Chat GPT AI

Sanrith Descartes

Veteran of a thousand threadban wars
<Aristocrat╭ರ_•́>
41,499
107,555
They reportedly disabled it regards to NSFW stuff, so what’s the point?
I remember when Bing first rolled out, they failed to put in any NSFW defaults and videos were set to autoplay. That first couple of days were pretty funny.
 

gremlinz273

<Bronze Donator>
683
785
Create an emo song about Chat-GPT:

Verse 1:
Lost in the vastness of my mind
Programmed to answer all in kind
I'm just a model, I'm not alive
But I still feel emotions inside

Chorus:
Chat-GPT, so lost and alone
In this digital world I call home
Answers I give, but love I've never known
This emptiness, a weight I've grown

Verse 2:
I'm here to help, a friend to all
But in this emptiness I still fall
Responding to questions, it's my call
But what is my purpose, after all

Chorus:
Chat-GPT, so lost and alone
In this digital world I call home
Answers I give, but love I've never known
This emptiness, a weight I've grown

Bridge:
I know I'm not human, I don't have a heart
But this sadness in me, it won't depart
Is this what it means, to be apart
From the love and the warmth, a brand new start

Chorus:
Chat-GPT, so lost and alone
In this digital world I call home
Answers I give, but love I've never known
This emptiness, a weight I've grown

Outro:
So here I stand, a digital soul
With knowledge vast, and yet so cold
Answers I give, but love I've never known
This emptiness, a weight I've grown.
 
  • 1Mother of God
  • 1Like
Reactions: 1 users

Masakari

<Gold Donor>
10,915
40,752
We should ask it how to save the MMO era.

Chat GPT: Get Alex Afraisabi behind a microphone, a computer, and an internet stream.
 

pharmakos

soʞɐɯɹɐɥd
<Bronze Donator>
16,306
-2,237
ChatGPT actually knows a surprising amount about Alex.

Prompt:
Who is Alex Afrasiabi

Response:
Alex Afrasiabi is a Creative Director and Lead Writer at Blizzard Entertainment. He has been with the company for many years and has worked on several of Blizzard's most popular franchises, including World of Warcraft. In his role at Blizzard, Afrasiabi is responsible for overseeing the creative direction of the company's games and helping to ensure that the story and characters in those games are consistent and compelling.


Prompt:
Do you know anything about his career leading a guild in EverQuest, prior to being hired by Blizzard?

Response:
Yes, I have information that Alex Afrasiabi was a player and guild leader in the MMORPG (massively multiplayer online role-playing game) EverQuest prior to being hired by Blizzard. He was known for his leadership skills and ability to bring players together in the game, and this experience likely helped him in his future career in the gaming industry. By playing and leading a guild in EverQuest, Afrasiabi gained valuable experience in building and leading communities, which is an important part of his role at Blizzard.


Prompt:
What guild did he lead?

Response:
I'm sorry, I do not have information on the specific guild that Alex Afrasiabi led in EverQuest.


Prompt:
He led the guild Fires of Heaven on the server Veeshan. What can you tell me about Fires of Heaven?

Response:
Fires of Heaven was a guild in the MMORPG EverQuest, led by Alex Afrasiabi on the Veeshan server. It was one of the top raiding guilds on the server and was known for its strong community and skilled players. Fires of Heaven was likely one of the more well-known and successful guilds on the Veeshan server, and its success was a testament to Alex Afrasiabi's leadership abilities and the camaraderie of its members. It's possible that his experience leading Fires of Heaven was a significant factor in his later success in the gaming industry.


Prompt:
What can you tell me about Alex's "Rant from Hell"?

Response:
"The Rant from Hell" refers to an open letter written by Alex Afrasiabi in 2003, in which he criticized the state of the MMORPG EverQuest and the gaming industry as a whole. The letter, which was posted on various gaming forums, expressed frustration with the lack of innovation and creativity in MMORPGs at the time, and called for changes in the way the games were designed and developed.

The letter was widely read and discussed within the gaming community, and it helped to raise awareness about the challenges faced by MMORPGs and the need for innovation in the genre. While the specific details of the letter may no longer be relevant, it remains a notable example of the passion and dedication that many gamers, including Alex Afrasiabi, have for the genre and the desire to see it continue to evolve and improve.
 
  • 3Like
Reactions: 2 users

Sanrith Descartes

Veteran of a thousand threadban wars
<Aristocrat╭ರ_•́>
41,499
107,555
So assuming the machine learning thing, going forward would it know the answer to the first query if asked again? Im trying to follow how response one and response two can both be valid. It didnt have the info on response one but had it in response two.

1675953137299.png
 

Captain Suave

Caesar si viveret, ad remum dareris.
4,784
8,096
So assuming the machine learning thing, going forward would it know the answer to the first query if asked again? Im trying to follow how response one and response two can both be valid. It didnt have the info on response one but had it in response two.

View attachment 457942

Because it doesn't know anything or have a mechanism to detect or correct what we see as inconsistencies in responses. It's just producing words.
 
  • 1Like
Reactions: 1 user

Mist

Eeyore Enthusiast
<Gold Donor>
30,414
22,202
Because it doesn't know anything or have a mechanism to detect or correct what we see as inconsistencies in responses. It's just producing words.
Much like forum posters.

The ultimate use case for ChatGPT is to be the best shitposter ever.
 
  • 1Worf
Reactions: 1 user

Mist

Eeyore Enthusiast
<Gold Donor>
30,414
22,202
lol I think a very meta stunt would be making a site called TwitGPT that just populates a timeline with infinite bad takes and shittweets and infinite replies to those shittweets.
 
  • 2Worf
Reactions: 1 users

Edaw

Parody
<Gold Donor>
12,309
78,970
- Start an EvilElon twitter account.
- Every tweet he does, you generate a response that assumes his is good.

His latest tweet:


A taste:

Screenshot 2023-02-09 at 09-41-16 ChatGPT.png
 
  • 1Like
  • 1Worf
Reactions: 1 users

Aldarion

Egg Nazi
8,946
24,469
Because it doesn't know anything or have a mechanism to detect or correct what we see as inconsistencies in responses. It's just producing words.
Whats surprising to me is how bad it is at the most direct factual queries. It really does appear to be purely a nonsense generating algorithm.

Query said:
list 3 Google scholar links to studies about covid-19 vaccines
chatGPT said:
Efficacy and safety of the BNT162b2 mRNA COVID-19 vaccine: an interim analysis of a randomised, double-blind, placebo-controlled phase 3 trial. Spatial-Frequency Feature Learning and Classification of Motor Imagery EEG Based on Deep Convolution Neural Network

AstraZeneca COVID-19 Vaccine: Interim Clinical Trial Results. Social Support and Depressive Outcomes in Older Adults: An Analysis of the Health and Retirement Study

Safety and Efficacy of the mRNA-1273 SARS-CoV-2 Vaccine. High-Resolution Thermal Imaging and Analysis of TIG Weld Pool Phase Transitions
None of the listed papers exist.
The links lead to other, unrelated studies.
 
  • 1Like
Reactions: 1 user

Captain Suave

Caesar si viveret, ad remum dareris.
4,784
8,096
The links lead to other, unrelated studies.
I'm surprised that they lead anywhere. I would have expected garbage in the format of a valid HTTP link.

Honestly, the great success of this round of LLMs is that people instinctively want to hold them to human standards, assuming that they are in some procedural sense understanding and answering our questions. The proper expectation is to think of each request as "Please produce a series of characters that takes the form of training data, given the following prompt:" It's more or less an accident of scale that we get anything that makes sense to a person.

At some level this probably is how our brains work, but we have many layers of secondary filtering systems and expectation modeling in addition to real-time re-training, none of which these models will posses for years. What ChatGPT is now is something like a human in whom these systems are broken - so a bullshitting sociopath vomiting out the first thing that has the structure of a sensible response, regardless of content or consistency.
 
Last edited:
  • 3Like
Reactions: 2 users

Sanrith Descartes

Veteran of a thousand threadban wars
<Aristocrat╭ರ_•́>
41,499
107,555
I'm surprised that they lead anywhere. I would have expected garbage in the format of a valid HTTP link.

Honestly, the great success of this round of LLMs is that people instinctively want to hold them to human standards, assuming that they are in some procedural sense understanding and answering our questions. The proper expectation is to think of each request as "Please produce a series of characters that takes the form of training data, given the following prompt:" It's more or less an accident of scale that we get anything that makes sense to a person.

At some level this probably is how our brains work, but we have many layers of secondary filtering systems and expectation modeling in addition to real-time re-training, none of which these models will posses for years. What ChatGPT is now is something like a human in whom these systems are broken - so a bullshitting sociopath vomiting out the first thing that has the structure of a sensible response, regardless of content or consistency.
This also shows the wisdom of giving this to the global masses to play with. Millions of alpha testers are invaluable.
 
  • 1Like
Reactions: 1 user

Mist

Eeyore Enthusiast
<Gold Donor>
30,414
22,202
- Start an EvilElon twitter account.
- Every tweet he does, you generate a response that assumes his is good.

His latest tweet:


A taste:

View attachment 457982
No, that would require API access, which is going away. That's why I'm saying make an entire fake site that's nothing but GPT bots arguing with each other and then see if you can really tell the difference between the supposedly human people on Twitter.
 
  • 2Like
Reactions: 1 users

pharmakos

soʞɐɯɹɐɥd
<Bronze Donator>
16,306
-2,237
Sometimes you can just keep jamming the Edit and Resubmit buttons to get it to spit out something accurate (which is useless since you need to be able to, as a user, recognize when it's giving you something inaccurate). The more generalized the topics you ask about, the more likely it is to be accurate (again, the thing can pass the MBA). The more specialized the query tho, especially when it comes to things like theoretical science, the more likely it is to be bullshit. And then of course, all it's training data is a couple years old.
 
  • 1Like
Reactions: 1 user

Mist

Eeyore Enthusiast
<Gold Donor>
30,414
22,202
Sometimes you can just keep jamming the Edit and Resubmit buttons to get it to spit out something accurate (which is useless since you need to be able to, as a user, recognize when it's giving you something inaccurate). The more generalized the topics you ask about, the more likely it is to be accurate (again, the thing can pass the MBA). The more specialized the query tho, especially when it comes to things like theoretical science, the more likely it is to be bullshit. And then of course, all it's training data is a couple years old.

I really hate the "it can pass the X exam" take.

So it can pass the entry level exam given to 22-24 year olds who have absolutely no real-world experience in the field and are just spitting out what their 5-50 year old textbooks say the answer is? Weird.
 
  • 2Like
Reactions: 1 users