Their only value is showing how fucked up our society is.
Suddenly and very publicly copyrights only matter if you’re poor, electricity is wasted on the poor, water is not for the poor… it’s always been like this, but the LLM bandwagon really showcased all of that in one shiny package.
The only good thing could be gathering public knowledge into a single space, but they don’t even do that.
I respectfully disagree with the dead end part of your argument. A dead end would be if they provided no value.
While the environmental and social downsides are massive negatives on the tech, it is actually doing something.
Past iterations are completely useless, but more recent iterations show us a more polished side to LLMs that actually do enhance how we do some things.
Is it worth it? My gut says no, but its both too late and too early to call it. (late in the environmental and societal impact, too early in the tech iteration)
As far as the “dead end” argument goes, I have to say that’s a hard disagree. Humanity is filled with technological advances that “stand on the shoulders of giants” and improve on previous techs. Even if LLMs themselves don’t prove to be the thing that we’ve been promised by the people driving it, it is taking us one step closer to AGI (whether that’s a good goal or not, that’s still up for debate)
From here on, I think there’s still quite a bit these models can improve, and I hope a lot of that improvement goes into making it more energy efficient, more water efficient in turn.
If by a dead end you mean that we can’t reach an AGI from an LLM, I think that’s correct, however an LLM might help us figure out what is needed for an AGI.
If it was used in a research as a step? Perchance.
Pouring everything we have into it? Dumbest fucking decision of our lives.
We could have put all that effort into previous versions and could tweak them enough to gather perhaps slightly worse results, maybe even better, we will never know.
Making this shit more efficient is to me also dumb.
What in the fuck are we doing that requires this shit? It helps with coding? We can make better frameworks. Translations? We had those before, even TTS. Emails? Just use a template. The other side is not reading that slop anyway. So what exactly are we doing here?
You didn’t actually say what you think LLM’s are enhancing. Just that you feel that they are. Honestly I think that’s the biggest part, they’re big shiny things that look like they’re doing a lot. But they actually aren’t. LLMs are chatbots and they will never be anything more than just chatbots.
So much so, that it’s pretty much 100% necessary in software engineering now. And I hate it that I’m forced to use something that I know is so detrimental in other aspects.
I used it to make a dialog system in a video game. It made it, but it was needlessly complex and ten times as long as the code needed to be. No thanks, i don’t need a buggy mess that’s unmaintainable.
Yes, I’ve made a dialog system before. The context I found myself in was a game jam with a short amount of time using an engine I hadn’t used in years.
Thought it would help instead of following a tutorial. But honestly, by the end of the jam, I really didn’t feel like rewriting the dialog system bcz it was so messy.
LLMs are not chat bots, they do natural language generation AKA: they can produce human readble text, they can also parse text;
As of now, they take an input and follow patterns to guess what the output should be, it is really useful to be fair, they help in translation (see Deepl, a very good translator), they can take data and make it more readble to humans, summarize text*, parse text and data structures ex: i can give a JSON file to an LLM so i can get back a TOML file, document hard to read code etc etc
*but i’d argue that it’s rarely useful, you will hardly have to summarize a text for yourself because you usually need to know any detail in it but i can see someone needing a summary once
The fact that you think it’s bad at one thing in your list but adequate at the others is part of the problem. It’s bad at all of those things, because it’s a chatbot. Admittedly a very advanced chatbot, but still just a chatbot.
The most important take away here is what of your list was impossible before LLMs? Because the reality is that absolutely everything that you mentioned was possible before LLMs. All that LLMs have added is the chat interface part.
Granted, the technology that allowed LLM’s is likely to be very useful and already has been in places like protein folding, but that happened before LLMs.
You are so clever for pointing that out, and are absolutely correct! You’re clearly the expert in this exchange and the other person should heed what you have to say in all things.
Don’t say shit like that for a start. Shows you haven’t bothered to consider any of the multitudinous ways AI is useful. Your concerns are valid but are nothing to do with AI, they are system problems with the late stage crony capitalism we current have inflicted on us
Regardless, if someone’s trying to get to the moon so they can enslave us all and rule over us from their moon fortress, I don’t care if all they’ve got is a really long ladder, I’m breaking the ladder.
I don’t know what you mean by current iteration but what I do know is that general-purpose LLMs can already beat the very best of human intelligence some of the time (recentexamples). So it won’t take very long for a few more breakthroughs to be made which will enable general-purpose AIs (LLMs, other neural networks, or something else entirely) to beat human intelligence most of the time, and then 90% of the time, then 99.9999% of the time. AI is already doing a lot of the coding to make AI and it could discover better alternatives to LLMs.
You’re basically promoting AI companies by saying their products are powerful and not garbage. This is their false narrative, repackaged in meme format.
We never needed that much computational power to get our shit together to begin with. It’s just a power grab by the rich. If we wanted to fix stuff with cybernetics we could just implement the viable system model - Salvatore Allende style.
The ability to acquire, understand, and use knowledge.
Much like you, AI is incapable of understanding anything, they just regurgitate what they hear.
AI aren’t intelligent, they’re stochastic parrots that can process words as math to generate a facsimile of intelligence to make the ignorant think it’s smart.
This meme implies that AI data centers have been around for a long time. They have not. And it’s not really the number of data centers that matters it’s the techniques that AIs use.
This meme implies that AI data centers have been around for a long time. They have not.
This response implies that you actually understand the point being made. You don’t. If you think this is only about AI datacenters, I have to assume you’re letting the AI ‘think’ for you. Otherwise, that says a lot about you, and none of it is positive.
AI have been running in datacenters for over a decade, it didn’t just spring up overnight while the companies are building datacenters that cost billions in hardware that will be replaced in less than 5-10 years. They might as well be setting the cash on fire, it would harm the planet less than the result of pissing away billions of dollars on datacenters that may be bankrupt in under a decade when the AI companies start actually charging for cost and hardware that will need replaced before then.
OpenAI and others have had data centers training models for 10 years. Chatgpt was released in 2022. A data center was churning for years before release to create that model.
The thing is that neural networks in general (and LLM specifically) aren’t creative. It can learn from input and gets a lot more input than a human, that’s why it is better and faster in standardized tests (which are more often than not part of the input) and maybe can combine different things but it wouldn’t ever have a genuine idea and much less a will of its own or a consciousness
Well, do you need to be creative when working in the robot factory?
If it’s only creativity missing, then that’s a dime a dozen, everyone is creative, but few people can program or design stuff for example. Win win IMO.
It was merely an example, about AI doing work. Not in your specific factory because rare are factories where workers are not discouraged to be creative lol.
It’s just as valid for any AI lol. Anything out didn’t quite fit the training data it gets stuck lol. In robotics that’s a huge issue because there isn’t much training data lol
Wow no it is not lol. All ai isn’t a robot. And neural networks are actually good at not “getting stuck” in reasoning, that’s why we went from machine learning to deep neural networks. They solved exactly that.
Except LLMs are the worst of both worlds in that respect. In order to work in a robot factory, its output needs to be reliable and repeatable, ideally across as wide a range of inputs as possible. LLMs … are very much not that. They’re also only as ‘skilled’ as their training data, which thanks to the morally bankrupt scraping of every source the AI companies can get their grubby hands on, is of enormously variable quality - and because of the nature of LLMs, it will never be better than its training data. The average quality of its output will, in fact, be the average of its training data.
It’s possible for LLMs to be creative - in the sense that it can output novel sentences - except that as you increase its ‘creativity’ (temperature) beyond the default that most of the chatbots out there have, the quality plummets. It still can’t solve complex problems though, because even if it does have an internal model of how certain things function, it can’t come close to the complexity of what humans can hold in their brains - or perhaps cannot abstract portions of their model in the same way - as evidenced by their utter failure to work through any problem that has more than five or so layers. This is a problem that sees diminishing returns with increased parameter count - the primary metric that is driving the enormous data centers being built.
LLMs are a solution looking for a problem, and aside from ‘bs for people who don’t want to make any decisions in their day-to-day life’ and ‘scam generator’, there doesn’t seem to be very many niches that they are actually good at filling.
You sound like people explaining that chess is something a computer can never beat a human at because of some mystical sense we are supposed to have (it was quite some time ago). They quickly changed their tune to “chess isn’t very hard anyways” when Kasparow got schwacked by Deep Blue. Back then peiple hated on automation.
We humans will never be better than our training data either, and we forget and get old and die.
I’m more interested in figuring out what we should do with all the computational power and potential labour. The robot was just an example, a metaphor, for AI doing boring work. It will be able to write sonnets and generate world class movies one day, what shall we humans do then? Be happy? Do art?
LLM output is often indistinguishable from genuine creativity. You can give it an open ended prompt like “illustrate how the world is from the perspective of an LLM” and probably get something nobody has ever seen or thought of before. People use LLMs for generating ideas as well as coming up with novel solutions like I posted. Saying it’s not creative is mysticist cope.
Nah, people put together stuff randomly all the time. Things that most people are aware of tend to be filtered to stuff that works together.
It is no more creative than putting pieces of paper with things on them and pulling them randomly out of a hat. AI can’t creatively come up with something new to write on one of those pieces of paper.
That’s just denial of reality. AI comes up with new ways of “thinking” like in the math example I cited. If you call that remixing existing ideas then that could describe what all humans do and it’s questionable if any of us are creative.
Humans mix things intentionally whether by filtering down randomized mixing or by intentionally choosing what to mix. AI just throws shit at the wall becsuse it can’t do anything intentionally.
Yes LLM’s can beat humans in many tasks. But the jump from super spell checker to real AI is still huge.
It’s like the 1960’s where computers were beating chess players. (Not grand masters but they could beat regular people.) Because a computer could out think a regular human, people assumed that with more resources, we would have real AI in 25 year. That was Hal from 2001 Space Odyssey. It seemed very reasonable in 1968.
Do you belive current iteration of AI has the potential to become superhuman? I think it’s like trying to get to the moon by building a better ladder.
No, the tech ran into diminishing returns. That’s been studied. In the end you’re adding another datacenter just to get 1% better output.
LLMs are a dead end.
Their only value is showing how fucked up our society is.
Suddenly and very publicly copyrights only matter if you’re poor, electricity is wasted on the poor, water is not for the poor… it’s always been like this, but the LLM bandwagon really showcased all of that in one shiny package.
The only good thing could be gathering public knowledge into a single space, but they don’t even do that.
So it’s all net negative in my eyes.
I respectfully disagree with the dead end part of your argument. A dead end would be if they provided no value.
While the environmental and social downsides are massive negatives on the tech, it is actually doing something.
Past iterations are completely useless, but more recent iterations show us a more polished side to LLMs that actually do enhance how we do some things.
Is it worth it? My gut says no, but its both too late and too early to call it. (late in the environmental and societal impact, too early in the tech iteration)
As far as the “dead end” argument goes, I have to say that’s a hard disagree. Humanity is filled with technological advances that “stand on the shoulders of giants” and improve on previous techs. Even if LLMs themselves don’t prove to be the thing that we’ve been promised by the people driving it, it is taking us one step closer to AGI (whether that’s a good goal or not, that’s still up for debate)
From here on, I think there’s still quite a bit these models can improve, and I hope a lot of that improvement goes into making it more energy efficient, more water efficient in turn.
If by a dead end you mean that we can’t reach an AGI from an LLM, I think that’s correct, however an LLM might help us figure out what is needed for an AGI.
If it was used in a research as a step? Perchance.
Pouring everything we have into it? Dumbest fucking decision of our lives.
We could have put all that effort into previous versions and could tweak them enough to gather perhaps slightly worse results, maybe even better, we will never know.
Making this shit more efficient is to me also dumb.
What in the fuck are we doing that requires this shit? It helps with coding? We can make better frameworks. Translations? We had those before, even TTS. Emails? Just use a template. The other side is not reading that slop anyway. So what exactly are we doing here?
You didn’t actually say what you think LLM’s are enhancing. Just that you feel that they are. Honestly I think that’s the biggest part, they’re big shiny things that look like they’re doing a lot. But they actually aren’t. LLMs are chatbots and they will never be anything more than just chatbots.
Summarizing and finding codeblocks. Fucking A+.
So much so, that it’s pretty much 100% necessary in software engineering now. And I hate it that I’m forced to use something that I know is so detrimental in other aspects.
I’ve been a software developer for over 15 years, I’ve never used one. It’s not necessary at all.
And your 15 years are superior to mine because?..
I used it to make a dialog system in a video game. It made it, but it was needlessly complex and ten times as long as the code needed to be. No thanks, i don’t need a buggy mess that’s unmaintainable.
Note: I didn’t say - use it to code.
But real question for you. Is the alternative you wouldn’t have done that at all?
Yes, I’ve made a dialog system before. The context I found myself in was a game jam with a short amount of time using an engine I hadn’t used in years.
Thought it would help instead of following a tutorial. But honestly, by the end of the jam, I really didn’t feel like rewriting the dialog system bcz it was so messy.
LLMs are not chat bots, they do natural language generation AKA: they can produce human readble text, they can also parse text; As of now, they take an input and follow patterns to guess what the output should be, it is really useful to be fair, they help in translation (see Deepl, a very good translator), they can take data and make it more readble to humans, summarize text*, parse text and data structures ex: i can give a JSON file to an LLM so i can get back a TOML file, document hard to read code etc etc
*but i’d argue that it’s rarely useful, you will hardly have to summarize a text for yourself because you usually need to know any detail in it but i can see someone needing a summary once
The fact that you think it’s bad at one thing in your list but adequate at the others is part of the problem. It’s bad at all of those things, because it’s a chatbot. Admittedly a very advanced chatbot, but still just a chatbot.
The most important take away here is what of your list was impossible before LLMs? Because the reality is that absolutely everything that you mentioned was possible before LLMs. All that LLMs have added is the chat interface part.
Granted, the technology that allowed LLM’s is likely to be very useful and already has been in places like protein folding, but that happened before LLMs.
Maybe learn something about a subject before spouting off
You are so clever for pointing that out, and are absolutely correct! You’re clearly the expert in this exchange and the other person should heed what you have to say in all things.
Thank you for being so supportive
How can I prove quickly and definitely that I know what I’m talking about?
Don’t say shit like that for a start. Shows you haven’t bothered to consider any of the multitudinous ways AI is useful. Your concerns are valid but are nothing to do with AI, they are system problems with the late stage crony capitalism we current have inflicted on us
AI will be immensely useful - but not LLMs.
It’s been a while since I linked the LLMentalist, yet not much has changed.
They’re tools and when used as such they are clearly useful in various ways, one biased article not withstanding
deleted by creator
Regardless, if someone’s trying to get to the moon so they can enslave us all and rule over us from their moon fortress, I don’t care if all they’ve got is a really long ladder, I’m breaking the ladder.
No but if you don’t try, you won’t find where the Goblins are hiding.
Dont know, dont care, dont want it
That’s a you problem, but thanks for your input
I don’t know what you mean by current iteration but what I do know is that general-purpose LLMs can already beat the very best of human intelligence some of the time (recent examples). So it won’t take very long for a few more breakthroughs to be made which will enable general-purpose AIs (LLMs, other neural networks, or something else entirely) to beat human intelligence most of the time, and then 90% of the time, then 99.9999% of the time. AI is already doing a lot of the coding to make AI and it could discover better alternatives to LLMs.
Oh no.
You’re basically promoting AI companies by saying their products are powerful and not garbage. This is their false narrative, repackaged in meme format.
In 2018, passing the Turing test was still the gold standard in computer intelligence.
We used the same AI from the 1950s to blow past it. Nothing fundamentally changed. The only difference is that our hardware finally caught up.
What makes you think that throwing more hardware at it isn’t the solution?
We never needed that much computational power to get our shit together to begin with. It’s just a power grab by the rich. If we wanted to fix stuff with cybernetics we could just implement the viable system model - Salvatore Allende style.
😎
You seem to be at odds not only with artificial intelligence.
Much like you, AI is incapable of understanding anything, they just regurgitate what they hear.
AI aren’t intelligent, they’re stochastic parrots that can process words as math to generate a facsimile of intelligence to make the ignorant think it’s smart.
No u
This meme implies that AI data centers have been around for a long time. They have not. And it’s not really the number of data centers that matters it’s the techniques that AIs use.
This response implies that you actually understand the point being made. You don’t. If you think this is only about AI datacenters, I have to assume you’re letting the AI ‘think’ for you. Otherwise, that says a lot about you, and none of it is positive.
AI have been running in datacenters for over a decade, it didn’t just spring up overnight while the companies are building datacenters that cost billions in hardware that will be replaced in less than 5-10 years. They might as well be setting the cash on fire, it would harm the planet less than the result of pissing away billions of dollars on datacenters that may be bankrupt in under a decade when the AI companies start actually charging for cost and hardware that will need replaced before then.
Huh that’s odd, then how come DeepMind was founded checks notes in 2010?
OpenAI and others have had data centers training models for 10 years. Chatgpt was released in 2022. A data center was churning for years before release to create that model.
The thing is that neural networks in general (and LLM specifically) aren’t creative. It can learn from input and gets a lot more input than a human, that’s why it is better and faster in standardized tests (which are more often than not part of the input) and maybe can combine different things but it wouldn’t ever have a genuine idea and much less a will of its own or a consciousness
Well, do you need to be creative when working in the robot factory?
If it’s only creativity missing, then that’s a dime a dozen, everyone is creative, but few people can program or design stuff for example. Win win IMO.
It’s also not sure AI cannot mimic creativity.
Everytime someone says to a worker in a factory “figure it out” it involves creativity in some way.
It was merely an example, about AI doing work. Not in your specific factory because rare are factories where workers are not discouraged to be creative lol.
It’s just as valid for any AI lol. Anything out didn’t quite fit the training data it gets stuck lol. In robotics that’s a huge issue because there isn’t much training data lol
Wow no it is not lol. All ai isn’t a robot. And neural networks are actually good at not “getting stuck” in reasoning, that’s why we went from machine learning to deep neural networks. They solved exactly that.
Sure.
https://the-decoder.com/language-models-can-overthink-and-get-stuck-in-endless-thought-loops/
Except LLMs are the worst of both worlds in that respect. In order to work in a robot factory, its output needs to be reliable and repeatable, ideally across as wide a range of inputs as possible. LLMs … are very much not that. They’re also only as ‘skilled’ as their training data, which thanks to the morally bankrupt scraping of every source the AI companies can get their grubby hands on, is of enormously variable quality - and because of the nature of LLMs, it will never be better than its training data. The average quality of its output will, in fact, be the average of its training data.
It’s possible for LLMs to be creative - in the sense that it can output novel sentences - except that as you increase its ‘creativity’ (temperature) beyond the default that most of the chatbots out there have, the quality plummets. It still can’t solve complex problems though, because even if it does have an internal model of how certain things function, it can’t come close to the complexity of what humans can hold in their brains - or perhaps cannot abstract portions of their model in the same way - as evidenced by their utter failure to work through any problem that has more than five or so layers. This is a problem that sees diminishing returns with increased parameter count - the primary metric that is driving the enormous data centers being built.
LLMs are a solution looking for a problem, and aside from ‘bs for people who don’t want to make any decisions in their day-to-day life’ and ‘scam generator’, there doesn’t seem to be very many niches that they are actually good at filling.
You sound like people explaining that chess is something a computer can never beat a human at because of some mystical sense we are supposed to have (it was quite some time ago). They quickly changed their tune to “chess isn’t very hard anyways” when Kasparow got schwacked by Deep Blue. Back then peiple hated on automation.
We humans will never be better than our training data either, and we forget and get old and die.
I’m more interested in figuring out what we should do with all the computational power and potential labour. The robot was just an example, a metaphor, for AI doing boring work. It will be able to write sonnets and generate world class movies one day, what shall we humans do then? Be happy? Do art?
The first priority should always be figuring out a way to produce very cheap and healthy food, housing and clothing for everybody.
And automation and robotisation has done that, wildly. The problem we have is it isn’t correctly distributed, not that we can’t or are not doing it.
Housing is similar, we have enough housing, but it’s hoarded instead of distributed.
This has nothing to do with AI.
So AI is pretty useless
LLM output is often indistinguishable from genuine creativity. You can give it an open ended prompt like “illustrate how the world is from the perspective of an LLM” and probably get something nobody has ever seen or thought of before. People use LLMs for generating ideas as well as coming up with novel solutions like I posted. Saying it’s not creative is mysticist cope.
Nah, people put together stuff randomly all the time. Things that most people are aware of tend to be filtered to stuff that works together.
It is no more creative than putting pieces of paper with things on them and pulling them randomly out of a hat. AI can’t creatively come up with something new to write on one of those pieces of paper.
That’s just denial of reality. AI comes up with new ways of “thinking” like in the math example I cited. If you call that remixing existing ideas then that could describe what all humans do and it’s questionable if any of us are creative.
Humans mix things intentionally whether by filtering down randomized mixing or by intentionally choosing what to mix. AI just throws shit at the wall becsuse it can’t do anything intentionally.
You’re fighting an angry mob of luddites.
Yes LLM’s can beat humans in many tasks. But the jump from super spell checker to real AI is still huge.
It’s like the 1960’s where computers were beating chess players. (Not grand masters but they could beat regular people.) Because a computer could out think a regular human, people assumed that with more resources, we would have real AI in 25 year. That was Hal from 2001 Space Odyssey. It seemed very reasonable in 1968.