WBAI.org online. 
We'll be extremely grateful that you do something like that and tell a friend or two about this radio station. 
Now stay tuned for Off the Hook. 
It is 7 p.m. 
and this is WBAI New York. 
It is 7 p.m. 
WBAI 99.5 FM in the New York area. 
It is just about 7 p.m. 
That means it's time for another exciting edition of Off the Hook. 
The telephone keeps ringing, so I ripped it off the wall. 
I cut myself while shaving, now I can't make a call. 
It couldn't get much worse. 
But if they could, they would. 
Bum Diddly Bum for the best, expect the worst. 
I hope that's understood. 
Bum Diddly Bum! 
Bum Diddly Bum Thank you. 
A very good evening indeed to you all. 
The program is Off the Hook. 
Rob T. 
Firefly here with you, and I am joined in this studio apartment studio by Gila. 
Hello! 
Hello. 
And out there in Skypeland, we should have Alex. 
Let's see if we can get you audible. 
Hello. 
Alex, are you with us? 
I cannot hear you, and that may be something to do with my setup here. 
I'll keep investigating that. 
Try to say something again. 
Okay, this isn't working, but I'll keep poking around and see if I can get this working. 
Voltaire is there as well. 
Can I hear you, Voltaire? 
Okay, I cannot hear Voltaire. 
This is fantastic. 
What we are doing, listeners, is we are steering the ship quite poorly, because Emmanuel and Kyle are away on assignment, and we are just troubleshooting some equipment here live on the air, which should be bringing Skypeland into our field of audio. 
Okay, something is definitely weird on this particular piece of hardware. 
So, Gila, how are you this week? 
I'm Jim Dandyfine. 
I'm sorry, I should probably be talking while you are trying to maneuver and finagle, but... 
We are the Hacker Radio Show. 
We are professional radio makers, and this is all quite fascinating. 
We are having a great time. 
All is well. 
And let's see here. 
Alex is saying hello. 
Alex is waving. 
If anybody... 
Alex, if you want to try, you know, fingerspelling or writing something down and holding it up to the camera, then I can repeat it. 
It'll be like on overtime, which, just for the record, we are not doing off-the-hook overtime tonight, which I think makes a lot of sense. 
Alex, I still can't hear you. 
Sorry. 
I don't know if you can hear us. 
We definitely cannot hear you. 
Well, this is what we do. 
We're flying by the seat of our pants, and this is what happens when we have to throw this stuff together. 
Now, we lashed this stuff up with kind of, I want to say... 
Shoestrings and gum? 
Yeah, something like that, yeah. 
And I think what we should be doing is looking right about here. 
Anyway, Gila, tell us about something that I think is of interest that you've been kind of looking at from the outside. 
There is the... 
There has been the Reddit blackout for a couple of days. 
Oh, yes, the Reddit blackout. 
First, describe what Reddit is. 
So, Reddit is a... 
Oh, what's the word I'm looking for? 
Reddit is a news aggregator, basically. 
It's kind of like a big old bulletin board, and it allows people to share pieces in smaller communities based around subjects of interest. 
And I think Reddit has gotten kind of a bad reputation in the past, because there has been some subgroups that have been built around things that you may not necessarily think are the most savory, shall we say. 
But it can be a place to gather, a place to talk about stuff. 
If, you know, people are sharing stories of, today I learned, today I learned this. 
You find fun things, obscure media, in the obscure media subreddit. 
And people have been using a lot of different apps to anything, anything at all. 
That's a no. 
So, people... 
Sorry, when you do that, the listeners can't tell what you're referring to. 
Sorry, we are still getting nothing from Alex. 
My apologies. 
Okay. 
I am continuing to troubleshoot this. 
While, Gila, you fill the listeners in on the Reddit blackout. 
So, a lot of people have been using third-party apps to connect to Reddit. 
And, a couple months ago, Reddit announced they were going to be shit-changing the API setup for the third-party apps. 
And, actually, the third-party app that I personally have been using, which is called Apollo, the developer has said it's going to cost $20 million. 
Ah, there you are. 
Alex! 
It's Apollo. 
We have repaired this live on the air. 
Number five. 
Look at that. 
Alex, okay, let's get these intros done seven minutes into the program. 
Alex, are you out there? 
Yes, I am indeed. 
Good evening, everybody. 
Glad to hear you, buddy. 
Voltaire, are you out there? 
Voltaire may or may not be out there. 
Voltaire, you're on mute. 
Okay, we're not hearing you at all, Voltaire, but I think that may be to something beyond my control here. 
We'll keep trying to see if we can get that working. 
I'll remove you from the conversation, Voltaire, and re-ad you, or if you could back out. 
Yeah. 
So, but yes, the Reddit blackout happened because Reddit management had been making something available called the API, API, which is basically the doorway into the system that third-party apps for accessing Reddit would use. 
API standing for Application Programming Interface, which I didn't know until right now. 
Nicely done. 
Thank you. 
And the API was always free for the third-party individuals, for the third-party clients to use, but Reddit basically decided to change that. 
They were going to, or they are going to, start charging heavily for the API access, and so that ends up costing when you add it up to quite a lot. 
The developer of Apollo had calculated it at about $20 million a month, I believe. 
No, a year. 
A year? 
It does say a year. 
Okay. 
Still kind of ridiculous. 
So what happened was that the users of Reddit did not take kindly to this. 
There was a whole bunch of drama involving Reddit's communication with the developers and going back and forth, and then Reddit's CEO accusing the developer of lying and blackmailing, and the developer, whose name is Christian Selig, actually posting a phone call recording and receipts showing himself to be in the right. 
So the users revolted. 
A bunch of subreddits, the communities on Reddit, the individual communities of interest, shut down. 
I think from anywhere from two days to indefinitely, a bunch of users stopped using Reddit, and it put such a weird strain on the site to not have activity that the site actually crashed a couple of times. 
It's true. 
And we're looking at an article from Vox here, and they're saying, so this temporary blackout shut down Monday and Tuesday, a lot of subreddits went dark, like 8,500 subreddits, apparently. 
There are a lot of them. 
And they're saying that as many as 6,500 remained dark today. 
So the intention had initially been to blackout Reddit for the first two days, a 48-hour blackout, but a bunch of subreddits have stayed gone. 
And so it'll be interesting to see where we go from here. 
Yeah, it's interesting. 
And I know I was a heavy user of Reddit. 
It was kind of strange to not use it all of a sudden. 
But it opened up actually a significant amount of my schedule for getting other things done. 
I don't know if I know the other two of you may be somewhat Reddit users. 
Yes, Alex. 
You know, I'm not. 
I want to ask, what is it that you're doing on Reddit? 
I mean, I know there's all these. 
What forums and things are you occupying over there, and how is this going to affect your life? 
Well, basically what Reddit is, is since the beginning of any kind of social communication online, there have been ways to create little communities for special interests. 
So you had Usenet, you had BBSs, and you had email lists. 
And then as the social web grew, you had web-based services, things like LiveJournal, things like Tumblr, things. 
Where does IRC fit into this conversation? 
IRC is Internet Relay Chat. 
So that's basically a chat room, and you can create a channel for anything. 
So it's basically a similar idea where it is a centralized space where you could find communities on any subject you want. 
And if there isn't one there, you can create it very easily. 
So, Alex, let me give you an example. 
I followed a subreddit called Delicious Compliance. 
So if you ordered a sandwich and said, I would like a lot of pickles on the sandwich, and the server gave you 100 slices of pickle and three extra cups of pickles just so you had enough pickles to make you happy, Delicious Compliance. 
Okay, yeah. 
I think I get it now. 
Yeah, this is definitely going to leave a big void in people's lives, it seems like. 
I mean, I was a Usenet user way back in the day. 
I really got into Reddit, the whole subreddit thing, I found to be always confusing. 
But Usenet, and I'm sure that many of our listeners will remember the addresses that begin with all.binaries.pictures. 
So, yeah, we'll get some mail about that, I'm sure. 
Yes, but it was the same, it filled the same void where you could start a Usenet news group about any subject you wanted, and other people could find it and join in. 
There were subreddits about, you know, lifestyle things, about cultural fandoms, you know, you're a Star Trek fan, you're a Doctor Who fan, whatever. 
And also, like, if you're left-handed, if you're a specific religion, if you're in a part of the country, there were New York City subreddits. 
So, basically, any kind of subject you can imagine to center some social interactions about. 
And, frankly, probably some that you couldn't. 
Yeah. 
Then that's, so, basically, it was the latest way to do that, the latest big popular way to do that. 
Yes, Alex? 
Yeah, and so what's the status of this now with the protests? 
Is it having any effect, or are we seeing any movement? 
Well, the protest happened. 
There was, there were, there were murmurings that it was a bit short-sighted for people to, when they started this protest, said, we're going to go off for two days, 48 hours. 
And so management was very clear with sticking to their guns and saying, well, this will blow over in two days. 
So there are, there are a certain amount of the protesters that are keeping it going, a certain amount of the subreddits that are still closed, a certain amount of users that still aren't using it. 
But I don't know if I'll go back to using Reddit. 
If I do, it probably won't be as much as I once was. 
But, you know, there, it, this is going to be interesting because it's where a lot of the web is for a lot of people. 
And it is also another example of, like we've seen with Twitter, one big company controlling, basically, a major chunk of the social web and using, using the weight that they can throw around to, to try and direct things the way they want to. 
And one thing I did not think about until this very second is a ripple effect that this may have. 
Because if some subreddits stay dark, if some users stay away, there is a whole series of places on the internet that exist solely to farm Reddit content and republish it in listicles and things of that nature. 
So what is this going to do to BuzzFeed? 
There are YouTube channels that just exist of farmed Reddit content. 
What's going to happen if these subreddits stay dark, if people stop using them, if this migration of users continues in the way that it has, what that will mean for other arms of popular culture? 
Yeah. 
So we'll see where this goes. 
And it's worth keeping an eye on. 
There was an active 2600 community on there. 
Shout out to the folks who run that and kept it moderated well and everything. 
But we'll see where that goes. 
But we don't want to spend the whole time here talking about one website. 
Well, if you're sure. 
Let's talk about a different website. 
Let's talk about – Alex, I know we talked about this a little bit last week, but there's been a little bit of development in the story. 
And Voltaire, I'm sorry, but we're still not having any success connecting with you. 
And I'm not sure what's up with that. 
Can you hear me now? 
There you are. 
Oh. 
Hooray. 
All right. 
Mike at Drag Race. 
All right. 
Well, tell your driver to take the car and pull it around and do whatever. 
But all right. 
Voltaire, it's good to hear from you again. 
You've been away for quite a while. 
But you've been steadily working, I must point out, behind the scenes of this program, helping to run the social media and everything else and sending in tons of stories and research. 
And so, you know, for those of you out there who miss Voltaire, he's still been plugging away all this time, just more quietly. 
But we've got you back on air now, finally. 
Thank you. 
Now in Berlin. 
You're now in Berlin? 
Between Berlin and New York. 
Okay. 
That covers a wide area. 
Where are you right now? 
New York. 
But then I'll be back to Berlin. 
Because if this was like some airplane service thing, then no. 
But getting back to what we were talking about, Alex, we were talking about the story of the lawyer who used ChatGPT to form his briefs. 
And his briefs got all wrinkled and so forth. 
And so the lawyer, Stephen Schwartz, this is a story from the Times on June 8th, appeared nervously upbeat, grinning while talking with his legal team. 
Nearly two hours later, Mr. 
Schwartz sat slumped, his shoulders drooping, and his head rising barely above the back of his chair. 
And the judge that he was in front of while he was doing all this slumping, Alex, you're familiar with this judge. 
Yeah. 
You know, this is an interesting coincidence, I guess, right? 
I mean, this is actually none other than the judge before him was Virgil Griffith. 
This is Judge Castell, Judge P. 
Kevin Castell, I believe, of the Southern District of New York. 
And as you know from the reporting and all the discussion about the Virgil Griffith case, Judge Castell is a very tough judge, extremely tough. 
And it just seems like the absolute wrong judge to pick to try to try out ChatGPT for writing your briefs. 
I mean, it's absolutely crazy. 
So I really think they could not have picked a worse judge to really try this on with. 
Now, speaking as a lawyer who I'm guessing is not using ChatGPT to write his legal briefs, you know, I'm not speaking to however you choose to do your job, but the idea of using ChatGPT to put together a legal briefs for you, I mean, it didn't work out too well for this guy. 
Yeah, it didn't. 
I mean, it's an interesting idea, right? 
Because, you know, and it also could depend on what you're using ChatGPT for with respect to a legal brief, right? 
I mean, some things might make sense, like, you know, organization of the arguments or something like that. 
If you have it generate an outline of what the brief should look like, that might be helpful. 
But then again, you are operating under the assumption that ChatGPT has ingested and knows how to output the proper legal standards for whatever the issue is in your brief as well. 
So you're expecting it to know and understand and to be able to regurgitate correct substantive law. 
So I think even that is questionable, even, you know, if you're relying on it for identification of certain legal elements. 
But, you know, that's kind of like just creating a skeleton of the brief. 
If you had ChatGPT actually create the briefs for you, that's an entirely different story where you are expecting it to do a lot of super complicated things. 
And just like, you know, just like what happened in this poor unfortunate lawyer's case, is that, you know, ChatGPT, when it's trying to cite sources, we know it's not great at that, and it starts to, quote, unquote, hallucinate. 
And, you know, I guess it's a real personification of these types of, you know, what are essentially, you remember when we had Cory Doctorer a lot, he was referring to these things as essentially predictive speech, right? 
So these very complicated predictive speech machines, you know, we're really personifying them with the ability to hallucinate now. 
But, I mean, essentially what's happening is, you know, and this goes back to an article I read in the Financial Times in the lunch with the FT section a couple of weeks ago. 
They had on, of course, I can't remember his name off the top of my head, but this was the author who wrote that famous article now about how ChatGPT or artificial intelligence is a blurry JPEG of the Internet. 
And that was a New Yorker article that came out just when ChatGPT was starting to take off. 
And this guy, in the Lunch with the FT article that I was reading, which I thought was quite good, he mentioned that he just thinks it's a really badly chosen term for a subject of computation. 
Artificial intelligence is just the wrong term. 
There's nothing intelligent about what's happening, and certainly not with the systems that are at play right now. 
And when asked, well, what would you rather call it? 
He said it would be more appropriately named applied statistics. 
That's what's happening here, right? 
You're ingesting a massive amount of content in text, and based on what's happened in the past, those statistics, it can predict what words would make sense going forward. 
And that's where these hallucinations come from. 
And that's why you should never, ever rely on ChatGPT to do things like cite legal precedent and understand it, because it just made it up based on the words that it had seen in the past. 
That's where these hallucinations come from. 
So, yeah, it's a long way of saying don't do it, I guess. 
Yeah, it behooves us to be aware of what exactly you're talking to when you talk to ChatGPT. 
And it is very compelling and very easy to slip into feeling like you're talking to some sort of intelligence. 
It's there in the name, artificial intelligence. 
It must be an intelligence. 
It is not. 
It is a very sophisticated idiot. 
And, you know, this is not your Knight Rider car ready to hold a complete conversation with you and help you solve crimes. 
This is, as Corey said, it's predictive text. 
It is autocomplete. 
It is guesses at something that's sentence-shaped, that's paragraph-shaped, that's legal brief-shaped, without actually having any of the content or motivation behind it that something written by a human would. 
And I don't know if you had a chance to look at this other story about the European Union. 
And I know you've got a little experience with the law over in Europe or bits of the world that used to be Europe. 
Not anymore. 
Yeah, not anymore. 
But the European Union took an important step today. 
This is another story from today's times. 
The European Union took an important step on Wednesday, that's today, toward passing what would be one of the first major laws to regulate artificial intelligence, a potential model for policymakers around the world as they grapple with how to put guardrails on the rapidly developing technology. 
And the interesting chunk of this is it is known as the AI Act. 
It's a draft law, which means it's a long time before anything actually legal might happen with it. 
It's going through deliberations and that. 
But it would put new restrictions on what are seen as the technology's riskiest uses. 
It would severely curtail uses of facial recognition software while requiring makers of AI systems like ChatGPT to disclose more about the data used to create their programs. 
And so this seems to be drawing out some of the complaints that we and folks in our community have had about all of this since the latest boom. 
And it's very interesting to see the European Union take the lead on this and start on what might become sort of an easily copyable policy if they get it through. 
I agree, and it's interesting, you know, if you think about, you know, how they're doing it too, right? 
So it's mostly on the basis of the potential use case for the artificial intelligence and done on the basis of risk, right? 
But how they assign those risks is going to be interesting. 
And obviously the level of regulation you have is directly proportional to the level of risk, right? 
So the more risk you have, the more regulation there's going to be. 
But, you know, what are the things, like, what's low-risk AI? 
You know, what are those things, what's that look like? 
You know, can we think of any examples of, you know, what would be low risk? 
Would low risk be, you know, something like, I don't know, employment decisions, right? 
You know, sorting through resumes, that would probably be something like medium risk, right? 
I mean, lives and livelihoods can depend on that kind of thing. 
You don't want to have them act in a too discriminatory way. 
I mean, self-driving cars, obviously, that's pretty high risk. 
But then the other thing is that's interesting, I think, with regard to the EU regulations, is that there will be some use cases that are just forbidden, right? 
Like policing with artificial intelligence, you know, making those types of decisions, or having a judge wheel, you know, by virtue of an AI. 
So, you know, yeah, I agree. 
Really interesting stuff. 
And, Voltaire, you've got an update on this story. 
Yeah, for the past few months, this has been drafted. 
There's been, although it's mostly been a good thing as an overall policy, there's been problematic language that would basically criminalize a lot of open source and free software, and the Free Software Foundation Europe, and Edri and other similar nonprofits have been lobbying against that. 
But it seemed unlikely that those worries would actually be taken care of. 
But surprisingly and thankfully that they just published today on the FSFE blog that they were able to put in safeguards of tech free and open source software so that it won't be, like, criminalized if it can be used for problematic AI. 
And it seems to have passed this stage. 
Yeah, well, the amendments, the FOSS protecting amendment passed. 
Ah, okay, important distinction. 
And the bit of this I find really also very interesting, as someone who does creative work of the types that a lot of people are talking about, the AI implications for, is the bit that, excuse me, the bit that requires makers of the systems like ChatGPT to disclose more about the data used to create their programs, their source material. 
And with a lot of copyright concerns and things of that sort being raised as the current sort of creative, quote-unquote, creative AIs out there that work on things like images or text, if they have to put out there what they're using as fodder for the system, as stuff for it to sample and sound like and look like and regurgitate, because there is the question of whether people whose text gets reused by these bots deserve some sort of ownership over that. 
Alex? 
Yeah, well, that's also being litigated right now in the federal courts. 
There's an ongoing case, and I don't want to speak about it too much because I have some connection to it, but it involves a band of artists, basically a putative class action of artists suing stability diffusion, one of the engines for these AI-generated art systems. 
And as well as a couple of other defendants that have used stability diffusion, like deviant art. 
And the allegations are that these images that were used to train stability diffusion included copyright materials of artists, and that ingesting copyrighted material for the purpose of training in artificial intelligence goes beyond what we would consider to be fair use under the copyright law in the United States. 
And it's a commercial use. 
And without having a license to use the copyrighted material, the argument is that they're violating copyright. 
Possibly millions amounts of times, right? 
So this latest version of the AI bill in Europe essentially requires the developers of AI systems to publish some form of summary of the copyrighted material that's used as training. 
And, you know, the issue that's going to come up here is that, well, what does summary mean? 
How detailed is a summary? 
Do you have to list specific artists? 
Or do you need to provide the individual songs? 
Is that the summary? 
Or can you give some kind of broad category of data that's been used to train the AI? 
So I think this leaves a lot of room for arguments and ambiguity as well. 
Yes. 
Voltaire? 
Yeah, I like for the topic of fair use and used by the incorporation of artistic works. 
And I think as a hacker community, this process is in an awkward position where there is a lot of, like, at least from my perspective, a lot of, like, knee-jerk, anti-corporate, because perspective where, like, wanting to take down all the biggest AI's because most of the big AI, like, TAPT and such are from big corporations. 
But I think that we don't, for a hacker protector, we don't want to throw out the baby with a bathwater and, like, actually end up weakening fair use rules. 
So I think that puts it into your position. 
Yeah. 
And the question of creative AI use is actually at the center of something that a lot of people who watch television or enjoy films or things of that sort are experiencing right now due to the current WGA strike. 
Absolutely. 
One of the major, there are a lot of sticking points right now in the WGA's search for a new contract. 
That's the Writers Guild of America. 
The Writers Guild of America. 
But that represents writers of television and film. 
So if you're wondering why they just decided that the fifth Avatar movie is going to come out in 2031, because it's already written and people are trying to stretch out what they can as long as they can. 
But one of the major points in discussion is how AI will be used going forward. 
Because if the idea is that you're the copyright holder and you can train AI on however many hours of, let's say, Star Trek already exist, and then instead of hiring writers, you can just have AI spit out new scripts for your existing IP, the writer said, we need to have ongoing discussions about how this is going to work. 
And the producer's organization said, yeah, we'll talk about it once a year. 
But if you've been following along, what we've been talking about with these technologies on this program, things change week to week, day to day. 
You can't go back to these conversations once a year and expect that to hold. 
Because the question is, where does, what produces content? 
Where does content come from? 
What, I saw a strike, I saw a picket sign somebody was holding that said, AI didn't have childhood trauma. 
I was entertained, I don't know if anybody else was. 
But, no, I think there's a very key piece about where do these things come from and who owns it and how does it get created? 
And, you know, can AI break a season? 
Yeah, Charlie Brooker, who, of course, created and writes the majority of Black Mirror. 
Ah, the notoriously easygoing. 
Yes. 
The guy who writes about how screwed up everything is thanks to misuse of modern technology. 
Lighthearted. 
Lighthearted was the word I was looking for. 
Lighthearted, exactly. 
There was an article that went by recently, you showed this to me, that he actually used ChatGPT to write a script for Black Mirror. 
And he had a look at the script that came out and he said, it's terrible, I'm glad my job is secure for at least a little longer. 
But this was a question that I had because if you ask ChatGPT or any of the other iterative AI products to write you something and you write it, you give it the prompt, it gives you whatever it is and you say, no, no, no, this isn't right. 
You correct its information. 
Are you then feeding the large language model in a way that's going to perpetuate this system regardless of what they trained it on? 
Absolutely. 
The enthusiast groups around these things, that you see people doing all sorts of things with these things. 
And one of the big subjects is tips on how to craft your prompt. 
Voltaire, you had something? 
Yeah, I certainly support the WGA writers and the strike and back and fully. 
But I think it's important not to, I think by playing, like the issue of AI is just one of the many issues that they have up for debate. 
I think it's important not to overplay it because I think that can end up playing into the hands of the film and TV companies. 
Because AI, I don't think it'll even like within like two decades, I don't think it'll be anywhere close to being able to replace writers. 
It was just like the quality of the output, like you mentioned about, it's just like not anywhere close to being able to, like, it'll make sense grammatically, but not, it'll be funny or have drama. 
But it's kind of like back 10 years ago, so we were at a point where people were, there's a moral panic about robots replacing Amazon workers. 
And then you had like Amazon having a puff piece on 60 Minutes where they had delivery drones and they were just supposedly going to replace all the workers in two years. 
And then it became like a deflection issue where people were worrying about robots instead of like protecting, like trying to regulate away robots instead of, like delivery robots, instead of crafting new regulations to protect the working conditions of Amazon warehouse workers. 
And we still have tons of warehouse workers with horrible working conditions. 
So I support it not to focus on the AI things that will never be able to replace writers within the next few years. 
Very true. 
And, you know, just brought up AI because I think it's more relevant to our wheelhouse here. 
But I think also, I don't think also I lost my train of thought. 
Sorry, Alex, what do you got? 
Yeah. 
Well, to build on the point that Voltaire was making, I think, you know, something we discussed a couple of months ago, maybe four or five months ago on this show is this notion. 
And it's more of a philosophical notion about artificial intelligence. 
And it comes from a philosopher by the name of John Searle, who is, I believe, still teaching over at Berkeley, I think. 
But in any event, John Searle wrote a lot about consciousness and artificial intelligence. 
And he believed that AI could not be made unless you can understand or replicate the causal powers of consciousness in the brain, right? 
That there couldn't be some other artificial intelligence or some other consciousness. 
The machines couldn't be conscious in that sense. 
And that there was a big difference between the simulation of consciousness, essentially what the Turing test tests for. 
You know, this old test came up by Alan Turing, essentially. 
It says, you know, if you think you're talking to a human when you're talking to a computer, then, you know, there's consciousness that doesn't really necessarily follow. 
But so there's a big difference between the simulation of consciousness or the simulation of intelligence and the duplication of it. 
The duplication would be the creation of real intelligence, essentially out of nothing. 
And so what we're dealing with now, again, to use the phrase that Cory Doctorow used on our show, is that, you know, we're dealing with, essentially, a predictive text engine here. 
That's simulation. 
That's clear simulation. 
And so I think Voltaire's absolutely right in that the things that chat GPT or these predictive text monsters are going to write are not going to be all that funny. 
Because what was said in the past, you know, it doesn't gauge for humor. 
You know, it ingests all of this text. 
How can it account for human humor without understanding that concept as a human? 
Absolutely. 
Number five is not yet alive. 
No. 
I think there's also a piece of this, though, and this is what I was thinking about earlier, is that I think it does fall on us to be maybe more discerning consumers of what's out there. 
And just because it is there, especially if it's, you know, gigantic tentpole AI, we don't all have to rush out to see it. 
Because the more we're able to say that we're looking for thoughtful human stories, and that's where the money is going. 
In the same way that I was adamant, adamant, that the weekend Are You There, God, It's Me, Margaret opened in the theaters. 
I said, we have to go see it this weekend, because I don't want to see another two hours of things getting blown up by robots. 
Sorry, people who enjoy things getting blown up by robots. 
But I think to find the humanity in everything we're looking for, and to find things that are not necessarily built on existing IP, and the new places we can go as a society, as a culture, if we remember the human behind whatever element it may be. 
Absolutely. 
One thing the current state of sort of the creative internet has really opened up for people is you can go out and find creative content by writers, artists, musicians, and so forth, who are doing things straight from their house, without the Hollywood machine behind it, or without the record industry machine behind it. 
And you know when you buy an album on Bandcamp or something, that the majority of that money is actually going to the artist, and just developing these sort of interests and wider sort of libraries based on things that people are putting out there themselves, sort of under their own steam, there is a lot we can do to, I think, support smaller creators, support creators who are taking risks, who are doing weird things. 
I have so many books on my shelf that would never in a million years get to the New York Times top ten, or featured at Barnes & Noble or whatever in the past, but they are things that people have made that I find I really enjoy, and there is worth in that. 
So, talking of choices we make as consumers, there's another story that I was really fascinated by. 
This is on a site called practical-tech.com, that's the URL. 
It's written by someone called Stephen J. 
Vaughan Nichols, and this story is how an Amazon Fire Kids tablet was allegedly used to stock a security pro. 
This story broke today on this site, and it involves a security researcher called Dick Morrell, who is known, I believe, as Cloud Guy. 
And this is somebody who is a security researcher, who has the chops to properly control access to their own devices. 
But the flaw in his Amazon ecosystem ended up being an Amazon Kids tablet, an Amazon Fire tablet for kids. 
That, you know, you've seen these, children play with them. 
They have big sort of rubbery cases to protect them from smashing and drool and whatnot. 
But one of these devices left his possession, and his account was still registered to it. 
And this became a way for somebody who later got hold of this tablet to spy on him, to drop in on his Alexa devices, to look at his Amazon content, to look at the things that his Amazon apps and devices and things like that could access. 
And this was somebody who basically did know better and still fell victim to this manner of stalking. 
And it's a fascinating story. 
We don't have time to fully get into it on this program, but I'll recommend people go to practical-tech.com and look for the story how an Amazon Fire Kids tablet was allegedly used to stalk a security pro. 
It looks so innocuous in its little rubber baby tablet bumper. 
Yeah, the twist with these is that they functionally have two accounts on them. 
They have the account that faces the kid-facing account, which is theoretically locked down, and so you can hand the device to a kid and they won't do anything dangerous or escape the sort of walled garden that you give them. 
But there is an adult's account also registered with it. 
And that adult's account is how somebody used this to get into Morel's info and past his safeguards. 
And a lot of folks may be saying, but you can go on Amazon.com and see your devices that are registered and deregister your devices remotely if you no longer have them. 
This tablet was not listed on that page. 
And this sort of thing looks likely to happen with this sort of thing, even if people check that, because Amazon's sort of control of the ecosystem is being lessened. 
They are taking away things that you could do to manage this strictly on the web because they want you to use their Alexa app for that sort of business. 
Yes, Alex? 
It's interesting, the point that you bring up, that the tablet has two accounts. 
I am quite familiar with the Amazon Fire tablet. 
And you really never see the other account until you need to do something configuration-wise, which is very rare, right? 
So you could be discarding this device or giving it away to somebody, and you completely forget that that account is actually behind it because you're never in it. 
You don't see it. 
And besides, it's not really your tablet anyway. 
It belongs to your kids. 
But the other aspect of this, too, is, you know, let's say you're a disused device that you're signed in, by which you were signed into iCloud or something, and all of a sudden comes online. 
Well, it might be flagged by Apple as a curious device, that there's something wrong here, or it might require you to authenticate that device from another device or something. 
So there are ways to see that. 
But in this particular situation, I imagine that because the Fire tablet was active and it was constantly signed into that particular account, it probably still had one of these persistent sessions. 
And it didn't look like there was any anomalous activity happening because it probably looked perfectly normal for that device to be connected and doing things on the account. 
So it makes it even more difficult to detect, in my opinion. 
Absolutely. 
And, I mean, these things are persistent. 
They stay logged in. 
They stay configured. 
I have a Fire tablet that is not a kid's Fire tablet, but it is my Fire tablet. 
It is our Fire tablet that we have kicking around this household that is actually attached to the chain of lashed-together technology I'm using to get this broadcast controlled and going and connected to the station. 
And this tablet is playing the music that you heard at the top of the show and that you'll hear later. 
And this is something that I had not switched on in a long time, but, you know, it switched right back on and everything went, you know, right back to working like it always does. 
And I didn't have to log in or anything. 
And so it's, you know, there are these things that have become such parts of our lives, the stuff that we have just sitting around. 
And it really, it really, we have to be aware of what happens when things can leave our physical possession like that. 
Absolutely. 
And the, the, the trail it can leave, the connections it still has to it, it can be very, very scary because it connects all over it. 
And what you, and you think about the kid account, you don't think about the adult account on it. 
Mm-hmm. 
And people have downloaded, like, you can access the data that Amazon keeps on you and it is clips of your voice. 
It is, you know, records, purchases, things like that. 
But it is also, like, these devices can be used to what they call drop in. 
And that's listen to what's going on wherever another Alexa device on the account is. 
And that's, that's Alexa appliances, like we've got, or, you know, off on a shelf somewhere, or it's a phone with the Alexa app on it, or a tablet with the Alexa app on it. 
Theoretically, it could also be the Ring doorbell. 
The Ring doorbell, especially, yes. 
The Nest, is it Nest? 
The Google, the formerly Google Nest. 
Right, but whatever smart thermostat is attached to this, it, you know, looking at this being posted on Mastodon by Dick Morrell, the engineer in question, who's saying, don't forget, this can be a huge part of the toolkit of a domestic abuser. 
And that's part of the reason why this is so scary. 
It's not just being spied on in one whole, which is scary enough in and of itself, but, oh my goodness, the next piece of it can. 
This seems to be a person who had ill intent in stalking him, and the data you could access with this goes all the way back to 2013 in this, in Morrell's case. 
So, yeah, interesting stuff indeed. 
I want to quickly get to, we, we love a listener email, and we received one from a, from a listener called Jan. 
Hello, Jan. 
Jan writes, first, thanks for the continued shows. 
You're welcome, Jan. 
Jan continues, I was wondering, how come you're still remote recording instead of going back to the studio? 
The answer to that is very simple. 
We can't all fit in the studio, not safely. 
There are still limits on how many people can be in the studio at the station, and our panel is just too big, too numerous. 
There are too many of us. 
So, uh, we're, we're still all connecting up from home, uh, with the help of our, our great friend, Reggie, back at the station, who, uh, helps us get, uh, get on, get on, uh, on the ground with that. 
Reggie, uh, if you're listening to this, we, we appreciate you, and we don't tell you that enough. 
And if you're not listening to this, tough. 
But, uh, yeah, so, we are, we are, uh, still doing what we're doing for the time being. 
Of course, we would love to go back to the studio, and, uh, I think when we can, you know, we'll, we'll, uh, see what happens then. 
Um, Jan also goes on to, uh, mention something we were talking about last week. 
Um, another remark I'd want to make is about the Gmail blue checkmark thing. 
While their implementation, blue checkmark, really, sucks, the thing they're using behind the scenes is an industry standard called BIMI in combination with other standards, DKIM, SPF, DMARC. 
Um, it would have been nice to dive somewhat into that instead of going about how bad it is. 
Um, well, we appreciate the further info, but, uh, I think our, our point still stands. 
We were, we wanted to talk about how bad it is. 
Um, it was a little underwhelming. 
Actually, um, excuse me, in, um, our last minute or two, we did actually get a listener email while we were on air, uh, responding to the Reddit discussion we were having at the beginning. 
So, just kind of put an envelope on all of this. 
Um, like to thank our listener, Alan, for mentioning, uh, here, he says, greetings off the hook. 
I hear you talking about the Reddit news on the show now, and I want to share this. 
I haven't heard anyone mention it in this context. 
Uh, sharing a story from TechSplore, saying that unpaid social media moderators perform labor worth $3.4 million a year on Reddit alone. 
Which is also just, I think, an interesting part of the discussion that we didn't get a chance to talk about. 
Absolutely. 
And we'll have to leave it there. 
Um, send us email, OTH at 2600.com. 
Uh, find us on Mastodon, off the hook at hackers.town. 
Um, and, uh, I'd like to thank Voltaire for, uh, popping in for, for a guest appearance again. 
Uh, thank you, Alex. 
Thanks for having me. 
Thank you, Gila. 
Um, thank you, Reggie. 
And, thank you, me. 
And we will all be back. 
Not next week. 
We are off next week. 
And we will not be doing off the hook overtime tonight, because we just don't have that box of tricks with us. 
Um, so, we will, uh, you will hear from us again in two weeks. 
We are preempted next week for special fundraising programming. 
Support WBAI. 
Give to WBAI.org. 
And keep doing what you're doing. 
For off the hook, this has been Rob T. 
Firefly. 
Good night.