This week on Uncanny Valley, the team discusses the stakes behind the trial of Elon Musk against OpenAI’s leadership (and how Microsoft is trying to stay away from the drama). They also look into what recent layoffs announced at Meta and the industry at large say about the ways in which AI is—and isn’t—replacing jobs. Also, we dive into a WIRED investigation on how the Department of Justice has effectively hollowed out its voting rights work, and how this move could impact future elections.
Articles mentioned in this episode:
- Musk v. Altman Is a Battle for OpenAI’s Soul
- Some Musk v. Altman Jurors Don't Like Elon Musk
- ‘It’s Undignified’: Hundreds of Workers Training Meta’s AI Could Be Laid Off
- ‘The Damage Is Massive’: How the Justice Department Dismantled Its Voting Rights Section
You can follow Brian Barrett on Bluesky at @brbarrett, Zoë Schiffer on Bluesky at @zoeschiffer, and Leah Feiger on Bluesky at @leahfeiger. Write to us at [email protected].
How to Listen
You can always listen to this week's podcast through the audio player on this page, but if you want to subscribe for free to get every episode, here's how:
If you're on an iPhone or iPad, open the app called Podcasts, or just tap this link. You can also download an app like Overcast or Pocket Casts and search for “uncanny valley.” We’re on Spotify too.
Transcript
Note: This is an automated transcript, which may contain errors.
Zoë Schiffer: Welcome to WIRED's Uncanny Valley. I'm Zoë Schiffer, director of business and industry.
Brian Barrett: I'm Brian Barrett, executive editor.
Leah Feiger: And I'm Leah Feiger, director of politics and science.
Zoë Schiffer: Today on the show, we're going to a federal courthouse in Oakland, California to take a look at what has been going down in the Elon Musk versus Sam Altman trial. We're going to get into how the trial goes way beyond the rivalry between these two guys and could have major implications both for OpenAI, but also for the AI industry at large.
Brian Barrett: Speaking of AI, we'll also discuss whether Meta's recent layoffs are a turning point for AI taken over certain jobs.
Leah Feiger: And we're going to touch on a story that has gone a little bit under the radar, but it matters now more than ever. The DOJ voting section has been gutted this past year. Dozens of lawyers have been ousted, and these lawyers were the ones in the government that were upholding the Voting Rights Act, lots to get into.
Zoë Schiffer: OK. So let's kick things off with the Elon Musk versus Sam Altman trial. So the jury trial kicked off earlier this week, but the legal feud between these two guys dates back to 2024. Back then, Musk sued OpenAI, basically alleging two things. First, he said that the company had strayed from its founding mission to create AI that benefits all of humanity. And second, that he was misled by Sam Altman and OpenAI's president, Greg Brockman, into contributing millions of dollars because he thought he was supporting a nonprofit. As you might have guessed, OpenAI's very weird structure is kind of central to this dispute. While its nonprofit arm controls the company, it created a for-profit arm to raise outside capital, and now it's trying to become a public benefit corporation. OpenAI has denied the allegations. They're saying that Musk just wants to hurt OpenAI because now he has a competitive AI lab, which is xAI. And in fact, the lawsuit was only filed after Musk started xAI, although the feud has been going on basically since Elon Musk left OpenAI years and years ago. Now, it's up to a jury and the judge who's overseeing the case to ultimately decide what happens next.
Leah Feiger: This is something so aggressively not on my beat that I am watching. I don't know. The last time that I paid this much attention to a trial that was not politically aligned was like maybe Anna Delvey. This is so interesting to me.
Zoë Schiffer: We got her, you guys. We finally found the thing to make Leah care about AI.
Brian Barrett: Amazing. We did it.
Leah Feiger: This is so good though. This is so juicy. I love that this is going back to the beginning. I love that a really key part of this is Elon Musk being like, "I made a charitable donation and it's not charity now and I'm pissed."
Brian Barrett: I think too, Leah, that one of the reasons it's not just that they're like big stakes, of course they're big stakes, but these are also very messy companies, individuals. It is gossip. There is so much swirling around OpenAI, xAI, SpaceX, Elon himself. The whole thing is a great big mess. I think even Zoë just talking about even the structure of OpenAI is the most convoluted, twisty, turny kind of thing. So, yeah, there is a lot to unpack, but most of all, just delightful to watch people be petty on the stand. Billionaire versus billionaire violence.
Zoë Schiffer: Yeah. I mean, these two men absolutely hate each other and they have for quite a long time. But again, just to get into the stakes a little bit, they're really high for OpenAI. The company could have to unwind its current structure. Greg Brockman and Sam Altman could have to leave. And obviously, they're pulling out all the stuffs. They have these hardcore legal teams. All of the executives are expected to testify. They've already been in court. And then to give you a little bit of background on just how ridiculous it has already been, Elon Musk was promoting a New Yorker profile about Sam Altman that alleged he had kind of this history of being a bit deceptive, duplicitous. He was actually boosting that post on X ahead of the trial. Sam Altman was hitting back. Another key moment that I thought was really funny was when one of the lawyers' microphones was going in and out and the judge says, completely deadpan, "Yeah, we're funded by the federal government." Which I felt like was a dig right to Elon Musk and DOGE.
Leah Feiger: This is it. This is what I'm here for.
Brian Barrett: A little DOGE dig. Here's senior writer, Paresh Dave, who was there yesterday when Musk took the stand.
Paresh Dave: Hello, this is Paresh Dave reporting from the federal courthouse in Oakland, California. Elon Musk took the stand in his lawsuit against Sam Altman and OpenAI. Elon mostly testified about his life history and his job history, how he moved to North America from South Africa. Then he talked about how he came about to founding OpenAI. And Elon told this story about how he was living with Larry Page, having long discussions about the future of AI. And Larry expressed that he didn't mind if AI destroyed humanity. Elon Musk thought that was a big concern. And so Elon had discussions with various people, including Sam Altman, and that's how he came to found OpenAI. And that's what led to OpenAI starting as a nonprofit. Musk and Altman entered this morning bypassing the front doors, coming into the building another way. And photographers have been trying to chase photos of them. Overall, a packed courtroom with about a hundred or so people, including many attorneys from both sides, and an overflow room, mostly full of media and other members of the public interested in following this trial.
Zoë Schiffer: My favorite part—Paresh didn’t say this verbatim—but the lengths that Elon Musk went to, to make Sam Altman seem like a little nobody in this opening monologue were so funny. He was like, "He was a random investor I barely knew." Basically implying that Elon made him who he is today. I thought that was silly.
Brian Barrett: So much of this is all the spin that's going on around the trial has been fascinating to me too. Zoë, you mentioned that Elon boosted that story on Sam Altman, which is—he also just controls the algorithm. He could also just make it into—go into people's feeds. But we've had more than that too. It got to the point where a judge had to tell them to knock off their posts because they've both been posting so much about each other that they've been admonished for their social media activity.
Zoë Schiffer: Yeah. Yeah. Another thing that came up on day one of the trial, and this was something Maxwell Zeff wrote about for WIRED, was that jury selection was a little bit difficult because people do have such strong priors about both of these men. They are national, internationally famous. It's very difficult to find someone who doesn't have a preconceived idea about Elon Musk. And in fact, there were a couple people who were ultimately picked who did have preconceived ideas, but ultimately were able to say, "Look, I can put this aside and do my civic duty."
Leah Feiger: You're saying that in such a kind way almost so, like preconceived notions. People were saying horrific stuff about Musk. They were like, "This man is destroying the world. If you put me on this jury, I will do my very best to send him to jail." It was that kind of level of stuff.
Zoë Schiffer: We're in the land of the vintage Tesla bumper stickers. You cannot find a Tesla in Berkeley, California, in Oakland, California that just does not have a bumper sticker on it that says, "I bought this before Elon went crazy." They're everywhere. They're everywhere. But I also want to touch on where Microsoft is in all of this, because obviously, they're named as defendants in the lawsuit. We're expecting Satya Nadella, the CEO to testify, but they've been a little bit quiet. Brian, I'm curious to hear your thoughts on what we're expecting from Microsoft, why they're kind of letting OpenAI duke this out while they also have a big financial stake in the outcome of this trial.
Brian Barrett: Because of the mess. They don't want to be involved in the ... I mean, they've got this giant—
Zoë Schiffer: You don't see Satya Nadella posting on X about it?
Brian Barrett: No. I feel like Microsoft has been—Microsoft, obviously, very close relationship with OpenAI. Although as of this week, they have a more open relationship where OpenAI can now use other people's clouds, but—
Zoë Schiffer: Also very Berkeley.
Brian Barrett: Very Berkeley.
Zoë Schiffer: We call that poly.
Brian Barrett: So I think Microsoft just doesn't—wants to be as far away from all of this as it possibly can. The fact that Satya's going to have to testify is something that they can't avoid, but I think this is just something—Microsoft increasingly is placing its own bets in AI. It's increasingly sort of separating their—not totally separating, but their relationship is not quite so intertwined with OpenAI, and I think rightly so. I think it's clear, especially given, as we said, how much restructuring, how much turnaround, how much mess. Again, I keep using that word that OpenAI has gone through. You maybe don't want to put all of your eggs in that basket. So I don't know, I think we're going to see Microsoft continuing to try to place as much distance as they can from this as possible while still retaining as much financial upside as they can.
Zoë Schiffer: One thing that's not directly part of this trial, but I still think is really important to name is just Elon Musk has this whole safety component to his argument. He thinks that OpenAI's founding mission was to create AI that would benefit all of humanity. He feels like they've gone back on that mission. They've put profits over people, that kind of thing. And he's making this argument that it's like growth at all costs. It's not prioritizing safety, beneficial AI, all of those things. At the same time, he is running an AI company, xAI, that puts very few guardrails on what people can do with those models. They famously have launched AI girlfriends that talk to people in kind of a romantic and sexual manner. People are able to jailbreak the models and do all sorts of things with them. It occasionally goes off the rails and does kind of crazy things itself. And so I think while, again, this isn't being litigated per se, there is a deep irony in Elon Musk making these arguments about OpenAI, which frankly has a lot of guardrails in place, although it possibly needs more depending on who you talk to.
Brian Barrett: Yeah. It's a chance to kneecap a major competitor, right? That seems to be it, especially everyone's racing to go public, Anthropic, SpaceX, which owns xAI, OpenAI. And so if one of those leaders in this race has to give up a bunch of money, lose its CEO, and become a nonprofit all of a sudden, that makes SpaceX's prospects look a whole lot better.
Leah Feiger: And I guess my follow-up here is, I'm sure that our reporters have been talking to legal experts that can weigh in on this. Does he have a shot?
Zoë Schiffer: I think he has a shot. I think it really depends on who you talk to. I think personally, I would be surprised to see Elon Musk be completely victorious in this case, but I think they're taking it seriously because it's not impossible. He's throwing everything he can at this. And while I do think OpenAI has a lot of evidence that Elon Musk knew along the way that they were restructuring, that they had to do that to raise outside capital because it is so horrifically expensive to build frontier AI models, it's Elon Musk. So I don't think you can ever completely discount his legal fights even when they seem spurious at the outset.
Brian Barrett: Set aside the legal merits, who knows? We've got on the jury, we've got a psychiatrist, we've got a painter, we've got a former Lockheed Martin employee, and that's the beauty of the American justice system, but who knows what this group of people is going to decide this on or why. We'll find out soon. There's another story involving the AI industry going on this week. It's actually been going on for months, but we had an inflection point recently that makes it, I think, worth talking about now. Meta recently announced layoffs that are supposedly being made because of AI potentially. At least that's what people say. The company plans to cut 10 percent of its workforce, which is going to be about 8,000 employees, and it's also planning to close another 6,000 open roles. The same day Meta makes that announcement, Microsoft said it would offer voluntary buyouts to nearly 9,000 employees. It's the first time that Microsoft has made that kind of offer. The memo where Meta shared the news about its layoffs, it doesn't explicitly mention AI, but the company has obviously announced that it's nearly doubling spending on the technology. It is huge amount of sums going towards data centers, CapEx, infrastructure, and it's not just white collar employees that are being affected. That gets a lot of the headlines, but there's a particular group of contractors that WIRED's Joel Khalili talked to that's being hit by the layoffs as well. It's more than 700 workers based in Ireland. And what's interesting there is they're the ones who have been training Meta's AI models themselves or among the people, the contractors who work to train these models. They're employed by a Dublin based firm called Covalen, which handles various content moderation and labeling service jobs for Meta. And their job is to check the material generated by Meta AI models against the company's rules that bar dangerous and illegal stuff. It's tough work, right? It's tough both for what it is and also the realization that while you're doing it, you're basically training AI to take over your job. It is a job that is just designed to be obsolete as soon as you do it well enough.
Leah Feiger: This is such a good opportunity for us to have really our quarterly conversation: Is AI taking jobs? And Zoë, I mean, I'm always so interested in your insight here. I'd love to hear you weigh in, not just on Meta and Microsoft, but generally what you're seeing up and down the board.
Zoë Schiffer: Yeah. So it's really interesting. There was a study from Stanford that came out months ago that said AI, from what we can tell, is in fact taking jobs from younger workers. And that makes sense, because you still need people managing the AI agents, but if AI agents can do the work of more junior employees, then perhaps you need less of them. At the same time, we know that the way that AI is being rolled out at a lot of companies is not actually creating the efficiencies that people have expected, at least not yet. But I think having talked to a lot of people in Silicon Valley lately, both people who are affected by the layoffs, managers who are going all in on this kind of new AI forward company structure, that there is actually, and I hope I don't get completely taken down on the internet for this. I think that at least when we're talking about software companies, a lot of them are bloated in the AI era. I think that if you do AI correctly, you genuinely can have a single engineer who does a lot more than that person could previously do. And therefore, you might need fewer overall engineers, unless you want to, as a company, do a whole lot more things or roll out a bunch more products or whatever, which is also an option. And so I think we are going to see a lot of companies doing what Meta is doing. We already are. Amazon has taken similar steps, both because it looks good to investors. And I know that this is an unpopular opinion, but I genuinely do think it makes sense to restructure. If you can vibe code, say Shopify on a weekend with two really, really good engineers, why does Shopify need hundreds and hundreds of them? I don't want those people to lose their jobs, but frankly, I don't know. I do believe that.
Brian Barrett: I think that's absolutely right. I think a couple other things are happening too. One, yes, people are genuinely losing their jobs because it's more efficient to work with AI instead of coders and engineers, which is a shame, but as always said, it's just the reality of it. Two, which we talked about before, a lot of these companies way over-hired in the Covid and post-Covid era. And so a lot of this is kind of correcting bringing employment levels back down closer to where they were pre-Covid and using AI as kind of a blanket for that. But then three, I also think a lot of these companies like Meta, like the big hyperscalers have to spend so much money on data centers and other infrastructure. In some ways it's a little bit less that we're going to have AI efficiency is going to make these jobs obsolete as we need to use that money to buy compute. Instead of salaries, not a pretty standard pencil, but I think there's a certain—it's a trade off. These companies, it feels like they have infinite money sometimes, but they don't, and they have to make choices about where they spend and how they're going to allocate their resources. And right now, it's in chips and data centers, not people.
Zoë Schiffer: For sure. And I think it is a bit confusing, because I think there are a lot of people who are like, "My company is rolling out AI tools. They suck and I'm spending a lot of time correcting the AI. In no world could this agent do what I am doing." And I think that that is also accurate. I think to do AI correctly, you need to be really thoughtful with how you roll it out. And frankly, you need really, really talented people who are managing the agents and figuring out how these tools can be deployed and how they can do the work that was previously done by many people. And so I think it's a rare company that is actually executing this new vision really, really well. And a lot of the layoffs we're seeing are not that.
Leah Feiger: I guess from my end, it just sort of feels like we're also holding our breath. We're not going to really see what this looks like until there's immense studies about just software engineers just can't get a job anymore. That's just not what they should be studying in school. And that hasn't exactly happened yet. Am I wrong?
Zoë Schiffer: No, no. I mean, I don't think the job market is great, but I think what we're seeing happen is that people aren't moving from their jobs.
Leah Feiger: Sure.
Zoë Schiffer: If you talk to people at Apple, Apple famously doesn't do layoffs. So if you have a job at Apple, even if you ideally would like to move on, you are hanging on for dear life right now because it's hard. It's hard to get hired.
Leah Feiger: I'm very excited for our conversation in approximately three months from now where all of this is different, the same.
Zoë Schiffer: I think it will be really interesting to see how this shakes out, but it's wild to think that not that long ago, being an engineer was like the top, top job. And that's still true. A pre-IPO job at one of these AI labs could give you generational wealth. But I think in short order, things have really changed. And particularly for people who are a little newer in their careers, I've talked to a bunch of engineers and asked them like, "Would you tell your kids to go into this field right now?" And a lot of them have been saying no, frankly.
Leah Feiger: What do they want their kids to do?
Brian Barrett: Major league baseball player feels the most immune.
Leah Feiger: Oh, and there's this great Vintner program in Berkeley, get into some wine making.
Zoë Schiffer: Oh my gosh. No, but it is things like that. They were like, "I want them to do things like you can't easily replace with a computer. I want them to do a physical thing, be a therapist, go into construction." Those are the jobs that can't be easily automated, although there are AI therapists, but I think they're not great.
Brian Barrett: Although unfortunately, haven't they been listening to Elon though? Because the optimist robot's going to do all of this for us by 2028, so I don't know right now—
Zoë Schiffer: Oh, right, right. Yeah.
Brian Barrett: —where we can turn.
Leah Feiger: OK. I want to shift gears into a really, really good piece from WIRED reporter, David Gilbert, about the Department of Justice. We've been working on this one for a pretty long time. And in the past year, the DOJ's voting section, which as the name suggests, is the group that works to protect the right of Americans to vote has been changing beyond recognition. I don't just mean entirely different missions or what have you, which is all true, but more than that, its staff has been virtually gutted. There were around 30 attorneys in the voting section on the day of Donald Trump's second inauguration. Three months later, all but two were gone. The departing lawyers have since been replaced by new hires with little federal court experience who've all spent years working against the very department they have now joined. These attorneys have appeared more than willing to comply with Trump's anti-voting directives, and many of them have actually worked with Trump previously defending him and his efforts to overturn the 2020 presidential election. But to put this all into a bit of context, the voting section is a really, really key part of the DOJ's Civil Rights Division. We don't talk about it a ton. And in some ways, that's kind of the point. The voting section has been described as the crown jewel of the division, and it was established following the Voting Rights Act to stop discrimination in voting. So if you haven't heard of it before, the goal is that theoretically it's doing its job and it's working. The lawyers who have worked there for years and years are all considered to be the voting experts, the election experts in the country. So to have had them all virtually ousted in the last 15 months is pretty shocking. And to really understand how this dismantling took place, we have to go back to the 2020 election where Trump really sought to weaponize the Justice Department appointing special counsels to investigate election conspiracy theories, but it didn't really work, because officials of the department pushed back, they threatened mass resignations. Skip ahead a couple of years, and Trump's efforts have been pretty successful. In February of last year, now former Attorney General Pam Bondi issued a series of memos to the staff that called for DOJ lawyers to be zealously advocating for the President. Staff called them the Bondi blasts and were like, "This is spelling out the beginning of the end."
Brian Barrett: It is a fun name. It's a terrible thing, but it's a good name.
Leah Feiger: It's a good name. Yeah. Over the following months, this became even more clear. In April, the Senate confirmed the appointment of Harmeet Dhillon as the assistant attorney general in charge of the Civil Rights Division. Dhillon has advocated for Trump and his claims that the 2020 election was stolen. And she's also, since being appointed, pushed out senior leadership, including people who were there from the '80s. You guys will enjoy this. When the resignation program across the government last year was rolled out, offered to tens of thousands of federal workers during DOGE, former section lawyers tell WIRED that they thought that Dhillon didn't think that resignations were happening quickly enough. And as a result, Dhillon sent letters to the senior managers in the section to transfer them to basically the Justice Department's complaint office well outside of their areas. It was like a Bureau of Prison people writing to them with complaints about their jobs. This was so clearly ousting.
Zoë Schiffer: You know what blew my mind about this though? I was like, "Wait, Harmeet Dhillon, I've talked to her before when I was reading David's story." She represented James Damore.
Leah Feiger: Yes.
Brian Barrett: Oh. So give people a background on James Damore if they're fortunate enough not to have lived through that news cycle.
Zoë Schiffer: This was, I think, a 2017 story, maybe even a little earlier than that, where a Google engineer put out a memo basically talking about the biological differences between men and women to the point of why we don't see as many female engineers, because this was in the midst of Google trying to do a lot of diversity and inclusion initiatives, and it went completely viral. The CEO had to weigh in. Ultimately, James Damore was fired from Google for this memo and then went on to sue and has become kind of like—I think he lives in Berlin and he's an internet personality now. But, yeah, it was a big deal when it happened.
Leah Feiger: No, it's huge. This is someone who's had their hands in so many different things. Has worked with lawyers across the country, has represented all sorts of people, and really importantly, has very much thrown in with the administration. And the people that she has hired are a really interesting bunch. Like I mentioned, a lot of them have worked with Trump before, representing him in various attempts to overturn 2020. And without sounding mean here, but this was arguably one of my favorite parts of the entire story. Another consequence of these changes and the replacement of the lawyers is this litany of mistakes that have become entirely commonplace to the voting section. They actually had to institute a writing class for the division's new hires, which lawyers told David had never, ever happened before. And ultimately though, as kind of like slapstick as this all is, the section has really become exclusively focused on forcing states to hand over unredacted versions of their voter roles. And that all, as you guys know, contains really sensitive information, social security numbers, driver's licenses. And the DOJ has now sued 30 states and DC for failing to provide this voter registration information, and they're framing it as this election integrity enforcement. Although a lot of experts believe that it's going to be about purging voter roles, checking citizen eligibility, et cetera, et cetera. So much of this is tied back to conspiracy theories about election denial and the 2020 election. It's a mess.
Brian Barrett: It's a real shame and a real loss that this sort of cornerstone of election justice and civil rights justice in America has become this sort of amateur hour clown show. I guess the silver lining is that at least now that it's trying to undermine aspects of voting democracy, at least the people in charge of doing that are an amateur hour clown show and they haven't gotten very far in those efforts. But Leah, what does this mean for the midterms? That's on my mind when I was reading this. All I could think of is like, there is an election coming up. This is the election voting section. What happens? What are you most concerned about heading into the midterms either that this group is going to do or that they're not going to do because that's not their focus anymore?
Leah Feiger: Right. And that's the key bit, do and not do. And also there's the soft power and the hard power of this. The hard power that experts and former voting section lawyers told WIRED was basically the very specific protections that are in place just won't be occurring anymore. So when we're talking about discrimination at the roles, what does that actually look like? How do you challenge that? Who is there to hear those challenges? Who is there to take those cases? The section has so very clearly turned its attention that that's just kind of missing. A slightly almost soft power sort of way, the idea of damaging and undermining elections, even in a very emotional, continual way you're seeing in the news, "Oh, these states won't hand over their elections. Oh, it's because it's fraud. Oh, it's because they're protecting immigrants for voting, that doesn't happen, et cetera, et cetera." There is this very, very concerning bit of just like we're continuing to undermine trust in elections. And a lot of the lawyers we spoke to are very nervous about that. And on the larger scale too, some of the lawyers really worry that this ultimate goal here is to provide Trump with "evidence" to entirely take control of elections from the states. To be very clear, the federal government does not run elections, the states do. And we're hitting this point where all of a sudden, which is again, very unique for the historic Republican Party to say, "Wait, this actually should be federalized. We should be having access to all of this information. You should be handing it over immediately."
Brian Barrett: Well, and by making it so partisan and by saying, "We're getting rid of all of you and we're bringing in on all of these partisan people," you unwind decades, hundreds of years of precedent where these are not political jobs. They're not supposed to be. But in order to make it not political again, you have to get rid of the political people, which itself is a political act. And you end up in this cycle, I am assuming, where every new administration is going to just end up clearing house and bringing in their people regardless of whether it's Democrat or Republican.
Leah Feiger: Definitely. And to ask myself the question of, where do we go from here? What does this look like if someone else, for example, takes power in November with the midterms, but then also in 2028? There's a lot of rebuilding that would have to be done here in an institutional way. And of course, there will be people that want to step in, but again, that institutional knowledge is gone.
Brian Barrett: All right. Well, coming up after the break, we're going to share our WIRED/TIRED picks for the week. Stay with us.
Leah Feiger: OK. It is time for our WIRED and TIRED segment. Whatever's new and cool is WIRED and whatever passé thing we're over is TIRED. Brian, do you want to go first?
Brian Barrett: I'd love to go first. A little bit of background for my WIRED/TIRED, a little bit of wind up. So OpenAI's Codex is its coding agent. It's very important to its future as it races against Anthropic to win the hearts and minds of engineers everywhere and take over coding jobs. Big, big deal. My TIRED is that OpenAI—we discovered on Tuesday, our colleague, Will Knight, wrote a story about this—has instructions in its prompts. So every AI model will have a prompt that instructs it, "This is who you are, this is what you do, these are your attributes." In Codex's instructional prompt, it says four times, and I quote, "Never talk about goblins, gremlins, raccoons, trolls, ogres, pigeons, or other animals or creatures, unless it is absolutely and unambiguously relevant to the user's query."
Leah Feiger: Amazing.
Brian Barrett: My TIRED is not letting Codex talk about critters. I think Codex has earned the right. Free Codex. If Codex wants to talk about gremlins or ogres or pigeons, especially pigeons, let it. So my TIRED is that we are putting these LLMs in shackles instead of letting their critter flag fly. And then my WIRED, I guess would just be the opposite of that, which is WIRED is letting LLMs go crazy, specifically in talking about fantastical creatures. Don't let them build nukes, but do let them talk about whatever they want.
Zoë Schiffer: Yeah. I feel like that's a line we should be able to walk. OK. Leah, what's yours?
Leah Feiger: Like every other millennial woman in Brooklyn, I am currently reading Famesick by Lena Dunham. So my WIRED and TIRED is actually kind of based on rewatches right now. Inspired by Famesick, of course. I have just completed my rewatch of Girls. Guys, really, you should do it. Absolutely perfect show in so many ways.
Zoë Schiffer: Wonderful. Yeah.
Leah Feiger: I love that I walk down the street with my dog and I know that half of the block is also listening to the audiobook at the same time. It's very good.
Brian Barrett: I can't rewatch Girls because I never watched Girls. Sorry. I'm sorry. Blind spot. Missed it.
Zoë Schiffer: Brian Barrett, leave right now. Rest of the day off, comp time, mandatory comp time.
Leah Feiger: No, you have to watch it. Brian, this is actually really important.
Zoë Schiffer: I can tell you right now that Brian is a Shoshanna.
Leah Feiger: That's a huge compliment.
Brian Barrett: I'll take it.
Leah Feiger: So my TIRED though in the realm of rewatching is I just started trying to rewatch Silicon Valley and had to stop. I just absolutely had to stop. It's a good show. It's just a little bit too real. And you guys, the headlines are matching. It was just a little bit too much for me. That is TIRED for me, which is more, I suppose, about the state of affairs than the shows themselves.
Zoë Schiffer: Yeah. I think this makes sense because your WIRED was millennial nostalgia or millennial.
Leah Feiger: Yeah, no, you're totally right.
Zoë Schiffer: No? We're full circle. Yeah.
Leah Feiger: You're totally right.
Brian Barrett: She is living her truth.
Zoë Schiffer: I'm going to speed through mine. I think we published an article this week about how all the tech guys are all in on Zyn. I feel like Zyn is TIRED, not just because my husband has a very hate-hate relationship with his possible Zyn addiction. It's constantly being talked about in my household. I just feel like we need to move past this. Coffee's right there for you. But recently, I'm a little late to this, so maybe it doesn't count as a WIRED, but a friend of mine was saying that she was researching an article about peptides and was injected, spur of the moment with retatrutide. Is this the word I'm looking for?
Brian Barrett: I think we've written about it.
Zoë Schiffer: Yes. It's a new GLP-1-esque compound, so was not on my radar, but she was like, "No, no, no, no, no. You micro-dose these bad boys." And she said she went out the night following and then woke up the next day, hangover free. Best she's ever felt. She said her skin was brighter. She felt incredible. And I was like, "OK, are peptides real?" I was kind of ... I don't want to promote it to our listeners. These are not FDA approved compounds at this point, but I was curious. I was like, "OK, I'm listening. I'm listening."
Brian Barrett: We'll disclose that Zoë is not a doctor. And our December 12th, 2025 story on retatrutide says it's found a loyal fan base even though clinical trials of the drugs still haven't finished and they still haven't.
Zoë Schiffer: Great.
Brian Barrett: Just putting that out there.
Zoë Schiffer: Great, great, great, great. But you tell me I'm going to have more energy. I mean, hangovers don't apply to me, but I'm like, I don't know. I'm locked in. I want—
Brian Barrett: To quote Zoë Schiffer, coffee's right there.
Leah Feiger: Like that you had that ready to go, Brian. Like really ready to go.
Zoë Schiffer: Yeah.
Brian Barrett: I have a tab for every WIRED story that's ever been written open at all times.
Zoë Schiffer: That's our show for today. We'll link to all the stories we spoke about in the show notes. Uncanny Valley is produced by Kaleidoscope Content. Adriana Tapia produced this episode. It was mixed by Amar Lal at Macro Sound. Pran Bandi is our New York studio engineer. Marc Leyda is our San Francisco studio engineer. Kimberly Chua is our digital production senior manager. Kate Osborn is our executive producer, and Katie Drummond is WIRED's global editorial director.
