a proper crownsourced dataset, hosted on a paid server managed by a community instead of a big corporation, I don’t see how ads are NOT getting in.
Yes, but that’s already the case. There are numerous Open Source’ish language models around that you can run on your own PC, no server required:
- https://old.reddit.com/r/LocalLLaMA/wiki/models/
- https://github.com/ggerganov/llama.cpp
- https://github.com/oobabooga/text-generation-webui
And some of them are getting pretty damn close to ChatGPT performance:
There is of course still plenty of work that needs to be done in letting LLMs interact with the outside world, use a webbrowser and stuff, but there are projects for that as well, e.g. AutoGPT. Just a matter of time until that stuff becomes good enough to be usable.
Thank you very much. My concern is rather in the direction of inserting ads or “promotional information” into the training material, much like SEO plagues search today. If the info is from the web it can still be malicious, even if you run your own LLM.
They’ll certainly try, though it’ll be quite a bit trickier due to LLMs providing direct answers, not just a list of sites. You can’t really sneak a product in there when it doesn’t actually fit the question. I think the bigger problem is just the lack of good information out there. Finding trustworthy reviews these days is getting really hard, most of the time all you have is the product description and some Amazon reviews, which even when done well, fail at providing how product X compares to product Y. No matter how smart the AI will be, that always leaves a ton of room for error and misinformation.
Hard to tell how things will end up. For the time being, LLMs are pretty much completely useless for product search, ChatGPT just doesn’t know enough and BingChat will just summarize the first three SEO-filled Bing search results. The deep knowledge LLMs have on Wikipedia-like topics is missing when it comes to products and services, and they can’t really do calculations either, so price information is almost always wrong. This will need some specific optimization.
I don’t know why you want to use an AI to purchase goods and learn about products. That’s what the current www is really really strong at. Lots of people are spending an awful lot of money to make that information really easy to discover, and popular search engines definitely prioritise that information.
Also, if an AI is to give you price and product information it’s going to have to be reading live web pages, which will of course be full of ads. SEO will become AIO/LLMO. There is no end to the time and money advertisers are prepared to pour into getting products in front of users. The irony is that you seem to want to view products and you have this weird perspective where you’re keen to avoid ads for products so that you can view marketing information about products without the ads.
It’s already fairly hard to tell without knowing some good websites or reading through to conclusions and using some common sense whether a review website is honest or biased. I don’t know why you think an AI with access to the Internet will filter out fake reviews and content crafted to lead you to specific products over others.
Also, downloading and configuring your own AI is unlikely to be the way the “AI revolution” comes. Amazon, Google, Microsoft, Apple and other mega corporations will be funding the “AI revolution” and will not sit idly by allowing their kingdoms to crumble.
The number of people who will be saved from the corporations that run the online world by open source grass roots AI will be smaller than the number of people who are saved by Linux from proprietory products and SAAS.
Yeah, everyone will get used to using an AI to interact with the web, but it will be freely supplied by a corporation, and I PROMISE you the enshitification of AI has been long planned before we even reach step one of making it awesome for the masses.
That’s what the current www is really really strong at.
You must be using a different WWW than I am, since product search for me is absolutely terrible. Even the simplest of queries can’t be answered, e.g. something trivial as “what’s the cheapest thing that matches query” fails due to some products coming different package sizes (e.g. 100g vs 1000g). If you want to buy a movie or game, and want to know about sequels and prequels, you have to go to Wikipedia to find out, since I have yet to see a single shop that organizes that well. Or try to find the equivalent of a product in another country where the original product isn’t available. Or try to search for the cheapest way to buy multiple product at once, taking shipping cost into account. Even just figuring out the size or what’s actually in the box is often impossible, I have yet to see another site that gives you a full CAD model of the products like McMaster-Carr.
Product search on the Web is utter garbage. I am kind of surprises that nobody ever put serious effort into making that work well. Googles product search is garbage and most other search engines don’t even have a specific product search. A product search engine that automatically bundles up information from different, shops, Youtube videos and comments doesn’t exist as far as I know.
Lots of people are spending an awful lot of money to make that information really easy to discover
Amazon deliberately puts sponsored products on top to make it harder to discover what you want. Some small shops put effort into it and let you search products according the specs, but that only works in that single shop, I have yet to see a search engine that can handle that across multiple shop and with any semblance of reliability.
Also, if an AI is to give you price and product information it’s going to have to be reading live web pages, which will of course be full of ads.
Yes, but that’s irrelevant as long as only the AI reads it. I don’t care what ads my adblocker reads either.
I don’t know why you think an AI with access to the Internet will filter out fake reviews
I am not looking for reviews, but for reliable and detailed product information. An LLM can help gather that information from multiple different sources and format it in a unified way. SEO has limited influence on that, as either the product has those specs or it has not, in which case the LLM should be able to find contradictions in the information and automatically write a letter to whatever consumer protection office is responsible for false advertisement.
Also, downloading and configuring your own AI is unlikely to be the way the “AI revolution” comes.
Given the way privacy is getting traction in the public consciousness, I wouldn’t be so sure. Look at how many people already use adblockers, around 40% or so, that’s quite a lot, many of them will be upgrading to some form of AI driven adblocking and information gathering sooner or later.
You know that a LLM is a statistical word prediction thing, no? That LLMs “hallucinate”. That this is an inevitable consequence of how they work. They’re designed to take in a context and then sound human, or sound formal, or sound like an excellent programmer, or sound like a lawyer, but there’s no particular reason why the content that they present to you would be accurate. It’s just that their training data contains an awful lot of accurate data which has a surprisingly large amount of commonality of meaning.
You say that the current crop of LLMs are good at Wikipedia style questions, but that’s because their authors have trained them with some of the most reliable and easy to verify information on the Web. A lot of that is Wikipedia style stuff. That’s it’s core knowledge, what it grew up reading, the yardstick by which it was judged. And yet it still goes off on inaccurate tangents because there’s nothing inherently accurate about statistically predicting the next word based on your training and the context and content of the prompt.
Yes, LLMs sound like they understand your prompt and are very knowledgeable, but the output is fundamentally not a fact-based thing, it’s a synthesized thing, engineered to sound like its training data.
deleted by creator
Its either that or extreme fragmentation and or de- informationalization
deleted by creator
Its not a problem if its removed by improved ai; it would be a transient fear that never manifests.
deleted by creator
deleted by creator
You never wanted to have the computer from StarTrek, the Holodeck or the Universal Translator? Modern AI provides a fundamental shift in how we can interact with data and allows us to do things that would have been impossible by classic means.
And it’s not like you can escape it anyway, phone cameras use AI, spell checkers use AI, mobile phone keyboards use AI, it’s already everywhere and we have barely started.
Its staying power will come in the form of replacing competent, expensive humans with crappy, cheap robots.
Unlikely to replace the “most” competent humans, but probably the lower 80% (Pareto principle), where “crappy” is “good enough”.
What’s really troubling, is that it will happen all across the board; I’m yet to find a single field where most tasks couldn’t be replaced by an AI. Used to think 3D design would take the longest, but no, there are already 3D design AIs.
I’m yet to find a single field where most tasks couldn’t be replaced by an AI
Critical-application development. For example, developing a program that drives a rocket or an airplane.
You can have an AI write some code. But good luck proving that the code meets all the safety criteria.
You just said the same thing the comment responding to did, though. He pointed out that AI can replace the lower 80%, and you said the AI can write some code but that it might have trouble doing the expert work of proving the code meets the safety criteria. That’s where the 20% comes in.
Also, it becomes easier to recognize the possibility for AI contribution when you widen your view to consider all the work required for critical application development beyond just the particular task of writing code. The company surrounding that task has a lot of non-coding work that gets done that is also amenable to AI replacement.
That split won’t work cause the top 20% would not like to do their day job clean up AI codes. It’s much better time investment wise for them to write their own template generation tool so the 80% can write the key part of their task, than taking AI templates that may or may not be wrong and then hunting all over the place to remove bugs.
Use the AI to fix the bugs.
A couple months ago, I tried it on ChatGPT: I had never ever written or seen a single line in COBOL… so I asked ChatGPT to write me a program to print the first 10 elements of the Fibonacci series. I copy+pasted it into a COBOL web emulator… and it failed, with some errors. Copy+pasted the errors back to ChatGPT, asked it to fix them, and at the second or third iteration, the program was working as intended.
If an AI were to run with enough context to keep all the requirements for a module, then iterate with input from a test suite, all one would need to write would be the requirements. Use the AI to also write the tests for each requirement, maybe make a library of them, and the core development loop could be reduced to ticking boxes for the requirements you wanted for each module… but maybe an AI could do that too?
Weird times are coming. 😐
I’m a professional programmer and this is how I use ChatGPT. Instead of asking it “give me a script to do big complicated task” and then laughing at it when it fails, I tell it “give me a script to do .” Then when I confirm that works, I say "okay, now add a function that takes the output of the first function and does " Repeat until done, correcting it when it makes mistakes. You still need to know how to spot problems but it’s way faster than writing it myself, even if I don’t have to go rummaging through API documentation and whatnot.
I mean that is exactly what programming is except you type to an AI and have it type the script. What is that good for?
Could have just typed the script in the first place.
It ChatGPT can use the API it can’t be too complex otherwise you are in for a surprise once you find out what ChatGPT didn’t care about (caching, usage limits, pricing, usage contracts)
Could have just typed the script in the first place.
Sure - but ChatGPT can type faster than me. And for simple tasks, CoPilot is even faster.
Also - it doesn’t just speed up typing, it also speeds up basics like “what did bob name that function?”
it’s way faster than writing it myself
I already explained.
I could write the scripts myself, sure. But can I write the scripts in a matter of minutes? Even with a bit of debugging time thrown in, and the time it takes to describe the problem to ChatGPT, it’s not even close. And those descriptions of the problem make for good documentation to boot.
Fashion designers are being replaced by AI.
Investment capitalists are starting to argue that C-Suite company officers are costing companies too much money.
Our Ouroboros economy hungers.C-Suites can get replaced by AIs… controlled by a crypto DAO replacing the board. And now that we’re at it, replace all workers by AIs, and investors by AI trading bots.
Why have any humans, when you can put in some initial capital, and have the bot invert in a DAO that controls a full-AI company. Bonus points if all the clients are also AIs.
The future is going to be weird AF. 😆😰🙈
If the AI is doing a better job at each of those things, why not let it?
That’s where we need to ask how we define “better”. Is better “when the number goes bigger” or is better “when more people benefit”? If an AI can better optimize to better extract the maximum value from people’s work and discard them, then optimize how many ways they can monetize their product to maximize the profit they get from each customer, the result is a horrible company and a horrible society.
In theory yes… but what do we call “doing a better job”? Is it just blindly extracting money? Or is it something more, and do we all agree on what it is? I think there could be a compounded problem of oversight.
Like, right now an employee pays/invests some money into a retirement fund, whose managers invest into several mutual funds, whose managers invest into several companies, whose owners ask for some performance from their C-suite, who through a chain of command tell the same employee what to do. Even though it’s part of the employee’s capital that’s controlling that company, if it takes an action negative for the employee like fracking under their home, or firing them, they’re powerless to do anything about it with their investment.
With AI replacing all those steps, it would all happen much quicker, and —since AIs are still basically a black box— with even less transparency than having corruptible humans on the same steps (at least we kind of know what tends to corrupt humans). Adding strict “code as contract” rules to try to keep them in check, would on a first sight look like an improvement, but in practice any unpredicted behavior could spread blindingly fast over the whole ecosystem, with nobody having a “stop” button anymore. That’s even before considering coding errors and malicious actors.
I guess a possible solution would be requiring every AI to have an external stop trigger, that a judicial system could launch to… possibly paralyze the whole economy. But that would require new legislation to be passed (with AI lawyers), and it would likely get late, and not be fully implemented by those trying to outsmart the system. Replace the judges by AIs too, politicians with AIs, talking heads on TV with AIs… and it becomes an AI world where humans have little to nothing to say. Are humans even of any use, in such a world?
None of those AIs need to be an AGI, so we could run ourselves into a corner with nobody and nothing having a global plan or oversight. Kind of like right now, but worse for the people.
Alternatively, all those AIs could be eco-friendly humans-first compassionate black boxes… but I kind of doubt those are the kind of AIs that current businesses are trying to build.
Thing is nobody will do that because once AI finds a way to spazz out that is totally unpredictable (black box) everything might just be gone.
It’s a totally unrealistic scenario.
They are working on mitigating the unpredictable “black box”.
Like making the AI explain their working method step by step. Not only does it make the AI more transparent, it also increases the correctness of whatever it types.
AI is still in development. It is good to list the problems you have, but don’t think those problems won’t be solved in the future.
People are already doing it, piece by piece, in all areas. As more AIs get input from other AIs, the chance of a cascading failure increases… but it will seem to be working “good enough” up until then, so more people will keep jumping on the bandwagon.
The question is: can we prepare for the eventual cascading spazz out, or have we no option other than letting it catch us by surprise?
Unfortunately everything AI does is kind of shitty. Sure you might have a query for which the chosen AI works well but you might as well not.
It you accept that it sometimes just doesn’t work at all sure AI is your revolution. Unfortunately there are not too many use cases where this is helpful.
I posit that in 80% of the cases, an AI working well even less than 50% of the times, is still “good enough” to achieve the shittier 80% of goals.
“I’ll have a burger with extra ketchup”… you get extra mayo instead… for half the price; “good enough”.
This is my worry as well
