I dont think we are in a bubble and I think all the media posts about it are just trying to make people sell their shares.
Sure there is no obvious profit yet but there will be. Once people start using AI in their phones and ask it questions, everything from baking to coding becomes way easier since its an interactive conversation, not a search result.
People will pay for that convenience since its a huge downside to not have access to it. Search results now seem very limited to me since I cant find out more about what its saying.
Literally the entire point of searching something is to open up the links and find out more about what it’s saying… Do you need your AI to come up with the search queries too???
The technology (at least with current methodologies) is flawed: that’s why people are warning of the bubble bursting. We can’t properly scale LLMs on our current grid in the same capacity as China. Our technologies are also incredibly energy-intensive compared to their technologies.
There is no intelligence, the hallucinations are likely fundamental, the cases of people being given dangerous or harmful advice are rising, human AI psychosis is a real concern, the sycophancy/bias confirmation is still present, and major actors in the AI space are existentially afraid of any form of regulation of the technology/industry (which does not signal confidence).
Also, it’s critical to factor in the whole copyright issue with training data… one domino is all it takes to collapse the whole thing.
Are you personally invested in the AI/LLM space? I’m wondering why you chose to engage with very few of my arguments. Is your account a troll account? If you’re not trolling: re-read. I will not engage further until you adequately address my points.
I was pretty clear: there is no intelligence. AGI is an absolute pipe dream and it will also be a far cry from actual intelligence if you look into it. Hallucinations won’t be fixed unless the technology evolves - adding more GPU power won’t be able to fix it.
The copyright theft is an extreme issue, regardless your hand-waving of it. Copyright law reform is not perceivably on the table. Major companies are caught red-handed stealing and these companies have no intention of compensating the rights-holders they stole from.
It’s less that you think differently, and more that people are questioning how much you’re thinking in the first place.
I’ve used AI enough to know that they struggle with basic concepts in things I understand well, which means I can’t trust it much on topics I’m unfamiliar with since the likelihood that it’s bad in my field (technology/programming) but good in others is drastically low. We regularly see how bad AI is in all kinds of fields, and the problems we see are all the same archetypical problems like hallucinations, basic misunderstanding of concepts, and it’s obsequiousness is out of control and regularly applied to incorrect user responses and corrections, and an inability to consistently follow basic directives.
I struggle to see how this isn’t a bubble given how all of the money being pumped into AI is predicated on it becoming something useful, and so far the main use I’m seeing out of LLMs is revenge porn, copyright infringement, mass market propaganda, and lowering the quality of output across tons of sectors because people think this shit is useful when it spits out trash.
I work in IT also and everyone I know use it constantly at work, daily. It has flaws and sometimes hallucinates, yeah. But humanity has never had something like this where they can ask questions and get replies that are often completely correct, just not always.
I dont think its a bubble, but thats the point of having a discussion about it. I think downvoting people because they like Ai is borderline retarded.
If you have read threads about this on hacker news, you will see both sides of this. People who like Ai and people who doesnt. But you have arguments instead of downvoting because people are actually smart enough to value the discussion.
I think you are overestimating the amount people will pay for convenience or cling to their old ways.
Did e-readers kill the bookstore? Some people will always prefer to cook out of a book or dive into docs to write code.
Or look at the modern streaming landscape. In the beginning there was basically Netflix and everyone was fine paying that monthly fee for the convenience of streaming basically everything. Now we have 20+ vendors all charging for some subset of content. And we have seen a corresponding loss in subscribers as people hit the limit of what they are willing to pay for convenience.
they’re giving out ai memberships and trials constantly to try to get ppl used to it paying for it, hasnt worked, ngl I generate images ocsssionaly I like seeing random mashups sometimes, but id never pay for it if its free and fast why not, I was never gonna pay anyone to do it and im not selling anything, dont get why ppl get all pissy when ai is used for self entertainment
I dont think we are in a bubble and I think all the media posts about it are just trying to make people sell their shares.
Sure there is no obvious profit yet but there will be. Once people start using AI in their phones and ask it questions, everything from baking to coding becomes way easier since its an interactive conversation, not a search result.
People will pay for that convenience since its a huge downside to not have access to it. Search results now seem very limited to me since I cant find out more about what its saying.
Literally the entire point of searching something is to open up the links and find out more about what it’s saying… Do you need your AI to come up with the search queries too???
Its not interactive. Did you read my comment?
If you dont understand why people are going to pay for that, I dont know what to tell you.
Thats literally the entire thing about Ai, you can have a conversation, not static text and links.
The technology (at least with current methodologies) is flawed: that’s why people are warning of the bubble bursting. We can’t properly scale LLMs on our current grid in the same capacity as China. Our technologies are also incredibly energy-intensive compared to their technologies.
There is no intelligence, the hallucinations are likely fundamental, the cases of people being given dangerous or harmful advice are rising, human AI psychosis is a real concern, the sycophancy/bias confirmation is still present, and major actors in the AI space are existentially afraid of any form of regulation of the technology/industry (which does not signal confidence).
Also, it’s critical to factor in the whole copyright issue with training data… one domino is all it takes to collapse the whole thing.
These things will evolve and change and improve, as always. When the first car was invented, there wasnt any proper roads for cars. Things change.
Copyright issues are not going to be an issue at all when entire countries are trying to get first to AGI. Nobody cares about that.
Hallucinations are flaws that will eventually be fixed, and the more gpu power that is available, the easier it will be to fix.
Are you personally invested in the AI/LLM space? I’m wondering why you chose to engage with very few of my arguments. Is your account a troll account? If you’re not trolling: re-read. I will not engage further until you adequately address my points.
I was pretty clear: there is no intelligence. AGI is an absolute pipe dream and it will also be a far cry from actual intelligence if you look into it. Hallucinations won’t be fixed unless the technology evolves - adding more GPU power won’t be able to fix it.
The copyright theft is an extreme issue, regardless your hand-waving of it. Copyright law reform is not perceivably on the table. Major companies are caught red-handed stealing and these companies have no intention of compensating the rights-holders they stole from.
Removed by mod
Removed by mod
It’s less that you think differently, and more that people are questioning how much you’re thinking in the first place.
I’ve used AI enough to know that they struggle with basic concepts in things I understand well, which means I can’t trust it much on topics I’m unfamiliar with since the likelihood that it’s bad in my field (technology/programming) but good in others is drastically low. We regularly see how bad AI is in all kinds of fields, and the problems we see are all the same archetypical problems like hallucinations, basic misunderstanding of concepts, and it’s obsequiousness is out of control and regularly applied to incorrect user responses and corrections, and an inability to consistently follow basic directives.
I struggle to see how this isn’t a bubble given how all of the money being pumped into AI is predicated on it becoming something useful, and so far the main use I’m seeing out of LLMs is revenge porn, copyright infringement, mass market propaganda, and lowering the quality of output across tons of sectors because people think this shit is useful when it spits out trash.
I work in IT also and everyone I know use it constantly at work, daily. It has flaws and sometimes hallucinates, yeah. But humanity has never had something like this where they can ask questions and get replies that are often completely correct, just not always.
I dont think its a bubble, but thats the point of having a discussion about it. I think downvoting people because they like Ai is borderline retarded.
If you have read threads about this on hacker news, you will see both sides of this. People who like Ai and people who doesnt. But you have arguments instead of downvoting because people are actually smart enough to value the discussion.
Removed by mod
Lol ok. :)
I think you are overestimating the amount people will pay for convenience or cling to their old ways.
Did e-readers kill the bookstore? Some people will always prefer to cook out of a book or dive into docs to write code.
Or look at the modern streaming landscape. In the beginning there was basically Netflix and everyone was fine paying that monthly fee for the convenience of streaming basically everything. Now we have 20+ vendors all charging for some subset of content. And we have seen a corresponding loss in subscribers as people hit the limit of what they are willing to pay for convenience.
they’re giving out ai memberships and trials constantly to try to get ppl used to it paying for it, hasnt worked, ngl I generate images ocsssionaly I like seeing random mashups sometimes, but id never pay for it if its free and fast why not, I was never gonna pay anyone to do it and im not selling anything, dont get why ppl get all pissy when ai is used for self entertainment