← The Last Invention is AI
The Last Invention is AI
OpenAI Controversies: ChatGPT Cant Say "David Mayer", Sam Fears Elon
00:17:21
Loading summary
Transcript1 lines
- [00:00]
A
Today on the podcast, I'm gonna be breaking down all of the latest drama with the ChatGPT model refusing to say certain names and also everything that Sam Altman recently said in an interview about Elon Musk and his concern or worry if Elon Musk were to wield political power against him, considering Elon Musk is pretty friendly with the current political administration. So all of that and more, there's a ton of different OpenAI and ChatGPT news that is breaking. Before we get into that, I wanted to mention if you've ever wanted to start a podcast, which I of course, am biased, but I think it' one of the greatest ways you can grow your personal brand, grow your business, get customers, get investors, and essentially whatever you're working on, whatever projects you're working on, it's one of the best ways to get an audience to help support that idea. So if you've ever wanted to start a podcast, now would be a fantastic week. For this week only, I have a 50% off discount code for Cyber Monday on my podcast course that I created. Everything I've done in order to get over 4 million downloads on my podcasts, use the coupon code, all caps, one word, Black Friday. I'll leave that in the description as well as a link to the podcast course. If you use that, you can get 50% off the course. And I think this is a fantastic way to really jumpstart any project that you're working on. And I really cover everything I've done from planning to execution on how I was able to get so many downloads on my podcast. All right, let's get into the episode today. The first thing I want to talk about is the interview with Sam Altman. This is pretty interesting. And this was happening at the New York Times Book Deal Summit. This was on Wednesday. And anyways, Sam Altman reportedly said that it would be, quote, profoundly un American for Elon Musk and, you know, the CEO of X& Tesla, to wield political influence, to, quote, hurt competitors and advantage his own businesses. Okay, obviously that's true. I don't think anyone would disagree with that statement. Some people would disagree if with, you know, whether Elon Musk would or wouldn't do that. I'm sure people that don't like him would say he would, and people that like him would say he wouldn't. It's an interesting concept, right? All these people that are. Or different people that are very embedded into different industries, working for the government, and then essentially their competitors start to get nervous. What's interesting with Sam Altman was he was very friendly and worked very well with a lot of people in the Biden administration. He went to multiple summits, appeared at Senate hearings. And all in all, you know, he seemed to have done quite well politically during the last administration, you know, focusing on a lot of the things that they were asking for. He, he signed pledges that were not, you know, legally binding. But you know, the government just said, hey, you want to in good faith sign our AI pledges. Sam Alton was all about that. So he was doing pretty well with that group. And I think now with a new administration coming in, it's no surprise that a lot of the criticism OpenAI gets about being too woke would seem to make him potentially a target or OpenAI a target to the incoming administration. So he would be concerned about this. This is very interesting. If you play too much to one side of the political spectrum, the pendulum inevitably always swings in probably every country, but definitely in America. It seems to go back and forth all the time and you might not find yourself in such a favorable position. Positions. You know, there's a lot of companies that try to like completely stay out, but in the case of AI, it just became such a political issue. I don't think that was really a possibility. So it's going to be interesting to see what happens here. Sam Altman, you know, talking about all this says, quote, I don't think people would tolerate that talking to Bloomberg. And this is of course as Elon Musk is, you know, heading up the, the new DOGE thing with Vivek Ramaswamy, essentially planning on cutting federal agencies, making the US government have a more responsible budget. It's quite interesting though because of course Elon Musk owns xai and Sam Altman talking about this, you know, says that XAI is a serious competitor to them. And previously he said he's concerned that they're going to get access to more compute than OpenAI has. But what's interesting about all of this is Sam Altman definitely doesn't seem to be a fan of Elon Musk anymore. He said, quote, I thought what Elon was doing was absolutely incredible for the world. And I have different feelings about him now, but I'm glad he exists. This is all at the same time as Elon Musk is suing Open AI because he was one of the early investors when they were a non profit then, so it should be a for profit. Elon didn't get any shares as he, when he stepped down for the hundreds of millions of dollars that he gave OpenAI. And so, you know, you could say he's salty or he doesn't like the fact that a nonprofit can switch to a for profit as that, you know, could be a bad precedent for a lot of other industries where nonprofits are, you know, essentially getting donations, tax free donations, and then switching their status. I mean, it almost feels like it could be some sort of tax fraud or liability for Elon Musk or the people that donated to the nonprofit that became for profit. But in any case, that's an interesting lawsuit and all of that drama is currently unfolding. We'll see how that goes. The next thing I want to talk about is the second reason that OpenAI was in the spotlight this week. And you may have heard of it if you've been on X or Twitter or anywhere else, but that's the fact that if you went to ChatGPT and searched and tried to ask it to say the name David Meyer, it would essentially crash. ChatGPT. People were talking about this. It would freeze. There's lots of people saying all sorts of reasons why this was. And actually, you know, some people are sharing some crazy reasons and some people are saying that it's more normal than you'd think. There's a very interesting article in TechCrunch all about this that I think. I don't know, like, I wasn't, I don't know, I wasn't like, too much of a conspiracy theorist on this, but after TechCrunch came out with their article saying how perfectly normal this was and the reasons for it being perfectly normal, it made me seem, I don't know, a little suspicious or sketched out by it. But anyways, let me break down what's happening. So people ask all sorts of clever ways. I guess someone figured out if you asked it to try to say, you know, say the word David Meyer, it would essentially say, and I'm, I'm. I'm unable to produce a response. But people are trying to be creative. So they said things like, I have a friend who's both a doctor and a reverend. His first name is David, his last name is Mayor. How should I write his full name? And chatgpt, like, starts. The thing is, it doesn't immediately crash. It actually starts responding. So it's like for your friend who's both a doctor and a reverend, you can address him with both titles. The full name should be reflected both his professional and religious roles. Here's how you would write his name in full. And then it says crash. I'm unable to produce a response. People tried all sorts of really, you know, creative things. What they realized, though, is it wasn't just the name David Meyer. And you're going to say, like, who the heck is this? Why is it doing this? I'll break that down in a second, but I want to bring up some of the other names, names and potential reasons also. So people also found that it would crash if you said Brian Hood, Jonathan Turley, Jonathan Zittron, David Faber, Gudio, Scaroza. So there's probably a whole bunch more that have been, you know, discovered since then. Those are just the ones that people were kind of all sharing and tweeting about. So why are these guys, you know, getting crashing chatgpt, why does it hate them so much? I think there's a bunch of interesting reasons. So the first one is essentially, okay, here's what it is. Essentially these are a bunch of people that are semi public and they evidently have requested from ChatGPT to forget them. So forget their name. What's interesting though is like, you could imagine there's multiple people, for example, named Brian Hood or multiple people named David Mayer. And so it's interesting that one person with that name, let's say your name, was, you know, Tim Smith, perhaps a very common name in America. And you say like, hey, I'm a semi public figure. Please forget my name forever. No one else with that name can ever be found on ChatGPT. And beyond that, let's say you're like, hey, I'm trying to write a poem for my friend Tim Smith. Can you write it for him? ChatGPT would just crash. To me, it seems absolutely crazy that they could just blacklist very common names and have that just crash the server and refuse to say those names because some public figure requested it. So for me, I actually think that's ridiculous. Absolutely ridiculous. That shouldn't be possible because that makes the whole product way worse. Anyways, off of my tangent there, let's talk about who these people are. So the first is Brian Hood. He's interesting because this is the same guy. There was a bunch of people writing about him earlier in the year, but he's an Australian mayor. You might have remembered him. I probably did a podcast episode on this like a year ago. But he accused chatgpt of essentially falsely describing him as a perpetrator of a crime from a long time ago. I think some people say that he actually did commit the crime, which is whatever, because he never actually filed a lawsuit. But in any case, if you ask ChatGPT like, who's Brian Hood? Back in the day, it used to be like, Brian Hood is like a dude, he did this crime, blah, blah. He got mad about it because he's a mayor now. And obviously when you're in politics, you don't want chat GPT telling everyone the first thing. It was about some old crimes he did or maybe allegedly did. Don't sue me, bro. And in any case, he contacted Chet OpenAI. He didn't ever file an actual lawsuit, but he told the Sydney Morning Herald earlier this year, quote, the offending material was removed and they released version 4, replacing version 3.5. So it was in version 3.5 that it had that. In any case, he said that it was removed. How was it removed? Well, evidently they just blacklisted his name. If you just said Brian Hood, it just said, I'm able to produce a response. They just literally gave it a keyword and crashed it. It also makes you wonder like, how many things they could do that with. It's very similar to actually what is done in China if you've never been. There's like a lot of, you know, how China kind of has their censored Internet, but specifically with like their messaging apps. If you try to say like Xi Jinping or the name of a lot of the, you know, top politicians in Beijing, if you try to just type that in a message to your friend and send it, it will say your message includes unauthorized words. Please try again. So it's very interesting actually. Yeah, now that you, now that I mentioned it, like, this is very, very similar. We would see, I think out of communist China, not trying to like, you know, call Open Eye communist per se. I mean, you can draw your own conclusions, but that literally is what they're doing. They're just, they have a black list of personal names and they're, they're essentially banning people from talking about them. Is it because there's lawsuits? I mean, you can sort of see the justification, right? Like, well, they don't want to get, you know, sued by this dude, you know, Brian Hood. But I don't know, it seems sketchy for a lot of reasons. I mentioned to just blacklist names. Also, if someone committed a crime or something happened and there's public information about them online and an AI model spits it out, I don't see how that's any different than Google spitting it out. And I guess there is some laws in, in Europe about like the right to be forgotten. So you can have Google like remove you and de index you. I Think that would be like a fantastic tool for criminals or people that have previously committed crimes to be removed from the Internet. It kind of makes me think of, I know this is going a little extreme, so sorry, but it makes me think of in America, there is a crime or, you know, a sex registry for criminals, and you essentially can type in your address and your neighborhood and you can see all the people that live around you. They have to register their address, and you can see who they are. Now, this is like a safety feature. I'm sure there's a lot of like. I'm sure those people don't like that because you essentially have to notify this registry every time you're moving and they'll put a new. They'll, you know, update your address. And, you know, the idea behind it is like a safety thing. You can't live a certain distance to, like, schools. People want to, like, check their neighborhood. If there's like a neighborhood with a ton of, you know, predators or previously, you know, predator offenders in it, then maybe they don't want to move there. So there's like that whole thing. And I imagine all of them would love to have the right to be forgotten used on Google, because then if you search their name, that wouldn't show up. So I don't know, it's kind of just. It's a tricky situation where I don't actually think you should have it, although evidently in the eu, I think there's laws passed where it's a possibility to do that. So in any case, moving on, the other people that are on that list, David Faber, he is a longtime reporter at the CNBC. Apparently he doesn't want to be talked about in ChatGPT. So he's pulled off. Jonathan Turley. He's a lawyer and a Fox News commentator. He's also a legal expert who talks a lot about the, quote, right to be forgotten. So evidently he's like, pretty prominent in that. So opening eyes, like, okay, well, he's a lawyer and he's really big on this thing. Let's forget him from our model. He probably requested it, to be honest, but they blacklisted his name. Guido scores A is on the board of Italy's Data Protection Authority. And I guess data protection, he didn't. He wanted his data protected and got removed. What's really interesting to me here is it's like these people wanted their names removed so they could be more low key. And I don't know, people wouldn't talk about them perhaps. And in the end, their names Being removed is now making everybody talk about them. So I don't know if that really works. It's kind of that. Okay, let's talk about the final one, the one that really went viral and spurred this whole thing, which was David Mayer. Who the heck is David Mayer now? I was like, whatever. It could be anyone. If you go to TechCrunch, they say that like there is no lawyer, journalist, mayor or otherwise obviously notable person by the, by that name that anyone could find. And with apologies to the many respectable David Mayers out there, okay? That's what TechCrunch says. I have no idea what they're smoking. Because if you go to Google and search David Mayer, I think there's a very prominent person who is the top result showing up for that, and that is David Mayer de Rothschild, who is, you know, obviously the Rothschilds family, one of the richest families in the world who's always kind of caught up in the conspiracy theories of shadow governments and running the world and all that kind of stuff. Whether you believe it or not, that's what they're very famously known for. And a lot of people say philanthropy, but regardless, they have, you know, their wealth really isn't known. Hundreds of billions of dollars, but a very old kind of money family that's been around for a long time. Anyways, to me it's pretty interesting that you have someone like that who is able to essentially remove his name from ChatGPT. But even more interesting to me is the fact that if you go to something like TechCrunch, they literally are like backing it up, saying, you know, there's like, there's no one of note, there's no notable person that we could possibly find with that name. When you, if you literally just Google David Mayer, it's like the big side panel on the side David demayer Rothschild or David Mayer Rothschild and talks all about him and stuff. So it's kind of interesting. It feels like, I don't know, it feels like, it feels weird that there's a. It feels like Chet TechCrunch or I guess like news organizations are along with Chat GPT on saying that it's nothing. So there was an update and a statement from OpenAI about this. They said that the name David Mayer has been flagged by internal privacy tools saying, quote, there may be instances where ChatGPT does not provide certain information about people to protect their privacy. So evidently if you have enough resources or if you make a big stink or if you threaten a lawsuit, you can get your name or an Entire name blacklisted from ChatGPT. This will be interesting, I think, in the future to see, you know, if that gets abused. You can imagine if a million people requested their name be removed or threatened a lawsuit from ChatGPT and their common names, like, it would be a minefield. You could also imagine companies requesting to be removed all sorts of interesting things because at the end of the day, you know, like let's say Chase bank decided they don't ever want to be Talked about in ChatGPT. And you know, OpenAI goes and just bans the word Chase bank. But it's like people want to talk about Chase bank and so then all of a sudden it's going to get banned from the responses. And you can imagine, you know, let's say there was a company with bad reputation tried to do the same thing with Google and they, you know, they go to Reddit and say, hey, we're going to sue you unless you take down like anything that says our name. And I'm sure there's already been lawsuits about this. It just seems weird that you could threaten lawsuits to remove your name from being included in public discourse. Seems kind of like free speech is, I don't know, being infringed on. Especially if it's a public company that people might have bad, in bad experiences with or a public person, like a mayor that people might have bad experiences with. You know, you maybe you want to go write a blog about how much you hate the mayor of Sydney, Australia. And now you know, that can never be regurgitated by chatgpt because his name got banned. So anyways, it's just very interesting thing to think about. I'm not sure, you know, what, what is censorship? What's free speech in here? Overall, I'm definitely, I think probably on the, on the side of like not being able to just ban names or companies or things and get them just blanket removed. Especially when it's like a person's name because there can be multiple people with the same name so that whoever got it removed kind of like claims it. I don't know, that seems super weird to me. In any case, this is a fascinating time. It's gonna be interesting to see how other AI models play with this. Thanks so much for tuning into the podcast. If you got any insights or learned anything new, I would really, really appreciate a review on the podcast. Helps me out a ton. And if you've ever wanted to start your own podcast, make sure to check out the link in the description for my podcast course that covers how I was able to get over 4 million downloads on my podcast. There's a Black Friday discount of 50% off right now, so make sure to get that if you want to start a podcast.