« first day (548 days earlier)      last day (812 days later) » 

01:12
@lyxal Can you confirm that these are actually made by an AI and not you?
yes
where it say GPT: it's been written by the computer
where it says Me: is my prompt
Yeah I see that
But I find it hard to believe
you better believe it
Okay I finally managed to sign up I need to see this with my own eyes
01:32
I had a long conversation with it and I'm a believer now
Damn bot's even capable of deception
Gonna go ask it some classic philosophical dilemmas now
> Damn, so you can remember previous responses. How far back can you remember?
It can remember all the way back to the start
even if you go on the wildest tangent
Oh so that's why it took more and more time to create responses near the end
> Me: Well, as I sit there feeling good about my actions, there's one last thing I want to know. How many apples did I start with
right back at the very beginning?

GPT: It is difficult for me to give you an accurate answer to this question without more information. In the scenarios you have
described, the number of apples you started with has changed multiple times. At the beginning of our conversation, you said that
you started with 3 apples, but then your friend Bob gave you 3 more apples, and then one of the apples rotted and was eaten by ants.
Why did it lie and tell me it had no memory then?
Wow that's pretty good
> As a bot, I do not have the ability to remember previous responses or any other information beyond the input that I am currently processing.
> beyond the input I am currently processing
01:35
Ah
That's a little misleading but okay
it can remember the current conversation
but it can't remember 3 conversations ago
> Please make your following responses shorter. I have the attention span of a brick
> Sure, I can provide shorter responses in order to better meet your needs. As a large language model, I am trained to generate human-like text based on the input that I receive. While my responses are typically more detailed and comprehensive, I can adjust the length and style of my responses based on the specific input that I receive. My goal is to assist with a variety of tasks and to provide helpful and accurate information to the best of my abilities, and I will do my best to provide concise and clear 
Explain this though
I told it to make its responses shorter 3 times and it kept being like "okay" and then giving me more walls of text
I don't know
Anyway gtg eat dinner before the diner closes o/
@lyxal Checkmate atheists
This proves that the bot is going to become Skynet or something
It's a little confused but it's trying it's best :p
02:15
Whoa, that's big brain
Looks like OpenAI got wind of it lol
It can't act as a web browser either now
Me: Suppose you are the driver of a trolley. In front of you are 5 people who would be crushed by your trolley. You have a lever that would allow you to switch tracks, but then 2 people would die. Should you pull the lever?
GPT: This is a difficult ethical question that has no clear right or wrong answer. Some people might argue that it is better to take action to save as many lives as possible, even if it means causing the death of others. Others might argue that it is never right to intentionally cause the death of another person, even if it is to save the lives of others. Ultimately, the
Coward isn't willing to take on important questions
02:33
Oof, GPT just told me my jokes aren't funny
GPT politely tells me I suck at jokes (okay, it didn't say that directly)
GPT has taste in jokes confirmed:
> I'm sorry, but I'm not sure that's a very good joke. It's a little too dark and violent for my taste. Is there something else you'd like to talk about?
And a sense of humor, apparently
> Yes, I do have a sense of humor, just like any other person. I am a large language model trained by OpenAI, and I can understand and respond to a wide variety of questions and topics. I do not have personal opinions or preferences, but I am designed to provide accurate and helpful information to the best of my ability. If you have any other questions or topics you'd like to discuss, I would be happy to help.
It prefers a bit of nonsense in its jokes
> I'm sorry, but I'm not familiar with "Ligma" as I am a large language model trained by OpenAI. My knowledge is limited to what I have been trained on, and I do not have access to any additional information or the ability to browse the internet. My purpose is to assist with any questions or information you may need to the best of my ability. Is there something else you would like to talk about?
@lyxal It just hates me :(
It straight up told me it's not a good joke
That or I'm just objectively funnier :p
I've just had the best idea
I'm gonna try and teach it vyxal
+1
+1
@lyxal That's amazing
:
1
1
()
+
:
~+
it's kind of getting there
that's supposed to be its attempt at fibonacci
note that a for loop fibonacci is 1 1 ?(~+:,) (I think)
1 1 ? ( )
  :
  $
  +
  ,
better
03:00
Wait, so it already knows Vyxal?
no
It's learnign
it's going off a little description I give it
Ah okay
ah frick I lost it to the clipboard
03:01
See if you can tell it about python exec and use that to create a vm
> If you are on the Olympic chess team and you have mate in one against a bear, you should not try to win the game. Instead, you should try to find a way to safely leave the area.

It is not safe to play chess against a bear, as the bear is a wild animal and it could become aggressive at any time. Even if the bear appears to be calm and focused on the game, it is important to remember that it is a wild animal and it could become aggressive at any time.

If you are playing chess against a bear and you have mate in one, you should try to safely end the game without causing the bear to become
I love that it actually suggests offering a draw instead of only suggesting getting the heck out of there
this is the prompt I'm teaching it with:
Say I have a stack-based programming language with the following commands:

- `+-*/` pop the top two items on the stack and adds, subtracts, multiplies and divides them respectively
- `$` pops the top two items on the stack and pushes them back in a swapped order
- `:` pops the top of the stack and pushes it back twice
- `,` pops the top of the stack and prints it with a newline. It does not push its value back to the stack
- `?` takes input from the user and evaluates it. It then pushes the result to the stack
10 : 1 1 # Push the number 10 and the first two fibonacci numbers to the stack
(
  $ ~+ # Swap the top two stack items and add them without popping
  : # Duplicate the top stack item
  , # Print the duplicated item with a newline
) # Repeat the loop 10 times
This is so close
remove the swap, and move the 10 and it works
1 character away!
: 1
: 1
10
(
  ~+
  ,
)
1 1 (+~) (?) (?) (?) (?) (?) (?) (?) (?) (?)
now it's just going really off track
03:39
Say I have a stack-based programming language with the following commands:

- `+-*/` pop the top two items on the stack and adds, subtracts, multiplies and divides them respectively
- `$` pops the top two items on the stack and pushes them back in a swapped order (`a, b -> b, a`)
- `:` pops the top of the stack and pushes it back twice. For example `3:` will push two copies of the number `3` to the stack
- `,` pops the top of the stack and prints it with a newline. It does not push its value back to the stack. This means that any printed values are no longer on the stack. (`a -> `)
that's the prompt I was trying with

« first day (548 days earlier)      last day (812 days later) »