Fable 5 is Back … But Why Was It Gone at All?

Summary

Hard Fork returns from break to unpack what Kevin Roose and Casey Newton call an “unprecedented” shift in how the U.S. government treats frontier AI. After the Commerce Department abruptly used export controls to pull Anthropic’s Fable 5 and Mythos 5 off the market on June 12 — reportedly after Amazon flagged a jailbreak to the administration — the restrictions were quietly lifted weeks later, with Commerce Secretary Howard Lutnick declaring the risks addressed. Meanwhile OpenAI was pressured to hold GPT-5.6 back to a government-approved list of partners. The hosts diagnose a “de facto licensing regime” with no published rules, no transparency, and no due process — AI, in Kevin’s phrase, “regulated by vibes” — and note the irony that the tech-right figures who once decried a Biden-era “licensing scheme” have now built exactly that. They also debate the China angle: whether yanking powerful American models off the market cedes ground to Chinese open-source models like Zhipu AI’s GLM 5.2, or whether it starves the very distillation pipeline those models depend on.

The middle segment is an interview with Dr. Dana Suskind — pediatric cochlear-implant surgeon, University of Chicago professor, founder of the TMW Center, and mother of eight — about her forthcoming book “Human Raised.” Suskind argues AI should enhance parenting, never replace the human connection that wires young brains, and offers two frameworks: HOPE (a values compass) and DETECT (a six-question checklist for vetting AI products). She likens AI to processed food — a spectrum from nourishing to ultra-processed “slop” — draws a hard line against AI companions and toys for young kids on precautionary grounds, praises Norway’s move to keep gen-AI out of early schooling, and warns that a “human-raised” childhood could become a luxury good while artificial alternatives become the “cheap calories of brain nutrition.”

The episode closes with a new recurring segment, “Against All Odds,” a prediction-markets roundup. The hosts skewer a Rick Rubin-scored Polymarket ad, recount the absurd “Donking of a Truth Machine” saga (a market resolved by whoever amassed the most UMA tokens over whether a caster said “donk”), cite a Wall Street Journal finding that paid Polymarket creator videos showed fake winning bets that would really have lost $166,000, and dissect NYT reporting that Mark Zuckerberg has dispatched a team to build a Polymarket-style app internally called “Arena.”

Highlights

”They have implemented a de facto licensing regime”

“Fast forward to today, they have implemented a de facto licensing regime. And because it’s the Trump administration, there are no known rules. There’s no transparency whatsoever. We have no idea what makes a model considered safe to ship or unsafe to ship. We’re truly just in this like authoritarian limbo.” — Casey Newton, 6:56

yt-dlp --download-sections "*6:56-7:50" "https://www.youtube.com/watch?v=pqin6_VaAZM" --force-keyframes-at-cuts --merge-output-format mp4 -o "de-facto-licensing-regime.mp4"

”AI is now being regulated by vibes”

“Yes, AI is now being regulated by vibes.” — Kevin Roose, 7:50

yt-dlp --download-sections "*7:50-8:30" "https://www.youtube.com/watch?v=pqin6_VaAZM" --force-keyframes-at-cuts --merge-output-format mp4 -o "regulated-by-vibes.mp4"

”They just moved from a default yes to a default no environment”

“I think they just moved from a sort of default ‘yes’ environment where, like, the assumption was if you trained a really good model, you could release it. And now I think they’re in a default ‘no’ environment where if you build a model that is more capable than Claude 3 Opus or GPT-5.6, your assumption has to be that the government of the United States is not going to let you release it.” — Casey Newton, 11:18

yt-dlp --download-sections "*11:18-12:00" "https://www.youtube.com/watch?v=pqin6_VaAZM" --force-keyframes-at-cuts --merge-output-format mp4 -o "default-no-environment.mp4"

The HOPE framework: enhance, don’t replace

“H is human connection is irreplaceable… O is owning your imperfections… P is protecting the early years… But lastly is E. If you’re going to use it, use it to enhance. Not replace.” — Dana Suskind, 29:09

yt-dlp --download-sections "*29:09-30:23" "https://www.youtube.com/watch?v=pqin6_VaAZM" --force-keyframes-at-cuts --merge-output-format mp4 -o "hope-framework.mp4"

”Human connection will become a luxury item”

“My concern is that artificial alternatives will become you know the the cheap calories of brain nutrition and you know human raised and human connection will become a luxury item. And we just can’t allow that to happen. Um every child deserves to have parents and caregivers who love them raising them.” — Dana Suskind, 45:59

yt-dlp --download-sections "*45:59-47:02" "https://www.youtube.com/watch?v=pqin6_VaAZM" --force-keyframes-at-cuts --merge-output-format mp4 -o "human-raised-luxury.mp4"

”You can gamble on the gambling mechanism”

“You can gamble on the gambling mechanism. It’s sort of turtles all the way down.” — Kevin Roose, 55:00

yt-dlp --download-sections "*55:00-55:43" "https://www.youtube.com/watch?v=pqin6_VaAZM" --force-keyframes-at-cuts --merge-output-format mp4 -o "gamble-on-the-gambling.mp4"

Key Points

• The June 12 export-control shutdown (1:01) - Commerce issued a directive barring foreign nationals from using Fable 5 and Mythos 5, forcing Anthropic to pull the models entirely because it couldn’t verify citizenship user-by-user
• The “trusted partner” was Amazon (2:22) - Andy Jassy reportedly alerted the administration after Amazon researchers found a jailbreak; one outside expert later called the concern “not a big deal”
• Project Glasswing halted (3:58) - Anthropic’s program sharing Mythos with partners to harden critical infrastructure stopped the moment export controls landed
• Controls lifted, reversal “inevitable” (4:25) - Lutnick’s letter said Anthropic addressed the risks; Fable 5 restored globally, Mythos 5 restored to a set of U.S. organizations
• Anthropic put the industry “on blast” (5:22) - Its blog post argued GPT-5.5, Kimi K-2.7 and other models could find the same vulnerabilities
• OpenAI pressured on GPT-5.6 (6:00) - Sam Altman told staff the government asked them not to release the model broadly, extending the pattern to a second company
• “Regulation by vibes” (7:50) - No published criteria for what makes a model safe or unsafe to ship
• The hypocrisy of the tech right (9:00) - Casey questions Marc Andreessen’s “convenient fiction” of a Biden meeting; David Sacks had said the private sector should be “allowed to cook”
• Donations bought little goodwill (11:00) - Kevin notes Greg Brockman’s reported $25M didn’t smooth GPT-5.6’s release
• Default “no” environment (11:18) - An overnight structural shift for the whole AI industry
• Dean Ball’s point on technical expertise (12:25) - Government needs experts who can run safety evaluations on unreleased models; right now it lacks them
• Marc Andreessen joins the Defense Policy Board (13:21) - Putting him in regular contact with Pete Hegseth
• The China angle (13:47) - Zhipu AI’s GLM 5.2 sparks “overheated” reporting that Chinese models have caught up
• “Show me the damn numbers” (16:05) - Casey’s two-tier model of AI: the frontier vs. “everything else,” including distilled Chinese models a step behind
• Could restrictions slow China? (19:14) - Kevin floats that cutting off American models starves Chinese distillation; Casey calls it a “ham-fisted tactic”
• Anthropic’s Alibaba complaint (19:54) - A letter alleging wide-scale distilling of Claude
• The silver lining (21:40) - Government finally acknowledging AI poses real risks, even if the first attempt is “clunky, ill-advised”
• Dr. Dana Suskind on wiring young brains (27:00) - For the first time, tech can mimic the human interaction that traditionally builds children’s brains
• AI as processed food (31:00) - A spectrum from whole-wheat-bread nourishing to Doritos-and-Twinkies ultra-processed
• A hard “no” on companions for young kids (33:04) - Precautionary principle: show it’s safe before it enters “the developmental sanctuaries of our kids”
• The DETECT method (37:29) - Design, Ethically trained, Troubles in children, Evidence, Confidentiality, Teaching values
• Applying DETECT to Cradlewise (38:07) - The Sam Altman-endorsed AI smart crib rates as “enhance, not replace”
• Claude aced the SPEAK test (43:41) - Claude scored top marks on the TMW Center’s child-development knowledge assessment — a resource, not a replacement for parents
• Against All Odds debuts (47:25) - A new prediction-markets segment
• “The Donking of a Truth Machine” (51:00) - A Polymarket dispute over whether a caster said “donk,” resolved by the Optimistic Oracle and UMA token holders
• WSJ: fake Polymarket bets (55:47) - Paid creator videos showed winning bets that were never placed and would have lost $166,000
• Meta’s “Arena” app (58:12) - Zuckerberg dispatched a team to build a Polymarket/Kalshi-style app, initially with fake points

Mentions

Companies

• Anthropic (1:01) - Maker of Fable 5 and Mythos 5; models pulled then restored; ran Project Glasswing and complained about Alibaba distilling Claude
• Amazon (2:22) - The “trusted partner” whose researchers flagged the Fable jailbreak; CEO Andy Jassy reportedly alerted the administration
• OpenAI (6:00) - Pressured to release GPT-5.6 only to government-approved partners
• Zhipu AI (13:47) - Chinese company behind the GLM 5.2 open-source model
• Alibaba (19:54) - Accused by Anthropic of wide-scale distilling of Claude
• TMW Center / University of Chicago (24:00) - Dr. Suskind’s center for early learning and public health; built the SPEAK assessment
• Miko / Luna (25:34) - AI-powered robots designed for kids
• Cradlewise (38:07) - AI smart crib that detects waking and starts gentle bouncing; endorsed by Sam Altman
• Common Sense Media (41:11) - Cited as possibly working on an AI-products rating/seal
• Polymarket (48:00) - Prediction market at the center of the Rick Rubin ad, the “donk” dispute, and the WSJ fake-bets investigation
• Kalshi (58:12) - Prediction market; its CEO pitched betting as a cure for Instagram addiction
• Meta / Instagram / Facebook (58:12) - Building a prediction-markets app internally called “Arena”
• The New York Times (0:44) - Kevin’s employer; broke the Meta Arena and “Donking” stories

Products & Technologies

• Fable 5 (1:01) - Anthropic’s model pulled June 12 over a jailbreak, restored globally weeks later
• Mythos 5 (1:01) - Anthropic’s most powerful model, shared via Project Glasswing; restored only to select U.S. organizations
• Project Glasswing (3:58) - Anthropic’s program sharing Mythos with partners to harden critical infrastructure
• GPT-5.6 / GPT-5.5 (5:22) - OpenAI models; 5.6’s release limited by the government; 5.5 named by Anthropic as able to find the same vulnerabilities
• Kimi K-2.7 (5:22) - Model Anthropic cited alongside GPT-5.5
• GLM 5.2 (13:47) - Zhipu AI’s open-source model at the center of the China-catching-up debate
• Claude (43:41) - Aced the TMW Center’s SPEAK child-development assessment
• HOPE framework (29:09) - Suskind’s values compass: Human connection, Owning imperfections, Protecting early years, Enhance not replace
• DETECT method (37:29) - Suskind’s six-question product checklist
• Alexa (33:44) - Cited as an acceptable smart device (playing songs) versus AI companions
• Optimistic Oracle / UMA token (53:18) - Polymarket’s disputed-market resolution system, weighted by token holdings
• Meta “Arena” (58:12) - Meta’s in-development prediction-markets app

People

• Kevin Roose (0:00) - Co-host; NYT tech columnist and parent of a four-year-old
• Casey Newton (0:24) - Co-host; author of Platformer; discloses his fiancé works for Anthropic
• Dr. Dana Suskind (24:00) - Guest; pediatric cochlear-implant surgeon, author of “Human Raised,” mother of eight
• Howard Lutnick (4:25) - Commerce Secretary; sent the letter lifting Anthropic’s restrictions
• Andy Jassy (2:22) - Amazon CEO; reportedly alerted the administration to the Fable jailbreak
• Sam Altman (6:00) - OpenAI CEO; told staff about the GPT-5.6 limits; separately endorsed the Cradlewise crib
• Greg Brockman (11:00) - OpenAI president; his reported $25M donation “didn’t buy a lot of goodwill”
• Marc Andreessen (9:00) - Cited for a possibly fictional Biden-era meeting; joined the Defense Policy Board
• David Sacks (9:00) - Former AI czar who said the private sector should be “allowed to cook”
• Dean Ball (12:25) - Former Trump AI advisor, soon-to-be OpenAI employee; argued government needs technical experts
• Pete Hegseth (13:21) - Official Andreessen will engage via the Defense Policy Board
• Rick Rubin (50:18) - Legendary producer behind the Polymarket ad
• Daniel Krishkovets (“Donk”) (51:00) - Pro gamer whose nickname anchored the disputed Polymarket market
• Mark Zuckerberg (58:12) - Directed Meta to build the “Arena” prediction app
• Tarek Mansour (1:01:43) - Kalshi CEO who framed prediction markets as a cure for Instagram addiction

Surprising Quotes

“The way to win is by becoming a political actor, right? Like, that is sort of the downside here, is that, like, now deploying models, if you’re an AI company, is a political process.” — Casey Newton, 10:48

“If you’re going to tell me right now that this is as good as a Mythos, I’m going to need to see the damn numbers.” — Casey Newton, 16:05

“You’ve got to show me it’s safe before I’m going to let it into the developmental sanctuaries of of our kids because the stakes are just way too high, uh, even higher than social media in my opinion.” — Dana Suskind, 33:04

“My like sort of half joke for like the the official marketing tagline of any prediction market should just be ‘betray your friends’. Because that’s the way to make money out of prediction market. You found out something you’re not supposed to know? Get on a prediction market.” — Casey Newton, 50:07

“The other guaranteed method for making money on PolyMarket, insider trading. Uh, but that one doesn’t play as well on camera.” — Casey Newton, 58:03

Transcript

Kevin Roose: 0:00 Well, we’ve had some big news out of Washington in the past few weeks. We are in the middle of something truly insane and unprecedented that is happening between the frontier AI companies, including OpenAI and Anthropic, and the federal government, which has decided all of a sudden, with basically no—no rhyme or reason, that it wants to start picking and choosing which customers are allowed to use the frontier models.

Casey Newton: 0:24 Yeah, when this happened, many listeners reached out to us and asked us if we would consider doing an emergency episode. And I’m happy to say that the emergency lasted so long that we’re able to get to this now on our regular schedule.

Kevin Roose: 0:36 Yeah, it’s sort of a rolling emergency. [laughs] It’s sort of like a permanent crisis, if you will.

Casey Newton: 0:42 Exactly, exactly.

Kevin Roose: 0:44 So, before we get into Fable, since it does involve AI, we should make our disclosures. I work for the New York Times, which is suing OpenAI, Microsoft, and Perplexity.

Casey Newton: 0:53 And my boy—oh, downgrade. And my fiancé works for Anthropic.

Kevin Roose: 0:56 Oh! [laughs] Let’s just explain what’s been happening, and then we can talk a little bit about what we think about it. So, the—the first sort of salvo in this dispute was on June 12th, which was the day that we officially left for our break. Very good timing by the US government. When the Commerce Department issued an export control directive shutting down access to Fable 5 and Mythos 5, the most powerful versions of Anthropic’s models. Fable, in particular, had just been released a couple days earlier, and basically the government said, no one is allowed to use this if you’re a foreign national, including an employee of Anthropic inside or outside the US, you are not allowed to use this model. That basically shut down access for all customers everywhere. Anthropic basically had to pull the model because it had no way to do sort of user-by-user citizenship verification. And as its reason for doing this, the federal government said, well, we received some notifications from a trusted partner that there was this jailbreak issue on Fable specifically.

Casey Newton: 2:18 And who was this trusted partner, Kevin?

Kevin Roose: 2:22 Well, we believe it was Amazon, which, based on some reporting from The Information, Andy Jassy, the CEO of Amazon, had gotten in touch with administration officials after researchers at Amazon had discovered this jailbreak, this way of getting supposedly restricted information that can be used in cyber attacks. Based on some reporting from the Wall Street Journal, we believe that Andy Jassy sort of tells the Commerce Department about this. They freak out and say, you know, we’ve got to get this model off the market.

Casey Newton: 2:41 Okay, so let’s zoom in as far as we can here, because I think it’s really important to get clear on, what was the thing that this model did that gave the Trump administration the heebie-jeebies? My understanding, Kevin, is that when other security researchers have reviewed what we believe the issue was here, it was a f— It’s a fairly standard back and forth and was the sort of interaction that you would expect a cyber defender to have with the model to try to get it to identify and fix a bug. So what do we know exactly about what was going on here?

Kevin Roose: 3:16 We don’t know exactly what Amazon’s researchers discovered that made them so concerned. One cybersecurity expert who reviewed Amazon’s work said in a blog post that these concerns were basically not a big deal, that this was something that defenders who are trying to patch systems from cyber attacks need, basically asking AI to fix, find and fix the bugs in a given file. And I guess Amazon’s researchers determined that that same technique could be used to exploit a piece of software or find vulnerabilities, which of course is the whole advertised purpose of giving Mythos and similar frontier models to cybersecurity defenders ahead of the general public.

Casey Newton: 3:58 Right, Anthropic had launched this program called Project Glasswing where it had shared Mythos with a relatively small number of partners in an effort to harden the defenses of a lot of the critical infrastructure in this country and in allies before similar capabilities were found in other models. And as soon as the government stepped in and said, you know, effectively we’re placing export controls on this technology, all that stopped.

Kevin Roose: 4:25 Yeah, and so it’s now been a couple weeks since these export controls went into effect. Last Friday, Anthropic and the Trump administration reached a sort of tentative deal allowing Anthropic to restore access to Mythos for some clients. And then on Tuesday night, news broke that the Commerce Department was lifting those export controls and restrictions on Anthropic’s models. Commerce Secretary Howard Lutnick sent a letter to Anthropic saying that the company had taken steps to address the risks associated with these models. And on Wednesday, Anthropic announced that Fable 5 is back, that it is again available to users around the world. Mythos 5 still not publicly available but is going to be restored to a set of US organizations and they will sort of coordinate with the government to expand that more broadly. Casey, what did you make of the government’s reversal here?

Casey Newton: 5:22 Well, I mean to me the reversal seemed inevitable. And it still concerns me how little was said about the process behind this. I was probably more interested in Anthropic’s response because they put out a blog post that effectively put the industry on blast and said that GPT-5.5, Kimi K-2.7 and several other models could find the same vulnerabilities that had resulted in Fable being banned. So effectively Anthropic said, ‘Hey, you know, we added another safeguard, but like really this is an industry thing. Don’t point fingers at us about this.’

Kevin Roose: 5:51 Yes, and speaking of other models, let’s talk about what is going on with OpenAI and GPT 5.6. So last week, as OpenAI… …they prepared to release their newest GPT 5.6 models, Sam Altman told the staff that they would not be rolling it out to the public right away. They would release it first to a limited group of partners that would be approved by the government and that the Trump administration had asked the company not to release the model more broadly than that. And he didn’t sound happy about it. This was clearly not the the way that OpenAI and Sam Altman had wanted to release GPT 5.6, but the way that he phrased it made it seem like this was sort of pressure that had been applied to him from the Trump administration. So this new sort of this new business that the government is in of telling AI companies, “you can’t release this model” or “you can only release it to a list of customers that we approve beforehand,” seems to be applied to at least two companies now, and this seems to be coalescing into something more of a broader strategy and not just a single targeted act of retribution.

Casey Newton: 6:56 Yes, and and what we have said before on this subject is that this is the exact state of affairs that the people who are now running AI policy in the Trump administration were apoplectic about during the Biden administration, right? They were warning, you know, they’re going to create a a licensing scheme and they’re they’re going to start picking winners and only a handful of the administration’s favored labs are going to be able to release any models. And that’s why we’ve got to get rid of this Biden executive order. Fast forward to today, they have implemented a de facto licensing regime. And because it’s the Trump administration, there are no known rules. There’s no transparency whatsoever. We have no idea what makes a model considered safe to ship or unsafe to ship. We’re truly just in this like authoritarian limbo where until a few people decide that, you know, Fable or GPT 5.6 are safe to use, uh, we’re out of luck.

Kevin Roose: 7:50 Yes, AI is now being regulated by vibes.

Casey Newton: 7:53 Yes.

Kevin Roose: 7:54 And that is a state of affairs that I think both you and I were worried about, which is why, you know, several years ago when these kind of pre-deployment testing regimes were proposed as part of these AI safety bills, we said, well, this is probably a good idea to have something like this in place so that you don’t just have this kind of slapdash, chaotic approach where like models with really scary capabilities start coming out, and then the federal government has to scramble and try to figure out what the models are even capable of and then take some, you know, fast action on that basis, rather than having something that is actually thought through and planned from the start.

Casey Newton: 8:30 Yes, and let me say, I think the government should be able to prevent the release of large language models. I do believe that. I even think that after a model has been released, if it does something really dangerous, I think the government should be able to come in and say, “hey, we actually need you to place a pause on that.” I just think that it should be done with a set of clear rules. I think the companies should have due process. I think there should be a process for fixing those models and getting them into public shape and getting some sort of consensus on when it can be released. You know, just sort of like basic good government stuff. And we are just so far from that world. the point that you made about the hypocrisy of the people, especially on the tech right, who made opposition to AI regulation and licensing their entire personalities several years ago. You know, Marc Andreessen has credited this, I think, non-existent meeting. I have tried to back up through reporting the facts of this alleged meeting that Marc Andreessen says happened with officials from the Biden administration, where they basically said, according to Marc Andreessen, that they were essentially going to pick winners and losers in the AI industry, and that their favored labs would be able to develop and release powerful models, and that no one else, including open-source developers, would be able to. Again, I have tried to back this up. I have tried to find, I’ve talked to many people who would have been in a meeting like this. No one remembers it happening like this. So, this may be a convenient fiction that Marc Andreessen has decided to tell. But this was their whole personality for years. They said, “We are becoming a socialist, Soviet-style state where the government’s going to tell private companies what they’re allowed to do.” David Sacks, the former AI czar, famously said that the private sector should be allowed to cook. And now we have, just a couple years later, a what you called a de facto licensing regime, and they’re picking winners and losers. They’re saying, “This company can release their models. This other company can’t.” They seem to have no clear criteria. They are going on these reports from random, you know, companies that are testing these models, and it does not seem like they have any interest in sort of formalizing that or making it any easier for AI companies to navigate.

Kevin Roose: 9:01 I mean, I want to return to… Yeah, I mean, you say they’re picking winners and losers; really, they’re just picking losers, right? They’re saying, “Uh, this model’s too good. You can’t release it. Sorry, loser,” right? You know, so there’s… it’s unclear what is the way to win here.

Casey Newton: 10:48 I mean, the way to win is by becoming a political actor, right? Like, that is sort of the downside here, is that, like, now deploying models, if you’re an AI company, is a political process.

Kevin Roose: 11:00 Well, look, I mean, you know, I would love to ask Greg Brockman over at OpenAI how he feels about his $25 million donation to the Trump people, right? Because that didn’t seem to buy them a lot of goodwill when it came to GPT-5.6. So, I don’t know that, you know, sucking up to the Trump administration will even get your model out the door right now.

Casey Newton: 11:18 It won’t get it out the door right now, but I think they are going to, if I had to predict, they’re going to have an easier time over at OpenAI releasing these models in the future. I think the big shift that is really weird right now that all the labs are trying to process is, I think they just moved from a sort of default “yes” environment where, like, the assumption was if you trained a really good model, you could release it. And now I think they’re in a default “no” environment where if you build a model that is more capable than Claude 3 Opus or GPT-5.6, your assumption has to be that the government of the United States is not going to let you release it, at least not right away and at least not to everyone. So, I think that is a huge shift in the sort of structure of the AI industry that has happened more or less overnight.

Kevin Roose: 12:00 Yeah.

Casey Newton: 12:00 I mean — and this is just another moment where I wish that we had a Congress that was taking this seriously. This is a clear moment where we need a legislative framework for how these companies should be regulated that does govern how and when models can be released and governs how and when they are taken off the market over safety concerns, right? We know what to do here, we just have a government that is not doing it.

Kevin Roose: 12:25 Yes. And Dean Ball, former Hard Fork guest, former Trump administration AI advisor, soon-to-be OpenAI employee, had a great post about all of this just a few days ago, where he made the point which I thought was reasonable, that like the government should have some role. It’s reasonable to expect that in the future, when these models are very powerful and have very scary capabilities, that the government will want to take an active role in deciding what can and can’t be sold. That seems reasonable.

Casey Newton: 12:54 Yeah. Yeah, and I think there’s a good argument that that moment is now. If the models are now capable enough and scary enough that the government has an interest in addressing it.

Kevin Roose: 13:03 Totally. But Dean made the point that you actually need technical experts in government to make that happen. You need people who are capable of evaluating and running the safety tests on unreleased models to determine what they are and aren’t capable of. And right now, at least according to Dean, we don’t have that.

Casey Newton: 13:21 Well Kevin, I have some great news. Because on Monday, none other than Marc Andreessen himself joined the U.S. Defense Policy Board, which will put him in regular conversation with Pete Hegseth and other government officials. And so now Marc can finally get to the bottom of this and demand that the government drop this de facto licensing regime.

Kevin Roose: 13:43 Do you think he’ll be running any of the safety evaluations himself?

Casey Newton: 13:46 I am terrified to find out.

Kevin Roose: 13:47 So there’s one other twist of this story, and it’s something I, I’m really interested to get your take on, which is the whole China angle. Because the other backdrop of these discussions between the U.S. government and the U.S. AI labs is that these Chinese models are getting pretty good. So recently we saw a new model out from the Chinese company Zhipu AI called GLM 5.2. You know, people are saying that this is, you know, one of the best open-source models out there. Um, there’s been a little bit of what I would consider overheated reporting that suggests that it’s as good or better than some of the frontier models from the AI companies in the U.S. I don’t think that is probably true, but it is at least in the sort of same tier of consideration. And so there are people who are saying, well, this is going to be a big problem because at the same time the U.S. government is pulling American models off the market that are very powerful, we now have these Chinese open-source models that customers, big businesses can use for much cheaper, that are just as capable and that are not in danger of being clawed back by the U.S. government. What do you make of this?

Casey Newton: 14:56 I think this is basically BS. Like, I do. Um, I think it feels like a lobbying tactic from people who want to get the American models back into production. I have seen no credible reporting that the Chinese models have kept up even when I’ve read the reporting that we’ve seen over the past week out of the Wall Street Journal and others. There aren’t numbers in there that would lend credence to it. What there are is a handful of people saying these models are really good. And look, I’m sure these models are pretty capable. I’m sure they’re more capable than the open source and Chinese models of a year ago. But we learned during the DeepSeek freakout that a reliable way to get attention in AI discourse is to wave your hands and say the Chinese are catching up, the Chinese are catching up, and everyone will pay attention to you. But when you look at how the Chinese are building these models, the theory just doesn’t hold because you have to keep in mind what these Chinese companies are doing is distilling American models.

Kevin Roose: 15:54 Yes.

Casey Newton: 15:55 They’re getting in there and they are scraping the responses of these chatbots and they’re using those to make their own chatbots. So the best Chinese models are kind of like American models once removed. And that is why they’re always going to be at least a little bit behind the frontier.

Kevin Roose: 16:05 Yes.

Casey Newton: 16:05 And let me say one more thing, Kevin, while I’m all worked up. Which is I’m increasingly coming to think about like the whole AI market in as like two systems. There’s the frontier and there’s everything else, okay? And if you’re if you’re truly at the frontier, we will know because you can see it in the benchmarks, you can see it in the revenue figures of the companies that are selling this stuff. That is where the actual value is. And then there’s everything else, which is the million little open source models, the distilled models, the Chinese models, and they’re just a step behind. And look at the revenues of those companies. There’s an insane amount of competition and it’s not as good as the best. And so, in my view, these Chinese companies that we’re talking about right now, they’re firmly in the everything else camp. That’s not to say they aren’t getting better over time because they are, but if you’re going to tell me right now that this is as good as a Mythos, I’m going to need to see the damn numbers.

Kevin Roose: 16:57 Right. I think that’s right, but I’m also like I’m not as quite as skeptical of this fear of China catching up as you are in part because I think these distillation tactics appear to work. You get a model that is almost as good as the original. If the Chinese companies are able to keep doing that, I think there’s a plausible case that they will shrink the time, the sort of gap between the American frontier and the sort of open source frontier. Maybe it’s nine months now, maybe in the future that becomes six months or three months. I think that sounds plausible to me. I also think that there are a lot of businesses right now who are really freaked out about what is happening with Fable and with GPT-5.6. And they are saying to themselves, well, if we were to sort of go with the latest American models at all times, not only are we paying a premium for that, but it could be yanked away by the government with no explanation and no due process and it would totally screw up all of our workflows and all of this. software that we’re building on top of those tools. So I- I did actually talk to someone recently who works at a- an American tech company, not one of the AI labs, but like a- a solid big American tech company, who was saying that they’re actually spending a lot more time with these open source Chinese models. Not because they’re as good as the frontier models, but because you can download them, you can run them on your own hardware, and they’re pretty good for a lot of tasks that don’t need the absolute capability frontier that you would get from a- a Mistral or a GPT 5 or 6. So I think that in this moment where there’s so much uncertainty coming out of Washington, I think a lot of companies that rely on these tools are going to take a hard look at the Chinese models and see whether they might be good enough for some percentage of what they’re doing.

Casey Newton: 18:43 Yeah, that makes sense. You know, like, the American economy has been so strong for so long in part because the- the government has mostly been safe, boring, and predictable. And the Trump government is not those things, right? We saw this during tariffs, right? Where like govern- the American businesses were having similar problems because they couldn’t predict how expensive their products were going to be on any given day based on what the tariff of the day was, right? And now we have the LLM restriction of the day, and it’s causing similar havoc in the economy.

Kevin Roose: 19:14 So I think one obvious link between the actions of the US government and the China subplot here is that maybe there’s a possibility that the US government restricting access to these frontier models from American companies actually doesn’t allow China to catch up. Maybe it slows China down, because if they are so reliant on distilling from, you know, Claude and- and the GPT series and other leading American models, maybe not having access to those models, um, will- will hurt them. So is there any sense in which you think this action by the Trump administration could actually widen the gap between the US capabilities frontier and the sort of Chinese open source frontier?

Casey Newton: 19:54 I mean, maybe, but that is an extremely ham-fisted tactic for accomplishing that goal. The United States government has many levers it can pull if it feels like Chinese companies are attacking its companies. Like for example, they could sanction companies that they believe have caught distilling. They can prevent them from selling their wares in America. They can get allies to do the same thing. We saw Anthropic the other day send a letter complaining that Alibaba has apparently been doing wide-scale distilling of Claude, and the hope there is that the government will intervene. So that is how these things are traditionally done. Uh, while it is true, I suppose, that this is going to make it harder for Chinese companies to distill the next models, there are still all of these other reasons why, at least for me, like, that’s the worst way to go about it.

Kevin Roose: 20:44 Right. Right. That makes sense to me. I mean, the argument that I would make against that point is like, what matters is not just the capabilities frontier of the internal models that the companies in each country have. What matters is sort of the deployment frontier, and for a society to really harness the benefits of AI, you have to be able to— They use the models, like you have to be able to deploy them inside big companies, you have to be able to deploy them inside consumer products, like people have to be able to use them. It is not just like which lab has in their basement the most capable model that nobody else can use.

Casey Newton: 21:15 Right, and all of this happened at a time when companies and government institutions around the world were using these very powerful models to improve their cybersecurity defenses, right? I mean, like that’s one of the reasons why this issue feels really urgent is that there was an ongoing project to protect the critical infrastructure of the United States and its allies and that was yanked away while at the same time we’re selling advanced chips to the Chinese to help them catch up.

Kevin Roose: 21:40 So I think there’s one silver lining to this story for me is that for years now you and I have been sort of hoping against hope that the US government would stop just sort of talking about AI as this unalloyed good that was just a normal technology that was going to supercharge the economy, that they would actually start to pay some attention to the risks of these powerful AI systems. That is happening now, for better or for worse. The government is now firmly aware that these models do pose risks. They are sort of thinking about cybersecurity, but I imagine they are also thinking about bio-risk and other types of risk. So I think that there is a sort of way to spin this whole saga as a clunky, ill-advised first step on a good path, which is like the government is waking up. These models are very powerful, they can be very dangerous, they’re going to get more powerful and more potentially dangerous as time goes on and so I think what we’re seeing now is sort of a fumbling attempt to kind of do something and I think they will come up with better ways of doing this over time. I hope, anyway, that they will. This is not sort of the permanent structure of the AI licensing regime in the United States, but at least they are not just pretending that this is just a normal technology anymore.

Casey Newton: 23:04 That is true, and I’m glad that it’s the case, but at the same time this same administration has been pushing to allow the export of more and more advanced chips to China, which will allow China to train more powerful models over time and they’re doing that at the same time as they are preventing good models from being used by our allies like Britain and other countries. So there is still an extreme level of incoherence in the administration’s position in my opinion.

Kevin Roose: 23:32 Well, you can’t win them all.

Casey Newton: 23:35 That’s true.

Kevin Roose: 23:37 Okay, so I’m very excited for our guest today, who is an expert on the intersection of parenting and AI. This is a topic that I’ve been wanting to explore on the show for a while now, in part because it’s just become a big part of my social life is talking with other parents about how we are or aren’t exposing our kids to technology, including AI and chatbots. So today we have on the show… So, our guest Dr. Dana Suskind. She’s a pediatric surgeon and childhood development expert. She’s a professor at the University of Chicago and the founder and co-director of the TMW Center for Early Learning and Public Health. She’s also an author and she has a book coming out soon called Human-Raised: Nurturing Connection, Curiosity, and Lifelong Learning in the Age of AI. And when I heard about this book, I thought this is just the person we need to tell us/me how we should or shouldn’t be using AI with our children.

Casey Newton: 24:33 Absolutely, because in addition to all of that, Kevin, Dr. Suskind is a parent to eight children who are now all grown up. So this is somebody who has a lot of expertise in this subject.

Kevin Roose: 24:46 Yes. So I think this is an important conversation today, but I think it’s about to become much more important because there are just a lot of AI products and tools aimed at young kids that are making their way to the market and are going to become, I predict, more popular in the next few years. So already we’ve seen Miko and Luna, which are two AI-powered robots that are designed for kids, and we’ve got all these sort of child-specific companions and chatbots with names like Askie and Elo and Hey Otto, and we’re just starting to see a lot more products using AI that are aimed at children. And I think for a lot of parents, that is creating a lot of anxiety because how the heck are we supposed to know which of these things are going to, you know, help our kids grow and develop and which are going to turn them into glorified slop cannons?

Casey Newton: 25:34 Yeah, and that’s what you don’t want is a slop cannon child.

Kevin Roose: 25:38 Yes, heaven forbid. But here’s what I’ll say, if your child is a slop cannon, you need to read this book.

Casey Newton: 25:44 It’s true.

Kevin Roose: 25:45 Let’s bring her in. Dr. Dana Suskind, welcome to Hard Fork.

Dana Suskind: 25:55 Thanks so much for having me.

Kevin Roose: 25:57 So, you’ve got this new book coming out, Human-Raised, which is about parenting and sort of child development in the age of AI. And I was so excited when I heard that this book was coming out because as the parent of a now four-year-old, I see this issue coming for me in a way that makes me quite nervous. So first of all, like, what made you want to pursue this topic as someone who’s long been interested in childhood development?

Dana Suskind: 26:25 So I’m a cochlear implant surgeon, and you may say, well, why exactly are you on this or even thinking about this issue? But early in my career, I started noticing profound differences in the outcomes of my own patients. So I would do the same surgery, same technology, same parents loving them, and the outcomes of the children were so profoundly different with some of them being able to develop language and others not. So that experience actually brought me into this incredible world of brain development and neuroscience that shows that human… connection is not just a nice to have, it is the foundations of how we learn and how we learn to be human. And so my whole career has been focused on that and supporting parents and caregivers in that important role. And then suddenly AI comes onto the scene. For the first time in human history, we have technology that can actually mimic that interaction that has traditionally wired our children’s brains. And now I’m like, oh, we need to probably step back and really think about what is happening in our children’s worlds because the choices that we make at this moment is going to determine what our species looks like. So that’s why I wrote the book and happy to be on the show.

Kevin Roose: 27:43 Dr. Suskind, can I run some scenarios by you from my own parenting journey and you tell me whether I’m being a bad father or not? So the two primary ways that I am using AI with my four-year-old right now, one is to just like answer random questions that he asks. We’re in like a big question phase right now. So the other day he asks me, “Where does wind come from?” And I realized I actually don’t know the answer to that.

Casey Newton: 28:02 Yeah, I actually think that’s like a hard question.

Kevin Roose: 28:06 Yeah. So I pull up an AI model and I ask and I get back this answer and it’s about like air pressure. And I’m like, okay, this is for a four-year-old, could you like, you know, dumb it down a little bit? And it gives me a great answer and we tell him the answer and he has the answer to his question. Now he gets to like repeat it to all his friends. That’s option one. By the way, on his playground, the kids are obsessed with where wind comes from. It’s a hot topic over there.

Casey Newton: 28:31 It’s been made quite popular in preschool.

Kevin Roose: 28:34 Yeah. So the other way is just like creating stories for him. So one of the things that these models are quite good at is like you give it some guidance, I want a story about a kid this age who has a best friend who’s a dinosaur and they go off and they play soccer together, or whatever it is.

Casey Newton: 28:54 I would do one where his best friend is the wind. I think that could really hit, I think land cleanly with him. But no.

Kevin Roose: 29:02 So are either of those examples of AI use in parenting of a young child going to harm my son?

Dana Suskind: 29:09 Absolutely not. So first of all, those are great examples. And you know, I put in this book, I tried to really also give grounded and practical frameworks. And one of the frameworks that I give parents to sort of think about AI in general, sort of an evergreen, I call it HOPE. H is human connection is irreplaceable. It is what we need to protect and double down on. But O is owning your imperfections. As I said, kids don’t grow through perfection, they grow through, I don’t know about you, but I was a very, I am a very imperfect parent, but now I know it’s really good for my kids. P is protecting the early years. That early brain development when 80, those first years when you get 85% of the physical brain being built, we’ve got to be extra careful about what we let in. But lastly is E. If you’re going to use it, use it to enhance. Not replace. You are really using this technology to enhance your relationship, fill in the gaps of knowledge that you may or may not have. And that’s a great way. I this—this book in my view is not at all anti-tech. I’m a cochlear implant surgeon, I build AI tech, I’m all about it. But how do we use it to deepen our relationships and support us in this important role?

Kevin Roose: 30:23 Yeah, I—that makes a lot of sense to me and it—it—it is compatible with, I guess, my own sort of vague optimism about AI in the lives of kids. And, you know, I see a lot of parents right now with a lot of anxiety about AI that seems like it is borrowed from this earlier era of like the screen time debates or the debates about sort of addictive social media platforms for kids. And to me it just feels fundamentally different what AI is—is—is as a technology and what it is capable for doing—of doing for children. Like, I just think back to my own childhood and like, I was a—I was a nerdy kid who loved to learn stuff and if I had had like a chatbot that could have taught me about the Bernoulli effect, um, in a way that was like enjoyable and entertaining to me, like I would have done that and I would have been happier than going to my, you know, World Book Encyclopedia and looking up that fact. Like, so I think for the curious kids out there, my sense is, and please tell me if you agree with this, like that this could be an incredibly enriching technology for them if it is designed well and if parents are sort of in the loop of how their kids are using it.

Dana Suskind: 31:00 I mean, 100 percent, I agree with you. I think the—I think to your point, this, you know, people go to the social media, um, as the best analogy and I get it, um, and I think—but I think a better analogy is actually processed food. You know, processed food is a continuum, right? There’s whole wheat bread which is processed but it’s still nourishing, and then you’ve got the ultra-processed variety, you know, the Doritos and, you know, Twinkies which are good at, you know, for us occasionally. But the truth is is that there—and in the same way, AI is a huge spectrum. I mean, not only being able to, you know, provide answers as a chatbot, but, you know, lifting invisible labors off of parents and teachers to diagnostics. I, you know, I come from the world of medicine where it’s transformative. I think people are really focused on the AI companions which I get, which I do—am concerned about in terms of crowding out that necessary human connection, but that’s a really—that’s a small part of what AI is. I mean, you can have socially assistive robots that help children with autism learn to read, you know, social cues to better connect with other humans. That’s an amazing example. So, I think to your point, it is a spectrum and we can’t, you know, lump it into just—it’s not a monolith.

Casey Newton: 32:57 I’m curious though, like where if anywhere are you like drawing the line?

Kevin Roose: 33:00 the bright lines. Like where are you saying like, you know, do this, don’t do that?

Dana Suskind: 33:04 Yeah. Um, that’s pretty easy right now. Um, we are at the very beginning of this tech revolution and AI companions in the, you know, AI toys that, you know, claim to be better alternatives to screen time, that feels like a pretty hard no. You’ve got to, you know, I think the thing that we can learn from social media is that, you know, we need to take a precautionary principle. The, you’ve got to show me it’s safe before I’m going to let it into the developmental sanctuaries of of our kids because the stakes are just way too high, uh, even higher than social media in my opinion.

Kevin Roose: 33:44 Yeah. What about like smart devices that are not meant to be plush toys but like my, you know, my kid loves talking to Alexa and and asking for, uh, various children’s songs that are, uh, intolerant to adults but that he loves to play on repeat. Like is that is that the same thing or is that different because it’s not pretending to be a companion or a stuffed animal?

Dana Suskind: 34:13 I mean, I think it’s it’s how how much. I mean, it’s a slippery slope but I think we can all agree that, you know, Alexa playing songs when you’re around or even if you’re not around is probably a fine thing. It’s at this point it’s more just the crowding out of the human interaction, uh, that I’m very concerned about.

Casey Newton: 34:36 Yeah. My my kid would otherwise be asking me to play the Blippi garbage truck song 400 times, so I’m I’m somewhat sympathetic to that.

Kevin Roose: 34:46 You know, you’ve mentioned that we’re pretty early in many ways in the development of this technology, but I also know that, you know, you’re really concerned with with some of the things that you’re seeing. Is this a moment where you feel like the government has a role to play in regulating the technology? And if so, what do you think that role should be?

Dana Suskind: 35:05 No, absolutely. I think we know in Norway they’ve actually, uh, set out a law that there’s no Gen AI in the school system in the early years. Um, and in the same way I think we’re seeing that in the US. I mean, in some ways this really mirrors the the industrial food revolution. When the food revolution happened, you know, initially there was no guardrails and eventually we started, you know, had the Pure Food and Drug Act and then eventually we got, you know, the little, uh…

Casey Newton: 35:39 The nutrition labels. Yeah.

Dana Suskind: 35:41 The nutrition labels. I mean, these are the things that, you know, are going to happen. It’s just we’re just at the very beginning but hopefully we’ll start seeing it.

Kevin Roose: 35:49 What what do you think of the Norway approach? Like does it seem wise to say like, ‘Hey, until we’ve done a little bit more research, we’re going to keep like kids under 16 from using ChatGPT in schools for any reasons.’ Does that does that seem… …wise?

Dana Suskind: 36:01 I mean I think it it’s taking a scientific approach. I don’t think they’re saying no never. I think they’re saying, ‘Wow, this is happening so fast. Uh, let’s take a step back.’ We know what works. I mean, look, if we didn’t know how to educate humans, um, I would get it, but I think they’re just being, you know, taking a prudent approach. And I’m a scientist who is about innovation. I just, you know, I think about what happens in medicine and we’re much more, you know, thoughtful when when we have a new drug on the market. It’s not like we have to say, ‘Oh, I wonder what that’s going what’s going to happen to my child if I give them this new medication.’ I’m not sure why, you know, this new technology, which is so incredibly powerful—it’s amazing, right? But we have to we have to acknowledge that it’s a powerful technology that we need to understand what happens to us. I think the story of AI is is as much about technology as it is about understanding what it does to humans. And I think that’s what Norway is saying.

Casey Newton: 37:12 Well, let’s try to give parents some practical advice. You suggest that when parents are trying to figure out whether a particular AI tool or product is suitable for them, they should ask themselves six questions. So, what what are the questions that parents should be asking?

Dana Suskind: 37:29 Sure, so because I’m a surgeon and I love acronyms, it’s called DETECT. So what does the DETECT method mean? D is design. E is, was it ethically trained? T is, were there any troubles with this technology in children? E is, what is the evidence? Does it really do what it’s supposed to do? Um, C is confidentiality, what happens to your child data? And lastly is T, what is it teaching from a value standpoint? And that DETECT method is used so that parents can quickly and easily figure out, is this something that they want in their child’s life?

Kevin Roose: 38:07 Alright, so I want to see if we can apply your DETECT framework to like an actual AI thing that is in the market right now. There is something called Cradlewise, which I learned about from a tweet by Sam Altman, the CEO of OpenAI, and um he has a baby, and he tweeted that they had bought, quote, ‘a lot of silly baby things that we haven’t needed, but definitely I recommend a Cradlewise crib and a lot more burp rags than you think you could possibly need.’ As for the Cradlewise, though, this is apparently a smart crib that uses AI to detect early signs of a baby waking up and automatically starts a gentle bouncing motion and soothing track uh to get them back to sleep. So if we were to use your your framework, uh doctor, how how would we rate a Cradlewise?

Dana Suskind: 38:54 Ha! All I can say is I wish it had been there when I had my—when I had my children. No. Um, so so what is it designed to do? In some ways it’s designed to please… Support the parents by helping the child sleep but uh so that’s number one E how is it trained I think it’s probably ethically trained is there any evidence that kids have had trouble?

Kevin Roose: 39:13 Not that I’m aware of but I haven’t so much as googled it yet but but that would be a thing to do would be to sort of investigate whether there’d been any problems with the device

Dana Suskind: 39:20 Problems yeah so and what the evidence were um and probably some researchers have looked to see does it keep kids asleep do they have any problems and I mean I think long and short of it I think that that feels very aligned with using technology to enhance not replace uh Sam Altman’s parenting allow him to have a good night’s sleep let the baby sleep so just hearing it very you know for the first time it sounds like a good thing

Casey Newton: 39:43 It seems okay I’d say I would take a bed that would actually gently bounce me and rock me back to sleep when I wake up in the middle of the night because whatever I’m doing now is not working for me

Kevin Roose: 39:51 You said to update the uh the firmware on your 8-sleep I I’m googling this right now and it seems like there have not been serious problems with the Cradlewise smart bassinet although it does say that they it has proprietary linens you cannot buy cheap third-party crib sheets because the mattress is a non-standard shape

Casey Newton: 40:13 You know Cradlewise was one of my favorite songs in the Sound of Music

Kevin Roose: 40:16 That was Edelweiss oh

Casey Newton: 40:17 Anyway

Dana Suskind: 40:18 Hey listen it’s either that or like me with my my late husband and I had three kids in our bed at one time and I don’t know if I had Cradlewise we might have had more sleep

Kevin Roose: 40:28 Yeah that’s that’s wild

Casey Newton: 40:29 That is heroic

Dana Suskind: 40:31 Yes seriously I know

Kevin Roose: 40:32 I think this falls for me into the category of like it it makes you a better parent it does actually make you a better parent to get six hours of sleep in a night

Dana Suskind: 40:40 It does

Casey Newton: 40:41 Here’s what I’m going to say do not overclock your Cradlewise so it starts shaking incredibly quickly even if it seems like it would be kind of funny don’t do it

Kevin Roose: 40:51 I mean I do think eventually we will need some kind of like a TV rating system for AI products um and I’m curious whether you’ve heard anyone suggest that or whether you have any ideas about how that could be implemented

Dana Suskind: 41:11 Yeah 100% actually I talk about the Good Housekeeping seal of approval which actually came out of the industrial food revolution and suggest that we actually need something similar and I’ve heard of different groups talking about doing it um common sense media I think is thinking about it uh and others maybe you all should be doing it

Kevin Roose: 41:37 The Hard Fork seal of approval we’re yeah

Dana Suskind: 41:40 Yes

Kevin Roose: 41:41 I mean I think one of the hardest things is that these models are are generally useful like they can do amazing things that are very good for kids they can teach them science they can teach them to read they can you know give them a creative outlet and they can addict them or become you know harmful companions or shout out the rest of their human friends because it’s just so fun to talk to the AI companion. So I just think like, I want the AI companies thinking more about parental controls and visibility, because I think parents should know what their kids are talking about. But I think as with screen time, I think there’s this sort of dangerous tendency to flatten every product into a single verdict, which is like, you know, it really matters what your kid is watching on their screen. It also really matters what your kid is talking to AI about. And I think the more that parents can be aware of and respond to that, rather than just saying like, this product is okay and this product is bad, the better off kids will be.

Casey Newton: 42:41 Well, and you know, Kevin, that actually brings up something interesting, which is that, you know, earlier during this interview, you brought up the Bernoulli effect. And I thought that I could imagine a lot of children listening to this and racing to their AI, you know, to find out what it is. So I thought I would take this opportunity as a human who wants human connection with children to share what the Bernoulli effect is. And what it is, and you guys already know this, but it’s, and honestly Kevin, it’s more commonly called the Bernoulli principle, so I don’t want to embarrass you, but it really is sort of known as the principle. And it states that for a fluid, such as air or water, flowing smoothly, an increase in the fluid speed is accompanied by a decrease in its pressure. So children, I hope that you heard that and have internalized that and are not going to ask AI about it.

Kevin Roose: 43:26 It’s what allows airplanes to fly.

Casey Newton: 43:27 Exactly. Didn’t you learn that in middle school science?

Kevin Roose: 43:30 Yeah.

Casey Newton: 43:31 Okay, let’s get back to Dr. Suskind here. We just like to do our little bits and just force our guests to suffer through them. Okay.

Dana Suskind: 43:41 Well, I’ll tell, you want me to tell you an interesting story related to knowledge? So, so in, in our center, we’ve built this really cool computer adaptive tool that measures what parents know about child development in all domains. It’s, it’s very predictive of what parents do and child outcomes. So we use it in lots of research. So recently, we haven’t published it yet, we had Claude take the, it’s called the SPEAK, to see how Claude would do. And so I’m going to ask you all, do you think Claude did really well, mediocre, or not so great on its knowledge of child development?

Kevin Roose: 44:18 Really well.

Casey Newton: 44:19 I would say well, yeah.

Dana Suskind: 44:22 You’re right. It aced it. Claude did incredible. Claude aced it. Claude aced it. Um, and so because we didn’t want the story to be, ‘oh, you know what, now Claude can parent,’ because that’s not the, the message. The message is Claude can obviously be a resource. So when parents have a question, oh, you know, XYZ about child development, it can confidently, they can confidently ask Claude and Claude will give a right answer.

Kevin Roose: 44:52 I mean, that tracks 100% with my own experience of these tools, which is that they are often more seasoned and more confident about parenting. advice than than I am and that they’ve been a very good resource for for me when I’m feeling stuck or I’m trying to deal with some like behavioral thing that I don’t know how to deal with or I just need a little backup I can I can just go to Claude or Gemini or ChatGPT and ask the question and usually the the answer it gives back is pretty good. Dr. Suskind I want to ask you about the the title of your book which is essentially a provocation as I read it that says that basically as AI tools become more normalized um this idea of being human raised this is going to become rare and maybe something of a luxury where like if you are born into a family with a certain level of security and privilege you will get sort of the human experience of having human parents and if you are not you will be raised by the machines. Am I reading that right? Um and is that what you think is likely to happen?

Dana Suskind: 45:59 Well I’m writing this book so that hopefully it will not happen. Um but yes. So in the same way as um you know a hundred plus years ago you never questioned if the food on your table was organic or farm raised and then suddenly you know fast forward with the rise of ultra processed food and suddenly organic was that for children with privilege. In the same way my concern is that artificial alternatives will become you know the the cheap calories of brain nutrition and you know human raised and human connection will become a luxury item. And we just can’t allow that to happen. Um every child deserves to have parents and caregivers who love them raising them. Not to say that AI can’t support those parents and support you know some of the development but I think we need to be very careful because we’ve seen this story before.

Kevin Roose: 47:02 Well Dr. Suskind thank you so much for coming on and for your great book Human Raised. Um I think this is a topic that a lot of parents are really interested in right now and so I predict um I predict it’ll do pretty well.

Dana Suskind: 47:14 Thank you so much.

Kevin Roose: 47:15 Thanks.

Casey Newton: 47:16 Thanks for coming.

Kevin Roose: 47:25 Well Casey one of the other stories that keeps on giving are prediction markets. This has been a source of nonstop um entertainment and scandal this year and so we thought we would introduce a new segment where we round up all the latest news and scandals about prediction markets and we’re calling that segment Against All Odds. Okay well Casey I think we should start this segment by looking at a recently released ad from the prediction market company Polymarket. And uh, I’m just going to play this for you.

Kevin Roose: 48:07 If you could ask one question, what would you ask? Can the US win it all?

Dana Suskind: 48:15 Est-ce que la France retrouverait la gloire ?

Casey Newton: 48:18 Que dirais-tu, football ou soccer ?

Kevin Roose: 48:21 Sénégal mouna choqué adouna bi ?

Dana Suskind: 48:22 ¿Pueden las tradiciones sobrevivir?

Casey Newton: 48:24 Wird Deutschland gewinnen?

Kevin Roose: 48:25 Would the world be better off in a hundred years? Sort of a Vanity Fair ad of gambling.

Casey Newton: 48:30 Mm-hmm.

Kevin Roose: 48:31 A series of questions appears on screens: ‘will borders matter in a hundred years’, ‘do aliens exist’, ‘will Messi win again’… ‘will tradition evolve’… Got a Kanye West song in the background, which is a choice. The people who never stop asking are the ones who find answers. Now Casey, did that ad make you want to open up an account on an offshore crypto-denominated prediction market and start gambling with your savings?

Casey Newton: 49:28 I, I find this ad so offensive. Like the way that it presents prediction markets as like, first of all, like grounded in concepts of like multiculturalism and togetherness, you know? It’s like, it it has this sort of almost like mystical feel of like if we could just all get in the same room and gamble together, maybe wars wouldn’t happen anymore. You know, I mean, like that’s the feeling. And I truly don’t get it at all.

Kevin Roose: 49:54 It’s very sad that they didn’t feature the actual customers of Polymarket who are like troops on active duty in war zones, or like middle schoolers…

Casey Newton: 50:03 Yeah.

Kevin Roose: 50:04 People in Bad Bunny’s Super Bowl entourage.

Casey Newton: 50:07 Yeah. I, my, my like sort of half joke for like the the official marketing tagline of any prediction market should just be ‘betray your friends’. Because that’s the way to make money out of prediction market. You found out something you’re not supposed to know? Get on a prediction market.

Kevin Roose: 50:15 Yeah.

Casey Newton: 50:16 Yeah. Go make some money. So that’s the ad I want to see, is is screw over someone close to you.

Kevin Roose: 50:18 Yeah, well, well, my second question about this is how much do you think they paid Rick Rubin and do you think they paid him in Polymarket credits or in dollars?

Casey Newton: 50:34 I’m, I’m going to guess it was in the millions, Kevin. Yeah.

Kevin Roose: 50:37 You don’t think Rick Rubin just did this one for free as a sort of testament to his commitment to craft and creativity and the artist’s dream?

Casey Newton: 50:45 No. Rick Rubin produced the Beastie Boys. He has not had to work in a very long time.

Kevin Roose: 50:51 So these ads from prediction markets are just inescapable. I was watching the World Cup on TV over the break and just like every… Every other ad is now for some kind of prediction market or gambling site. It is wild. But there’s also been more than advertising happening in prediction markets. I want to talk about this incredible story that David Siegel at the New York Times recently wrote called “The Donking of a Truth Machine,” which is about a very silly but serious and high-stakes controversy on Polymarket, the very company whose ad we just saw. And it all boils down to this seemingly crazy question: Did a guy in a video say the word donk? So the backstory here is that there is a professional gamer named Daniel Krishkovets, whose nickname is Donk, and back in April, there was an active betting market on Polymarket about whether during the broadcast of this video gaming tournament, someone would say the name Donk. Now crucially, Donk was not participating in this video gaming tournament.

Casey Newton: 51:58 Because if he was participating, it would seem very likely that at some point someone would say the word donk.

Kevin Roose: 52:03 Yes. So this is the kind of sort of silly market that often happens, you know, will this word be uttered during a CEO’s earnings call, or will the White House press conference say the word, you know, some word? And people bet on this, it’s, you know, it’s a time-tested form of gambling on these websites. But this particular market, the donk market, went sideways because there was this very long, seven-hour-long gaming tournament stream, and it appeared that Donk was not mentioned. But then somebody on this market was like, wait a minute, at one point the commentator doing the play-by-play on this gaming stream appeared to stumble over the word “don’t” and it sounded like “donk” to this person.

Casey Newton: 52:58 Now could you please, no, here’s what I’d like you to do. I’d like you to say a sentence where the word don’t—say a sentence with the word don’t but make it sound like donk.

Kevin Roose: 53:11 Donk look back in anger.

Casey Newton: 53:13 Okay, very good, thank you. You can go off now.

Kevin Roose: 53:18 Okay. So this started a very fierce argument about whether this commentator had actually said donk or whether this was just a slip and whether an accidental donk is still a donk for the purposes of resolving this. And so this big fight occurs and David goes into his—into this fight in very amusing detail in his story. But to boil it down, basically whenever there’s a contested market where the resolution isn’t clear on Polymarket, there is a system called the Optimistic Oracle where basically the people who hold this special crypto token called UMA can vote on disputed outcomes.

Casey Newton: 53:56 Just—at the end of the sentence, I just wanted to jump out a window. I just have to say that. Just learning… All of that information, I truly want to forget I’ve ever learned that information. None of those words are in the Bible. Truly.

Kevin Roose: 54:06 So basically the way this works is the more of these UMA tokens you have, the more voting power you have in these disputed markets. And basically, this is their way of sort of trying to inject some sort of a fair resolution criteria into these markets where it’s not exactly clear who is supposed to win and who is supposed to lose.

Casey Newton: 54:29 And the fair criteria is whoever amasses the most UMA tokens is the arbiter of truth. What a great system!

Kevin Roose: 54:40 Yes, it’s a crazy way to run a market. You’re essentially like giving the power to resolve these markets and determine who gets paid out or who doesn’t to the richest people on the platform, or at least the people who have spent the most money on this meaningless crypto token. Anyway, so this one resolved in predictable fashion, which was that a company called UMA Rocks amassed a bunch of these tokens and announced that it was going to vote no on this question of whether ‘donk’ was uttered on this gaming stream, which led to a bunch of follow-on no votes. So the ‘no’s’ won out in the end, but it was sort of a rigged resolution.

Casey Newton: 54:54 So there’s even like a form of gambling in how this thing gets resolved, is what it sounds like.

Kevin Roose: 55:00 Yes. You can gamble on the gambling mechanism. It’s sort of turtles all the way down.

Casey Newton: 55:05 I, when I was very—I don’t know what it was, but when I was young, I just sort of decided that gambling was not going to be for me. And to watch the way the world has transformed over the past 20 years to turn absolutely everything into gambling, I just continuously feel like I’m losing my mind.

Kevin Roose: 55:21 Yes. Because here’s the thing: you will lose money. Statistically, you will lose. This is not going to turn out well for you. And yet, like now, some meaningful portion of our economy is now just people who have decided to lose money that way.

Casey Newton: 55:31 Why don’t you lose money on journalism? Buy a subscription to a publication. At least you’ll get something in return other than suffering.

Kevin Roose: 55:35 But you won’t get the satisfaction of amassing UMA tokens in order to resolve the criteria for the market on whether ‘donk’ was uttered during a video game broadcast, and that’s what really counts in today’s day and age.

Casey Newton: 55:43 That’s true.

Kevin Roose: 55:44 So what else is going on in Against the Odds, Casey?

Casey Newton: 55:47 Well, speaking of Polymarket, Kevin, would you believe we have another Polymarket scandal?

Kevin Roose: 55:49 I’d believe it.

Casey Newton: 55:50 So a couple weeks ago, there was this great investigation in the Wall Street Journal about Polymarket, and specifically their strategy for social media marketing. They have done for a while now these social media videos where they’ll pay a creator to film a video of themselves making a bet on Polymarket. And the Journal actually went back and looked at those bets to determine, A, whether they were real, whether these bets actually happened on Polymarket, and B, whether, if they had been real, the people making these bets would have won or lost money.

Kevin Roose: 56:29 Oh, I can’t wait to find out.

Casey Newton: 56:32 So they analyzed more than 11…

Kevin Roose: 57:00 100 videos from 10 creators and they’ve found that, although 70% of these videos showed a bet being placed, none of these bets were real. These were all fake bets. 118 of the videos showed the creators winning and the video suggested that these people would have won almost $900,000 in total. But in reality, if those bets had really been placed, these same people would have lost more than $166,000.

Casey Newton: 57:27 So it’s like, I mean, in some sense it’s like every ad you’ve ever seen for like a casino where it’s just like people celebrating winning.

Kevin Roose: 57:36 Right, just making it rain, tossing $100 bills in the air.

Casey Newton: 57:40 Yes, but it turns out that you do not always win on PolyMarket and you are not always happy and tossing money everywhere.

Kevin Roose: 57:44 But I love that, first of all, this was a really great story by the Journal. And second of all, what I love about it is that it shows the actual sure-fire way to make money on PolyMarket which is to make a video as a creator in which everyone is lying about everything, and then you’ll really rake it in.

Casey Newton: 58:03 The other guaranteed method for making money on PolyMarket, insider trading. Uh, but that one doesn’t play as well on camera.

Kevin Roose: 58:12 Okay, we have one more prediction market story this week. This one was broken by our colleagues at the Times, Mike Isaac and David Yaffe-Bellany. This one is about Meta. The headline is Mark Zuckerberg directed Meta to create a prediction markets app. Mike and David report that after seeing the success of prediction markets, Mark Zuckerberg got the bright idea to build his own. Two people with knowledge of the matter at Meta said that Zuckerberg had recently dispatched a small team to create a smartphone app similar to PolyMarket and Kalshi, where instead of wagering real money the app would rely on a kind of video game-like fake money point system. But the company, they reported, had not ruled out the eventual use of real money betting. The app is internally called Arena and would be functioning independently from Meta’s existing social networking apps. Casey, what did you make of the Meta Arena story?

Casey Newton: 59:04 So this was a great scoop and I think the truth is I don’t know how big of a reaction I should have to this because Meta does stuff like this all the time. Whenever a hot new thing comes out, they spin up their own version of it. Sometimes it works like with Instagram stories or with Reels, which was its answer to TikTok. Many, many other times it just like completely fails and flops. Um, I don’t know what is going to happen here because as we’ve just been talking about, these prediction markets are becoming extremely popular. And I imagine that the version of it where you can’t lose real money will be less popular weirdly than the one where you can lose all your money. Um, but I suspect that if it is successful, one, they will absolutely do real money wagering. I think they’re probably salivating over that prospect. And two, it will come to Meta’s social networking apps and you will see, you know, little carousels for this in your Instagram and your Facebook.

Kevin Roose: 59:54 I love this, like imagine you’re on Instagram and you see a post of your friend with his new girlfriend and right below it there’s like a…

Casey Newton: 1:00:00 market for like how long will this relationship last? Totally. Totally announcing your engagement and then it’s like, what’s the two-year odds on this one gang?

Kevin Roose: 1:00:11 Yeah, no, that is so perfect and they would absolutely do it. You know, the ethos at Meta, and I know this from speaking to them, is they believe that if people are doing it, then we should respect that they’re doing it and let them do more of it and encourage them to do more of it. And they almost don’t care what that thing is. But if you want to make a bet, if you want to make endless bets and there’s a chance you’ll talk about it on Meta properties where they can show you an ad and maybe even take a cut of your prediction, they’re absolutely going to do it.

Casey Newton: 1:00:33 I just love that this is a company that has been mired in lawsuits about the addictive nature of their platform, especially for young users.

Kevin Roose: 1:00:43 And is currently mired in dozens of these lawsuits around the country. It’s like, you know what sounds like a really great next focus area for us? Gambling. Gambling, the famously addictive behavior.

Dana Suskind: 1:00:52 I was laughing though because there was a story, you know, after you guys had the scoop that came out and said that, you know, some Meta employees were really concerned, you know, they’re really disheartened to learn that their company was exploring a prediction market app of all things. And I was like, what? Oh, is this somehow going to interrupt you from the important work of sending push notifications to 13-year-olds in the middle of the night telling them to check Instagram? Do we really need to protect that from the prediction market?

Casey Newton: 1:01:26 Yeah.

Kevin Roose: 1:01:27 Yeah.

Casey Newton: 1:01:28 Yeah, I you know, if kids you know spend too much time on these prediction markets, Kevin, they might not have time to get an eating disorder. So we really gotta watch the tradeoff.

Dana Suskind: 1:01:35 Wait, wait, you are laughing but that is literally something that the CEO of Kalshi recently said.

Kevin Roose: 1:01:40 Wait, really? What’d he say?

Dana Suskind: 1:01:43 Tarek Mansour the CEO of Kalshi recently literally compared gambling on Kalshi and other prediction markets as the cure to Instagram addiction. He said the way to think about it is the more that people spend time on Kalshi and prediction markets, the less they’re spending time brain rotting on Instagram. And Mark Zuckerberg heard that and he said, ‘porque no los dos?’

Casey Newton: 1:02:18 The cure to one addiction, Kevin, another addiction.

Dana Suskind: 1:02:23 Anyway, I’m sure this will end well.

Kevin Roose: 1:02:29 Mm-hmm.

Dana Suskind: 1:02:30 Yeah, well, everyone involved I would just, you know, start your preparing for your testimony to Congress that you’ll be giving eventually. Probably not soon, but eventually.

Kevin Roose: 1:02:47 Oh.

Casey Newton: 1:02:49 And that’s something you truly can bet on Kevin, but I wouldn’t use real money.

Kevin Roose: 1:02:54 All right, that is against all odds, our new and possibly final installment of our prediction markets roundup.