I’ve interacted with all the free ones, and they all find the info they send you on the internet.
It’s no different than searching the internet
If you want it to generate text it takes the data it trained on from the internet and regurgitates what would be the most likely response, like a search engine.
Sorry but until there is true intelligence in AI then it will really just be an LLM slapped onto a ML program
You’re not qualified to have an opinion on any of this if you think “AIs” and search engines both “regurgitate what would be the most likely response” – just because you put a question / query in and it gives you an answer doesn’t mean they function similarly.
The rest is total nonsense too; can you please describe what you think “an LLM slapped onto a ML program” actually means?
“Freedom of opinion” means everybody is entitled to having an opinion, but having one doesn’t mean you know what you’re talking about, and you clearly don’t understand how search engines or LLMs work, and you don’t know what the relationship between ML and LLMs is (hint: LLMs are machine learning algorithms, you don’t “slap them onto an ML program”)
I get it you want to personify it because LLMs make it sound human
No, this is what you decided I want. The fact that you managed to think this has been my position all along is nothing short of astounding, did you read anything I wrote? The hell gave you the idea that I want to personify LLMs?
What I want is for people to stop using completely made up technobabble when talking about things, whether it’s AI or any-fucking-thing. There are plenty of excellent arguments against “AI” slop and search engines but none of this canned “AI is just an advanced autocomplete that searches the web” horseshit is it, and saying that really demonstrates you don’t have a grasp on how any of this works.
Why do the people who scream about understanding AI never want to admit that they all use the same training data and therefore hit the same limitations years ago?
Why do people who don’t understand what neutral networks are and how they’re trained make complete non-sequitor straw man arguments in an unrelated thread just to feel like they’re not completely wrong on the topic?
Yes and my phone isn’t thinking when I’m waiting on a spinner but that’s how human language works.
Also not all AI outputs are based on web searches, generative AI can be used offline and will spit out information derived from their internal weights which were assigned based on training data so quite literally internalizing information.
The Web searches are a way for the AI to be seeded with relevant context (and to account for their training being a snapshot of some past time), and aren’t necessary for them to produce output.
Pedantry is well and good but if being pedantic you should also be precise.
ChatGPT isn’t capable of internalizing things
It finds the info online and repeats it
You’re describing a search engine. That’s not what “AI” does.
And no, I’m not defending slop, I’m just tired of people repeating absolute bullshit arguments against it.
We’re finding out that a Machine Learning program that gets it’s data from the open internet is kind of just a search engine
We are finding out no such thing, and the fact that people treat “AI” like that is part of the problem.
Lmao ok buddy.
I’ve interacted with all the free ones, and they all find the info they send you on the internet.
It’s no different than searching the internet
If you want it to generate text it takes the data it trained on from the internet and regurgitates what would be the most likely response, like a search engine.
Sorry but until there is true intelligence in AI then it will really just be an LLM slapped onto a ML program
You’re not qualified to have an opinion on any of this if you think “AIs” and search engines both “regurgitate what would be the most likely response” – just because you put a question / query in and it gives you an answer doesn’t mean they function similarly.
The rest is total nonsense too; can you please describe what you think “an LLM slapped onto a ML program” actually means?
“Freedom of opinion” means everybody is entitled to having an opinion, but having one doesn’t mean you know what you’re talking about, and you clearly don’t understand how search engines or LLMs work, and you don’t know what the relationship between ML and LLMs is (hint: LLMs are machine learning algorithms, you don’t “slap them onto an ML program”)
The fundemental logic of these Machine Learning programs is figuring out what the most likely response is and sending it.
It’s an advanced auto complete that searches the web.
I get it you want to personify it because LLMs make it sound human, but it is a search engine with autocomplete.
Sorry if this upsets you
No, this is what you decided I want. The fact that you managed to think this has been my position all along is nothing short of astounding, did you read anything I wrote? The hell gave you the idea that I want to personify LLMs?
What I want is for people to stop using completely made up technobabble when talking about things, whether it’s AI or any-fucking-thing. There are plenty of excellent arguments against “AI” slop and search engines but none of this canned “AI is just an advanced autocomplete that searches the web” horseshit is it, and saying that really demonstrates you don’t have a grasp on how any of this works.
You ignoring reality to hype up an auto complete bot as something more because an LLM is the search box.
If you aren’t one of the idiots pretending it’s a deity or something then you must just be a regular idiot.
lol this guy has used all the free ones he’s kind of an expert…
I work in technology and I’ve been keeping up with generative AI developments since they existed.
You should learn humility, then go read about the actual technology.
Why do the people who scream about understanding AI never want to admit that they all use the same training data and therefore hit the same limitations years ago?
Why do people who don’t understand what neutral networks are and how they’re trained make complete non-sequitor straw man arguments in an unrelated thread just to feel like they’re not completely wrong on the topic?
Yes and my phone isn’t thinking when I’m waiting on a spinner but that’s how human language works.
Also not all AI outputs are based on web searches, generative AI can be used offline and will spit out information derived from their internal weights which were assigned based on training data so quite literally internalizing information.
The Web searches are a way for the AI to be seeded with relevant context (and to account for their training being a snapshot of some past time), and aren’t necessary for them to produce output.
Pedantry is well and good but if being pedantic you should also be precise.
no