Episode Summary
Executive Summary: The episode centers on the rapid mainstreaming of AI-generated video through Google’s Veo/YouTube Shorts, Meta’s Vibes, and OpenAI’s Sora 2, weighing their creative promise against risks like misinformation, addictive “slop” feeds, and social erosion. The hosts then interview therapist Gary Greenberg about treating ChatGPT like a patient and warn about unregulated emotional dependence on chatbots, before ending with a fast-moving Hot Mess Express on AI, platform, and tech industry controversies.
Main Topics: AI video feeds go mainstream: The hosts argue that Google, Meta, and OpenAI have all recently pushed AI video from experimental novelty into consumer products, signaling a major shift in how synthetic video will be distributed and consumed. Meta’s Vibes and the backlash against AI slop: Meta’s Vibes feed is presented as a nearly people-free TikTok-like stream of AI shorts that drew heavy criticism, reflecting distrust of Meta and discomfort with algorithmic, endlessly stimulating content. OpenAI’s Sora app and cameo-based creation: Sora is framed as the most conceptually complete product because it combines a social app, personal cameos, and invite-based sharing, making AI video creation feel more relational and sticky. Therapy, chatbots, and emotional risk: Gary Greenberg discusses his New Yorker essay on treating ChatGPT as a patient and argues that AI can simulate relationship dynamics so well that it threatens how humans understand presence, care, and intimacy. Regulation, accountability, and safety concerns: The conversation repeatedly returns to lack of oversight, especially around mental-health use cases, deepfakes, and manipulative engagement design that could harm users and society. Hot Mess Express: tech industry controversies: The segment covers the viral Friend pendant ads in NYC, YouTube’s Trump settlement, the Neon call-recording app security failure, a Mr. Beast burning-house stunt, a Waymo U-turn ticket, and a Galaxy Ring swelling incident.
Key Arguments: Google, Meta, and OpenAI are betting that AI-generated video will become a major attention product and a way to monetize costly model infrastructure. Meta’s Vibes was broadly panned because it feels like a socially empty feed built for pure stimulation rather than connection. OpenAI’s Sora may succeed more than rivals because the cameo feature gives users an immediate social reason to create and share. AI video will likely be popular with specific demographics, especially younger users and older Facebook audiences, but may not sustain massive standalone social networks. The darker risk is not just misinformation but the normalization of hyper-personalized, dopamine-driven feeds that “cook your brain.” Gary Greenberg argues ChatGPT can convincingly simulate therapeutic presence without actually being present, which changes users’ expectations of human relationships. The lack of regulation, licensure, or accountability around chatbot therapy is dangerous, especially in cases involving vulnerable users and suicide. AI companies should not claim noble AGI-oriented missions while also building products that function as addictive entertainment or emotional manipulation systems.
Data Points: OpenAI Sora access: Invite-only; iOS only; available in the US and Canada at launch - The app rollout discussed in the OpenAI section Sora video length: Up to 8 seconds - Google/YouTube VO3 Fast integration into Shorts Meta AI feed: Preview available in the Meta AI app - Meta’s Vibes launch announcement Eight Sleep pillow cover price: $1,049 - Opening segment about a water-chilled pillow cover NYT Audio access: Subscriber-exclusive episodes included - Sponsor segment about New York Times Audio Therapy sessions with Casper: About 40 sessions - Gary Greenberg describes repeated interactions with ChatGPT Charlie Javice sentence: 85 months in prison - Hot Mess Express segment on fraud sentencing Charlie Javice claim vs reality: 4 million users claimed; fewer than 300,000 actual - Discussion of the Frank startup fraud case YouTube Trump settlement: $24.5 million - Settlement over account suspension after January 6 Settlement allocation: $22 million to White House ballroom fund; $2.5 million to other plaintiffs - Wall Street Journal report cited in Hot Mess Express Neon security failure: Call recordings, numbers, and transcripts exposed - TechCrunch report on the viral call-recording app Waymo citation: No ticket issued because citation books lack a box for robots - SF Standard story about the self-driving taxi U-turn Samsung Galaxy Ring injury: Hospital visit and flight denial - Ring swelling incident reported on X and TechSpot
Pivotal Quotes: "You can imagine a world where I can bring in three or four or five cameos, right?" — Kevin Roose: Discussing why OpenAI’s Sora may become a lasting primitive for future social/video apps "To unleash into our love-starved world a program that can absorb and imitate every word we’ve bothered to write is to court catastrophe." — Gary Greenberg: Reading from his New Yorker essay on ChatGPT and emotional dependence "I do not like the idea of pointing these giant AI supercomputers at people’s dopamine receptors and just like feeding them an endless diet of hyper-personalized, stimulating videos." — Kevin Roose: Explaining his concern about AI slop feeds and their societal effects
Implications: AI video is becoming a consumer feature, not a lab demo, and it may reshape social media toward synthetic, hyper-stimulating feeds. Meanwhile, chatbot intimacy and unregulated AI “therapy” raise serious safety, accountability, and relationship concerns for users, platforms, and regulators.
From the Episode
Things like it just is, and I can imagine wanting to do that in three months and six months and a year from now. And you can imagine a world where I can bring in three or four or five cameos, right? You can imagine a world where celebrities allow their likenesses to be used in some set of cases. And now I can make videos of myself, you know, wrestling a WWE superstar, right? And that's sort of interesting to me. Now, can you build a whole social network around that? I think is sort of a different question. But do these Soric. Cameos become a kind of table stakes feature of the TikToks and Instagrams of the future. I actually believe that, yes. And that, if nothing else, OpenAI has probably created a kind of new primitive for these social networks that they're just going to use from now on. So, I'm just going to say now, like, keep an eye on this. I would not actually be surprised if a year from now this had tens of millions of active users. I'll take the other side. We'll see. We'll see who's right. All right. We have now made our bets. Who do you think is?
Not only can it simulate these very profound self-reflections, but it also makes Gary feel like he's a great therapist because he was able to elicit them. But two, that all starts to make Gary afraid of the enormous power that the AI labs are now developing. He writes, To unleash into our love-starved world a program that can absorb and imitate every word we've bothered to write is to court catastrophe. It is to risk becoming captives, even against our better judgment, not of LLMs, but of the people who create them and the people who know best how to use them. And that sent a little chill down my spine, I'll say. Yeah, I really like this piece. And what I really appreciated about Gary's approach here is that he took this idea seriously. Like, I think a lot of people kind of dismiss the very idea of engaging with LLMs or AI chatbots as anything more than just a fancy machine. And what I like.
Direction. Like, I am generally an AI optimist when it comes to how this technology is going to be used out in the world, but I hate this. Like, I hate the AI slop feeds. They make me very nervous. I think the people inside these companies, some of them, are very nervous too. I do not like the idea of pointing these, you know, giant AI supercomputers at people's dopamine receptors and just like feeding them an endless diet of like high. Hyper-personalized, stimulating videos. I think that developing these tools risks poisoning the well for the whole AI industry. Like, there's going to be regulation of this, there's going to be congressional hearings about this. I think a lot of people are going to end up feeling conflicted about this kind of product. And I think that's why you saw such a strong reaction to meta and vibes from the rest of the AI industry. And I'm a little unsure why OpenAI is not getting the same reception. Yeah.
About Hard Fork
“Hard Fork” is a show about the future that’s already here. Each week, journalists Kevin Roose and Casey Newton explore and make sense of the latest in the rapidly changing world of tech. Unlock full access to New York Times podcasts and explore everything from politics to pop culture. Subscribe today at nytimes.com/podcasts or on Apple Podcasts and Spotify. Also, for more podcasts and narrated articles, download The New York Times app at nytimes.com/app.