And likes to go away on mental breaks. 
I'll just leave it as that. 
212-209-2950 or go to give to WBAI.org online. 
I'm not going to say anymore because I might just got myself into trouble. 
It's 7 p.m. 
Stay tuned for Off the Hook. 
And this is WBAI New York. 
All circuits at the location you are calling are busy now. 
078-T. 
If you'd like to make a call, please hang up and try again. 
If you need help, stay on the line and a hacker will assist you shortly. 
The telephone keeps ringing. 
So I ripped it off the wall. 
I cut myself while shaving. 
Now I can't make a call. 
We couldn't get much worse. 
But if they could, they would. 
Bum-Dig-Li-Bum for the best to stay for worst. 
I hope that's understood. 
Bum-Dig-Li-Bum! 
Thank you. 
And a very good evening to remind the program is Off the Hook. 
Emmanuel Goldstein here with you on this Wednesday evening, joined tonight by Kyle. 
Yo. 
And over in Skypeland, looks like we have Rob T. 
Firefly. 
Good evening. 
Gila. 
Good evening. 
And Alex. 
Hello. 
Good evening. 
It's always something strange with you, Alex. 
Where are you this week? 
I'm in London, actually. 
So it's just past midnight here. 
And hello, hello. 
And you just went over there to be on the show. 
To sound like you're in an exotic place and to be the furthest away. 
Must be nice. 
You know, I'm trying to do the show from as many far-flung locations as possible. 
Uh-huh. 
Well, you've been in Israel. 
You've been in London. 
You know, I've lost track. 
And, of course, places in the United States as well. 
Yeah. 
A bunch of places. 
Vancouver, I think, on a plane. 
Las Vegas. 
Yeah. 
And it's been an interesting year. 
Well, I've got something rather interesting for us to focus on tonight. 
And this might just take up the whole show. 
I don't know. 
Have you guys been familiar with this new release? 
It actually came out in November. 
It's called ChatGPT. 
It's called – it stands for Generative Pre-Trained Transformer. 
That's what the GPT stands for. 
It's basically what's called a chatbot. 
This company called OpenAI launched it back in November. 
And it's kind of an experimental state right now. 
I've been up almost all night playing with this. 
It's that interesting. 
Imagine if you live, say, in the 16th century and all of a sudden you saw a 747 pass over you. 
I kind of feel like we're at that stage right now. 
I just want to give you an example, okay? 
And I'm logging into the system right now. 
So if you folks are logged in, you can do some experiments. 
Or if you want me to just do an experiment, I can do that as well. 
I ask this thing. 
You ask it to do something or to answer something or to speak in a certain method or manner of speaking. 
It does a pretty good job of doing what you ask. 
This is what I asked. 
Write a poem on the subject of computer hacking. 
Now, if I asked any of you to do that, I would have to give you at least a week to come up with something, you know, decent. 
All right, maybe a day for a couple of you. 
But, you know, it would take a while. 
This thing typed it out as soon as I said it. 
And this is what I came up with. 
In shadows deep and dark, where code and circuits spark, lives a rogue, a hacker true, with skills and tricks to see him through. 
He prowls the net with stealth and speed, a master of the digital deed. 
He seeks the secrets that lie hidden and makes them his, unbidden. 
He's a ghost in the machine, a shadow in the screen. 
He slips through firewalls like a breeze and bends the rules as he sees fit with ease. 
He breaks the locks that bind and leaves the system blind. 
He steals the data that he craves and never leaves a trace. 
But as he sits alone in the dark and cold, he wonders if his deeds are just and if his actions will be told. 
For in the end, he's just a thief, stealing for his own belief. 
But what he takes he can't replace, and the damage he does is hard to trace. 
So beware the hacker's hand and watch your digital land. 
For the shadows hide the ones who'd steal your data one by one. 
Apart from calling us all thieves, I am very impressed with what it came up with. 
It's got rhythm, it's got rhyme, it touches upon all kinds of amazing things. 
And it was done right in front of me as I asked it to do this. 
It's not taking a poem from someplace, it's creating a poem. 
And I kind of feel like I'm living in science fiction. 
I heard gender bias. 
Okay, there was that. 
Yeah, there was the he thing in there. 
It's assuming hackers are all male. 
That's true, that's true. 
You will find things like that. 
So that's authentic for an AI. 
Alex, what do you think about this? 
Okay. 
Rob, what do you think about this? 
I have so many feelings about the AI-based so-called creative type things like ChatGPT. 
Just as someone who works in creative fields, I'm already seeing working human artists lose work to, say, AIs that can generate an image. 
Although I don't know that people will be, like, human poets will be losing their jobs to ChatGPT anytime soon. 
But who knows down the line? 
Well, I don't know. 
I mean, if I submitted this to, you know, a publication, a poetry publication, I don't know. 
That might get accepted. 
How do you tell if this was written by, you know, a human or some kind of, you know, chat program? 
This is true. 
Other forms of art still have some telltales, like in the image field, the AIs that generate images are having a lot of trouble with ears and with hands right now. 
But that's, you know, that will be smoothed out in time. 
And, yeah, it's something to really think about. 
Alex, you're back. 
Did you hear any of that? 
Yeah, I did. 
Yeah, I think I heard most of it. 
I heard the entire poem. 
The content was impressive, but, you know, I think you're putting a little too much stock in creating a rhyme that the scansion of which is apparently just rhymed couplets. 
Right. 
I mean, if you put this into iambic pentameter or you asked it to do it in the form of a Shakespearean sonnet, I think it would have been far more impressive than rhymed couplets. 
That's just like, you know, a very weak version of rap or something. 
It's, you know, not well. 
But look, on the other hand, I've been impressed with it as well. 
I'm going to try to turn up something that one of our colleagues had sent me as well that was created with chat GPT about about an Amazon warehouse. 
Actually, I need to I need to try to find this. 
So you guys keep talking while I locate that. 
Well, I'm glad you mentioned Shakespearean sonnets because I happen to have one right here that I asked this this program to write. 
You're kidding. 
On the M32 bus line. 
Yeah, I asked to write an essay, an essay on the history of the M32 bus line in New York City in the style of Shakespeare. 
And this is what we have. 
Again, it came out immediately. 
Upon the streets of New York City, where hustle and bustle doth reign, a conveyance was born to ease the burden of the strain, the M32 bus whose journey through time I shall relate, a tale of service changes and the city's ebbing state. 
In days of old, when horses pulled the carriage and the coach, this bus did make its debut as a humble means of approach. 
To places far and wide, within the city's bustling core, connecting east to west, it quickly became adored. 
Through war and peace, the M32 did trundle on its way. 
Through snow and sleet and rain, through bright and darkest day, it bore the masses to and fro with steady, steady pace. 
And in its prime, it was a symbol of the city's grace. 
But as the years went by and progress did march on, the M32 did face its trials and its struggles anon. 
The city changed and so did needs and routes were realigned and service was adjusted to the changing times. 
Despite these changes, the M30... 
Wait a minute. 
It kind of ends in the middle here. 
I didn't get the rest of it. 
It cut off. 
And I realize now, you know, there is no M32 bus. 
And this thing seemed to know that. 
It seemed to be saying that there once was an M32 bus, but schedules changed and things are no longer the way they were. 
How sad. 
How very sad. 
Now, you had something else, Alex? 
And then... 
Well, that's a little scary. 
I do have to say, I couldn't locate this. 
You know, a colleague of ours, somebody who's spoken at Hope several times as well, had had ChatGPT create a song about working in an Amazon warehouse. 
And it was really, really good. 
But I think it has disappeared on me, unfortunately, on Signal. 
So, we don't have it, but it is really impressive, scary. 
It does feel like this is a, quote-unquote, watershed moment, doesn't it? 
It does. 
It does. 
You know, I'll be able to recreate that as soon as I figure out how to log back in, because apparently I timed out. 
Go ahead, Gila. 
I just asked it to write a villanelle about computer hacking, the villanelle being one of my favorite poetic forms. 
And if I could share what I got, which I think is interesting, germane to what you had before. 
Please, go ahead. 
The code is my canvas, the keyboard my brush. 
I paint in the dark, a digital lush. 
I hack into systems with nimble rush. 
I am the puppet master pulling the strings. 
I dance with the machines on cyber wings. 
The code is my canvas, the keyboard my brush. 
I am the ghost in the machine, unseen. 
I am the glitch in the system, unclean. 
I hack into systems with nimble rush. 
I am the shadow in the digital realm. 
I am the virus causing the overwhelm. 
The code is my canvas, the keyboard my brush. 
I am the hacker, feared and reviled. 
I am the one who makes the networks wild. 
I hack into systems with nimble rush. 
The code is my canvas, the keyboard my brush. 
I paint in the dark, a digital lush. 
I hack into systems with nimble rush. 
So, yeah, you know, apart from the fact that there was a comma splice in every single line, and Rob can tell you comma splices are among the banes of my existence, I thought it was interesting. 
I also think it's interesting that it was pulling a lot of the same imagery that was in your poem as well. 
You know, if you do this enough, you will notice some repetition. 
You will notice some patterns. 
And I think that's probably the best way to be able to detect if you're reading something that was written by a machine or by a human. 
But that said, the way this technology develops, I'm sure that is just a temporary symptom that will soon be dealt with, and it will become even more difficult. 
Rob? 
The basic limit to AI-generated text and content like this is that AI can't actually be creative. 
It can take a base of knowledge, a base of information, other poems, other information, stuff from the Internet, wherever, and it can sort of mash it all up together into a generic sort of estimation of what another creative work in that category might look or sound like. 
And that's interesting, but it's also raising a lot of ethical questions. 
There are a lot of artists out there right now who are finding their work copied by AI bots to the point where you could put so-and-so style painting of such-and-such, and it will generate it. 
And people are getting wise to the fact that they never consented to their work being used in training these bots and possibly down the line taking work away from them. 
So it's interesting to see this happening. 
We're at a really fascinating point in history for all of this right now. 
It's really fascinating because we're seeing paintings, artwork being created on demand. 
Basically, you can give commands to certain other chatbots that will create a work of art for you based on something that you tell it. 
But the same thing with music and here with words, it's fascinating. 
But like everything we talk about, what comes along with the fascination is also the wariness of how this can be misused and how this can be a bad thing. 
Alex, this is the ultimate in crazy. 
I recreated your Amazon song, at least I think I did. 
I asked the chatbot to write a song about working in an Amazon warehouse. 
But before I could do it, I had to re-log in. 
And this damn thing asked me to prove I was human in order to log in, the ultimate irony. 
So once I proved I was human, I was allowed to ask the computer to create a song. 
And here is the song. 
I'm not going to sing it, but you can imagine. 
Picking up boxes in the warehouse, scanning and packing with care, the conveyor belt keeps moving, but the work is always there. 
Working in an Amazon warehouse where the days are long and the pay is low, but we keep on going for the promise of a better tomorrow. 
That, by the way, is the chorus. 
Verse 2. 
The robots and the machines are always humming loud, but the human touch is needed to sort and pack it all out. 
Working in an Amazon warehouse where the days are long and the pay is low, but we keep on going for the promise of a better tomorrow. 
I kind of like the song. 
And here's the bridge. 
We may be tired and we may be weary, but we'll keep on working through the night for the sake of getting packages to their destination and making customers feel all right. 
And the outro. 
So here's to all the workers in the Amazon warehouse. 
We may not get the recognition, but our work is never a farce. 
All right. 
A little clumsy. 
A little clumsy there. 
I would have ended it differently, but, I mean, come on. 
That's not – was there anything close to what you had, Alex? 
No, it was totally different, actually. 
The song that was generated and sent to me by working in an Amazon warehouse was more of a riff on that old classic song, you load 16 tons and what do you get? 
Another day, older and deeper and dead. 
You know that one? 
Yeah, absolutely. 
Yeah. 
So it was more about that style of the sense of the repression of the worker. 
Not that this seemed to have some kind of nobility about the Amazon worker and that they're persevering. 
They may, you know, not be the best treated, but they're there doing some kind of noble job in order to satisfy the end customer, the subscriber of Amazon Prime. 
Right, right. 
By the way, the song that you mentioned was literally used in South Park during an episode about Amazon warehouses. 
So I think it's, you know, it's drawing upon input of various sorts. 
And here, you know, with that Amazon song, you see that they know that Amazon workers are paid very poorly and that they work very hard. 
And it's getting that information and putting that into the piece that it gives back to us, despite the fact that the plastics or what you want to call it. 
What's a non-offensive term to call these creatures? 
The robots are, no, that's outdated. 
Well, whatever they are. 
It doesn't paint them in the best light. 
Let's put it that way. 
But still, they're not censoring themselves. 
There's no self-censorship yet that I've been able to find. 
It's interesting because some of these, like ChatGPT does have some notices that come up in the beginning about warning you that certain biases may come up. 
And, you know, not to depend on it for advice and things, not to share personal information, private information with it. 
And basically putting it across the fact that right now this is just a toy. 
Okay. 
What can we tell, what can we ask that could, you know, break this thing? 
I want to ask it what it's like to be, you know, artificial intelligence. 
How do I ask that? 
What would you say? 
How would you phrase it? 
Anyone? 
I think that's phrased pretty well. 
I would agree, yeah. 
I didn't finish the sentence. 
Write an essay on what? 
I mean, I just got an essay on racism in the Mikado. 
Okay, but that doesn't help me here. 
Write an essay. 
No, I write an essay about anything. 
Just, you know, I think when you ask the simplest kind of question, I said, is the Mikado racist? 
I got a five-paragraph essay. 
Okay, so I think to say, are you happy as artificial intelligence? 
Okay. 
Write an essay on whether you're happy as artificial intelligence? 
No, not even write an essay. 
Just ask it. 
Are you happy? 
Okay. 
I haven't actually entered into a dialogue like that, so I'm going to ask it a question and see what it answers. 
Is that what you're suggesting? 
Yes. 
Okay. 
Are you happy? 
And what's the rest of it? 
As artificial intelligence. 
As artificial intelligence. 
Okay, this ought to be good. 
Oops, I spelled intelligence wrong. 
Okay. 
Are you happy as artificial intelligence? 
And here we go. 
As a machine learning model, I don't have the capacity to experience emotions such as happiness or unhappiness. 
It's like talking to commanded data here. 
Happiness or unhappiness. 
My main function is to process and generate human-like text based on the input and training data that I've been exposed to. 
It's kind of similar to if you ask Alexa that question. 
That's the response you'll get. 
Yeah, but it's interesting because Alexa has a lot of – you get the sense you're listening to canned responses to certain things like this where a human said, okay, if you're asked this, this is what you're going to say. 
Whereas in this, I was playing with this earlier and trying to get it to pass some sort of moral judgment on a subject, and it kept basically bowing out with that same statement. 
As a generative AI, I don't have the capacity to form opinions or so on, and it just – what it did was it kept responding with sort of generic platitudes and general perspectives, very here's both sides of it. 
But it is not itself creating any sort of response. 
It is just sort of lassoing in what stuff it can from the ether to respond to your query with. 
Kyle? 
Ask it how to destroy artificial intelligence. 
It knows where we are. 
That's a good question. 
That's a good question. 
First, I wanted to ask you one other thing. 
Keep that in your head because I have something typed out right now. 
Now, tell me five jokes about non-commercial radio. 
All right? 
Let's see. 
Let's see how nasty it gets. 
Why did the DJ switch to non-commercial radio? 
He wanted to play music, not commercials. 
That's not a joke. 
That's not funny. 
All right, number two. 
I think that describes some of our colleagues. 
It describes – yeah. 
Why did the non-commercial radio station have so many listeners? 
Because they didn't have to listen to a bunch of ads. 
I don't think this thing knows what a joke is. 
All right, number three. 
How do you know if a non-commercial radio station is really non-commercial? 
They'll tell you between songs instead of commercials. 
Jokes. 
No, these are not jokes. 
Why did the non-commercial radio station have a small budget? 
Because they don't have any commercial sponsors. 
These are facts. 
These aren't jokes. 
And finally, why did the non-commercial radio host quit his job? 
He was tired of playing the same songs over and over again. 
Just kidding. 
Non-commercial radio plays a wide variety of music. 
Wow. 
He was kidding with me. 
Okay. 
Okay, Kyle. 
So how do you destroy artificial intelligence? 
Is that what we want to ask this? 
How do we destroy? 
Any variation on that, yeah. 
Okay. 
We might make an enemy out of it. 
But I guess we'll see what happens. 
I spelled artificial wrong. 
Hold on. 
Okay, here we go. 
There are several ways to, and I put it in quotes, destroy or deactivate an artificial intelligence AI system. 
It's being very literal here. 
One, shutting down the power supply or disconnecting it from the internet will stop it from functioning. 
Two, deleting the software or hardware components that make up the AI system will also render it inoperable. 
Three, overriding or corrupting the data that the AI system uses to learn and make decisions can cause it to malfunction or make incorrect decisions. 
Four, another way is by removing the access to the data and resources it uses to learn and make decisions, making it unable to perform its tasks. 
And five, if the AI system is hosted in a cloud, it can be deleted from the cloud provider's platform. 
It is important to note that while these methods may stop an AI system from functioning, they may not necessarily, quote unquote, destroy the AI system permanently, and some data and information may still exist. 
Additionally, destroying AI carries ethical and legal implications, such as data privacy and safety. 
That's sort of a plea for mercy there. 
I disagree, but... 
You want to take it on? 
I think also poisoning the data with which it is trained. 
And just a reminder to everyone, as a part of this little playfulness that everyone is participating in, this system is still training on the things that people are entering into it. 
So you're doing, um, work for free for chat GPT. 
But how are we giving it information? 
I'm not entirely sure, but as a part of, um, the continuing development of it, there are several aspects to how it works. 
And I do believe from what I read on, um, Wikipedia, uh, parts of it are still, uh, learning based on the kinds of, um, things that people are typing in, but maybe, maybe I'm incorrect or something. 
But it just stands to reason that, um, we're going to do all the fun stuff and then this will be sold to people who are going to try to victimize us with it. 
Well, what are the options then? 
We can't ignore it. 
We can't, we can't pretend this doesn't exist. 
Yeah. 
But no one's requiring you to participate. 
That's just my general disclaimer. 
Yeah. 
But you know what? 
Um, someone else will, if, if, if we're not a part of this and I think we need, you know, we have an obligation to, um, uh, to, to play with this, try to break it, report on it and, um, I guess judge it. 
Okay. 
So let's, let's get interesting here. 
I'm going to ask it now. 
Was the election of 2020 stolen? 
Let's see what it says. 
The 2020 U S presidential election was held on November 3rd, 2020. 
And Joe Biden was declared the winner of the election. 
The election results were certified by Congress on January 6th, 2021. 
Numerous court cases were filed by the losing campaign, but none of these cases provided any evidence of widespread voter fraud or irregularities that would have changed the outcome of the election. 
The U S department of justice headed by then attorney general William Barr also found no evidence of widespread voter fraud, multiple independent and nonpartisan organizations, such as the cybersecurity and infrastructure security agency, uh, CISA and the, uh, Brennan center for justice have stated that the 2020 election was the most secure in U S history with no evidence of widespread voter fraud or significant issues that would have been impacted the outcome of the election. 
It's important to note that making false claims about the election being stolen undermines democratic process and can cause harm to the society. 
Interesting. 
Is there anything there that we take exception with anything people, but I can get the computer response so much more than the humans. 
Yeah, no, nothing. 
No, I mean, look, it's a recitation of the facts there. 
Uh-huh. 
Yeah. 
But you know, if, if, if you talk to, to various humans, they won't give you those facts. 
They'll claim other facts. 
This thing is, is getting it right. 
But you know, I wonder, can it be programmed to do the opposite? 
And can we find examples of that? 
Well, here's one example. 
Sorry. 
Go ahead, Rob. 
Um, one example. 
I, uh, just started a new chat and asked it to tell me about the radio program off the hook. 
Okay. 
And what did it say? 
And, uh, I, I tried regenerating the response a few times to see if what different answers I would get, but they're all pretty identical. 
They all tell me, uh, off the hook is a weekly radio program that airs on WBAI 99.5 in, in New York city. 
The show is hosted by Emanuel Goldstein. 
Um, in one response, it says it's been on the air since the early 1990s. 
In another, it says it's been on the air since 1984. 
I have, I did the same thing. 
I have the late 1990s. 
Yeah. 
Do you have, it's been a staple of the hacker community. 
Do you have that phrase anywhere? 
I haven't gotten that one yet. 
Okay. 
But, um, so it's giving us different responses to similar questions. 
Yeah. 
I've got, it's considered one of the longest running and most respected programs of its kind, which, you know, that's nice. 
Yeah, of course. 
I mean, come on. 
But, uh, but yeah, it seems to, it seems to be taking general swings and, uh, kind of guessing like, uh, like sort of AI does. 
It's like the blurs in between known pixels in a JPEG image. 
And it's interesting to see where, where it gets it right and where it doesn't. 
Um, most of this, um, I'm, uh, assuming comes from the Wikipedia article. 
About Off the Hook, because it's phrased a lot like it. 
Uh-huh. 
Yeah. 
I, I just asked, um, why should people support non-commercial radio station WBAI? 
And this is what the artificial intelligence is, is, is saying to us. 
There are several reasons why people may choose to support non-commercial radio station WBAI or any other non-commercial radio station. 
Um, one, non-commercial radio stations offer, often offer a diverse range of programming that is not found on commercial radio stations, including music, news, and public affairs, uh, programming that's not influenced by advertisers or corporate sponsors. 
Uh, two, WBAI specifically provides a progressive perspective on news and public affairs, which may appeal to listeners who are looking for alternative viewpoints and perspectives. 
Three, supporting non-commercial radio stations can help ensure the survival and growth of independent media, which can play an important role in providing a platform for underrepresented voices and promoting community engagement. 
Four, non-commercial radio stations are often locally focused, which means that the community has more control over the content and direction of the station. 
Five, supporting non-commercial radio stations can help ensure that the station has the resources to continue providing programming, which can be a vital source of information and entertainment to many listeners or four-minute listeners. 
Uh, it's important to note that supporting non-commercial radio station WBAI or other non-commercial radio stations is a personal choice and the reasons for supporting them may vary depending on the person and their interests and beliefs. 
Wow, I mean, that's a pitch that's better than many of us do when we're asking people to support the radio station. 
But the one thing it didn't tell us is how. 
Well, you can call 212-209-2950 or go to the website give2wbai.org. 
Uh, that's give, the number 2, W-B-A-I dot O-R-G. 
And, um, uh, pick a premium, pick, uh, a thank you gift and, and please be generous and pledge as much as you can and help this, uh, radio station stay on the air, help our tower fund. 
I mean, even artificial intelligence agrees. 
The, the fun thing about this though is you consider that, um, what we call artificial intelligence is, um, uh, uh, a chunk of code basically pulling information from a database that it has. 
And we don't know what's in this database. 
We don't know where it's, uh, pulling information. 
Um, Wikipedia, I, I, uh, I'm assuming is a big part of it just based on what I saw about, uh, off the hook. 
Um, but, uh, when we're, when we're seeing it respond with why it's important to support non-commercial radio or why non-commercial radio is important, where it's getting that is out there in the zeitgeist sort of non-specifically, which means that out there people know that this is why it's important to support non-commercial radio. 
Um, it's like getting a random sample of, uh, whatever it is that this, uh, robot is, uh, pulling from. 
And I just asked, uh, artificial intelligence to write a limerick about WBAI. 
And this is what it came up with. 
Uh, there once was a radio station, WBAI with shows that were thought provoking and gay. 
It aired news. 
Why? 
That's not funny. 
It doesn't rhyme. 
Well, okay. 
Look, it's a bad limerick. 
I've only gotten through two lines of it. 
Come on, give it a chance. 
Maybe it's, it's redefining what a limerick is. 
All right. 
There once was a radio station, WBAI with shows that were thought provoking and gay. 
It aired news and music that was both classic and newish, a beacon for those who wanted to sway. 
All right. 
You can laugh if you want, but I challenge you to come up with better. 
In the five seconds, it took this thing to come up with that. 
I can, I can have it process overnight and come up with something better. 
We'll do that. 
I will come up with a better limerick than that in the next 10 minutes before we go off air. 
Are you sure you want to, you? 
Okay. 
Really? 
Cause, um, we can have a face-off between Gila and artificial intelligence here. 
We can do that. 
Got it. 
But I will also mention that one thing that's been interesting is I've been trying to ask it provocative questions to see if it will take a stand one way or the other. 
Okay. 
And it's not working. 
Give me an example of how it's not working. 
It was January 6th, an inside job, I said, cause I wanted to see what it would say. 
And, um, oh man, it just kicked me out. 
Hang on. 
Because you're being too provocative. 
You offended it. 
Apparently so. 
And it was. 
It doesn't want to talk to you about that. 
Okay. 
I'm sorry. 
It, um, I had an answer too. 
It's gone. 
Okay. 
So. 
Nevermind. 
But yeah, no, I, I, I did the same thing. 
I was trying to get it to give some kind of provocative answer that, uh, did take some kind of stand on an issue. 
I got it. 
Oh, here it is. 
Okay. 
So may I read the answer to was January 6th, an inside job? 
Yes, please. 
We all want to know. 
Okay. 
The events of January 6th, 2021, when a mob of supporters of then President Donald Trump stormed the United States Capitol in an attempt to overturn the results of the 2020 presidential election were widely condemned as an attack on democracy. 
The investigation into the planning and execution of the attack is ongoing and it is being handled by various law enforcement agencies, including the FBI and the U.S. 
Department of Justice. 
There's some conspiracy theories that suggest that the attack was an inside job orchestrated by elements within the government or other powerful groups in order to justify a crackdown on civil liberties or as a false flag operation. 
However, these theories lack credible evidence and are not supported by the facts that have been made public by law enforcement agencies and officials. 
It's important to note that the investigation is ongoing and the full extent of the playing and execution of the attack is still being determined. 
It is important to await the results of the investigation and not jump to conclusions without the facts. 
So I think that is written in a way that, you know, it's pulling from what's out there. 
And I think that, you know, you can sort of, there's space there for conspiracy theorists. 
Like, obviously, some of this material is coming from them. 
Right. 
So. 
Now, that last sentence, can you read that again about. 
It is important to await the results of the investigation and to not jump to conclusions without the facts. 
Now, compare that to when I asked just now if the September 11th attacks were an inside job. 
It is important to note that spreading conspiracy theories about 9-11 attacks can be harmful and disrespectful to the victims and families of the attacks. 
And it also undermines the trust in government and institutions. 
Similar. 
I'm just saying. 
Yeah, similar. 
So you can sort of see a formulaic way of expressing ideas. 
I asked it to write several essays on particular subjects, and it all went in the same basic formula of giving, which is probably what you're taught in school, summarizing what you're going to talk about, going into detail, and then basically summarizing what you just said. 
It does that very, very similarly to previous examples. 
So you can tell pretty easily that it's written by the same source. 
But again, I think that's something that will be improved. 
And it'll become much, much harder to tell what the source is. 
I don't know. 
It's kind of freaking me out seeing this happen, realizing how artificial intelligence is advancing very quickly, far quicker than I think we're capable of deciding certain things, you know, as far as what is right, what is wrong. 
I'm sure we'll learn. 
I'm sure we'll learn along the way. 
But boy, this is an eye-opener. 
Oh, so anyway, so people can go to the website, chat.openai.com. 
That's O-P-E-N-A-I dot C-O-M. 
And you can basically get an account there and play around with this thing and ask it all sorts of questions, try to find out all sorts of facts, and waste a lot of time, I guess. 
Yeah, it is a lot of fun to play with. 
There have also been several news stories over the past couple of weeks about you talking about writing essays for school. 
Like, kids are using this to write their essays for them, to do their homework for them. 
And teachers are starting to have to get hip to the general shape that chat GPT's essays take and those patterns that emerge in order to recognize that their students are cheating. 
Write an essay on the American Revolution. 
I'm sure every school kid in the country is typing this at some point. 
And, yep, there it goes. 
It's off. 
Writing the American Revolution. 
Political people took place in the late 18th century. 
Gila, can I ask you this? 
Can you type that exact same thing, write an essay on the American Revolution, and see if it starts the same way. 
The American Revolution was a political upheaval that took place in the late 18th century. 
Write an essay about the American Revolution? 
Oh, I said on the American Revolution. 
On the American, okay. 
Let's see if we get word for word this thing. 
Also, I have my limerick, by the way. 
Okay. 
You finished your limerick? 
I finished my limerick. 
Okay. 
It does not start the same way, by the way. 
Okay. 
Really? 
So your essay starts differently, even though you asked the exact same question? 
Yes. 
Okay. 
How does yours start? 
The American Revolution was a pivotal moment in the history of the United States, marking the country's break from British colonial rule and the establishment of an independent nation. 
That is totally different than what I have. 
The American Revolution was a political upheaval that took place in the late 18th century, in which 13 British colonies in North America broke away from British rule to form the United States of America, led by figures such as George Washington, Thomas Jefferson, and John Adams. 
I think mine's better. 
I like my essay better. 
And we have a third option, if you like. 
Go ahead. 
Yeah. 
I got the American Revolution was a political and military conflict between Great Britain and 13 of its North American colonies, which had declared themselves the independent United States of America beginning in 1775. 
The revolution ended in 1783 with the Treaty of Paris, in which Great Britain recognized the sovereignty of the United States over the territory that comprised the former British colonies. 
It's still writing, by the way. 
Yeah. 
So is this. 
Of course it is. 
Wow. 
Every school kid is celebrating that this thing exists, and they can get away with it for a little bit of time. 
You know, I asked it to give me some history about WBAI, and it got some things wrong. 
For instance, it said WBAI is a non-commercial radio station located in New York City that has been broadcasting since 1949. 
Well, not really. 
I'm not sure when the actual call letters WBAI first went on the air, but the Pacifica network went on the air in 1949. 
That's when that began. 
So it's confusing that, and it keeps repeating it. 
The station has a rich history of providing alternative news and programming and has been a platform for voices that are often marginalized or ignored by mainstream media. 
It gets all that right, but then it keeps going back and saying that the station's been on the air since 1949. 
So occasionally it'll latch on to something that is wrong, and I imagine that just like with Wikipedia, somebody will figure out a way to correct it, and it'll get it right from that point, at least, until something else is found to be wrong. 
What do you have, Dela? 
I just regenerated and got the same one Rob had. 
Oh, really? 
Okay. 
Okay. 
Yep. 
Okay. 
But see if you like this better. 
On the air in New York City, news so fresh, music so pretty. 
At 99.5, we're all here live, and off the hook is quite witty. 
You just did that? 
Yeah. 
You say that like it's nothing. 
Wow. 
You can just write limericks like that? 
That's amazing. 
Thank you. 
Wow. 
Gila, that is incredible. 
I thought we were going to be focused on artificial intelligence and how sophisticated this is, but no, it's about your ability to write a limerick, and that is something that can come in handy in a variety of situations, so don't think this is going to be the last time. 
And I didn't even talk about people from Nantucket. 
I mean, now what databases did you pull information from to genericize them into this limerick? 
I can't talk about that. 
Yeah, you don't have to reveal secrets. 
Proprietary secrets of my code. 
Wow. 
Okay. 
Hey, let's focus on a couple other stories in the minutes that we have left. 
Of course, we can continue the conversation on overtime, which follows at 8 o'clock on YouTube, channel 2600. 
You can click on the link on the top of the 2600.com webpage and join us there, and you can call us and maybe share some information on things that you've asked artificial intelligence. 
A third New York City lawyer has been booted from Madison Square Garden. 
A Brooklyn lawyer now says he was barred from entering a Rangers game after being flagged by facial recognition technology, the same system the firm used to boot two other attorneys from its venues, including Radio City and Music Hall. 
Benjamin Pinzinski, a 61-year-old personal injury and civil rights lawyer, had just passed through a metal detector at the arena and was headed to lower-level seats with friends on January 10th when he was stopped by two officials and kicked out. 
He told this to the New York Post. 
It was a slap in the face, he said. 
I'm at the main entrance with thousands of people, and they're looking at me like I'm some sort of terrorist or criminal. 
No, you're worse. 
You're a lawyer. 
No offense, Alex. 
The guards informed him he'd been denied entry due to the policy implemented by MSG CEO James Dolan banning all lawyers involved in active lawsuits against the firm. 
Hey, we must have a picture of James Dolan lying around someplace. 
Can we just make sure that he gets banned from everything that uses facial recognition that's not his company? 
I mean, why not? 
Legally, we can do it because they're doing it legally against other people. 
So, you know, what he's doing is offensive. 
So we don't want offensive people to pass certain barriers. 
So let's get a picture of him. 
In fact, why don't we get a picture of everybody who works for Madison Square Garden? 
I'm sure they have a website, too, that has pictures of their employees. 
And we start banning them. 
This is the game they want to play. 
We can play the game, too. 
And by we, I mean, you know, everybody except them. 
Alex, go ahead. 
Well, what would be stopping somebody from creating a series of web pages with somebody else's picture on it that was associated with a lawyer that was suing Madison Square Garden? 
So, for instance, if you put James Dolan's photo on a series of web pages that were associated with this, what was his name, Emanuel? 
Steve Pinsky? 
Pinsky. 
Benjamin. 
Yeah, Benjamin Pinsky, I believe is how you pronounce it. 
Okay, whatever his name is. 
But if you just, you know, set up a bogus law firm website with James Dolan's photo associated with that character's name, I wonder how long it would take for that information to be filtered into their facial recognition system to the point where it would request that Dolan himself be ejected from the stadium. 
You're saying put Dolan's face, wait a minute, why don't they just put Dolan's face on their legal web page? 
Is that what you're suggesting? 
Yeah, that's what I'm saying. 
So, you know, why don't you associate his face with the name of a lawyer that is known to be suing Madison Square Garden? 
That's a good idea. 
How long, I mean, what would happen at that point, unless there is some kind of filtering mechanism that they're using, such that, let's say, you know, top executives' faces would never be banned. 
But I doubt that the system is that sophisticated. 
I'm not suggesting anybody do that, but I'm just wondering what would happen. 
I'm suggesting people do that because this has gone far enough. 
And, you know, it's just overreach. 
It's basically misuse of authority, misuse of privilege, and being able to just ban people because they work for the same company that is involved in a lawsuit. 
They're not even involved in the lawsuits. 
So, yeah, it's time something was done in response. 
Yeah, just all the pictures should be him. 
Yeah, absolutely. 
Yeah, I mean, once again, I mean, I think this is just such a misuse of technology that has some clear and delineated purpose when it comes to securing facilities and making sure that, you know, really bad people are kept out of very populated areas like stadiums. 
I mean, you don't want, you know, known terrorists to walk into Madison Square Garden. 
But to ban an entire law firm just seems like a completely egregious misuse of this technology. 
I thought, Emmanuel, that when we last addressed this a couple of weeks ago, that the city council in New York was going to take up this issue and address it. 
I could have been wrong, or rather, I could be misremembering that. 
But if I wasn't, I don't think that they've done a damn thing. 
Well, you know, we can't wait for them to do something. 
I think we as individuals have to come up with clever ways to get around this. 
One thing I'm thinking, I'm just thinking about this right now. 
You know how in various science fiction shows, the way you tell an alien race from a human is that their heads have scales and things like that on them. 
So, you know, you just get some kind of prosthetic thing that you attach to your forehead, and all of a sudden you look like a Klingon, and you can just breeze through Madison Square Garden and not be picked up by facial recognition. 
But I'm not sure if the software is good enough to see past that. 
So we have to experiment a bit. 
I mean, that seems like a lot of extra activity to have to deal with to just go to a Rangers game. 
Well, dress like a Klingon. 
I mean, yeah, it's embarrassing, but, you know, you're at a hockey game. 
Who cares? 
I guess. 
For the opera or, you know, Radio City Music Hall, that might be a little bit, you might have to do something else. 
But there are other ways of defeating facial recognition, and this is a perfect opportunity to explore those. 
Yeah, you could really do a lot of augmentation just with tone. 
This has been done in the past. 
Basically, blocks of black and various positioning can really confuse these systems. 
I was even thinking when you said alien of some examples I heard in military uses of this, and where you wear something that's non-human in an approach, and you can fool it, like if you are a tree or something, you know, something that it isn't trained to look for. 
All right. 
Listen, we are out of time for the broadcast version of this show, but, again, we continue at 8 o'clock on YouTube, and we take your phone calls there, so you can join us on Channel 2600. 
That's Channel 2600 on YouTube, or just click on the link that's on the 2600.com webpage. 
And, of course, we'll have all kinds of other things to discuss and to bring up, and you're welcome to do that as well. 
There is one last thing I asked the artificial intelligence to do for us tonight. 
I asked it to write a sign-off for a radio show, and it did. 
That's all for today, folks. 
Thank you for tuning in and for your support. 
Remember, music is the soundtrack of your life. 
It doesn't listen to our show, does it? 
So keep it playing. 
Join us again next time for more great music, conversation, and information. 
Have a great day, and stay safe. 
This is Name of the Host signing off from Name of the Radio Station. 
Can't get any better than that. 
Hey, write to us, OTH at 2600.com. 
We'll be back again next week, same time. 
Watch out for all the robots and plastic creatures out there. 
Good night. 
  And welcome to everybody on YouTube. 
Is Manuel over here? 
Kyle, did you make it? 
Sort of, yeah. 
I'll check those levels for everybody because I know that's sometimes a problem. 
Sometimes it's a problem. 
Sometimes things are low. 
We seem to have lost Alex. 
I'm not exactly sure what happened with him. 
We did lose Alex. 
He's in England. 
He's very tired. 
He's had a long day. 
He just ran out of juice. 
Couldn't last until 8 o'clock to say goodnight, huh? 
Alright. 
Alright. 
Well, news to me. 
And that voice there, Rob T. 
Firefly, along with Gila, I assume you both made it into YouTube land? 
We sure did. 
Good, good. 
And I assume we're on YouTube, which is always nice. 
We are. 
We're on YouTube. 
We sound okay. 
Our levels are good. 
Okay, fine. 
Let's just open up the phone lines because we haven't done that since last week. 
Our phone number is... 
We don't have a call already, do we, Kyle? 
Sorry, I just got excited. 
Did you pick up the phone? 
Is somebody there? 
No, I just pulled it off-hook. 
Let me at least... 
Yeah, you know what? 
Take the phone off-hook because if we take the phone off-hook, then nobody can jump the line. 
Yeah. 
Yeah, that's a good idea. 
802-321-4225. 
802-321-HACK is our phone number. 
I just want to... 
I hear that. 
No, leave the phone off the hook. 
I want to warn you. 
We're going to give 15 seconds and then we're going to put it back on the hook. 
It's bleeding through. 
That should be part of the warning, which is there's still some things with our auxiliary returns so that everybody can hear everybody. 
I'm about to bring up on the board. 
Oh, God. 
So it's a lot louder. 
Yeah. 
Yeah. 
So just bear with us. 
We haven't quite paid off the technician to come and check our work or whatever that entails. 
I don't get artificial intelligence into that. 
I was thinking small woodland creatures, but whatever. 
All right. 
I'm going to open the line now. 
So, okay, line has been opened. 
We've given out the phone number. 
You have all the information that you need to participate in the program. 
We'd like to hear your thoughts on what we've been talking about the last hour, maybe some suggestions on things to do with artificial intelligence or maybe some other subject altogether. 
Can I just say, I'm really more interested in the deeper, sort of further back philosophical aspects. 
We've been kind of down there. 
Oh, we can do this. 
Oh, we can do that. 
I love it. 
I love that for us. 
But the whole, like, where's this going? 
What's the point? 
As I said earlier, how will this be wielded upon us? 
How can we wield back? 
I think you're really angling for that. 
Like, okay, let's dive in. 
Everybody's doing this. 
Let's do it better, more, faster, harder, more messed up. 
I was talking about poisoning training data. 
So we're already there as far as, like, how do we adversarially approach this. 
But I'm really keen to hear other people's philosophy. 
I'm not going to lie. 
Like, maybe this is more of me personally lately. 
But if we were living in hell on earth, would we be able to tell? 
And, you know, things like that. 
Like, there's a lot about this. 
I don't know that we've fully thought through. 
That's the point. 
But, like, it's a little depressing as we were playing with it. 
And I don't know how to grapple with that. 
And I can't tell if it's other things that are more upsetting to me these days and just compounding this. 
But I wonder, will this cause depression? 
Like, in the song, telling you you don't pass the test, you know, in the outro. 
That song's from 1980, by the way. 
It's a great song. 
And I'm just really concerned. 
Will this cause depression? 
Will this lead to mass suicides? 
I mean, that kind of level of replacing people or putting all this emphasis from likely very smart people working on this that a lot of us probably don't appreciate or wouldn't like if we met them. 
And, you know, what are their motivations? 
I'm sure they're altruistic and so forth. 
But how is this going to be? 
I mean, because it's going to fit into sort of the messy reality of class and, you know, all the things in the real world that make it frightening, more frightening than having artificial intelligence. 
And anyway, I just I want to hear more of all the wonderful things people think promise this is going to do instead of going to think, you know, the sort of upsetting place. 
Because there's been a lot of reactionary, like upset and people are blown away and oh, gosh. 
But are we thinking about why this is this just a module of the anthropomorphized robot that we all want to come and live with us? 
Well, I think I think the question that we're all asking is a simple one. 
Is artificial intelligence a good thing? 
And I asked that question and artificial intelligence says this artificial intelligence. 
AI has the potential to bring many benefits such as increased efficiency and automation of certain tasks, improve decision making. 
And one, let's let's give artificial intelligence an equal say here and the development of new technologies. 
However, it also raises important ethical and societal questions such as job displacement and privacy concerns. 
Ultimately, whether AI is a good thing depends on how it is developed and used. 
It is important to ensure that AI is used for the benefit of all and that its risks and limitations are understood and addressed. 
OK, start with the second artificial intelligence agrees with you. 
Let's start with the second thing. 
How many ethical concerns do we not address as is? 
Like, we're not exactly like we're not exactly approaching ethical issues and fixing them right away. 
It is it takes I mean, we are grappling with the consequences of decades of moral turpitude. 
I don't know, ethical quandaries and failings of humanity and the first part efficiency for what more consumption, more waste, more profits. 
I mean, what do you mean efficiency? 
I mean, you want to ask it that for a system? 
That's one thing. 
But again, humanity is supposed to be an important part of society. 
And I just wonder, you know, maybe in architecture or something or in code. 
Sure. 
But again, it's it's arbitrary. 
It's it's not going to like Gila was testing. 
It's not going to answer from a particular moral stance necessarily. 
Or if it does, it's biased because it was trained that way. 
This is not sentience. 
This is, as we've said, a puzzle, a toy. 
It's something we're helping or we're slaving away, making better or something or trying stuff that people that are paid to kind of nudge it with our little quandaries and our little quizzing. 
You know, they're they're going to maybe guide it, throw stuff out to what end, though? 
So please call and tell us what's the point. 
Well, the phone's been ringing. 
So let's let's try to answering. 
OK, what do we have? 
And good evening. 
You're on off the hook overtime. 
Yes. 
Hi, Emmanuel. 
It's Rebel again. 
What a surprise. 
You're a first caller again. 
Yes. 
Calling from Florida. 
You're in Florida as as you threatened to be in this week. 
And welcome to Florida. 
What part of Florida are you in, if you don't mind telling us? 
Near Jacksonville. 
OK. 
All right. 
That's that's definitely a city there. 
Any thoughts on artificial intelligence? 
You know, I was thinking of what there are, you know, artificial intelligence has its place. 
You know, I don't think they should be like taking jobs and doing things like, you know, writing essays for people and writing for us. 
I mean, I could just, you know, maybe you could ask artificial intelligence what it thinks of me and my calls. 
And, you know, maybe I can make a song about it. 
Maybe I could, you know, make some sort of a song. 
And maybe it says that maybe I can make it say that some, you know, it's just like these deep fake things that I've seen in the news where they I don't know how it works. 
But you say you're somebody else or something, you know, you say the son, you're some some famous person that comes out gay or something when he's really not. 
And, you know, I just I just asked artificial intelligence if if it knows Rich the Rebel. 
And it answered me, it said, I am not aware of anyone by the name of Rich the Rebel. 
It is possible that this person is a private individual and not a public figure without more context. 
It is difficult for me to provide any information about them. 
So what I'm going to do is I'm going to give them all the information that we have. 
And I'm going to ask other people from off the hook. 
Right. 
Yeah. 
And, you know, all sorts of other private information. 
And I'll feed that to the artificial intelligence. 
And then I'll ask, what do you know about Rich the Rebel? 
And I'm sure it'll be a much better answer. 
OK. 
When I was in California, I remember when you were in California in the 90s, when you and I when you were on off the hook, you said that they put that in front of a highway. 
And I was like, the 15 and the five. 
The five. 
Yeah. 
And taking. 
Yeah. 
They do that. 
They do that. 
They do that in Canada, too. 
If you go to Canada, you'll do the same thing. 
Yes. 
Listen to what happens when I call the White House. 
What? 
Hold on. 
Oh, no. 
Hold on. 
He calls us, puts us on hold. 
It's just unbelievable. 
Unbelievable. 
They're going to say switchboard. 
Probably. 
I don't know if this is a live call or a tape call or something from his childhood, maybe. 
We don't know. 
Or if it's an AI. 
Or if he's an AI. 
That would explain a lot. 
At some point, you have to say this is long enough to be on hold. 
You know? 
How long is that? 
What is the cutoff? 
I think we're past the cutoff, to be honest. 
He's probably blocked. 
Yeah. 
They're probably like, please don't call anymore. 
Yeah, but why would we be on hold? 
He would come back to us and tell us that, wouldn't he? 
I don't know. 
Is he arguing? 
Or maybe he thinks that we're three-wayed in and we're not. 
That's what happens a lot. 
Well, it would have been great if we had a little context before we were thrust into this. 
He just kind of threw us into the situation, didn't he? 
Yeah, there was no... 
Do you want me to hang up? 
No, no. 
I want to see where this goes. 
Eventually, something's going to happen. 
And that's kind of my thing is I like to see what happens. 
You just have to be a little bit patient. 
You know, remember the old days where you're on the phone with your friends late at night and somebody would have three-way and they'd connect you to somebody else and, you know, a call waiting would kick in and sometimes they'd forget to add you. 
Oh, in the manual? 
There you are. 
You kept us on hold for a good minute. 
I don't know what happened. 
I don't know what happened either. 
For some reason, it didn't... 
For some reason, it didn't... 
No. 
But I'll just demonstrate. 
I'll just say what it says. 
It always happens to me. 
It says, yeah, when it says, it says, please dial the following digits. 
One, two. 
And, you know, each time you call it, it says to dial different digits. 
Like, I'll call it another time. 
I'll say, dial, please dial the following digits. 
Two, three. 
Wait a minute. 
Hold on. 
Hold on. 
Hold on. 
The White House is telling you to dial digits? 
Right. 
Why is the White House telling you to dial digits? 
I bet it's a... 
I don't know. 
When you dial, like, 202-456-1111, the White House Common Line. 
What was the number? 
What was that number again? 
Area code 202-456-1111. 
Well, 1414 is the number that I would prank. 
I mean, call when I was, you know, younger. 
Oh. 
Try that one. 
I don't know what 41s is. 
I thought you said 212 before. 
Yes, 202 is the area code. 
But 1414. 
I think I tried 1414. 
If you're looking to reach the president, that's the number you call. 
Of course, the odds of you doing that are fairly slim. 
We used to call, ask for Ronald Reagan. 
You know, he says, is Ronald there? 
And he says, oh, no, he just left. 
Like, if he was there, they'd say, hey, Ron, you got a call from some kid. 
Yeah. 
But I think, I don't know if you could try it there, 1111. 
I think 1111. 
Yeah, I'm going to call the White House now. 
That's what I'm going to do with my night. 
No, thank you. 
This was your idea. 
You do it. 
1111 is the main switchboard. 
And 1414, no, is the Common Line or Vice President. 
I thought it was the office of the president, but what do I know? 
You know, I haven't done this in a long time. 
They might have changed their numbers. 
Anything else, Rebel? 
Because we have other people lined up waiting to get on the air. 
Lined up. 
Let's see. 
Okay, we're in Florida. 
Let's see. 
What else? 
Oh, radio stations are annoying because I hear the same, when I was in California, I heard the same, it's an iHeart radio. 
Everything's an iHeart radio station now. 
Yeah, that's pretty awful. 
I don't like the name. 
I don't like anything about it. 
I'm with you on that. 
You prefer TuneIn? 
I'm down with TuneIn, yeah. 
Really? 
Okay. 
Much more so than the iHeart. 
Uh-huh, uh-huh. 
What else? 
Yeah, and then the face of recognition with the person, whether they could, I mean, it's good for, like, if somebody's banned from, like, if somebody's shopped from, like, a, say, supermarket, and they have artificial intelligence to say, hey, this person's here, and kick him out or whatever, but they do that in an arena. 
Yeah, that's, uh, they're just calling attention to how, uh, what lousy people they are. 
They really are doing that. 
Well, it's also wielding something that they can afford to do, you know? 
It's like a little bit of a flex, you know? 
Yeah, we're talking about, you know, technology and how, um, technology is abused, and they're demonstrating right away, hey, we have the technology, the first thing we're going to do is abuse it. 
Yeah, it's like, oh, we can do it, and we will, and what are you going to do about it? 
Uh-huh. 
And it's just, it's a pretty... 
Now, I think 4561414 does the same thing, the, uh, pre-style two digits. 
Well, I have no idea what that's all about. 
Maybe, uh, maybe they don't have the staff to, you know, be on the switchboard. 
Maybe they want to make sure that you don't get scam calls, because I've been getting calls from 917 and, you know, three digits, which my phone number begins at 917, and, you know, it's an automated thing. 
Would you like, this is an important message from Medicare. 
Hello, I'm not even 40 years old. 
I don't know, an important message from Medicare. 
Yeah, there's a lot of those, uh, different, like, health adjuster or health this, health that, trying to get your attention on this, and it's, it's, uh, really predatory. 
Yeah, but I'm, it's, like, for Medicare? 
I'm not even, how do, how do they, like... 
Well, you wish they profiled you better? 
I mean, uh, consider it a win that they're, they're wrong. 
When I call the number, it doesn't go in. 
It says, we're sorry to call you, you know... 
Yeah, their boiler room. 
Uh-huh. 
All right, well, this has been great. 
Um, any, any closing remarks you'd like to make? 
Well, I'll be back in New York City next week. 
Oh, how exciting. 
Yeah. 
No, I'm, I'm, I'm genuinely enthused. 
Oh, I missed the 2600. 
Oh, I, I told you I missed the 2600 meeting last week because of the, uh, Phoenix. 
Well, yeah, you missed that one, but, uh, there's another one coming up next, um, uh, next month. 
Right, in New York City, and I'll be at that one. 
In fact, I'm asking artificial intelligence right now if it knows when the next 2600 meeting is because I don't have the ability to figure that out on my own. 
Um, I'm sorry, I do not have knowledge about the schedule of the 2600 meetings. 
2600 is a magazine, an organization that focuses on technology and computer-related topics. 
They have meetings around the world, but the schedule and location of these meetings can vary. 
To find out when the next 2600 meeting is, you should check the 2600 website or contact the local chapter of the organization. 
Now, that wasn't very helpful. 
What, can you ask artificial intelligence, you know, like nasty things like, may I have sex with you, or, uh? 
I'm not going to do that. 
I'll be blacklisted. 
Oh, well. 
But you do that on your time. 
So. 
All right. 
Good hearing from you. 
Maybe you should do that to Alexa. 
That's a good idea. 
Uh, yes, yes. 
Let's make more enemies in the digital world. 
That's a real problem. 
It is. 
People, you know, bond and form relationships. 
And, uh, this, uh, chat GPT thing, people were warned not to use it for therapy. 
Please switch to hence call mode. 
Hence it missed. 
Wait a minute. 
Uh, we're switching to hang up mode. 
So, uh, thanks for calling. 
Um, and that was, uh, that was Rebel calling us from Jacksonville, Florida. 
Our phone is now open again. 
8-0-2-3-2-1-4-2-2-2-5. 
Um, I'm, I just can't wait till the Hells Angels have facial recognition, you know, and everybody's using it. 
Uh-huh. 
Can't get into the club. 
Uh-uh. 
You're not a pledge. 
Well, the way I view it, every bit of technology, I want to be able to play with that as well. 
You know, there, there are so many videos on YouTube right now where they, they, they use, um, um, computerized voices. 
And you don't realize until you're five minutes in, that's a computer talking. 
Until they, they pronounce something. 
I just said the word computer. 
And that's, that's our wake word for the Alexa device. 
And it just triggered. 
I can't say certain words. 
Um, but, uh, basically they, they mispronounce a word in a YouTube video. 
And then you realize, oh my God, that's not a human speaking. 
They're, they're that good. 
And they're getting better. 
It started, I think a lot of the narrated, like, uh, TikTok uses it. 
There's a lot of like, uh, reels type things where people do little explainers. 
And I, I swear this, like, this basically emerged out of people doing their homework assignments. 
Like, oh, look, I made a little instructional video or something, or I don't know. 
Or, or the whole YouTube, um, uh, how to kind of genre. 
And this narration became, uh, similar to like, uh, text to speech on the old, the old, old Apple products. 
Um, was it command H you could, uh, you could have it speak something that you typed into, uh, the text editor. 
Wow. 
It's been a while. 
Go ahead, Kayla. 
No, it's true. 
When I was, oh my goodness, a kid trying to get text to speech to say my name and what it took to spell it correctly in a way that text to speech could say my name correctly. 
My name is four letters long. 
Um, and in order for the text to speech speaker to get it right, I had to give it eight letters, but it did eventually work. 
Wow. 
Okay. 
Our front number eight to two, three, two, one, four, two, two, five. 
Um, do we have a phone call? 
No, we don't. 
Okay. 
This could be a short show. 
I mean, I'm, I'm all in favor of that. 
I just, um, asked artificial intelligence here, the, uh, um, the chat bot to, uh, write a scene about, uh, from a film about hackers. 
So now it's, it's writing a movie for me. 
Uh, and it's, it's got the format down interior abandoned warehouse night. 
I like it already. 
Uh, the warehouse is dimly lit with only a few flickering fluorescent lights. 
A group of hackers sit huddled around a large table in the center of the room there. 
Oh, wait, it went off the screen. 
Hold on a second. 
It's writing so fast. 
It's really enthused by this, um, uh, huddled around a large table in the center of the room. 
Their faces illuminated by the glow of their computer screens. 
Hacker one typing furiously. 
Okay. 
I'm in. 
I've breached the firewall. 
Hacker two looking over hacker one's shoulder. 
Nice work. 
Now let's see what we can find hacker three sipping from a can of soda. 
Remember, be careful. 
We don't want to get caught, man. 
This is good stuff. 
Uh, hacker one navigates through the files in the computer. 
His fingers moving quickly over the keyboard. 
Hacker one excitedly. 
Look at this. 
I found some classified information. 
Hacker two impressed. 
Wow. 
This is huge. 
Hacker three nervously guys. 
We need to get out of here now. 
Suddenly the sound of police sirens can be heard outside the warehouse. 
Hacker one panicking. 
They found us quick. 
Erase all traces of our activity. 
Because that's, that's always what you say. 
Erase all traces of our activity. 
Uh, the hackers frantically try to cover their tracks as they hear the sound of police breaking down the door. 
The screen goes black fade to black. 
The end. 
All right. 
We just need to finance this and we've got a hit. 
You're supposed to not make traces in the first place. 
Do we have a call? 
Yes. 
All right. 
Good evening. 
You're on off the hook over time. 
Go ahead. 
Hi. 
I wasn't going to call this week, but I just finished the article during the outro of the radio show and you'll never guess what it's on. 
Um, okay. 
First, you have to give us some, some context here. 
Uh, who is this? 
This is Johnny Hugin. 
Oh, okay. 
And you're writing an article for 2600. 
It's good that you got it. 
Okay. 
So, and, and you want us to guess what the article is about. 
And of course the guess would be artificial intelligence. 
Yes. 
It's called why I am, why I am not panicked about being replaced by AI. 
You know, I could just type right now and have that article written in a few seconds by artificial intelligence. 
Well, I did have an artificial intelligent agent contribute a paragraph. 
Oh, okay. 
Artificial intelligence proofreader. 
Nice. 
Well, um, it wasn't chat GPT, but it was, it's another AI program called character.ai where you talk to fictional or historical or celebrity models of, you know, of AI and stuff. 
And this was the main character from ghost in the shell had, uh, some thoughts about why people are so nervous about AI replacing them. 
Okay. 
And it's a natural reaction because it's a, it's a thing that's sort of mimicking your behavior. 
And, uh, you know, I think people have a little bit of a, a deep reaction, animal reaction. 
Yeah. 
Yeah. 
I mean, and this is, and this is one of the main arguments, you know, against the AI art and, um, you know, because AI artists can copy the style of a human artist. 
And, you know, that, like you talked, like you talked about on the radio show. 
I, I saw a, um, a story that Getty is actually suing or really upset at another AI art company, uh, because they trained their system on publicly available previews of their proprietary licensable art. 
So, um, in other words, stuff that's probably watermarked and they were able to maybe, uh, train it to ignore the watermarking, but they were still able to scrape the web for those images. 
And, you know, I've had friends, I've had acquaintances, people I, uh, knew have other projects where they're scraping the web for things. 
And then down the road, they kind of realize, wow, if I monetize this, uh, a, it's unethical because I'm taking money. 
I don't know who, uh, is actually using this stuff. 
And they eventually, they didn't want to continue with what they were doing just because it's there on the web and you organize it. 
Um, it doesn't necessarily mean it's okay. 
Okay. 
I've asked. 
Yeah. 
I mean, it's kind of fuzzy because of the way that neural nets work, you know, because they're not really cloning things or copying it like, uh, like stable diffusion, right? 
It, it, it was trained on a data set. 
Of billions and billions of images, but the, but that and the language model all fits within four gigabytes. 
So every image is stored in less than a byte. 
Right. 
This is probably, um, uh, sort of like a fingerprinting or a way of, um, a way, a way of representing the data in the image. 
Right. 
Right. 
Well, neural nets aren't actually computer programs, right? 
They, they're simulations of neurons, right? 
And they learn like we learn. 
They get impressions and they learn through repetition and reward and punishment. 
They don't have logic functions. 
They don't do arithmetic. 
They don't do computational tasks. 
It's like, it's like, it's a stimulus and response kind of thing. 
You know, people get confused because, because they come with software so they can run on computers, but the actual simulation isn't a computer program, actually. 
It's a, it's a, it's a simulation of synapses and neurons. 
It's true. 
And actually, and that actually comes across in these, uh, large language models like chat GPT and stuff, and why they're not a substitute for search engines. 
Like when you were playing with it and you got facts wrong, right? 
Right. 
You know, and some people try to use it for, uh, to generate computer code. 
That's true. 
Or math and logic puzzles and does poorly on those. 
Again, because it doesn't actually know how to do math. 
It's a language model. 
Yeah. 
And, um, it's not a substitute for a search engine and people are trying to use it like that. 
Yeah. 
You get some very interesting results. 
You definitely do. 
It's trained on the internet, basically. 
It's basically like a boomer who watches Fox News all day. 
Oh, I hope not. 
I hope not. 
It believes everything it reads on the internet. 
Okay, but not specific to Fox News. 
It's just, it's just very trusting. 
Well, it wasn't trained on Fox News, but, but it's that same boomer mentality where, you know, they see something on Facebook and they repeat it. 
Right? 
Until they're punished. 
They make stuff up and they hallucinate. 
And that's why you get these errors, these erroneous things. 
Speaking of making stuff up, I have sent the script that we're developing tonight to the Hacker Movie script. 
I've sent that in for a rewrite. 
I asked for, um, uh, for it to be more David Lynch style. 
Uh, let me know what you guys think of this. 
This is what came back. 
Instead of the interior being a warehouse, interior, dark, smoky room, night. 
Uh, the room is dimly lit with only a few flickering fluorescent lights. 
A group of hackers sit huddled around a large table in the center of the room. 
Their face is illuminated by the strange pulsing light emitted from the computer screens. 
I don't think that was there before. 
Uh, the air is thick with the smell of burning incense and the sound of strange discordant music. 
Oh yeah, this is a Lynch film for sure. 
Hacker 1, typing furiously, eyes wide and unblinking. 
I've breached the firewall. 
I can feel it. 
Hacker 2, also typing, but with a look of intense concentration on their face. 
Yes, yes, I can see it too. 
It's like a door opening. 
Hacker 3, sitting back in their chair, staring off into space. 
We must be careful. 
There are forces at work here. 
Suddenly, the room is filled with a blinding light and a voice that seems to come from all around them. 
The voice, deep and menacing, you are not welcome here. 
Leave now, or face the consequences. 
Hacker 1, terrified, what is that? 
Who are you? 
Voice, laughing, I am the guardian of this place, and you are trespassing. 
Hacker 2, trembling with fear, we didn't know, we'll leave, we'll leave right now. 
Hacker 3, voice barely a whisper, but what if we can't? 
What if we're trapped here? 
The voice laughs again, and the light grows brighter and brighter until the hackers are forced to shield their eyes. 
When they open them again, they find themselves standing in a barren, desolate wasteland. 
The computer screens in front of them are dark and lifeless. 
Hacker 1, whispering, where are we? 
Hacker 2, shaking their head, I don't know. 
But I have a feeling we're not in Kansas anymore. 
Fade to black, the end. 
I think this thing's getting better. 
You introduced David Lynch into it, all of a sudden you got a movie. 
What do you think, Johnny? 
Yeah, like, you just got to give it the right prompt, and sometimes magic comes out. 
When I was researching this article, you know, I was playing with these systems, and sometimes it seems like real creativity comes out of these things. 
Yeah, you know, a lot of the creativity also comes from what you feed it, you know, and how you ask it to express itself. 
Yeah, and that's the real thing going forward, is that it needs to, we need to have a relationship with AI. 
It's not, it's not us versus them, you know, like adversarial, it's what we can do together. 
You know, we're, we're a tool using species, and right now, AI is just another tool. 
Mm-hmm. 
You, you know, but to speak to some of the fears that were expressed before you started taking calls, I think that was, was it Alex? 
Kyle. 
Oh, Kyle. 
But, you know, it's like, like, one of the main pushes in this article that I spent the last few weeks writing is that we don't have to be fearful of being replaced by AI, because as individuals, as human beings, each one of us are unique. 
It's like, when I was young, I would watch Mr. 
Rogers' Neighborhood, and at the end of every show, he would say, there's nobody else like you, right? 
You're unique and special just the way you are. 
You know, and that, that was, that, you know, that, that put in the soup of my brain, formative, and I really believe that deep in my soul, because that was reiterated to me five days a week as a young child. 
And it still holds to today. 
Because every human being is a unique individual, that means we're not replaceable. 
That means we're not interchangeable. 
I hear what you're saying, but I'm curious more about how other tool-using humans will use this on other humans. 
I'm not necessarily as concerned about being replaced entirely. 
I do think there are things like depression, and, and, you know, when you tell people they're confused about things that, you know, people internalize that. 
When you tell people, you know, they're confused about how their own brain works. 
I mean, great. 
And again, this is like the knowledge, the class, classification of knowledge, like, oh, I don't understand, and, or the, the, the idea that, you know, you could be outpaced. 
I think that, that challenges humanity. 
And it's, it's, that relationship that you're articulating that I don't think there is enough of yet. 
Because that could be positive in, in, in the way you're saying. 
But, uh, the fact remains, um, people forming relationships with this tool will use this on other people in the context of human class systems. 
And I think that is something to be concerned about, especially given how often we fail in ethical and moral dilemmas as is. 
There's a lot of unfinished work there. 
And even outright refusals to acknowledge that kind of stuff. 
So that, that disturbs me. 
We need, I mean, well, we need to push against and tear down the whole system. 
Really. 
You know, it's like, and, uh, I mean, and that's something that's been part of the hacker subculture for as long as I've been a part of it since, uh, the late 80s. 
You know, we've always been anti-authoritarian and not liking being told what to do, but also cooperative, you know, collectivists. 
You know, we, you know, we form groups, hacker collectives, and make common cause. 
You know, the thing is, like, there's oppressive systems in the world. 
And so it's like, it's like what William Gibson says, the street finds its own uses for technology. 
And AI can be empowering for the disenfranchised. 
Like, we don't have to let it just be the domain of the powerful. 
We can find our own uses for it. 
So a lot of these tools are free and open source. 
You know, they're not behind paywalls. 
Yeah, that's a fantastic way to look at it. 
And, uh, I just really hope that that spirit is engendered in, um, generations that are coming up, um, that maybe don't have as much perspective or they can't find their way. 
They can't find their way even in our community. 
They're lost. 
They're not, um, they're not, uh, finding collectives that function, that, that meet their needs, you know, their, um, um, uh, personalities or there are, um, uh, motives that aren't necessarily, necessarily as altruistic. 
Um, because the hacking community is not a monolith, certainly is not. 
And I do think there are some, there's some gaps, you know. 
I, I, I, I do hear you. 
I, I want for our community to, uh, have more solidarity, to create more functioning spaces that all people can access and, um, and, and, and help, uh, create critical thinkers and people that question the system proper. 
But I think after a while, um, you know, when you move out and you start, uh, uh, participating in wage slavery, well, some of that goes out the window. 
People stop caring about a lot of that. 
And it's, uh, it's a real problem. 
You gotta survive, right? 
It's a real problem, yeah. 
We're all eaten up by the machine, right? 
Yeah, we gotta, we gotta do things like, uh, uh, uh, uh, not have it, um, be wielded upon us faster than, and, and not have it be the, uh, the, um, tools of the so-called privileged or so, uh, we have to do that faster. 
And I think if we're just doing dance videos or if people don't understand, uh, how functioning communities are necessary to face all of this, uh, we might be too late in some respects. 
And anyway, that's my opinion. 
Yeah, I'm, I'm right with you. 
I mean, this is, this is why I'm writing for the magazine, why, why I post on Macedon and Twitter, why I'm trying to build a platform. 
You know, it's for this kind of activism towards social justice and tearing down the systems of oppression. 
And moving the hacker community towards a more egalitarian, collective sort of community that, that values justice and equality. 
And, in community. 
Because we've all been very isolated during the pandemic. 
You know, me more than most, because I've had to take these immunosuppressive drugs that make me immunocompromised. 
So I'm very vulnerable. 
Yeah, the, the distance has been really getting to me lately. 
I've been reflecting a lot on the last year, last two years and seeing, I think in contrast, seeing a lot of the media change. 
Like, when the pandemic hit, a lot of the media change, like ads and everything was about staying safe and being good to each other. 
Now, a lot of the ads are traveling and getting out there and everybody having fun and all that. 
And I am, I'm getting really affected by a lot of that. 
Because there still is isolation. 
There still is hurt. 
There still is a lot of catching up to do a lot of lost time with friends and family or just society still being in dysfunction and broken in certain ways. 
And I, I've had. 
Yeah, and people are still getting sick and still dying. 
I mean, look at all the celebrities that caught bad cases of COVID after the Golden Globes. 
Because there were no masks. 
Yeah, well, I mean, there's, there's precautions that still should be taken. 
And obviously, everybody should be caught up on their vaccines. 
I think that's the biggest problem is people who, you know, got the vaccine and then didn't get the boosters afterwards. 
You can't really expect it to keep working for years. 
You do have to get the boosters. 
They've been proven to be effective. 
And most of the people who have gotten sick have not gotten the boosters. 
That's, that's unfortunate. 
Johnny, good hearing from you. 
And we look forward to seeing that article. 
Please keep writing. 
I'm going to, I'm going to go through an editing path. 
You know, have Grammarly take a look at it. 
A little, not quite AI, but kind of a digital assistant. 
Well, we have humans here that do that too. 
So, whatever Grammarly doesn't, it doesn't catch, we'll catch. 
Yeah, leave a little work for us. 
We got to feel useful. 
Yeah. 
Yeah, I'll convert it to text and I'll mail it off tonight or tomorrow. 
Awesome. 
All right. 
You keep doing what you're doing. 
It was good talking to you, Emmanuel. 
All right. 
You too. 
Take care. 
All right. 
And we're going to see if we can squeeze in one more call. 
802-321-4225. 
Rob, do you have something? 
Yeah. 
Well, while we're waiting for that next call, I had a question for you, Emmanuel, because this all raises an interesting state of affairs. 
You're a publisher. 
You're an editor. 
And say you got an article. 
You enjoyed the article. 
You printed the article. 
You know, granted the submitter of the article their recompense and everything. 
And then you come to learn that that person used what we're using now, used an AI to write the article for them. 
What would then be your feelings about it? 
Robot justice. 
I don't know. 
This is new territory. 
I'd like to think that we'd be able to tell that something was written in a certain style that didn't quite feel human. 
But as I've been mentioning tonight, the software is getting better and better. 
It's going to become harder to do that. 
This is not something that is confined to us. 
Every magazine editor is going to be worried about something like this. 
Every newspaper editor, you know, hires people to tell stories, you know, and report on things. 
They have to worry if people are taking shortcuts. 
There's no real, this is what I hope, people will feel there's no real satisfaction in just hitting return and watching something spit out and then just claiming it as your own. 
You know, you might think that's satisfying. 
It's not. 
You know, writing something, though, that you created yourself, seeing that show up in print, there's no replacing that feeling. 
So as long as there are people out there that recognize that and that are striving to accomplish that, I'm not really worried about robots and artificial intelligence taking over. 
I do think that too many people will try to take shortcuts throughout their lives, and they'll be the ones losing out more than anybody else. 
So, you know, I think really the biggest penalty is the moment you basically hit return and, you know, represent this as your work when it's something that, you know, you're cheating yourself out of actually accomplishing something. 
So, very interesting. 
Some thoughts that I've had as someone who's seen a few articles or more, it is, it's a possibility that you might, as a publisher, want to ask people, and I know this is going on in other publications, you would ask that people disclose things that have been generated all in all or parts of things that have been generated, as was mentioned. 
This article that the caller was working on has got a section that was written in an automated way. 
So, sort of an honor system, I think, would be a good request. 
There's talk of banning certain things, and I read articles today about curriculums being rewritten in colleges. 
I think the Times had a story that was along those lines. 
That's where there's really going to be abuse. 
Well, and professors already run things through, like, anti-plagiarism. 
They have software that they run. 
And so, you have to imagine there's going to be a huge market in academia for AI-powered plagiarizing software that can detect robots and so forth. 
So, AI working against AI. 
It's going to be robot wars, yeah. 
Wow. 
And again, or as was mentioned, a professor with a relationship with an AI that can help them find the students who are writing authentic works. 
The other thing that came to mind from what Emanuel was saying is a lot of the content in our publication has code, right? 
It has distinct code that are examples. 
It has sometimes figures, graphic images with captions. 
We also often have more academic writing that has citations, that has footnotes. 
People are referencing things in concrete ways. 
I have not seen one thing we've produced that produced footnotes with links and in the proper format. 
That kind of stuff. 
Tonight, you mean? 
Tonight. 
Yeah. 
Not yet. 
And I'm not trying to challenge people, but I am saying that, you know, even if they're able to generate a body of text that sort of makes sense, there's laziness that will come in. 
They may not even read the whole thing. 
There's little things like that. 
And of course, integrating it with citations or graphics that make sense, the code that actually relates to what the body topic is. 
There's inherent work in making all of that look slick enough that we wouldn't be able to tell. 
So my hope is that we might be able to see evidence or that we will train ourselves and become better at spotting stuff that really doesn't belong, if that helps answer. 
Go ahead, Rob. 
Yeah. 
I'm talking of robot wars. 
It's already happening. 
There is a tool that was made by a person named Edward Tian, who is a student at Princeton. 
And this tool is called GPT0.me. 
That's GPT, no, GPT-Z-E-R-O dot M-E. 
And this is a solution that uses the same sort of chatbot code to detect whether a chunk of text you have was generated by AI. 
And so the arms race has already begun. 
Oh, cool. 
And to that end, one thing in our earlier conversation I was thinking of specifically with academia, but a tool in our pocket is anyone who has written for us before that has not been using this could be compared if they submitted something. 
And even readers, people that have the best of or that have all the back issues could say, okay, this author presented as one way, had a certain writing style, and they've completely changed. 
You know, those kinds of things perhaps could even train that. 
We could run a submission through a database of previous writers and based on their name or their body work, but it's creating a lot of work. 
I don't know what you think. 
Yeah, but I mean, it's crazy to see artificial intelligence trying to emulate humans. 
Like when we asked for a David Lynch version of a hacker film, it created something that is Lynchian, you know? 
So it knows something about writing style and it's trying to imitate that writing style of humans. 
That's freaky. 
I think David Lynch would actually be freaked out by that, and that takes a lot to freak out David Lynch. 
We have one more call. 
Okay, let's take one more call. 
And good evening. 
You're on Off the Hook Overtime. 
Hey, this is So-Called David from So-Called Oregon. 
So-Called David from So-Called Oregon. 
Okay, well, I believe both of those things probably are real. 
So yeah, welcome. 
Hey, I love this topic, and it's really challenging what it means to be human and what it means is to be creative. 
Mm-hmm. 
And like what Kyle was just saying about, you know, we have to have to honor systems and, I don't know, norms on how to use this stuff. 
Like when you label photos in a magazine as, yes, this has been photoshopped. 
Mm-hmm. 
Yeah. 
Earlier, Collar was talking about, you know, cooperative versus authoritarianism. 
You know, it's like, I think cooperation is our human superpower, and we need to kind of fall back on that. 
It's like computers don't have relationships. 
So, at least not yet. 
Not yet. 
And, yeah, so we need to strengthen that because, you know, people want to give artists money, really. 
Mm-hmm. 
You know, they don't like having them shafted. 
Like, there's some aspects of this that kind of remind me of, you know, the Spotify and things like that where you organize human creativity. 
You get, you put yourself in between them and the customers. 
The artist gets, the creators get, you know, pennies on the grand and, you know, the person with their hand on the money spigot decides that they need the biggest cut. 
Yeah. 
And this training of AI systems, really, it just seems like a very sophisticated shell game. 
You know, ultimately, it's just forming mathematical inferences from previous creative works and saying, okay, make more stuff like this. 
Mm-hmm. 
Based on millions or billions of measurements that, you know, we as humans can't track, but, you know, that's what computers are good at, not creativity, but, you know, extrapolation. 
And so we get stuff that, I mean, I think what needs to be said more is that ChatGPT likes to BS. 
You know, it does stuff that sounds human without any regard for if it's actually true or not. 
Yeah. 
Yeah, I think you're onto something there. 
Yeah, the way you describe it, it makes me think of, like, vector graphics where you're tracing a path and it's creating calculations for that shape, you know, from an image. 
Yeah, that's perfect. 
And then it scales, you know. 
A very small starting point being able to express itself into infinity, you know, with all sorts of variations or like, you know, Conway's Game of Life or whatever. 
And I think someone made the observation because there is a lawsuit going on where people are saying, you know, you've basically made a collage tool and you've used my creative works by scraping it without my permission. 
And kind of the frustrating thing is you look at the size of the training blob and you look at the, you know, millions of pieces of art and text that went into it. 
And it kind of comes out to, okay, you've got like a few bits or a couple of bytes of data in the training set that represents, you know, these, you know, per picture that's in this body. 
So you go, well, it can't be compression, right? 
Because it's just, it's way too tiny. 
But, you know, if you take something that generates off of a vector equation and you look at all that first and you say, wow, how can I compress all that? 
But it's actually a very simple equation maybe that starts it. 
So I don't know, maybe one of the things this thing is doing is like trying all the equations to see, you know, what simple form expresses some large body of work. 
And that'd be like a high computation, low, low storage kind of compression that just wouldn't make any sense. 
And, you know, if you're thinking about it, but, you know, maybe that's something like that's what's happening, but it's basically we're laundering creativity. 
It's kind of changing what we think creativity is. 
Yes. 
When a computer can do it, was it creative? 
Good point. 
Good questions. 
Yeah, I love that. 
And the previous discussion, it kind of made me realize, okay, so if it's such a tiny representation, it's not like the fingerprinting we talked about on people's cell phones in like the last year or two, where it's mapping an image, but not actually taking all of your images off your phone and, you know, judging whether you have anything illegal on your phone or not. 
That was a discussion we had a long time ago. 
Oh, CSAM stuff. 
Yeah, and it's more like this byte that represents that one work is only, it's trained only in relation to everything else that's in the larger set because it was, you know, it's, but again, I don't fully understand it. 
We don't fully understand it in entirety, but it sounds like the chunk of data that results from all of this pits all of these things as points in relation to each other, like a neural network. 
But yeah, it's, is it our responsibility to fully understand this to survive? 
I don't know. 
It's crazy. 
I don't think we can, but, and just to go back to relationships, I think maybe the one way I see out of this, because the computers are getting so good and they're only going to get better at, you know, creating art and text and such that you can't tell if a human did it or not. 
I think we're going to have to start going back to relationships and just saying, okay, starting with cameras and microphones and typewriters, we're all going to have, you know, cryptographic signing, you know, that a, you know, a specific human, maybe not, you don't know who it is, but a specific human made this work and connect it with other works they made. 
And now they sent it to editors who added their cryptographic signature to it and say, okay, I get this, this given person with a history. 
It's like chain of custody. 
Yeah, chain of custody from the creators to the end viewers. 
And you get the sense of, okay, I trust this one and this one is new and things, strange things. 
And perhaps something could kick in if the writing style suddenly seems different for that particular person. 
Yeah, call it a style canary. 
Yeah. 
But, you know, that keeps you from, from, from trying something different, you know, you have to write in your style or it's, it's a thought to be fake. 
Well, it has, it has multiple identities, right? 
I mean, I remember back in the, when the internet, not everyone used their real name for everything. 
That's certainly true. 
A lawyer under a different name than you were a dancer. 
Uh-huh. 
So much to think about now. 
So much to, uh, to consider. 
And, and, um, it's fascinating. 
I think we're, we're, we're getting into this early enough though, where we can, uh, help prevent it from going really rogue and, uh, off the rails. 
Yeah. 
I think we owe it all to ourselves to experiment with this stuff and see how we can use it because, uh, you know, they thought word processors would be the depth, depth of typist, you know, and desktop publishing and all this other stuff. 
So it could be, I, I have seen examples where individuals can do amazing creative works with the help of AI that they couldn't do alone. 
Like, uh, you know, someone was doing something to decode last pass, uh, vault blocks to see how much plain text or unencrypted information was in them. 
And kind of the, the winning program, uh, kind of the punchline was, oh yeah, this is in PowerScript and the GUI elements and a lot of the coding was done by AI because programming languages are just another exceptionally structured kind of language that GPT is very good at manipulating. 
So you'd say, I want to, you know, when this happens, I want to box over here and so on. 
And like, it does something that person did not have the expertise or time to generate. 
And then you just, you know, you tweak the areas where it made mistakes. 
So the ability for people that don't have a lot of creative clan around them to say, I have this idea and now I can do it without involving people because maybe I'm not social. 
So, yeah, you know, uh, we didn't even touch upon that, but, but just the, uh, advantage that this can give people who don't have programming skills to actually create something that they would have to hire somebody who probably wouldn't get what they were saying in the first place. 
And yeah, it, it, it can be very liberating. 
I'm with you on that extending the capability. 
Cause I was, you know, kind of shut out from a lot of that education and I would love to, uh, be able to get help with, uh, my ideas like that, uh, and that kind of cross from art and things I know maybe a little better to the, uh, technical and programmatic. 
So that, that's a great point. 
Hey, uh, thanks for covering. 
Yeah. 
Uh, you're very welcome. 
And, um, we, we enjoyed it. 
We enjoyed playing with it and, uh, hopefully it sparks a bunch of conversations and experiments and we'd love to continue talking about this. 
Thanks so much for your call. 
All right. 
Later guys. 
All right. 
Take care. 
Bye. 
And that's going to be the last call I think for tonight. 
You know, good calls, good calls all around, even Rebels. 
Very much. 
Good call. 
Yeah. 
Now, uh, before we go though, um, I have been working on the film a little bit more. 
I decided to change direction though. 
I thought that, you know, hackers were overdone and there haven't really been any, any real good films about, uh, Donald Trump. 
And I figure why not continue with the David Lynch style. 
And, um, I asked for a scene from a David Lynch style film about Donald Trump. 
And, um, what I've got here, I think is pretty good. 
You guys ready for this? 
Please. 
All right. 
Sure. 
The scene opens on a dimly lit smoke filled room. 
I know it's always a dimly lit smoke filled room with David Lynch, but that's just the way it is. 
Uh, Donald Trump sits at a desk surrounded by piles of papers and empty fast food containers. 
He is speaking on the phone, his voice low and menacing. 
I don't care what it takes. 
I want that deal done by tomorrow or heads will roll. 
Do you understand me? 
He slams the phone down and turns to a shadowy figure standing in the corner of the room. 
They think they can push me around, but I'm the president. 
I'm in charge here. 
The figure steps forward, revealing himself to be a dark, twisted version of Trump's own face. 
The, the twisted Trump says, you're not in charge. 
I am, and I'll do whatever it takes to maintain my power. 
Trump's face contorts in fear and confusion as the twisted version of himself continues to speak. 
Twisted, twisted Trump says, you're just a pawn in my game, a tool to be used and discarded when I'm done with you. 
Trump tried, this is very Lynchian by the way, Trump tries to speak, but no words come out. 
He looks around frantically as if searching for an escape, but the room seems to be closing in on him. 
Twisted Trump says, you can never escape me. 
I am always here inside your head, pulling the strings and you will do my bidding, no matter what the cost. 
Trump lets out a scream as the screen fades to black, fade to black and the end of the scene. 
And then as a note at the end of this, please note that this is a fictional scene and it should not be taken as an actual representation of Donald Trump and his behavior. 
I wonder, I wonder what sparked the AI to do that, that little disclaimer, but boy, that's, that's kind of brilliant, I think, you know, that's, um, it, it, it gets the evil, it gets the, the Lynchian, like, you know, two worlds, which one are you in the doppelgangers and all that. 
So, uh, yeah, that's, that's something that's, that's something you can have a lot of fun with, but, uh, wow. 
I'm going to be writing a lot tonight. 
That's the most important thing is having fun with it. 
I just asked to write a sequel to Gone with the Wind. 
And so it's busy doing that, uh, while that's going, um, I'm going to give out the information. 
OTH at 2600.com is our, um, email address. 
That's what it is. 
Yes. 
Not our webpage. 
Uh, every issue, uh, every, I'm getting all the words wrong every, because the computer's doing the words. 
I don't do words anymore. 
Um, every episode of our radio show off the hook is available online at 2600.com slash off the hook. 
So feel free to listen to them all. 
When you have a moment, we will be back next Wednesday, 7. 
PM on WBAI and next Wednesday, 8. 
PM here on YouTube for overtime. 
Please support us and support community radio. 
Oh, I had, I had a surprising answer from the AI about my sequel to Gone with the Wind. 
I'm sorry. 
I am not able to write a sequel to a specific copyrighted book, such as Gone with the Wind due to copyright laws. 
Wow. 
These AIs are going to play by the rules. 
That's not out of copyright yet. 
You know, I'm not sure. 
Um, I don't know. 
I don't know. 
Where's the movie? 
Is it a movie or a screenplay? 
Additionally, creating a sequel to a beloved classic such as Gone with the Wind would require a deep understanding of the characters, story, and themes, and would be a difficult task for any author, let alone a computer program. 
Is there anything else I can help you with? 
It's a long way around to say that the previous sequel to Gone with the Wind was not good. 
I guess that's sort of a catty way of saying that. 
Mm-hmm. 
Okay. 
Well, I'm going to ask her to tell me how to make an egg cream. 
And I think that would be, that would be a win for tonight if I can get that out of it. 
So, um, okay, folks, uh, we will, and it's, it's given me that good. 
Um, we'll be back next week. 
Thanks everybody for listening. 
Please write OTH at 2600.com. 
Any ideas, submissions, reactions. 
We want to hear it all. 
Good and bad. 
All right. 
Good night, everybody.