WIRED Uncanny Valley Podcast Episode Transcript (Rewritten)
This week on Uncanny Valley, WIRED’s tech and politics podcast, hosts Zoë Schiffer, Brian Barrett, and Leah Feiger break down how disinformation campaigns and clashes between the AI industry and the U.S. Department of Defense have moved straight to the heart of the escalating U.S.-Iran conflict. The team also digs into growing insider trading allegations and ethical controversies roiling prediction platforms like Polymarket and Kalshi, unpack how Paramount outmaneuvered Netflix to land a historic merger deal with Warner Bros. Discovery, and share forward-looking predictions in a brand-new segment.
Articles referenced in this episode:
X Is Drowning in Disinformation After U.S. and Israeli Strikes on Iran
How Journalists Are Reporting From Iran Amid Total Internet Disruptions
Anthropic Fights Back After U.S. Military Labels It a ‘Supply Chain Risk’
A Former Top Trump Official Takes on Prediction Markets
Everything Larry and David Ellison Will Control If Paramount’s Warner Bros. Deal Closes
You can follow the hosts on Bluesky: Brian Barrett @brbarrett, Zoë Schiffer @zoeschiffer, and Leah Feiger @leahfeiger. Reach the team by email at [email protected].
How to Tune In
You can stream this week’s episode directly via the audio player on this page. To subscribe for free and get every new episode delivered automatically:
On iPhone or iPad, open the Apple Podcasts app, or tap this link to subscribe.
You can also find Uncanny Valley on Spotify, or search for the show in third-party apps like Overcast or Pocket Casts.
Transcript
Note: This transcript was adapted from an automated version and may contain minor errors.
Brian Barrett: Hey everyone, it’s Brian. Over the past few weeks, Leah, Zoë, and I have loved stepping in as your new hosts, and we want to hear from you. If you enjoy the show, please take a minute to leave a review on whatever podcast app you use — it really helps new listeners find us. You can always send questions or feedback to [email protected], too. Thanks for listening, and let’s get into the show.
I’m really excited to be recording this in-person in New York with Leah Feiger today. Zoë, you’re still joining us remotely.
Zoë Schiffer: That’s right! Brian, should we tell everyone why you’re in New York?
Brian Barrett: Absolutely we should. Last night, Leah was honored as Journalist of the Year at the Front Page Awards.
Zoë Schiffer: She even got a physical trophy and everything!
Brian Barrett: She did. And Zoë, you didn’t get to attend, but I did, so I have extra context. Our editor-in-chief Katie Drummond introduced Leah, they put together a video highlighting her incredible work from last year, and she gave such a lovely acceptance speech.
Zoë Schiffer: That makes me so happy!
Leah Feiger: It was such a nice night. To be clear, though, this award is really for the entire WIRED team. I didn’t even mention myself once in my acceptance video — this is all WIRED’s win.
Brian Barrett: C’mon, this is your award, and you earned it.
Zoë Schiffer: Welcome to WIRED’s Uncanny Valley. I’m Zoë Schiffer, WIRED’s director of business and industry coverage.
Brian Barrett: I’m Brian Barrett, executive editor.
Leah Feiger: And I’m Leah Feiger, senior politics editor.
Zoë Schiffer: This week, we’re unpacking the rapidly escalating conflict in the Middle East, starting with how the AI industry has increasingly aligned itself with the U.S. Department of Defense. We’ll also break down the latest controversies hitting prediction markets, and walk through what the proposed Paramount-Warner Bros. merger means for media and journalism.
Leah Feiger: Let’s start with Iran. Things have moved nonstop since the U.S. and Israel launched a coordinated military strike on Iran Saturday, and Iran has responded with attacks on U.S. bases across the Gulf. Escalation has happened incredibly fast.
[Archival audio]: Iran Supreme Leader Ayatollah Ali Khamenei has been killed in today's joint attack by the US and Israel.
[Archival audio]: Iranian officials say airstrikes hit an elementary school Saturday, killing more than 160 people, mostly children.
[Archival audio]: US embassies across the region are now telling Americans to shelter in place—
Leah Feiger: We all worked through the weekend on this story, and I was still stunned by how quickly disinformation became a core part of this conflict. WIRED reviewed hundreds of posts on X, many of which racked up millions of views, pushing false or misleading claims about where strikes happened and how large they were. Our colleague David Gilbert reported on dozens of specific examples, and the range of bad content was wild: AI-generated fakes, video game clips passed off as real combat footage, even posts misidentifying which country was being attacked. To me, it’s a mix of rampant disinformation infrastructure and just general chaos around a fast-moving breaking story.
Brian Barrett: Yeah. For me, the disinformation itself is less surprising than how little urgency there is to fix it — though I guess I shouldn’t be shocked anymore. Every major breaking event has the same pattern: the same recycled old video game clips getting passed off as new footage, over and over.
Leah Feiger: It basically writes itself at this point.
Brian Barrett: It really does. And on top of that, X gutted almost its entire public safety team. They rely on Community Notes to flag bad posts now, but by the time a Community Note gets added, that post has already been seen by millions of people. And the note lives below the original post anyway — most people see the fake content before they ever see the correction, so it barely slows down the spread.
Zoë Schiffer: This is the end result of years of intentional product and policy choices at X. This is what you get when you make the platform hostile to journalists, lay off almost your entire fact-checking and content moderation teams, and offload fact-checking to Community Notes. Community Notes work well for small, low-stakes stuff, but they’re completely useless during fast-moving breaking news. X also pays creators for traffic, which incentivizes people to post hot takes as fast as possible, no matter if they’re actually true. So we shouldn’t keep acting shocked when this happens — we know exactly how we got here.
Leah Feiger: That’s 100% right. WIRED has had incredible coverage of how Iranian journalists, activists, and ordinary citizens are fighting to get footage and information out of the country, right? Which makes you stop and ask: who is this disinformation even for? It’s just meant to sow more chaos and muddy the waters, right? Whatever the goal, X is an absolute cesspool right now. It’s already nearly impossible to trust any numbers or facts coming out of the region, between the disinformation and the near-total internet blackout inside Iran. It’s one mess on top of another.
Brian Barrett: A lot of this disinformation is driven by blue-check accounts, which can monetize their content on X now. So a lot of it really is just for clicks — that’s the ugly truth, even though the phrase is usually thrown at journalists. And to your point about internet access: only 4% of Iranians have active internet connectivity right now. So almost the entire global narrative about what’s happening in Iran is being shaped off-platform from journalists who left X, but politicians are still there. We’ve already seen multiple high-profile cases where legitimate politicians commented on fake stories as if they were real, and that shapes public opinion. Public opinion matters a lot right now, because this is a war that was never authorized by Congress, and it could easily spiral into a much larger regional conflict.
Leah Feiger: It doesn’t look like this is ending anytime soon, if Defense Secretary Pete Hegseth’s recent comments are to be believed. CNN has reported that more than a thousand people have been killed in fighting across the Middle East from these strikes, including multiple U.S. service members. This touches every core area WIRED covers, too: trade, energy, data centers — all of it is going to be upended by this. I don’t see us stopping talking about Iran anytime soon.
Brian Barrett: It’s interesting to look at all the downstream ripple effects. Our colleague Molly Taft, who covers climate, wrote that oil and gas prices have spiked already. The Strait of Hormuz isn’t officially closed, but it’s effectively closed after the Iranian military warned vessels away, and that has effects all down the line. Fertilizer prices have skyrocketed, too — the Middle East produces a huge share of the world’s fertilizer, and this is peak fertilizer season for U.S. farmers right now, heading into spring planting. Molly has that story up on WIRED this Wednesday.
Leah Feiger: Perfect. Just what we needed. There are so many knock-on effects, I have to shout out CNN’s world team, which has been doing unbelievable minute-by-minute, hour-by-hour updates. I have it open in a corner of my screen all day, and they’re digging into every detail, not just the strike updates. For example, there’s currently an 18-hour traffic jam in Lebanon as people try to flee the country. These are real, human impacts that don’t always make the top headlines. But yeah, between fertilizer prices, all these other ripple effects, the whole world is being pulled into this downward spiral. This morning, I saw travel influencers already talking about whether it’s even safe to go to Europe this summer. This has already hit mainstream pop culture — people are sharing that old Sex and the City meme where Sarah Jessica Parker says “What do you mean World War III? What about me?” The fact that this has entered everyday public conversation this fast really drives home that this isn’t ending anytime soon.
Zoë Schiffer: Within this conversation, we have to talk about the AI angle, because all of this is unfolding right after the Department of Defense struck (and almost broke) deals with top AI companies. Last Friday, OpenAI finalized a deal with the DoD, right as Anthropic was in a head-to-head fight with the department over its terms. Anthropic wanted specific conditions written into the contract, including a ban on using its tech to surveil U.S. citizens and a ban on using it to build fully autonomous weapons. The DoD refused to add those conditions. Then, Saturday evening — the same day the Iran strikes started — Sam Altman hosted an AMA on X where he admitted the OpenAI-Pentagon deal was rushed, and that the optics were bad, to put it mildly. But he ultimately defended the decision, saying the deal was meant to de-escalate tensions between the AI industry (and specifically Anthropic) and the DoD.
Leah Feiger: Our conversation from last week feels so prescient now. We said things could get really bad, and just days later the strikes hit, and it’s so much worse than we even thought. Can we talk more about Altman’s AMA?
Zoë Schiffer: Yeah. It’s so interesting to watch this play out. We talked last week about optics and branding, and yet again, Anthropic has come out of this looking better. I’ve been thinking about this through the lens of AI talent recruitment, which sounds silly on the surface, but there’s an incredibly fierce war for top research talent between all the major AI labs right now. Anthropic keeps positioning itself as the ethical, level-headed alternative, while OpenAI keeps fumbling these high-stakes moments. Whether you believe Altman’s explanation or not, OpenAI looks sloppier here, and less anchored by a clear set of core values. That actually matters a lot when you’re competing to hire the best research talent.
Brian Barrett: The public perception angle is really interesting to me, because it feels off in a lot of ways. The narrative now is that Anthropic is the “good,” liberal-leaning AI company, right? But first of all, Anthropic’s technology was already widely used in the initial strikes on Iran, they’re just phasing it out over six months.
Leah Feiger: I think we’re giving Anthropic way too much credit for their PR win here. They were already part of this. Their big ask was just “don’t let our tech kill people without a human clicking the button first” — that’s the entire fight.
Brian Barrett: Even that conversation, isn’t the hesitation mostly because the tech isn’t reliable enough yet, not because it’s inherently wrong?
Zoë Schiffer: That’s exactly right. They’re basically saying “our tech isn’t fully reliable yet.” But what’s interesting is that Anthropic was able to get in the door for these DoD contracts in the first place because their models run on Amazon’s secure, cleared servers — Amazon was already FedRamp authorized, so Anthropic could access classified contracts that OpenAI didn’t have access to. I’ve also talked to people inside OpenAI who push back on the idea that this deal is some huge sea change for the company. Remember, OpenAI originally had a blanket ban on all military use, but they’ve been reworking that policy over time. People there say that when they were a tiny lab, a blanket ban made sense, but now they have to define what counts as off-limits. Is a blanket ban on all uses, or is it OK for the Pentagon to use ChatGPT to summarize emails, just not to build autonomous weapons? They’ve had to work through those questions. But because OpenAI originally branded itself as “building AI for the benefit of all humanity,” any tweak to that policy gets called out as hypocrisy, and they’re getting more and more sensitive to that criticism.
Leah Feiger: We talked earlier about how tech workers at Google and OpenAI have been circulating open letters calling for clearer limits on military and DoD partnerships after the Iran strikes. When it comes to recruiting and retaining talent, do these fights actually matter?
Zoë Schiffer: 100% yes. It’s a great question, but if you talk to AI researchers — not all, but a huge number of them, many of whom came from academia and are pretty idealistic — they don’t want anything to do with military use of AI. They want their company to take a firm stand against involvement in autonomous weapons. It’s almost impossible now to join a top frontier AI lab and avoid all military work, because every lab is chasing these government contracts. But there’s a large group of researchers who are disgusted by the idea that their cutting-edge AI work could be used to kill people, and they don’t want to be part of that. That absolutely matters for retention.
Leah Feiger: Do you think they’ll actually quit over it?
Zoë Schiffer: We’re already seeing people quit OpenAI to go work at Anthropic over this.
Brian Barrett: Yeah, that’s absolutely true. And you have to remember, the AI talent market is unique here. If you’re a top researcher, you’re already vested, you’ve got enough money to never work again if you want. You’re not trapped at a company you hate. There are a dozen labs that will pay you millions to come work for them, so the financial barrier to leaving is basically zero.
Zoë Schiffer: Right, to be clear: we’re not talking about people quitting and giving up a comfortable life. They’ve already made generational wealth. They can quit a job that pays them millions to go to another job that also pays them millions. The world is open to them.
Brian Barrett: I want to circle back quickly to a point Zoë made earlier about OpenAI’s early branding pushing them into this corner. I think the DoD’s interest in AI is also directly a result of the messaging AI leaders have put out there for years. If you go around saying from day one that you’re building something more powerful than nuclear weapons that will irrevocably change the entire world, of course the U.S. military is going to want a piece of it. If someone said they were building a nuclear bomb in their garage, of course Pete Hegseth and the DoD would be on the phone immediately. Whether you think AI is actually that powerful or not, that’s the message leaders have been projecting, so it shouldn’t be a surprise that it’s come to this.
Zoë Schiffer: That’s right. It’s also worth clarifying that when we talk about fully autonomous weapons, we already have fully autonomous weapons deployed right now in Ukraine — they’re small drones, they don’t run on large language models, they use simpler models. So the conversation now is just about integrating more advanced AI into these systems. It makes sense that there’s hesitation to give the Pentagon carte blanche when the tech is still less reliable, but that’s exactly what the DoD is demanding from contractors right now.
Leah Feiger: Watching the Pentagon respond to this whole fight has been wild. They very much see this technology as theirs. The whole narrative of “you built this in America, so it belongs to us” is so clear. Watching this play out publicly on X and in press conferences has been crazy. For example, Emil Michael has just been going on public rants about this. Zoë, you had a good joke about that, right?
Zoë Schiffer: Yeah, Emil Michael should turn off the read receipts on his LinkedIn messages — people can see when he views their messages from his phone lock screen.
Brian Barrett: That’s a pro tip for Emil if you’re listening, which you should be, since you’re in the middle of this fight.
Leah Feiger: Emil is the former Trump official leading the charge against Anthropic right now. He has deep ties to the tech industry, and he’s the Pentagon’s top enforcer going after anyone who pushes back against their policies. The fact that he’s making this such a public fight, and can’t even get basic opsec right, is worth noting. This isn’t just a message to Anthropic — it’s a warning to any other AI company that would dare question DoD policy.
Brian Barrett: Let’s shift to something that’s been top of mind for a lot of people lately: have you guys checked your Polymarket or Kalshi portfolios lately? How’s that going?
Leah Feiger: I haven’t put any money down on our Survivor winner bet yet, but I’m getting there.
Brian Barrett: Joking aside, prediction markets have become such a huge part of online discourse lately, and they’re increasingly tied to big global events. To stay on the Iran thread: there’s already massive betting on whether the war will escalate. Right now, one of the biggest markets on Polymarket is “Will the Iranian regime fall by June 3
