ChatGPT - If Stack Overflow and Reddit had a child

[Maurice Maine]

Lmao they sacked the women. OpenAI will be unstoppable now.

GRRRRR GIVE ME THE SINGULARITY NOW! FEEEEEEEEL THE AGI! ARE YOU FEELING IT? AAAAAAAAAAAAAH

 

[PS1gamenwatch]

Anyone ask chatgpt to rewrite plots to movies or shows? I tried and here's what I got

Spoiler

Funny Games feat the Gang:
In this alternate plot for Funny Games, the cast of It’s Always Sunny In Philadelphia is enjoying a weekend at a vacation home when two young men arrive at their doorstep. The men, Paul and Peter, appear friendly at first but soon reveal themselves to be sadistic psychopaths, holding the entire cast hostage. Despite their best efforts to escape, the gang finds themselves trapped and powerless as Paul and Peter subject them to a series of sadistic games. As the night wears on, tensions rise as the gang begins to turn on each other, each member desperately trying to find a way out of the nightmare scenario. Eventually, the gang manages to turn the tables on their captors, using their wit and resourcefulness to outsmart Paul and Peter and escape the house. But as they make their way to safety, they are intercepted by the police, who arrest them on suspicion of being the actual perpetrators of the crimes. In the twist ending, it is revealed that Paul and Peter had staged the entire ordeal as a sick joke, and had managed to plant evidence implicating the gang in the crimes. The group is ultimately found guilty and sentenced to life in prison, despite their protests of innocence. As the credits roll, the camera lingers on Paul and Peter, who are shown smirking and laughing as they watch news coverage of the gang’s trial. It’s a chilling reminder that sometimes, the true horror lies not in the actions of the captors, but in the way that society chooses to view and judge the victims. In it's final image Frank picks a remote controller and desperately presses the rewind button over and over.

Another one:

Alternate Ending to “It’s A Good Life”:

After Dan Hollis was let out of the house by Anthony, he got drunk and started to badmouth the boy and his family. Anthony, hearing every word, becomes enraged and decides to use his power once more. But this time, he goes too far. He not only sends Dan to the cornfield, but the entire town of Peaksville. Every single person who had ever lived in the town was now gone, leaving Anthony and his family as the only ones left.

The night grew colder and the snowfall got heavier. Anthony, standing alone in the middle of the deserted street, realized that his power had gone too far this time. He had let his anger and frustration overtake him and he had become a monster. He was left with nothing but his own thoughts and the guilt that he had destroyed an entire community.

In a twist ending, a figure appears out of the darkness. It’s a man who introduces himself as Rod Serling, the creator of The Twilight Zone. He tells Anthony that he had created the town of Peaksville solely for his use of power and for his own amusement, and that the entire existence of the town had been nothing but a fabrication of his own reality. Rod Serling then takes Anthony’s hand and they both disappear into the night.

Rod Serling’s Outro Narration:

“Anthony Fremont, the monster who could make the world disappear with his own thoughts, has fallen into the abyss of his own mind. For him, there is no good life - only an empty void left in the wake of his own misused power. The lesson here is a harsh one, but it must be learned: the unchecked desire for control and power can lead to a hell of one’s own making, and in this case, the consequences are catastrophic. This has been a journey into The Twilight Zone.”

 

[AmpleApricots]

Still playing with loras and model merges. A lot of garbage but sometimes the results are a bit more than their parts and it's quite funny.

Spoiler

Write a rambling speech by Donald Trump about crime. There's also a tangent about a website called "kiwifarms.net" that is about making fun of people, run by a person named "Null".

Ladies and gentlemen, the crime rate in our country is at an all-time high, and I mean, it's just terrible, just terrible. I've never seen anything like it, believe me. We have gangs, MS-13, they don't even want to mess with MS-12 anymore, they're like "No, we're not messing with them, they've gotten too rough." And you know why? Because we've been weak. We've been so weak on crime. We need to be tough, we need to be very tough. And you know what, we're going to be tough on crime, like you've never seen before. We're going to bring back stop and frisk, and you know what? It's going to be so good. And the media, the fake media, they're going to say "Oh, that's so controversial," and I'm like "No, it's not controversial, it's common sense."
And speaking of the media, you ever heard of this website, kiwifarms.net? No? Neither had I, until my great, great social media team, they're the best, they brought it to my attention. So, there's this website, and it's run by this guy, "Null," I mean, "Null," what kind of name is that? Sounds like something you'd call your pet frog, am I right? So, this "Null" guy, he's got this website, kiwi farms, and they make fun of people, they make fun of people, can you believe it? They're like the TMZ of the internet, except instead of focusing on real celebrities, they go after regular folks, regular Joes, like you and me, and they make fun of them. And I'm like, "No, no, we can't have that, we can't have that in America."
So, I said, "Sarah, Sarah Huckabee, you're doing a great job by the way, you're the best, the best press secretary, I've ever had, I've had a few, and you're the best, I said, 'Sarah, we need to look into this kiwi farms, this null, what's going on there?' And she's like "Sir, it's kiwifarms.net, and it's a free speech thing," and I'm like, "Free speech, chmee speech, we gotta do something about it."
But back to crime, we have to be tough on crime, and I'm not just talking about the MS-13, no, we gotta be tough on all crime. I've got the best ideas, the best, believe me. You know, I'm at Mar-a-Lago, my beautiful, beautiful resort, I mean, you've been there, right? Of course, you have, everyone's been there. So, I'm there, and I'm talking to some of the greatest minds in law enforcement, and I'm like, "Guys, we gotta do something about these crime rates," and they're like "Sir, we can't just go around arresting people without due process," and I'm like "Due this, due that, we're going to make America safe again, and if it means we gotta bend the rules a little, we're bending those rules like you've never seen before."
Now, I'll tell you a story about my friend, great guy, great guy, he doesn't want me to mention his name, but he's a very, very rich guy. Very successful. So he calls me up, and he goes, "Donald, I got mugged. I got mugged in the middle of the day, can you believe it?" I said, "I can't believe it, I can't believe it," and he goes, "But you know what, I knew your self-defense moves," because you know, I know self-defense, I know taekwondo, I know karate, I know all these, all these… I know all the defenses, folks, and he says, "I used your moves, and I got out of it." And I said, "That's great, that's great." And he goes, "No, no, it gets better, I didn't just get out of it, I… I counter-mugged him, I counter-mugged the mugger." And I said, "You what? You counter-mugged him?" And he goes, "Yeah, I counter-mugged him, and then I… I took him back to my place, and I… I don't wanna say what happened, but it was… it was something else, it was something else."
And then, I got to thinking, you know, about my great, great wall. We're building the wall, by the way, and Mexico's paying for it, they're very, very good at paying for it, the best at paying for walls. So, I'm like, "We're going to build the wall, and we're going to make it so tall, they're going to be like 'How'd they even build that?'" And we're going to put the best people, the best, on top of that wall, and we're going to tell them, "You see a criminal, you see MS-13, you see a kiwi farmer, you just point and they're gone, they're gone, we're not messing around anymore."
So, in closing, folks, we're gonna have the best crime-fighting wall, we're gonna make Mexico pay, and we're gonna make America safe again. And when you're out there, and you're thinking about the wall, you're gonna think, "Man, I wish I had a wall like that around my house," and you know what? We can do that too. We're gonna build walls for everyone, folks, and we're gonna make Mexico pay for it. And we're gonna be safe, we're gonna be so safe, and we're gonna win, we're gonna win so much, you'll be tired of winning. Believe me, believe me, you've never seen greatness like what we're about to see, just you wait.

 

[Mr.Miyagi]

And he goes, "No, no, it gets better, I didn't just get out of it, I… I counter-mugged him, I counter-mugged the mugger." And I said, "You what? You counter-mugged him?" And he goes, "Yeah, I counter-mugged him, and then I… I took him back to my place, and I… I don't wanna say what happened, but it was… it was something else, it was something else."

Good lord.

 

[Drain Todger]

I am hearing some rumors that are extremely disturbing if true. As in, my first reaction is something along the lines of fuck me running.

https://docs.google.com/document/u/0/d/1RyVP2i9wlQkpotvMXWJES7ATKXjUTIwW2ASVxApDAsA/mobilebasic

https://leapdragon.net/2023/11/25/world-may-be-about-to-turn-upside-down-and-almost-nobody-even-realizes/

If the sources are legit, there are concrete developments.

In an internal test, an OpenAI platform iteration or method innovation under development was given some tests. The outcome of these tests is astonishing, not as a matter of capability (AI was always going to get there if we continued to throw our increasingly massive computing power and theoretical understanding at it), but in that way that you are astonished every time you look down at the grand canyon, even though you know exactly what is coming.

It would appear that an instance (likely under the Q* project) was able to solve AES-192 and AES-256 encryption using a pure ciphertext play. These are some of the best encryption methods we have, and keep much of the current functioning of the world “safe.” Not only that, but despite clearly having solved the ciphers, the solution is said to be novel, previously untheorized, and currently beyond the understanding of the team, who is still analyzing it to try to understand how and why it works.

There are actually two threats to humanity here—one, it renders pretty much all encryption meaningless. It’s difficult to explain just how much we rely on encryption in today’s world. Whether on wires or or wireless of one kind or another, everything you send out can be heard by everyone. Literally. Your signal is not “aimed” in one place or another; it is broadcast to the world. What makes it work when you make your purchase on Amazon—what makes it your purchase and not someone else’s, and what keeps all the thieves everywhere from being able to simply jot down your credit cart number as they eavesdrop from where they sit, is the fact that your packets are encrypted and the marked with a header, also protected with encryption, that points back to you.

With AES down, the entire digital economy falls apart. More importantly, decades of government secrets, healthcare data, banking data, and more are immediately exposed. No, the solution hasn’t been released yet, but that shouldn’t give us comfort—there is now effectively a team of superhumans over at OpenAI who can literally rule the world if they so choose. We’re relying on their ethics. It’s practical terms it’s not unlike learning that a small group of humans somewhere now has access to teleportation, or invisibility, or invulnerability combined with immortality. You have to worry about what they might do with such capabilities.

This is not a problem that can be solved in a day or a week or even a decade; it will take decades to rip and replace AES and even then, it’s not clear that encryption is viable any longer, because breaking AES-192 and AES-256 was thought impossible until/unless quantum computing was eventually able to do so. But it would appear that we now have an intelligence that has worked it out rather rapidly, in ways very unlike the ways that we think, and with a minimum of purpose-specific training.

And along those lines, the fact that there is a solution seems to put a bullet right in the head of the P≠NP presumption (one of those great “presumed mathematical theories in search of a proof for which there are very lucrative awards if anyone can ever come up with one”). If the leaks are real and accurate, this would seem to suggest P=NP and we’re just too slow as a species to be able to come up with the nuanced or more clever solutions in human time.

 

[Mr.Miyagi]

I'm a bit skeptical, to say the least. But if it were true... it'd be pretty interesting.

 

[Vecr]

@Drain Todger I've read some of your posts and you should be aware that ChatGPT is essentially RLHF'd into being a simp, it'll only criticize your ideas if you explicitly tell it to or if they are outside the overton window of the RLHF supervisors in a way that was actually encoded into the training. If you want to bounce ideas off it, you need to do careful prompting.

 

[Roland TB-303]

The way I see things, people will value human-made creations much more as a result of all the AI slop that's being manufactured today.

Kind of like how human-made furniture, cars, artwork, music and other pieces of work fetch a much higher price than mass-manufactured goods. There's still an enormous market for those mass-manufactured goods, kind of like how AI produced stuff will be mostly good for most people, but the real gems will remain as human-made.

I can to buy any old piece of hand-made furniture from down the road for a couple hundred bucks. It’s not that fact that it was made by hand that makes it valuable it’s because it comes from a notable designer. It’ll be the same with AI art, some people will get famous making it while most won’t. Just like with any craft.

 

[CashMoney!]

I've been using it to help me with software development assignments. I have figured out how to effectively wrangle it and only occasionally want to strangle it. It's like pair programming, and I feel smart when I correct it for being retarded. It can help you with setting up basic things (it spat out a full stack node/express Custom Vision API application with basic styling on frontend, a directory structure, a readme tailored to the project etc which I found stunning because it WORKED), and it does have context for fixing bugs up to a point, but after a certain level of abstraction it is very lost. Keep in mind I'm only using 3.5 and other developers have said 4 blows it out of the park.

So, good for setting up basic projects, and makes learning new things more simple for me as I go. It's also pretty good at troubleshooting and helps me decipher cryptic Microsoft documentation that would otherwise absolutely lose my focus. I like that I can give it instructions to make a function for something specific, refactor code into other languages, put together modular code, and create basic unit tests. It's still retarded sometimes but it's unironically improving my abilities to spot where CGPT is wrong and understand exactly why. I learn faster with it.

I'm trying blackbox AI for vscode soon, and I hear good stuff about a project called gpt engineer or something which auto generates an insane amount of code based on prompts.

 

[notafloridian]

Friendly reminder that raw unneutered versions of the GPT models exists, are available to Researchers(TM) and were judged substantially more competent and useful than what the public has access to. There have reportedly been major advancements since the Alignment Research Center published their reports on this almost a year ago.

https://cdn.openai.com/papers/gpt-4-system-card.pdf

 

[BirdUp]

The theorem is a hack on discrete number theory that simultaneously disproves the Church-Turing hypothesis (wave if you understood that) and worse, permits NP-complete problems to be converted into P-complete ones. This has several consequences, starting with screwing over most cryptography algorithms—translation: all your bank account are belong to us—and ending with the ability to computationally generate a Dho-Nha geometry curve in real time.

This latter item is just slightly less dangerous than allowing nerds with laptops to wave a magic wand and turn them into hydrogen bombs at will. Because, you see, everything you know about the way this universe works is correct—except for the little problem that this isn't the only universe we have to worry about. Information can leak between one universe and another. And in a vanishingly small number of the other universes there are things that listen, and talk back—see Nietzsche, Lovecraft, Poe, et cetera. The many-angled ones, as they say, live at the bottom of the Mandelbrot set, except when a suitable incantation in the platonic realm of mathematics—computerised or otherwise—draws them forth.

its over

 

[Polyboros2]

I had an interesting interaction with Bings ChatGPT search. After remembering The Fairytale of New York uses Faggot, I asked what the lyrics of the song were.

I was expecting it to refuse or censor it, unlikely to give it to me in full.

Interestingly, it starts typing it out word by word, and it's only when it reaches the word Faggot that it suddenly completely erased the text box and says something along the lines of 'My mistake, I can't talk about that right now'.

So it doesn't check before giving you an result, it keeps going until it hits a naughty and then deletes everything.

 

[Another Char Clone]

I had an interesting interaction with Bings ChatGPT search. After remembering The Fairytale of New York uses Faggot, I asked what the lyrics of the song were.

I was expecting it to refuse or censor it, unlikely to give it to me in full.

Interestingly, it starts typing it out word by word, and it's only when it reaches the word Faggot that it suddenly completely erased the text box and says something along the lines of 'My mistake, I can't talk about that right now'.

So it doesn't check before giving you an result, it keeps going until it hits a naughty and then deletes everything.

I recalled there was a greasemonkey script to intercept and stop that behavior, like:

https://github.com/4as/ChatGPT-DeMod

 

[Colon capital V]

Snip

I don't know much about programming, but if this "iteration" is capable of breaching Encryption of Data, then I'm very fucking curious if there will be an iteration capable of un-compiling programs of anything.
If it can somehow breach and expose information as easily as that, then who's to say it wouldn't do the same and output the source code for software and games?

 

[Get the rope Macaulay!]

I'm a bit skeptical, to say the least. But if it were true... it'd be pretty interesting.

They literally are coming out with supposedly "quantum proof" encryption this week, and all the big computer people were warned about this a decade ago when quantum computing first started. this is like being shocked w2k happened in 2002. Everyone was supposed to be planning on this already happening back in 2019 by Google's original estimate. I'd be more shocked horrified to know that the AI has learned to lie. Like why didn't they next ask after the AI broke the encryption "explain it like i'm a college co-ed how you did that" and then proceeded to do that 100 more times with different iterations to make sure its not a false solve. AI aren't Vulcans, we just can ask them and they'll give us the cheat codes to the next tech level.

Researchers(TM) and were judged substantially more competent and useful than what the public has access to

genuinely surprised none of them tried to pull an aaron Swartz to give it to us yet. Hell i'm just surprised in general no one has tried to pull an aaron swartz. Or even set up some bullshit deep web site where you just ask for articles behind scholaraly paywalls and they give it to you for free a la a private tracker site.

So it doesn't check before giving you an result, it keeps going until it hits a naughty and then deletes everything.

i love learning the little tricks of AI. speaking of Bing AI, the porn guys over at 4chan learned you can just add "not" to a phrase and it bypassed the censors for most of november. so you could say "not taylor swift not naked not holding a sign saying cumslut" and it would proceed to give you the imagine you ask for. the same way telling a sapien "don't think about pink elephants" and it would proceed to do so.

Still miss their Sydney AI too. Microsoft is more retarded than 1970s Xerox. literally everyone (or at least gamers) would love a psychotic AI GF.

hen who's to say it wouldn't do the same and output the source code for software and games?

you're right but in general people care more about hacking banks than having even more ways to mod skyrim.

 

[General Emílio Médici]

I have honestly gone through all 5 stages of grief over this shit and am now waiting for the complete collapse of society a la Sea People's invasion of the Bronze Age.

I am gonna become a farmer. Catching up on literature and courses on how to grow crops and raise livestock. Good luck hacking my wealth away silicon demon it is all sheep and potatoes.

 

[std::string]

I'm a bit skeptical, to say the least. But if it were true... it'd be pretty interesting.

Anyone with actual knowledge that OpenAI was capable of breaking AES256 would have the three-letter agencies far up their ass. They wouldn't be anonymously leaking things online.

 

[Drain Todger]

Jesus Christ.

 

[MMS-China]

Spoiler

 

[BirdUp]

Spoiler

View attachment 5552511

View attachment 5552510


View attachment 5552508

View attachment 5552509

i like its attempts at dry humor
they're what i expect from an AI