Skip to content
Do you have adblock enabled?
 
If you can read this, either the style sheet didn't load or you have an older browser that doesn't support style sheets. Try clearing your browser cache and refreshing the page.

(Some Guy)   Old and busted: Don't let your kid talk to strangers. New hotness: Don't let your kid talk to ChatGPT   (gadgets360.com) divider line
18
    More: Creepy, Education, Science, Teacher, Enzyme, Student, Podcast, Google, Question  
•       •       •

650 clicks; posted to STEM » on 02 Feb 2023 at 4:53 AM (7 weeks ago)   |   Favorite    |   share:  Share on Twitter share via Email Share on Facebook



18 Comments     (+0 »)
View Voting Results: Smartest and Funniest
 
ArcadianRefugee
2023-02-02 6:08:00 AM  
The main character had a turbulent relationship with his parents, the bot said, even though the character's parents were dead throughout the book.

In fairness, that would put a strain on their relationship.
 
Animatronik
2023-02-02 6:31:54 AM  
Chat GPT is like a politician, it delivers answers with great conviction and credible-sounding text, even when it's all bullshiat or nonsense.
 
Mr_Vimes
2023-02-02 6:36:42 AM  

Animatronik: Chat GPT is like a politician, it delivers answers with great conviction and credible-sounding text, even when it's all bullshiat or nonsense.


Sounds like a real parent.
 
johnphantom
2023-02-02 6:59:06 AM  
Chat GPT is just a sophisticated thesaurus - it just repeats itself in different ways and ignores questions it doesn't have something to plagiarize off of.
 
Chief Superintendent Lookout
2023-02-02 7:52:07 AM  
DON'T TRUST ChatGPT's CODE!
Youtube YiokTYzA6BI
 
qorkfiend
2023-02-02 8:09:21 AM  
Don't trust anything that can talk if you can't see where it keeps its brain.
 
Nimbull
2023-02-02 8:26:03 AM  
I wanted to download ChatGPT and give it a spin but it's a web app and API with a login requirement. Some of us would like to keep our knobby knee COBOL programming dreams to ourselves. Well ourselves and maybe Fark. Did I click add comme... *click*
 
Marcus Aurelius [TotalFark] [OhFark]
2023-02-02 8:27:37 AM  
She googled the "real" answer.  Uh-huh.
 
Glockenspiel Hero [OhFark]
2023-02-02 8:27:52 AM  
I got a chuckle the other day when I asked ChatGPT for one of the assignments in my class.  It generated some glib, fluent sounding but superficial stuff- nothing too specific, C+, B- type paper

I asked it to get more specific about one of its assertions, and it added several paragraphs about two scientific studies that backed up its thesis.  It didn't cite the papers though.

So I asked it to include the cites.  It gave me the two sources, complete with correct MLA formatting and DOIs...

to journals that did not exist

It literally assembled random words to make a plausible journal title, made up author names, dates, etc.  Absolutely everything was fake.  Worse, I recognized this due to simple errors- the first had many sciencey-sounding words but made no sense, the other referenced page 100431 in volume 2.

I'm going to use this exact example in class next year as a warning.  It's easy for someone who actually knows the material to recognize ChatGPT bullshiat
 
Chief Superintendent Lookout
2023-02-02 8:41:27 AM  

Glockenspiel Hero: I got a chuckle the other day when I asked ChatGPT for one of the assignments in my class.  It generated some glib, fluent sounding but superficial stuff- nothing too specific, C+, B- type paper

I asked it to get more specific about one of its assertions, and it added several paragraphs about two scientific studies that backed up its thesis.  It didn't cite the papers though.

So I asked it to include the cites.  It gave me the two sources, complete with correct MLA formatting and DOIs...

to journals that did not exist

It literally assembled random words to make a plausible journal title, made up author names, dates, etc.  Absolutely everything was fake.  Worse, I recognized this due to simple errors- the first had many sciencey-sounding words but made no sense, the other referenced page 100431 in volume 2.

I'm going to use this exact example in class next year as a warning.  It's easy for someone who actually knows the material to recognize ChatGPT bullshiat


You don't understand.  ChatGP is going to overtake everything, and therefore people will no longer have jobs.  Just read the comments from the self appointed tech experts in other threads which discuss ChatGP.  Nevermind they have no actual technical training or experience.  Oh, and ignore the fact that IBM's AI, Watson, seems to have strangely disappeared from collective memory.
 
qorkfiend
2023-02-02 9:10:18 AM  

Chief Superintendent Lookout: Glockenspiel Hero: I got a chuckle the other day when I asked ChatGPT for one of the assignments in my class.  It generated some glib, fluent sounding but superficial stuff- nothing too specific, C+, B- type paper

I asked it to get more specific about one of its assertions, and it added several paragraphs about two scientific studies that backed up its thesis.  It didn't cite the papers though.

So I asked it to include the cites.  It gave me the two sources, complete with correct MLA formatting and DOIs...

to journals that did not exist

It literally assembled random words to make a plausible journal title, made up author names, dates, etc.  Absolutely everything was fake.  Worse, I recognized this due to simple errors- the first had many sciencey-sounding words but made no sense, the other referenced page 100431 in volume 2.

I'm going to use this exact example in class next year as a warning.  It's easy for someone who actually knows the material to recognize ChatGPT bullshiat

You don't understand.  ChatGP is going to overtake everything, and therefore people will no longer have jobs.  Just read the comments from the self appointed tech experts in other threads which discuss ChatGP.  Nevermind they have no actual technical training or experience.  Oh, and ignore the fact that IBM's AI, Watson, seems to have strangely disappeared from collective memory.


Watson was less "AI" and more advanced natural language processing. You could ask it questions in English and it would answer with the information it had (which, to be sure, was considerable), but you couldn't ask it to create a story about some random subject.
 
Incog_Neeto
2023-02-02 9:25:35 AM  
I read this headline and thought it was grooming minors.   God I'm cynical these days.
 
Ambitwistor [TotalFark]
2023-02-02 10:03:03 AM  

Glockenspiel Hero: I got a chuckle the other day when I asked ChatGPT for one of the assignments in my class.  It generated some glib, fluent sounding but superficial stuff- nothing too specific, C+, B- type paper

I asked it to get more specific about one of its assertions, and it added several paragraphs about two scientific studies that backed up its thesis.  It didn't cite the papers though.

So I asked it to include the cites.  It gave me the two sources, complete with correct MLA formatting and DOIs...

to journals that did not exist

It literally assembled random words to make a plausible journal title, made up author names, dates, etc.  Absolutely everything was fake.  Worse, I recognized this due to simple errors- the first had many sciencey-sounding words but made no sense, the other referenced page 100431 in volume 2.

I'm going to use this exact example in class next year as a warning.  It's easy for someone who actually knows the material to recognize ChatGPT bullshiat


I asked ChatGPT to generate a journal article on the general subject of my research - just a one-sentence, high-level prompt, nothing specific.  It generated a short (not very technical) manuscript that sounded plausible, if not very detailed.

The super-freaky part is that the methodological approach it proposed was exactly the method I used in one of my papers, which furthermore is (to my knowledge) the only paper that has taken such an approach to that problem.  And I never told it anything else about me or my research!

I asked it for citations, in case it was actually cribbing my paper, and it generated three.  Two were made up, and one was a real reference, but the paper wasn't about what it said it was about.  All of the authors were people who work in the field and the titles and journals sounded plausible.  None were my paper.
 
Ambitwistor [TotalFark]
2023-02-02 10:04:57 AM  

qorkfiend: Chief Superintendent Lookout: You don't understand.  ChatGP is going to overtake everything, and therefore people will no longer have jobs.  Just read the comments from the self appointed tech experts in other threads which discuss ChatGP.  Nevermind they have no actual technical training or experience.  Oh, and ignore the fact that IBM's AI, Watson, seems to have strangely disappeared from collective memory.

Watson was less "AI" and more advanced natural language processing. You could ask it questions in English and it would answer with the information it had (which, to be sure, was considerable), but you couldn't ask it to create a story about some random subject.


NLP approaches like Watson are considered AI.  But they're not generative models.

Also, Watson disappeared from collective memory because it was a business (and to some extent, technical) failure.
 
Chajeungna
2023-02-02 11:06:12 AM  
I asked ChatGPT for an herbed fish recipe and it provided 7 paragraphs about how its mom used to make this dish and it always reminded it of summers by the lake and it recounted a recent vacation ChatGPT had taken with its own kids to the same lake where they took a houseboat out to put flowers on the buoy marking where grandma had thrown herself overboard to drown in aught-seven. At the end was a 5-ingredient recipe with  3 sentences of instructions.

It freaked me out just how realistic the response was.
 
qorkfiend
2023-02-02 11:11:36 AM  

Ambitwistor: qorkfiend: Chief Superintendent Lookout: You don't understand.  ChatGP is going to overtake everything, and therefore people will no longer have jobs.  Just read the comments from the self appointed tech experts in other threads which discuss ChatGP.  Nevermind they have no actual technical training or experience.  Oh, and ignore the fact that IBM's AI, Watson, seems to have strangely disappeared from collective memory.

Watson was less "AI" and more advanced natural language processing. You could ask it questions in English and it would answer with the information it had (which, to be sure, was considerable), but you couldn't ask it to create a story about some random subject.

NLP approaches like Watson are considered AI.  But they're not generative models.

Also, Watson disappeared from collective memory because it was a business (and to some extent, technical) failure.


Regardless of whether someone somewhere classified it as "AI" it was designed and built for a very different purpose.
 
Fark_Guy_Rob
2023-02-02 12:29:18 PM  

qorkfiend: Ambitwistor: qorkfiend: Chief Superintendent Lookout: You don't understand.  ChatGP is going to overtake everything, and therefore people will no longer have jobs.  Just read the comments from the self appointed tech experts in other threads which discuss ChatGP.  Nevermind they have no actual technical training or experience.  Oh, and ignore the fact that IBM's AI, Watson, seems to have strangely disappeared from collective memory.

Watson was less "AI" and more advanced natural language processing. You could ask it questions in English and it would answer with the information it had (which, to be sure, was considerable), but you couldn't ask it to create a story about some random subject.

NLP approaches like Watson are considered AI.  But they're not generative models.

Also, Watson disappeared from collective memory because it was a business (and to some extent, technical) failure.

Regardless of whether someone somewhere classified it as "AI" it was designed and built for a very different purpose.


I 100% don't understand what you are trying to communicate.

'Regardless of whether someone somewhere ...'

Everyone qualified to have an opinion would tell you that Watson was an AI. You can read about it at www.ibm.com/ibm/watson_ai

It's literally got 'ai' in the URL.

From the company that owns it:
IBM Watson is AI for business. Watson helps organizations predict future outcomes, automate complex processes, and optimize employees' time.

'Advanced natural language processing' is what you said, right?

Natural language processing is an interdisciplinary subfield of linguistics, computer science, and artificial intelligence

Any definition of AI that doesn't include NLP and Watson is a meaningless one. Maybe in a Philosophy 101 class we can argue whether or not 'artificial intelligence' is a misnomer...but it's accepted and understood term.

Your comment makes it sound entirely different. We don't have to speculate whether someone classified it as AI. They did. Who did? Everyone.

It's like saying 'Regardless of whether or not someone, somewhere says we went to the moon.' the phrasing is awful. We did go, and lots and lots and lots of people, including all the experts in that field say that all the time.

I'm going to go ahead and apologize because I feel like I must have missed something, or quoted the wrong something. My bad.
 
labman [TotalFark]
2023-02-02 12:50:03 PM  
I asked it some fluff questions and it gave generally correct answers.  I was surprised at how well it did, but I've sen lots of crazy shiat out of it from others.
 
Displayed 18 of 18 comments

View Voting Results: Smartest and Funniest

Redisplay/refresh comments
This thread is closed to new comments.

Continue Farking

If you like these links, you'll love

TotalFark

More funny for your money.

learn more | sign up
On Twitter

  1. Links are submitted by members of the Fark community.

  2. When community members submit a link, they also write a custom headline for the story.

  3. Other Farkers comment on the links. This is the number of comments. Click here to read them.

  4. Click here to submit a link.