What do you guys think about AI and stuff like ChatGPT? I know there are some people who are against it, especially on social media (maybe I really need to get off there because there's a lot of toxicity). I personally think that it's helpful because I use it for drafting work emails, such as when I need help to write with a professional tone and giving me ideas. I confess that I also use it as my kind of therapy sometimes, like kind of talking to a friend. I know that most people who are against it usually talk about invasion of privacy, the danger of it (afraid of it giving us misleading information, giving dangerous ideas, etc), art and idea stealing, making people anti-social and lazy, things like that. And it's like this whole AI thing just gets better everyday, the technology just gets more and more sophisticated. I can't really imagine how far it will go. I do think that there should be some boundaries, and we need to use it wisely and not abuse it.
ChatGPT and AI
- Thread starter puffymilk
- Start date
I think it has its uses, and despite generally disagreeing with tech bros, i like their overall ambition with AI. I like the sound of a world where none of us are obligated to work to earn our keep, because most jobs are done by AI-piloted robots. And it's a future we could realistically reach at some point. But i think in their eagerness to reach that future, they've kind of overlooked that we aren't at that point yet, and we're instead in kind of an awkward transitional phase.
The tech isn't good enough to reliably handle a lot of the things being thrown at it. It frequently hallucinates information that either isn't true, or doesn't have anything to do with the topic at hand. For example, i was googling something about Call of Duty Ghosts recently, and Google's AI overview gave me information related to Ghost of Tsushima instead. So it was telling me i could decide the ending to a Call of Duty game by writing a haiku. It's a silly example, but also a good demonstration, cos any human who knows anything about CoD could immediately tell what a ridiculous claim that is.
Despite the name, there's no actual intelligence in AI, so it relies on tools we build into it to sort and aggregate data, and determine correlation and causality. And as it stands, those tools are woefully insufficient for most purposes. We're likely several decades and numerous very large discussions away from it being ready. So for the time being, it's mostly a fun gimmick, with the potential to someday be a genuinely useful tool. I suspect its bad reputation is more related to techbros trying to push it as the replacement for useful services and actual people's jobs, before it's really ready for primetime. And there are reasonable considerations about the ethics of software that is essentially founded on a basis of plagiarism, and gives rise to people putting in even less effort to pumping out meaningless, soulless product for profit.
So all that being said, while i would personally recommend an actual therapist over AI, i know that in a lot of places, that's usually pretty expensive, so if you find genuine benefit from talking to AI, i say why not. Yeah, it's true there's no privacy in your conversations with it, but i think privacy is kind of a dead concept in this day and age anyway. If you can verbally express your interest in buying a new [insert product here] within vague proximity to anything with a microphone, and start getting targeted ads for that product, i think it's fair to say we're past the point we can hide anything from the corporate hellscape.
The tech isn't good enough to reliably handle a lot of the things being thrown at it. It frequently hallucinates information that either isn't true, or doesn't have anything to do with the topic at hand. For example, i was googling something about Call of Duty Ghosts recently, and Google's AI overview gave me information related to Ghost of Tsushima instead. So it was telling me i could decide the ending to a Call of Duty game by writing a haiku. It's a silly example, but also a good demonstration, cos any human who knows anything about CoD could immediately tell what a ridiculous claim that is.
Despite the name, there's no actual intelligence in AI, so it relies on tools we build into it to sort and aggregate data, and determine correlation and causality. And as it stands, those tools are woefully insufficient for most purposes. We're likely several decades and numerous very large discussions away from it being ready. So for the time being, it's mostly a fun gimmick, with the potential to someday be a genuinely useful tool. I suspect its bad reputation is more related to techbros trying to push it as the replacement for useful services and actual people's jobs, before it's really ready for primetime. And there are reasonable considerations about the ethics of software that is essentially founded on a basis of plagiarism, and gives rise to people putting in even less effort to pumping out meaningless, soulless product for profit.
So all that being said, while i would personally recommend an actual therapist over AI, i know that in a lot of places, that's usually pretty expensive, so if you find genuine benefit from talking to AI, i say why not. Yeah, it's true there's no privacy in your conversations with it, but i think privacy is kind of a dead concept in this day and age anyway. If you can verbally express your interest in buying a new [insert product here] within vague proximity to anything with a microphone, and start getting targeted ads for that product, i think it's fair to say we're past the point we can hide anything from the corporate hellscape.
Last edited:
My personal view is that we're still in the early days of AI. Though huge strides have been made, we're only just beginning to tap onto it's potential. Consequently and as already alluded to here, the limitations and potential for error should or must even be recognised.
If you ask me, concerns about AI should be more focused on what humans intend to do with it rather than AI itself.
If you ask me, concerns about AI should be more focused on what humans intend to do with it rather than AI itself.
I'm not looking to get into a fight with anyone, I just want to say that I honestly don't see how using ChatGPT or some other AI is any more useful for looking things up than the Internet. Let's say I have an ingredient in my fridge I want to use up. Sure, I could ask ChatGPT, but I could also go to Allrecipes or the BBC cooking website and I'd get reviews and feedback as well as a wide selection of recipes, with pictures.
Someone in my family uses AI to plan their holiday itinerary. I don't understand it. Why would you allow a machine to decide what you're going to do on each day of your holiday? Half the pleasure of going on holiday is researching the place you're visiting and making plans that are unique to you.
Someone in my family uses AI to plan their holiday itinerary. I don't understand it. Why would you allow a machine to decide what you're going to do on each day of your holiday? Half the pleasure of going on holiday is researching the place you're visiting and making plans that are unique to you.
I have kind of mixed feelings about all the AI tools.
On one hand, it helps me at my work a lot. Im a coder, so these tools are extremely helpful, or even almost a mandatory requirement these days. I mean it the way that one could do a fine job without any, but then the other ppl in the job market can do the same with half the time used, so not using them is a severe handicap.
On the other hand, im dreading all the bad shit that comes along them. Like the internet turning into a hellscape of propaganda and misinformation bots, ai generated ads, ai generated images, ai generated articles, ai tracking your every movement and who even knows what else...
It looks like there is not much use fighting back this technology. Its like the wheel, you either develop your own or see your competitors take you over. I guess we just have to wait and see what happens...
On one hand, it helps me at my work a lot. Im a coder, so these tools are extremely helpful, or even almost a mandatory requirement these days. I mean it the way that one could do a fine job without any, but then the other ppl in the job market can do the same with half the time used, so not using them is a severe handicap.
On the other hand, im dreading all the bad shit that comes along them. Like the internet turning into a hellscape of propaganda and misinformation bots, ai generated ads, ai generated images, ai generated articles, ai tracking your every movement and who even knows what else...
It looks like there is not much use fighting back this technology. Its like the wheel, you either develop your own or see your competitors take you over. I guess we just have to wait and see what happens...
It is much like most innovations in the modern internet era.
AI will probably be an huge net profit for humanity (given all the research, products and new business that AI will create out of nowhere) but society isn't prepared for the consequences of it.
But artists and other creative professions will be absolutely ripped off by AI. But that happens since the internet exists, and most artists are pretty resigned by it.
But when you read the american CRS report on AI, they just claim that AI is protected by fair use and that is because they don't make their profit directly from their AI creations but from their licenses.
AI will probably be an huge net profit for humanity (given all the research, products and new business that AI will create out of nowhere) but society isn't prepared for the consequences of it.
But artists and other creative professions will be absolutely ripped off by AI. But that happens since the internet exists, and most artists are pretty resigned by it.
But when you read the american CRS report on AI, they just claim that AI is protected by fair use and that is because they don't make their profit directly from their AI creations but from their licenses.
I’m not so worried about skynet, as in AI growing too powerful and deciding that the human race has to go.
What really, really, really worries me is people using AI to make decisions, e.g., about traffic signal timing. This is destructive because the results of those decisions are fed back into the model, creating a self reinforcing feedback loop that is relatively insensitive to learning from failures.
This is not theoretical. There is real life field data that show that machine learning suffers from the exact same cognitive biases as the human who built the model. In other words, a model is not somehow omniscient and the people selling it as such are more dangerous than any machine that could be invented.
What really, really, really worries me is people using AI to make decisions, e.g., about traffic signal timing. This is destructive because the results of those decisions are fed back into the model, creating a self reinforcing feedback loop that is relatively insensitive to learning from failures.
This is not theoretical. There is real life field data that show that machine learning suffers from the exact same cognitive biases as the human who built the model. In other words, a model is not somehow omniscient and the people selling it as such are more dangerous than any machine that could be invented.
Ummmm, I hate to break it to you, but that ship sailed so long ago that it would have crashed into an iceberg by now, except that we melted all the glaciers. There are many factors that cause this (propaganda elevation) but the unintended consequences of monetizing attention are definitely coming home to bite us in the metaphorical butt.
I appreciate that CHatGPT can do this and I wouldn't judge anyone who uses it for learning. But... If it didn't exist, don't you think you would have learned how to use Excel in some other way? Maybe from a fellow human?
I know that it is getting harder and harder to find humans to connect with for help. And I can see how AI is a solution to that problem. But it may be the case that all this tech is what has been isolating us from each other in the first place. Interacting with other humans, asking for help, offering help, communicating meaningfully with each other: these things are hard and for many people they don't come naturally. We have to learn them, and as AI becomes increasingly effective as both friend and mentor - a friend and mentor who requires no interpersonal skills from us whatsoever (it's not like AI shuts itself down if you're too rude or needy or demanding) - our incentive to learn how to deal with one another will become less and less.
I know that it is getting harder and harder to find humans to connect with for help. And I can see how AI is a solution to that problem. But it may be the case that all this tech is what has been isolating us from each other in the first place. Interacting with other humans, asking for help, offering help, communicating meaningfully with each other: these things are hard and for many people they don't come naturally. We have to learn them, and as AI becomes increasingly effective as both friend and mentor - a friend and mentor who requires no interpersonal skills from us whatsoever (it's not like AI shuts itself down if you're too rude or needy or demanding) - our incentive to learn how to deal with one another will become less and less.