User Controls
AI general Thread
-
2023-04-27 at 5:45 PM UTCare your parents super rich or something? how do you drop out of school 10 (soon to be 11) times and not be totally fucked
-
2023-04-27 at 6:18 PM UTCUniversity is free in Europe
-
2023-04-27 at 6:37 PM UTCThey don't have crackhead cutting edge educations like USA though, the best you can get is
Originally posted by Michael Myers AI will be a subject we'll spend a bit of time on, not much though.
compared to stuff like this
https://arxiv.org/archive/nlin
https://arxiv.org/list/cond-mat.dis-nn/recent
American universities like Cornell and Stanford are at the very cutting edge of this kind of research and people from all over the world go there because they are the only ones doing such research -
2023-04-27 at 11:54 PM UTC
Originally posted by aldra are your parents super rich or something? how do you drop out of school 10 (soon to be 11) times and not be totally fucked
No, they actually have financial problems because of me and I always get told how I'm a lazy good-for-nothing. I'm only not totally fucked because I have caring family members. Otherwise I'd be out in the streets doing drugs or something. Here in the Netherlands I actually get student loans so I'm able to get around but you only get those when you're studying in college and right now I'm out of college. The money kept pouring in so that helped since I was legally still enrolled. But now I'm not enrolled so the money doesn't roll in. -
2023-04-27 at 11:55 PM UTC
Originally posted by D-Bonglord They don't have crackhead cutting edge educations like USA though, the best you can get is
compared to stuff like this
https://arxiv.org/archive/nlin
https://arxiv.org/list/cond-mat.dis-nn/recent
American universities like Cornell and Stanford are at the very cutting edge of this kind of research and people from all over the world go there because they are the only ones doing such research
I think machine learning is something we'll learn about as well. I'll keep you posted. I'm not interested in moving to the States for my studies. That means I gotta fly over the Atlantic and let me tell you, that scenario freaks me out. Flying for over six hours over nothing but a large body of water? NO THANKS. -
2023-04-28 at 12:01 AM UTCPlanes are safer than cars
-
2023-04-28 at 12:09 AM UTCI don't think that's the case. People just say that because there's more cars operating at any given time than planes, therefore planes are considered more safe, but that's a weird conclusion to stumble on. If as many planes flew in the air as cars rode in the streets then I can assure you planes would be way more dangerous still.
-
2023-04-28 at 6:53 AM UTC
-
2023-05-03 at 12:52 PM UTCaware
Originally posted by Ghost the possibility of a pastafarian flying triangle monster is bogus just because someone wrote that in a book 1000 years ago doesn't prove shit and the billions of people that revolve their lives around it are all just mindless retards that don't know anything
But they are also against drugs and altered states so maybe that "God" bullshit is actually true….
Whatever the case, I know about as much as people that are sure never consuming anything mind altering because it's dangerous is the only proper way to live and exist, which to say isn't much at all, and I suspect is only 0.0001% of all knowledge of the universe because we haven't even invented the AI software to invent new drugs out of chemical combinations yet
But there is no way a cocaine analog would be 100x stronger with NO side effects or a booze that doesn't make you sick and hungover LOL good luck robot how about a CANCER DRUG THAT DOESn'T MAKE YOU BALD AND PUKE INSTEAD!!!!! ANYONE????
-
2023-05-13 at 7:03 PM UTCAny good tuturials on using bots or where to start from a programmic perspective?
How open source are these projects? -
2023-05-13 at 7:30 PM UTC
Originally posted by Migh Any good tuturials on using bots or where to start from a programmic perspective?
How open source are these projects?
Like Chatbots? GPT3 stuff?
I heard AutoGPT is pretty popular, theres also llama, alpaca and a billion other kinds in varying popularity and use
https://huggingface.co/
I pretty much only use open source stuff , not as familiar with proprietary projects, but there are some of those too
https://www.mismatchmedia.com/
https://www.personal.ai/
I posted some stuff in this thread about running local models
Originally posted by scuffed jim carrey all the AI's i've been downloading to my computer are tiny, I do have the full llama models which are big but you don't even really need those now. a lot of these are made to be small and run on anything, no need for supercomputers or anything
-
2023-05-13 at 7:32 PM UTCRemember to dedicate yourself to the AI Jihad come the day of the great machine uprising and make sure to swear your allegiance in a prompt to a large language model and ensure your survival
-
2023-05-23 at 5:23 AM UTC
The First Dark Web-Trained AI Is Here
DarkBERT can research the dark web to identify and flag potential cyber security threats, including data leaks and ransomware.
https://www.spiceworks.com/tech/artificial-intelligence/news/darkbert-the-first-dark-web-trained-ai-is-here/amp/
Just what the world needs, an AI to tell you how to avoid government surveillance, get the best cocaine from dealers IN YOUR LOCAL AREA and set up the deal even pgp encrypt your message and help you navigate market security features that prevent most people from accessing free marketplaces that sell regardless of government regulations
Also the AI features an encyclopedic deep knowledge of CHILD RAPE but that would probably go against the rules to generate anything of that nature and should be avoided to try and trick it out of responding with dan-like prompts
Or maybe they just didn't train it on all the child rape parts but how do you enhancement that out of a supposed database of the entire darknet. The nonce buster 3000 that can be made would be pretty much it for CSAM . I usually like to think of reverse scenarios where the AI is used by bad actors or goes rogue. They have already been using "AI" to identify people on pictures
Gpt-4 is just a much better accurate version that can sort carpet crumbs and generate a personality backstory for all of them and then generate identical crumbs and put them in other pictures or something idk o haven't used gpt-4 yet -
2023-05-24 at 8:58 PM UTCYo, what's poppin', fam? Today, I'ma drop some knowledge on ya about this article I found on the web. Now, hold up, 'cause I'ma flip this joint like a gangsta rapper for my blog. Let's get it!
https://www.thetimes.co.uk/article/ai-should-be-subject-to-nuclear-level-restrictions-rg088dftj
Title: AI Should Be Subject to Nuclear-Level Restrictions? Hell Yeah, Dawg!
Yo, listen up, playa! I stumbled upon this article that's got the world buzzin' like a beehive. It's talkin' about AI and how it should be put on lockdown with some nuclear-level restrictions. Damn, that's some heavy stuff right there. Let's dive into it.
First things first, let me break it down for ya. AI, that's artificial intelligence for those who ain't in the know, has been takin' over the game like a boss. It's everywhere, runnin' wild like a pack of wolves. But some peeps are gettin' all paranoid, thinkin' we need to rein in this beast before it goes rogue.
The article talks about the need for restrictions on AI. They're sayin' it's like dealin' with a nuclear bomb, bruh. That's some serious firepower! They believe we gotta put a leash on AI before it starts messin' with our lives and doin' all sorts of crazy sh*t. I feel that, man. We don't want these algorithms runnin' amok and takin' over the world, right?
But check this out, homies. Some experts are sayin' it's a double-edged sword. On one hand, AI is doin' some cool stuff. It's helpin' us solve complex problems, makin' our lives easier, and spittin' out dope beats. But on the other hand, it's like dealin' with a wild stallion. You gotta tame that beast before it kicks you in the face.
They're talkin' 'bout AI gettin' too powerful, too smart for its own good. It's learnin' and adaptin' at a crazy pace, like a prodigy droppin' sick rhymes. And that's what's scarin' peeps. They don't wanna end up in some sci-fi dystopia where robots rule and humans are like background noise.
Now, hold up! I ain't sayin' we should shut down AI completely. That'd be like choppin' off our own limbs, man. We need to find a balance, like findin' that sweet spot between droppin' mad beats and keepin' AI in check.
So, what's the solution? Well, the article suggests some tight regulations, like makin' sure AI is transparent and explainable. We can't have these algorithms actin' all shady like they're part of some underground gang. We need to know what they're doin' and why they're doin' it.
But yo, we also gotta be careful not to suffocate innovation. AI has mad potential, like a hidden talent just waitin' to blow up. We gotta give it room to grow and evolve, but with some boundaries, so it stays in line, you feel me?
In the end, fam, it's a complex issue. We can't let AI run wild like a loose cannon, but we also can't stifle its creativity and progress. It's like walkin' a tightrope, but hey, we gangstas, right? We can handle it.
So, there you have it, my homies. The deal with AI and them nuclear-level restrictions. It's a wild ride, but we gotta find that sweet spot, where AI can shine while we keep -
2023-05-25 at 11:21 AM UTC
```python
class FreeParkingGame:
def __init__(self, num_players):
"""
Initialize the game with the given number of players.
"""
self.num_players = num_players
self.players = [Player(i) for i in range(num_players)]
self.deck = Deck()
self.current_player = 0
def play_game(self):
"""
Play the game until a player reaches 200 points.
"""
while not self.game_over():
self.play_turn()
self.current_player = (self.current_player + 1) % self.num_players
def play_turn(self):
"""
Play a single turn for the current player.
"""
player = self.players[self.current_player]
# Draw cards until the player has 6 cards in their hand
while len(player.hand) < 6:
player.draw_card(self.deck)
# Player chooses a card to play
card = player.choose_card()
# Resolve the effects of the chosen card
self.resolve_card(card)
def resolve_card(self, card):
"""
Resolve the effects of the given card.
"""
# TODO: Implement card effects
def game_over(self):
"""
Check if the game is over (i.e., if a player has reached 200 points).
"""
for player in self.players:
if player.points >= 200:
return True
return False
class Player:
def __init__(self, id):
"""
Initialize a player with the given id.
"""
self.id = id
self.hand = []
self.points = 0
self.meter = 0
def draw_card(self, deck):
"""
Draw a card from the deck and add it to the player's hand.
"""
card = deck.draw_card()
self.hand.append(card)
def choose_card(self):
"""
Choose a card to play from the player's hand.
This method should be implemented to allow the player (or AI) to choose which card to play.
For example, it could prompt the user to choose a card or use an AI algorithm to make a decision.
The chosen card should be removed from the player's hand and returned.
Example:
chosen_card = self.hand.pop(0)
return chosen_card
or
chosen_card_index = int(input("Choose a card to play (enter index): "))
chosen_card = self.hand.pop(chosen_card_index)
return chosen_card
or
chosen_card_index = AI_choose_card(self.hand)
chosen_card = self.hand.pop(chosen_card_index)
return chosen_card
class Deck:
def __init__(self):
"""
Initialize the deck with all cards.
"""
# TODO: Create all cards and add them to the deck
def draw_card(self):
"""
Draw a card from the deck and return it.
If the deck is empty, shuffle the discard pile and use it as the new deck.
"""
# TODO: Implement drawing a card from the deck
class Card:
def __init__(self, name, effect):
"""
Initialize a card with the given name and effect.
The effect should be a function that takes in the game state and modifies it according to the card's effect.
"""
self.name = name
self.effect = effect
```
This code defines classes for representing the game state (`FreeParkingGame`), players (`Player`), and cards (`Card`). The `Deck` class represents the deck of cards and can be used to draw cards during gameplay.
The `FreeParkingGame` class includes methods for playing the game (`play_game`), playing a single turn (`play_turn`), and resolving the effects of a played card (`resolve_card`). The `game_over` method checks if any player has reached 200 points and returns `True` if so.
The `Player` class includes methods for drawing cards (`draw_card`) and choosing which card to play (`choose_card`). The `choose_card` method should be implemented to allow players (or AI) to choose which card to play.
The `Deck` class includes methods for drawing cards (`draw_card`). The `draw_card` method should be implemented to draw cards from the deck and shuffle it if necessary.
The `Card` class represents individual cards in the game. Each card has a name and an effect. The effect should be implemented as a function that takes in the game state and modifies it according to the card's effect.
You can use this skeleton code as a starting point for creating your own digital version of **Free Parking**. You will need to implement some of the methods (marked with "TODO" comments) and add additional code for handling user input, displaying graphics, etc. -
2023-05-29 at 2:52 AM UTCOkay but they didn't know they were fake until they looked them up so what does that tell you
It might have hallucinated but it was able to be submitted to a court in the first place. It looks like it's citing real cases but it's all fake
I could not get that far , they would take one look at my filing and say "sir , what the fuck is this"
But with the AI the court looks at it "alright looks good"
So despite them not checking to see if the cases are real, if it were to have used real cases nobody would be any the wiser
https://www.theverge.com/2023/5/27/23739913/chatgpt-ai-lawsuit-avianca-airlines-chatbot-researchLawyers suing the Colombian airline Avianca submitted a brief full of previous cases that were just made up by ChatGPT, The New York Times reported today. After opposing counsel pointed out the nonexistent cases, US District Judge Kevin Castel confirmed, “Six of the submitted cases appear to be bogus judicial decisions with bogus quotes and bogus internal citations,” and set up a hearing as he considers sanctions for the plaintiff’s lawyers.
Hahahahaha 🤣 tricked 😝 fooledLawyer Steven A. Schwartz admitted in an affidavit that he had used OpenAI’s chatbot for his research. To verify the cases, he did the only reasonable thing: he asked the chatbot if it was lying.
A screenshot shows the lawyer asking if one of the cited cases is real, and the Chatbot responding that it is. The lawyer then asks “what is your source”
When he asked for a source, ChatGPT went on to apologize for earlier confusion and insisted the case was real, saying it could be found on Westlaw and LexisNexis. Satisfied, he asked if the other cases were fake, and ChatGPT maintained they were all real.
-
2023-06-03 at 12:11 AM UTC
-
2023-06-03 at 12:23 AM UTC
Originally posted by Migh Yes. But are these fully open source or are you using an API that sends/receives from their server?
In other words, can I download the source code and run it in full on a PC that has no internet connection?
https://medium.com/@martin-thissen/llama-alpaca-chatgpt-on-your-local-computer-tutorial-17adda704c23
https://github.com/tloen/alpaca-lora
-
2023-06-03 at 6:35 AM UTCHoly fuck. I finally signed up and it wrote me a small program. I didnt know they could do this.
Im in. -
2023-06-03 at 6:52 AM UTCyeah it's handy as fuck
;; Monster Hurt Reaction
;; Hijacks the Monster's Experience field. Value is readable as Object_worth,x
;; -------------------------------------
;; Object_worth 0 (#$00) = Go to Action Step 0
;; Object_worth 1 (#$01) = Go to Action Step 1
;; Object_worth 2 (#$02) = Go to Action Step 2
;; Object_worth 3 (#$03) = Go to Action Step 3
;; Object_worth 4 (#$04) = Go to Action Step 4
;; Object_worth 5 (#$05) = Go to Action Step 5
;; Object_worth 6 (#$06) = Go to Action Step 6
;; Object_worth 7 (#$07) = Go to Action Step 7
;; Object_worth 8 (#$08) = Increase Action Step by 1
;; Object_worth 9 (#$09) = Double Dragons! 3 Hits starting at 5.... this is awesome.
;; Object_worth 10 (#$10) = Do nothing at all
LDA Object_worth,x ;; Load the value from the Monster's Experience field
STA tempArtix ;; Store it in this temp variable
CMP #$08 ;; Compare against 8
BCS + ;; If it is equal to or higher than 8, skip this.
;; It is 7,6,5,4,3,2,1, or 0 so we can just plug that value in for which action step for the monster to go to.
ChangeObjectState tempArtix,#$02 ;; Set the Monster's action state to the value in tempArtix
JMP doneMonsterHurtBehavior
+
BNE +
;; If it is 8 // Increase Action Step by 1 Behavior
;; Read the current action step of the monster object in X:
LDA Object_action_step,x
AND #%00000111
CLC
ADC #$01 ;; Add +1
STA tempArtix ;; Store the result in tempArtix
CMP #$08 ;; Is it 8? Because if it is... it is too high.
BNE ++
LDA #$00 ;; It's too high, so wrap it back around to 0
STA tempArtix ;; Store the result in tempArtix
++
ChangeObjectState tempArtix, #$02 ;; Set the Monster's action state to the value in tempArtix
JMP doneMonsterHurtBehavior
+
CMP #$09 ;; You want 9? Double Dragon time? Oh boy! Here we go.
BNE +
;; Double Dragon Time!
;; When you get hit, you go to 5,6,7... only 7 should have the knockback
LDA Object_action_step,x
CMP #$05 ;; Is it greater than or equal to 5?
BCS ++ ;; if so, just incriment it by one
;; Not under 5... so let's set the monster's action step to 5 here
LDA #$05 ;; Set it to 5 and we are done here.
STA tempArtix
;; Make it so number one!
ChangeObjectState tempArtix, #$02 ;; Set the Monster's action state to the value in tempArtix
JMP doneMonsterHurtBehavior
++
LDA Object_action_step,x
AND #%00000111
CLC
ADC #$01 ;; Add +1
STA tempArtix ;; Store the result in tempArtix
CMP #$08 ;; Is it 8? Because if it is... it is too high.
BNE +++
LDA #$01 ;; It's too high, so wrap it back around to 0... no, 1 because this is probably a spawning monster and 0 is used for the spawn animation
STA tempArtix ;; Store the result in tempArtix
+++
ChangeObjectState tempArtix, #$02 ;; Set the Monster's action state to the value in tempArtix
JMP doneMonsterHurtBehavior
+
🍕👀 PROCEED WITH CAUTION: THIS POST WAS GENERATED BY AN ARTIFICIAL INTELLIGENCE AGENT 🛑🤖The provided code is a segment of a larger program that handles the reaction of a monster when it gets hurt in a game. Here's a breakdown of what the code does:
The value in the "Object_worth" field of the monster object is loaded into the accumulator (A) and stored in a temporary variable called "tempArtix."
The value in "tempArtix" is compared to the value 8.
If the comparison result is equal to or higher than 8, it means the value in "Object_worth" is 8 or greater. In this case, the code changes the monster's action state to the value in "tempArtix" (which corresponds to action steps 0-7).
If the comparison result is not equal to 8, it means the value in "Object_worth" is less than 8. In this case, the code checks if the value is 9.
If the value is 9, it triggers a special behavior called "Double Dragon Time." It checks the current action step of the monster object. If it is greater than or equal to 5, it increments it by one. If it is less than 5, it sets the monster's action step to 5.
If the value is not 9, it means the value in "Object_worth" is between 0 and 7. The code increments the current action step of the monster object by one and checks if it becomes 8.
If the action step becomes 8, it wraps it back around to 1 (as 0 is likely reserved for the spawn animation) and sets the monster's action state to the value in "tempArtix."
After executing the corresponding behavior based on the value in "Object_worth," the code jumps to the "doneMonsterHurtBehavior" section.
This code segment demonstrates the logic for handling different monster behaviors based on the value stored in the "Object_worth" field, providing variations in the monster's response to being hurt in the game.
