Which remind me, should AI be used?
@Raspberry_yoghurt , thank you for your informative and useful post. And it helps to highlight one of my other points: one of my fears is that some might blindly accept AI's creations as unquestionable, since it comes from "a powerful AI" and therefore must be "the science."
In the field of translation, it is easy for a skilled person like yourself to tell when AI has screwed up. But in some other fields, although AI can apparently still screw up, it will be much harder for all but the most expert to even tell that there's been an error (in some cases, at least).
I don't oppose AI -- and I think it will lead (and already has led) to some very useful things. But the "garbage in / garbage out" nature of even simple programming continues to scare me. And if AI feasts upon terrabytes of writing (scientific papers, economics journals, newspapers, blogs, and the like) and then draws a conclusion from that writing, its conclusion will only be as good as the writing it uses to train itself. And if that writing is flawed -- because the writers' own understanding is imperfect or colored by some ideological or practical bias -- it can lead, in a completely unintended way, to imperfect or biased conclusions.
Yet "because it comes from a super computer" (I can hear the crowd say in my mind's ear) it might largely be treated as "beyond debate." We already have far too much that is considered "beyond debate." I'd hate to see the scope of the "unquestionable" be enlarged further.
REAL science -- of any sort -- QUESTIONS and is never too sure of itself. And if science itself must remain open-minded, lesser disciplines, like politics, for example, should likewise remain free from utter, unyielding certainty.
In the field of translation, it is easy for a skilled person like yourself to tell when AI has screwed up. But in some other fields, although AI can apparently still screw up, it will be much harder for all but the most expert to even tell that there's been an error (in some cases, at least).
I don't oppose AI -- and I think it will lead (and already has led) to some very useful things. But the "garbage in / garbage out" nature of even simple programming continues to scare me. And if AI feasts upon terrabytes of writing (scientific papers, economics journals, newspapers, blogs, and the like) and then draws a conclusion from that writing, its conclusion will only be as good as the writing it uses to train itself. And if that writing is flawed -- because the writers' own understanding is imperfect or colored by some ideological or practical bias -- it can lead, in a completely unintended way, to imperfect or biased conclusions.
Yet "because it comes from a super computer" (I can hear the crowd say in my mind's ear) it might largely be treated as "beyond debate." We already have far too much that is considered "beyond debate." I'd hate to see the scope of the "unquestionable" be enlarged further.
REAL science -- of any sort -- QUESTIONS and is never too sure of itself. And if science itself must remain open-minded, lesser disciplines, like politics, for example, should likewise remain free from utter, unyielding certainty.
no i dont
what did the AI algorithm learn from -_-, it learned from data from humans, if it wasn't for doctors identifying diseases, how would the ai learn to do this, if as you said ai replaced doctors, what will it learn from if there are no doctors lmao
an example of this is why chat gpt, stackoverflow lost so much usage after it came, when new technologie comes that chat gpt has no data on, everyone back to stackoverflow ;)
also in the chat gpt example -_-, it produces human sounding text, human sounding text!= correct information, u will understand what i mean if you used it before to code something lmao
idk why you are all exaggerating the impact of AI, but it is nothing more then a tool that can make actions easier as by definition AI is training an algorithm to perform a task usually performed by human and that by learning from them thus if it replaces humans what will it learn from lmao
also this is just a hype, people working intensively on it... like many things that were hyped in the past and forgotten :) (:
an example of this is why chat gpt, stackoverflow lost so much usage after it came, when new technologie comes that chat gpt has no data on, everyone back to stackoverflow ;)
also in the chat gpt example -_-, it produces human sounding text, human sounding text!= correct information, u will understand what i mean if you used it before to code something lmao
idk why you are all exaggerating the impact of AI, but it is nothing more then a tool that can make actions easier as by definition AI is training an algorithm to perform a task usually performed by human and that by learning from them thus if it replaces humans what will it learn from lmao
also this is just a hype, people working intensively on it... like many things that were hyped in the past and forgotten :) (:
@Noflaps said in #24:
> @Raspberry_yoghurt , thank you for your informative and useful post. And it helps to highlight one of my other points: one of my fears is that some might blindly accept AI's creations as unquestionable, since it comes from "a powerful AI" and therefore must be "the science."
>
> In the field of translation, it is easy for a skilled person like yourself to tell when AI has screwed up. But in some other fields, although AI can apparently still screw up, it will be much harder for all but the most expert to even tell that there's been an error (in some cases, at least).
My guess is that it's probably the same in other fields.
My cousin is an engineer and tried to use chatGPT - but it makes so many mistakes he says it's easier to just do it yourself.
IDK if this will last though, because I guess mathematics is something the AIs should be able to learn.
> I don't oppose AI -- and I think it will lead (and already has led) to some very useful things. But the "garbage in / garbage out" nature of even simple programming continues to scare me. And if AI feasts upon terrabytes of writing (scientific papers, economics journals, newspapers, blogs, and the like) and then draws a conclusion from that writing, its conclusion will only be as good as the writing it uses to train itself. And if that writing is flawed -- because the writers' own understanding is imperfect or colored by some ideological or practical bias -- it can lead, in a completely unintended way, to imperfect or biased conclusions.
I think most people quickly find out the AI makes bad mistakes. They also sometimes say super weird thngs.
ChatGPT for instance always say that everything regarding art is subjective.
So if you ask it "Who was the best drummer in Queen?" it does mention Roger Taylor, but also says it's subjective. But he was the drummer lol. It's silly to say it's subjective if the actual only drummer in the band or the other guys that didn't play drums were the best drummer.
> @Raspberry_yoghurt , thank you for your informative and useful post. And it helps to highlight one of my other points: one of my fears is that some might blindly accept AI's creations as unquestionable, since it comes from "a powerful AI" and therefore must be "the science."
>
> In the field of translation, it is easy for a skilled person like yourself to tell when AI has screwed up. But in some other fields, although AI can apparently still screw up, it will be much harder for all but the most expert to even tell that there's been an error (in some cases, at least).
My guess is that it's probably the same in other fields.
My cousin is an engineer and tried to use chatGPT - but it makes so many mistakes he says it's easier to just do it yourself.
IDK if this will last though, because I guess mathematics is something the AIs should be able to learn.
> I don't oppose AI -- and I think it will lead (and already has led) to some very useful things. But the "garbage in / garbage out" nature of even simple programming continues to scare me. And if AI feasts upon terrabytes of writing (scientific papers, economics journals, newspapers, blogs, and the like) and then draws a conclusion from that writing, its conclusion will only be as good as the writing it uses to train itself. And if that writing is flawed -- because the writers' own understanding is imperfect or colored by some ideological or practical bias -- it can lead, in a completely unintended way, to imperfect or biased conclusions.
I think most people quickly find out the AI makes bad mistakes. They also sometimes say super weird thngs.
ChatGPT for instance always say that everything regarding art is subjective.
So if you ask it "Who was the best drummer in Queen?" it does mention Roger Taylor, but also says it's subjective. But he was the drummer lol. It's silly to say it's subjective if the actual only drummer in the band or the other guys that didn't play drums were the best drummer.
@sdkman, "just to be clear" (as many political hacks like to say) I do NOT think AI will make all doctors, lawyers, engineers or the like disappear.
It will almost certainly NOT replace such people entirely -- at least not in this century.
The question is -- will it replace SOME of them. And will that be a good thing or a bad thing. Well, for anybody replaced it won't feel very good.
AI IS competing with real human beings for work. I don't think that can be seriously doubted. It won't replace ALL humans, of course. I don't see AI crawling under a sink any time soon and replacing the pipes.
But entry-level knowledge workers are already living in a competitive hell in some areas. Will AI help them or replace them?
If I were young, plumbing and electrical work would be looking to me increasingly like a smart choice. Or maybe dental work. Nobody wants AI in their mouth, I suspect. lol.
It will almost certainly NOT replace such people entirely -- at least not in this century.
The question is -- will it replace SOME of them. And will that be a good thing or a bad thing. Well, for anybody replaced it won't feel very good.
AI IS competing with real human beings for work. I don't think that can be seriously doubted. It won't replace ALL humans, of course. I don't see AI crawling under a sink any time soon and replacing the pipes.
But entry-level knowledge workers are already living in a competitive hell in some areas. Will AI help them or replace them?
If I were young, plumbing and electrical work would be looking to me increasingly like a smart choice. Or maybe dental work. Nobody wants AI in their mouth, I suspect. lol.
Alien Illegals
These things could probably do assembly-line work ...
www.youtube.com/watch?v=tF4DML7FIWk
www.youtube.com/watch?v=tF4DML7FIWk
This topic has been archived and can no longer be replied to.