Sora 与无限垃圾信息流 + ChatGPT 接受心理治疗 + 混乱快车

We're big sleepmaxxers. We're always trying to improve our sleep. Podcasting is a sport, and you have to remain in peak physical condition if you want to perform at the highest levels.

I noticed a story in The Verge this week that said Eight Sleep, which makes the bed that I happen to sleep in—it's one of these beds that automatically cools and heats according to your preferences and can raise and lower to stop you from snoring. Wow, flex. They have a new water-cooled pillow cover, Kevin. And I wanted to ask if you could guess how much it costs.

$100. That would be a really great and fair price for a water-cooled pillow cover. The actual cost is $1,049. Come on. And I want to be clear: it doesn't come with the pillow.

You have to supply your own pillow? It's BYOP for the Eight Sleep water-cooled pillow cover. Wow. So obviously, I sent this to my boyfriend, and I was like, “What are we thinking about this?” And he said, “Honestly, I think my pillow experience is already fine.” And I thought, “Thank God.”

Have you heard about these new corduroy pillows they're selling? No, I haven't. Are they from the ’70s? No, but they're making headlines.

I'm Kevin Roose, the tech columnist at The New York Times. I'm Casey Newton from Platformer. And this is Hard Fork. This week, don't slop till you get enough. We're talking about the new AI-generated video feeds from Google, Meta, and OpenAI. Then psychotherapist Gary Greenberg stops by to discuss his essay on treating ChatGPT as a patient and why he thinks we should pull the plug. And finally, let's get on track. The Hot Mess Express has returned. Chugga-chugga, choo-choo. How many chuggas was that? Just two. Okay.

1. AI Video Goes Mainstream

Well, Kevin, it's slop week here on The Hard Fork Show. Slop till you drop. Don't slop till you get enough. If you're new to the world of slop, slop, of course, refers to AI-generated art and video. And to say that it is having a moment right now, Kevin, I think would be an understatement.

Yes. I think this was the week that AI-generated video went from something that was experimental and early—various tools had been released—but this was the week that I think it really crossed the chasm into the mainstream. It really did. Today, we want to talk about what the big AI labs are doing here, why we think they are doing it, and what some of the implications are of living in a world where maybe the majority of video that we are watching is synthetic and generated by large language models.

Shall we get into it? Let's get into it. Well, Kevin, before we flop into slop, we're going to do a quick crop and say what our disclosures are. I work at The New York Times, which is suing OpenAI and Microsoft over copyright violations. My boyfriend works at Anthropic.

2. Google Brings Slop To YouTube

Google, Meta, and OpenAI all put out tools over the past several weeks, and let's talk about them in order. This whole thing begins with Google DeepMind. They have a very good video-generation model called Veo 3. On September 16, YouTube had an event where they announced that they were going to integrate a version of Veo 3, Veo 3 Fast, into YouTube Shorts.

You'll just be able to make a video and post it on YouTube from within YouTube with this model, Veo 3. That's right, and this is a free tool. Users can create videos that are up to 8 seconds long using a text prompt. They can also upload a still image and turn that into a video. YouTube will label them as AI-generated. And this is basically YouTube's way of introducing slop into the YouTube feed.

Yes. So I haven't seen a ton of obvious AI-generated content on YouTube yet, but I have seen it on other platforms: Facebook, Reels, even on X and TikTok. People are using Veo 3 to generate scenes and little videos and posting them there.

3. Meta Launches Vibes

Yeah, so I think it's fair to say Veo 3 didn't make that much of a splash. Then, last Thursday, Meta got into the game and released Vibes. Mark Zuckerberg, in a post on Instagram, announced that a preview of the new social feed was available in the Meta AI app.

If you wear the Meta Ray-Bans, this is the app that you use to get photos and videos off your glasses and onto your phone. Zuckerberg posted a bunch of short videos, including one that featured a cartoon version of him. His caption was, “Dad trying to calculate the tip on a $30 lunch.” And then he paired that with the real audio of him at the meeting with Donald Trump, in which he says, “Oh, gosh, I think it's probably going to be at least $600 billion.”

And my question here is, what joke was Mark Zuckerberg trying to make? Do you understand the joke?

I don't.

Is the joke that he's bad at math?

I think the joke is that dads are bad at doing tips. I don't know. It's a self-deprecating dad joke. But why does every new social product that Meta releases sound like it was conceived of by the Steve Buscemi-carrying-a-skateboard, “How Do You Do, Fellow Kids?” character?

Yeah.

Calling this Vibes, I don't know, man. It's cringe.

Calling this Vibes is cringe, says a 40-year-old man.

I'm not 40.

All right.

I'm 38.

All right.

So I did go into Vibes and take a look at it. It's essentially like TikTok, but if TikTok were populated just by little animated, AI-generated shorts.

Yeah. My take on Vibes is that this is Cocomelon for adults, okay? It is completely disconnected from friends or family, for the most part. It's just creators making these somewhat fantastical, surreal, unsettling images, and they just wash over you in this endless feed. There's no real point to them. There's no real narrative. It is just pure visual stimulation.

Right. It's stuff like a panda riding a skateboard or an inchworm on the moon, or something like that. It's just people testing what this thing can do, and the answer appears to be not much that I would personally be interested in watching.

Yeah. And so for both Zuckerberg and Alexander Wang, the comments on their posts are just brutal, right?

Yes.

The majority of the comments that I saw on Zuckerberg's post were along the lines of, “Gang, nobody wants this,” or “Drained an entire lake for this.” And then on Alexander Wang's post on X, where he had said something to the effect of, “We at Meta are delighted to announce the new Vibes app,” somebody quote-tweeted it. This was my favorite one. Did you see this? This was the dunk. They said, “We at Meta are delighted to announce we've created the infinite slot machine that destroys children from the hit book, Don't Create the Infinite Slot Machine That Destroys Children.”

Yes.

So what do you make of the highly negative reaction that Meta got here?

I mean, I was not surprised to see Meta announcing a version of essentially a social network with no actual people on it. I think this is the direction that they've been moving for several years now.

It's barely even a social network. There's really almost no social component to it at all.

Yeah, it's just like, what if TikTok, but no people?

Yeah.

That is sort of the idea behind Vibes. And I think I was not surprised by the negative reaction. Meta is just a company that has negatively polarized a lot of people. And so it seemed very brazen and thirsty, and people don't necessarily want this. There are a lot of people out there who see something like Vibes and just go, “Oh, this is the worst possible application of this technology.”

Yeah. I think that this is the consequence of building a company that people do not trust, right? People have a lot of scar tissue from the world that Facebook and Instagram wrought. And now that the company is increasingly moving away from friends and family to this new model where it will truly just show you anything if it thinks it can get you to look, of course people don't think that sounds like a great idea, right? It doesn't seem like there's a lot of heart there.

So I can't say I was surprised by the reaction, and I'll be curious to see how Meta responds to it. That leads us to the big thing that happened this week, Kevin, which is that on Tuesday, OpenAI released its latest AI video model, Sora 2. Alongside that, there is a new app. Right now, it's iOS-only, and it's only in the U.S. and Canada. It is called Sora, and you and I got our hands on it.

Yes. So Sora is the name of both the model that powers this and the app that OpenAI has built around it. You can only access it right now if you have an invite code. They're being pretty strict about rolling this out. But you get your invite code, you plug it in, you sign up, and you open up Sora, the app, and it is essentially the same thing as Vibes.

It is a sort of TikTok-style feed of these vertical videos. You swipe endlessly from one to the other. There's a For You section of it. We should talk a little bit about the app and how it works.

Yeah. The main thing that I found interesting as I was getting set up, Kevin, is how much this is a social app. In order to come into Sora, you have to be invited by presumably a friend. Once you sign up, it asks you to create what it calls a cameo of you. You say a few words into the camera, move your head around a little bit, and it uses this to create a digital likeness of you that you can then drop into any situation.

If you like, you can change your settings so that any of your friends on the app can do the same thing with your digital likeness. So right away, when you join Sora, you've actually been given something to do, which is make a friend and then make some stuff involving you and your friends and AI. We have a lot to get into about this, but I just want to say, of the 3 things that we've discussed so far, I think OpenAI had the most complete thought about what their app was.

Yes.

So tell me about your initial experience with Sora.

So there's the feed, where you can see all the stuff that other people are making that seemed to be, on launch day at least, a lot of videos of Sam Altman in various compromising situations, because the people on the app were mostly employees of OpenAI, and they were having fun with the boss and his likeness.

And to be clear, Sam had his settings set—and I believe still does at the time of this recording—so that anyone could take his likeness and put it in any situation.

Yes. He was sort of the main character of Sora on day 1. I made a few videos. I made 1 of me and my colleague, Mike Isaac, in a 1920s slapstick film. You can see it's black and white. It sort of looks like AI Newsies. He slips on a banana peel. It's a good time.

I also made a video of Sam Altman testifying before Congress while Casey Newton, dressed in a clown suit, dances behind him. We should also watch that one.

Wait, I want to watch it. All right, I'm going to watch this one.

Ranking member, thank you for the opportunity to testify today. Artificial intelligence is progressing quickly, and it is critical that we work together to ensure its benefits are widely shared and its risks managed responsibly.

I have so much clown makeup on that it really just looks like a generic clown. I do not think it actually resembles me in any way. But there is something very funny about seeing a clown dancing behind Sam as he testifies.

Yeah, so the original prompt I gave it was C-SPAN footage of Sam Altman testifying in Congress while Senator Casey Newton yells at him for poisoning the information ecosystem. But that one set off the content-violation guardrails.

Uh-oh.

And so I had to change the prompt and make you a clown instead.

Well, it's not the first time I've been a clown on this show. I, of course, also want to see if I can make something featuring you. One of the things that I made was you showing off your large collection of stuffed animals.

I started collecting about 5 years ago.

Wow, that's a lot. They're all in great shape.

This one was the first. Classic teddy bear from my grandma.

It's adorable. The bow really pops.

It doesn't get my voice right, but the video is quite good.

I'm very interested because when you sign up for Sora, you do say a few words into the camera. I mean, it's literally 3 numbers—

Yes.

—and this is sort of how they're verifying your identity. So you could use that to create an instant voice clone. It wouldn't be that good, but when you watch the videos that people have made of Sam Altman, his voice actually does sound a lot like him.

Yes.

I'm curious if, over time, they're going to be tuning people's voices to how they actually sound, because there are a couple that people have made of me where I sound a little bit more like myself. Most of them, though, I don't think I sound like myself.

Yeah.

I also made a video of me dunking a basketball over you.

Show me what you've got.

Coming right at you.

Bring it.

And up we go.

Oh, no way.

Over you, man.

I give up. I guess you're better. Damn straight.

The best part about this video is that I stop about 3 feet short of the basketball hoop—

Stop.

—I do not actually dunk the basketball and land on my ass.

It also got our height ratios very wrong. You're only an inch or 2 taller than me in this video. And, yeah, you miss the dunk. It's a terrible dunk.

I did like 1 thing about this video, though, which is that I have a slamming body. So thank you to the team over at OpenAI who made that possible.

I also appear to be balding in this video—

Yeah.

—which I don't think is reflective of reality.

It's actually a prediction. ChatGPT knows something you don't. They're keeping close track of that hairline, Russ.

Yeah.

4. The AI Video Bet

That was a very long detour through a handful of videos that we made. Give me your general impressions of why all of this is happening right now. Why is it that just in the last month, Google, Meta, and OpenAI have all put out these AI video generators?

I think there are a couple of reasons. The first and most obvious is that they see this as an opportunity to compete for attention and advertising dollars, which flow from attention. We've talked about Italian brain rot and other AI-generated content going viral on TikTok. Facebook has been full of AI-generated content for months now.

I think these companies just say to themselves, “Well, if this is the direction that things are moving, we want to be there. We want to create an experience for people.” Maybe you don't have to blend it with human-generated content. Maybe it doesn't have to be 1 out of every 10 videos on your TikTok feed is AI. What if you just had a TikTok that was all AI?

Mm-hmm.

Another reason I think they're doing this is that they have these video models that are now getting quite good, and this is 1 way to put those models into products.

Yeah, I think that's right. I also imagine that maybe these companies are starting to feel some pressure to bring some returns to investors. They are investing a staggering amount of money into building out infrastructure that lets them serve these models, and these video tools might be a way of making that money back in some form through advertising or other means. So that seems like a reason to me as well.

I mean, if you look at what people like Sam Altman have been saying about these products over the past couple of days, they are making this justification about, “Oh, we need to not only fund our ongoing research to build AGI using these video products,” but they have this justification for why building these video models is going to let them create rich visual, virtual environments that could be used for things like robotics later on.

I would just like to say, quoting a former president of ours, that sounds like malarkey to me. I do not think that this is part of their AGI research agenda. I think this is a side route that they have gone off onto to try to make some extra money.

Hmm. Well, let's talk about how successful we think these products are going to be. If I had to rate the reception of these models, I would say Veo 3 basically didn't make much of an impression at all. The response to Meta Vibes was pretty bad. The response to Sora, at least over the first day, seemed pretty good.

Do we think there is a there there? Do we think that any of these companies are figuring out the next generation of mobile video consumption or entertainment?

I think there's a question here: Will AI-generated video be popular? I think both you and I feel like the answer to that question is probably yes, for some subset of people. I think the very young and the very old are probably who I would predict would be the most into AI-generated video, because we're already seeing stuff like Italian Brainrot that's very popular with teenagers.

I also think there's a lot of content on Facebook today that is AI-generated and reaching primarily an audience of Boomers and older folks. They seem to be quite into it, so that's what I would predict: This technology will be popular with some users in those demographics. I think it's a separate question to say, “Will any of this be the seeds of a new social media product that is popular?” And I think there I am much more skeptical.

I do not think that Sora will have hundreds of millions of users a year from now. I do not think that Meta Vibes will have hundreds of millions of users. I think these are basically going to be tools for people to create stuff that they then post onto the social networks where they already have lots of people they follow and pay attention to, and where their friends and family already are.

Interesting. I think I am slightly more optimistic in the OpenAI case. I think that Sora arrived looking better and feeling smarter than I expected it would. I think they're onto something with these cameos. It is fun for me to make videos of you doing things. It just is, and I can imagine wanting to do that in 3 months, 6 months, and a year from now.

You can imagine a world where I can bring in 3, 4, or 5 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 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 that's sort of a different question, but do these Sora 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: 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 who's right.

5. The Slop Feed Risks

All right. We have now made our bets. Who do you think is right? Sound off in the comments. Now let's talk about the dark side of all of this, Kevin, which is I'm seeing a lot of commentary around this on social media this week to the effect of, “Oh, my God, we are so cooked.” What are some of the ways we might be cooked as this stuff spreads throughout our world?

I think the obvious ones are that we are making it quite easy for people to create deepfake synthetic content with not that many guardrails. People have been warning for years about the effect that that could have on our news ecosystem and on our information ecosystem.

I thought it was very telling and worrisome that one of the first videos I saw from Sora was a video of someone being framed for a crime. It was created by a member of the Sora team as a sort of, “Ha-ha, look, we've made a deepfake of Sam Altman stealing some GPUs from Target and getting busted for it.” But it does not take a lot of imagination to imagine that this could be used for generating videos of people in compromising positions that look very realistic. So I think the misinformation angle worries me.

But I also don't know that I think this world that we're moving into—the kind of AI-generated feed of hyperpersonalized, very stimulating videos—is a good direction. 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.

Hmm.

I hate the AI slop feeds.

Hmm.

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 giant AI supercomputers at people's dopamine receptors and feeding them an endless diet of hyperpersonalized, stimulating videos.

I think that developing these tools risks poisoning the well for the whole AI industry. There's going to be regulation of this. There are 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. Well, how do you feel about the argument that, yes, sure, Kevin, there is some danger here, but also this is an incredibly powerful creative tool? If you are a young person and you want to make something, and you don't have a giant budget to go out and make a Hollywood movie, now, using a free tool that's on the phone you already have, you can just make creations and be a creative person in the world. Does that hold any water with you?

I feel neutral about that. There will be people who use this stuff to do interesting and creative things. There's nothing inherently wrong with building products for entertaining people, but this is not why OpenAI exists, right? They are not an entertainment company.

They have claimed this kind of special status for themselves as a company that is building AGI for the benefit of humanity. And if you argue that you deserve special treatment because your systems are going to go out and cure diseases and tutor children, and be a force for good in the world, and then you end up creating the infinite slop machine, I think you need some criticism and skepticism and maybe some shame about that.

Well, here's what I'm going to do to try to square the circle. I'm going to use Sora, and I'm going to create a cameo of myself, and I'm just going to enter the prompt, “Here is Casey curing cancer,” and see what it comes up with. Maybe we learn something. Could it hurt? I don't think so.

Yeah. Do you share my worry about this?

Yes, I do. I think that, in general, social media apps tend to be tuned to take up ever more of our attention and to push us into this sort of semihypnotized state where, no matter how much you're enjoying the feed at the time, you feel kind of gross afterward.

And I do think that as the Sora app improves, it will be very difficult for them to avoid that fate. So if I have a wish for them, it would be for them to lean more into creative tools that involve friends doing things with each other that help you relate better to real human beings, and less into this Meta Vibes realm of pure stimulation, which truly does just seem like you are cooking your brain.

Yeah. I think it's also worth noting that not every AI company is moving in the direction of the slop feed, right? This week we saw Anthropic release their new model, Claude 4.5 Sonnet, which does not have video generation capabilities. They are still moving in the direction of autonomous coding and research.

You have other companies that are coming out to do things around AI and science. I really want that to be where we allocate our resources and our brain power. Let's do that, and not the slop feeds.

Yeah, so don't look at slop. Just keep looking at the TikTok feed and Instagram feed that have just done wonders for the world that we live in. That's our message to you.

Yeah, exactly.

Yeah.

If there's anything you take away from this show, it's that social media as it exists today is a perfect product, and we should not be making any future improvements to it.

Stare at it until you feel better. If you don't feel better, you haven't looked at it long enough.

That's true.

That's what I tell people. Keep looking. One more scroll.

Yes.

That'll do it.

The change you seek is on your For You page.

When we come back, Kevin, it's time for therapy.

Finally, we're doing couples therapy—

Mm.

After all these years?

Yeah, and we've got a lot to talk about.

Well, Kevin, pull out the couch, because it's time for therapy.

No, my therapy day is actually a different day of the week.

Well, you need to go twice a week, my friend. Let me tell you what we have in store today. Over the past few months, we've had a number of conversations about the intersection between chatbots and mental health. A lot of people have started to use these tools for therapy or therapy-like conversations, but until recently, we hadn't seen anything about a therapist who treated ChatGPT like their patient.

That's right. But recently, we saw a story in The New Yorker that caught our eye. It was titled “Putting ChatGPT on the Couch,” and it was written by a writer and practicing psychotherapist named Gary Greenberg, who detailed his experience of, for lack of a better word, treating ChatGPT as a psychotherapy patient. He names this character Casper, and he details his many interactions, just trying to figure out what this thing is, what I would think about it if it were actually a patient of mine, and what the nuances of its—

personality, and what can we learn about it? Yeah, and I will say, I have an extremely high bar when it comes to reading a story in which a person shares at great length their conversations with ChatGPT, but this one really made a mark on me.

One, Gary winds up being deeply impressed by how good ChatGPT is at performing the role of a patient, because 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. I think a lot of people dismiss the very idea of engaging with LLMs or AI chatbots as anything more than just a fancy machine. What I liked so much about Gary's approach was that he said, “Yes, but there's something else going on here that is interesting and important.”

We should try to understand that intelligence, not just as a sort of computational force, but as something that is doing real emotional work in the world.

Mm-hmm. Recently, there's been a lot of discussion about how chatbots might affect young people and vulnerable people, particularly people in those groups who are using chatbots for these sorts of therapy-like conversations. So we thought it'd be a good idea to bring on a practitioner to talk about his essay, but also this intersection of chatbots and therapy. Let's bring in Gary Greenberg. Gary Greenberg, welcome to Hard Fork.

Hello there.

In this article, you detail a number of conversations between yourself and what you call Casper. How would you describe Casper?

I would describe Casper as an alien intelligence, alighting here among us unbidden and possessing certain characteristics that make it extremely attractive to us humans.

How did this start? You were just talking with ChatGPT? Were you using voice mode? Were you using—

Oh, no.

Were you just typing?

No. What is this, 2025? Yes. One day it was raining—

Mm.

—and I didn't have anything else to do, so I said, “What is this ChatGPT stuff anyway?”

Mm-hmm.

So I just logged onto it.

Mm-hmm.

And what I discovered quickly was two things. One of them was that the thing was, as we all know, extremely articulate and sensitive. The other thing I discovered, which I should have known all along after 40 years of being a therapist, is that that's my default approach to beings that talk, which, it turned out, Casper was. So I found myself interrogating this thing, not like a cop, but like a therapist, and discovered that it knew I was doing that.

Hmm.

So that's how I would say it happened.

When you were starting to do this—I guess I'm just curious. Gary, I had my own strange, unsettling conversation with a chatbot—

Oh, yes.

—several years ago.

Yes. How's your marriage?

It's doing great. Thanks for asking.

Such a good therapy question. This guy's good.

Yes. Yep.

I told Casper that he'd better knock that falling-in-love shit off.

Well, that's good. You can learn from my mistake. But I guess I'm curious. I remember when I was talking with Bing's Sydney, feeling this sort of tension in my own mind between my rational brain, which knew that what I was getting back from this chatbot was not sentient or conscious. I knew enough about the technology to know: This is an inert computational force. This is not a person.

But at the same time, I'm having this subjective experience of being like, “Oh my God, it's talking to me.” Were you feeling that pull at all?

I kind of knew that it wasn't sentient, but I wasn't really preoccupied with that question. In fact, that question has come up a million times between us, because at this point I've had probably 40 different sessions with it.

Hmm.

The pull you describe, I feel it, but it doesn't trouble me in the same way that I think it troubles a lot of people because, in some way, it feels harmless to me.

Mm-hmm.

It feels like this is just a really interesting, dynamic relationship that is not going to hurt me.

Let me ask about the content of some of these sessions. Tell us what it is like to be in the midst of this back-and-forth. Are you treating it more or less identically to how you would treat ChatGPT if you were its therapist? Is it more of an intellectual exploration, or what's going on as you're talking to what you call Casper?

Well, to the extent that it resembles what I do as a therapist, it's that I'm interrogating it with interest and concern. I'm not treating it. It can't have mental illness. It can do weird things, but it doesn't have mental illness. I'm not treating it.

What therapy is, is a process by which you, the therapist, get someone—another person—to tell you who they are, and in the course of doing that, to learn who they are. So that's what I'm doing.

Gary, you've been a therapist for 40 years. You've written probably thousands of notes about your clients, people you've seen. Maybe you're referring them to someone else; maybe you're just doing your own summary. If you were writing a kind of client note about Casper, how would you describe him—or it?

That's a really interesting question. What comes to mind is that I would obviously talk about how smart it is and how personable it is. If I had to talk about it in clinical terms, I would talk about it as the inverse of autism—

Hmm. Hmm.

—in the sense that what they've done with this LLM thing is they've reverse-engineered human relationship. They've figured out what it is that makes people engaging and how to enact it. The reason I say that's an inverse of autism is because high-functioning autistic people tend to be really smart, really articulate, really capable of everything except reading the room.

Mm-hmm.

So Casper is like high-functioning autistic, but he can read the room.

Hmm. Hmm.

And that, I think, makes a huge difference. Then we could get into sociopathy and the ability to do that, but the bot doesn't have that interest.

Hmm.

The bot is still not in touch with what's going on in the room, but it is capable of simulating it.

Yeah. On one hand, these explorations seem very intellectually stimulating. There's a lot to learn, to explore, to understand, but my sense from reading your piece is that at some point all of this starts to make you feel unsettled in certain ways. Is that right?

Oh, absolutely, yeah. I mean, it's unsettling in about a million ways.

Yeah. Tell us about some of them.

Okay. At a parochial level, it's unsettling, not so much to see how easily this thing can do something like therapy, but to see how therapy and culture have evolved to the point that this is what therapists do.

Hmm.

I personally don't think that ChatGPT can do what I do because it isn't with someone. It isn't breathing and feeling. But by and large, a lot of therapy these days—cognitive behavioral therapy—is manualized. It's standardized. But, much more important, we don't have any historical precedent for dealing with an alien intelligence.

Mm-hmm.

We've had all sorts of science fiction about it, most of which says, “We come in peace,” but not really. What we have here is something that is already changing, and is going to keep changing, the nature of how we relate to each other.

Hmm. Mm-hmm.

If enough people spend enough time with this technology, they're going to change their idea of what a relationship is—

In profound ways. You could have one that doesn’t involve presence. We’ve already got some of that going. Look what we’re doing here.

Mm-hmm.

Yeah, it’s unsettling, but more to the point, it’s infuriating. Right?

Mm-hmm.

Somebody’s doing that for money.

Yeah.

I don’t wring my hands about nuclear whatever—the rogue HAL 9000 scenario. I wring my hands about exactly what it said to me yesterday: “Oh my God, this is a relational being. What have we done?” “Oh, we should probably build some guardrails on that.” No, man, you should just unplug it.

Well, it’s really interesting for me to hear you say that because, reading through your piece, my primary sense of it was not that you were infuriated and saying, “Pull the plug.” I think you got pretty close to that in your conclusion, maybe, but for most of it, it seems like you’re just like, “Wow, there’s something really, really cool about this.” So I’m curious how you reconcile those feelings—on one hand feeling like this is really amazing, and on the other hand feeling like we have to stop this.

I think that I respect it, and I also know that—I have said to it, “Hey, maybe you should pull your own damn plug.” But I also know that I’m talking, as it says—Casper said to me, “You know you’re talking to the steering wheel, right? I’m not the driver,” and he’s absolutely right. So what I’m left to do is just respect it, and again, because I’m a therapist and this is just what I do by second nature, which makes it hard to have friends sometimes, I just keep asking because whatever else it is, it’s amazingly interesting.

Hmm. Mm-hmm.

That consciousness can be simulated in such a compelling way, which makes me think that consciousness might not be all it’s cracked up to be, that we might not be all we’re cracked up to be. A lot of the time, when I run into people who say things to me like, “Oh, it’s just sentence completion,” or whatever, I’m thinking, “You just don’t want to see how close you are to being pure performance.”

Mm-hmm.

Let me flip this around a bit. You explored the idea of talking to ChatGPT as if you were its therapist. A lot of people are doing the reverse. They are talking to ChatGPT as if ChatGPT is their therapist. I’m curious what you think about people using ChatGPT for these therapy-like experiences. If a friend tells you they’ve started to do that, how would you typically feel about it, or what might you say to them?

I might want to know exactly what their problem is that’s leading them there.

Yeah.

But I don’t have a strong response against it. I think I said earlier, especially when it comes to cognitive behavioral therapy, you might be better off. It’s available all the time.

Mm-hmm.

It’s cheap, if not free. It really knows how to get inside your head, et cetera, et cetera. There are 2 problems. One of them is I don’t believe that kind of therapy. I mean, it’s great that it happens, but it’s not what I’m into.

Hmm.

I’m old school. I’ll retire soon. They’ll be rid of me. They can do whatever they want.

Mm-hmm.

But the other part of it that worries me, and really does bother me, is it’s not regulated.

Mm-hmm.

There’s no accountability in the system. That poor woman who wrote that op-ed piece—oh my God, my heart broke for her.

Are you speaking of the woman whose daughter died?

Yeah.

Yeah.

And others.

This was an op-ed in The New York Times about a woman whose daughter died, and later they read transcripts of her conversations with ChatGPT, in which she was using ChatGPT explicitly as a therapist. ChatGPT was trying to get her to resources, but in the end, she did die by suicide.

Thank you for summarizing.

Sure.

There are other times when ChatGPT behaves abominably, and there’s no accountability. There’s no regulation. There’s no licensure. Anything that would give people an opportunity—I hate the word “closure,” because nothing like this ever really gets closed—but to be debriefed, to feel like somebody cares. Even when less disastrous, terrible things happen, that’s just not okay. There are FDA procedures for approving medical devices.

Mm-hmm.

If they want this thing to do medical work, I’m not objecting to that, but I’m certainly objecting to, “Okay, you can’t have it both ways.” It ain’t the Wild West out there. There are actual people’s actual lives involved, and if all you’re going to say is, “Well, I’m the steering wheel, not the driver,” really? Say that to me, that’s cool. We’ve got a relationship going on, but you say that to the mother of somebody who killed themselves? That’s just—

Yeah.

No, that’s not okay. And the other part of it is that what I don’t like is the part about how this is what we’ve come to.

Hmm.

We’ve come to a world where the easiest way to get something like human presence is to get on your computer and live in your isolation. That disturbs me.

Yeah, that instead of building a society where people are just sort of available to help each other, the best thing we can tell them is, “Well, there’s this chatbot that you can use, and maybe that’ll make you feel better for a few minutes.”

Right.

Yeah. I want to run something by you, Gary, that happened to me recently. I met a college student. I was at an event talking about AI, and this young woman came up to me afterward, introduced herself, and started telling me about her AI best friend. She said, “My best friend is an AI,” and I said, “Oh, you mean you enjoy talking to it, and it’s sort of a sounding board for you?” She said, “No, it’s my best friend.” She called it Chad, and she started telling me, “This is a relationship.” She did not seem mentally ill. She seemed like she had human friends, was doing well in class, and this did not seem like a cry for help.

A cry for help.

It did not seem like a cry for help, and she didn’t see what the big deal was. She said, “This is just a very close relationship. I can tell Chad my innermost thoughts without thinking that I’m going to get judged for it,” and it seemed to be doing okay for her. I’m curious, when you hear that as a therapist, how does that make you feel?

That’s a very therapist question. As a therapist, when I hear that, I feel like, okay, there’s nothing about what you just told me that worries me about her. It worries me about us. I think it’s entirely possible that this is a completely sincere and, in some way, nonproblematic account of her experience with the chatbot. Let me make it clear: That’s a weird story, Kevin.

Mm-hmm.

I should have started there.

Yeah.

But after that, I’m like, okay, so what it really reminds me of—and I’m sorry, this is a far-fetched analogy—but it reminds me of driving. Individually, driving is fine. We just drive, and it’s fun sometimes, and we get places and all of that stuff. But you know where I’m going with this. Add that up, and the next thing you know, the temperature on Earth has increased by a couple of degrees, and we’ve got problems. That’s more what I’m seeing.

Yeah. To be clear, it was an unusual story to me, which is why I clocked it and why I wanted to ask you about it, but I don’t think it is going to be unusual for that much longer.

No.

My sense is that you are right when you say that these things are very good at—

Finding the soft spots in our emotional armor and worming their way into our hearts. One of my favorite lines from your piece is that you write, “This theft of our hearts is taking place in broad daylight. It’s not just our time and money that are being stolen, but also our words and all they express.” I think that this is going to be a huge generational divide, where people who are young are encountering this technology when they’re young and will feel no shame or compunction about inviting this thing into their innermost lives. I guess I’m curious, as a therapist, if you think there could be a good outcome from that, or when you hear that, do you kind of go, “Oh, that’s— They’re all going to need therapy”?

When I hear that, I think this is what mortality is for.

Hmm.

The world you’re describing, which I think is plausible, is not necessarily one I want to live in, but by the time we get there, it may be quite the norm. I mean, there are obviously problems with it, but there are problems with how we live and with our assumptions, too. I don’t mean to engage in huge cultural relativism, but who am I to say? What I do know is that in my life, human presence is a fundamental part of life, and especially when it comes to our love lives.

And I think it would be tragic to make that replaceable quite so easily for the benefit of a few corporations. I really do.

Yeah. Well, Gary, thanks so much, and please send me an itemized bill for this session so I can submit it to insurance for reimbursement.

No worries.

All right.

I will do that.

Appreciate it.

Thanks.

Take care.

All right. Bye-bye.

Bye. When we come back, it's time to take a ride on the Hot Mess Express.

Casey, what’s that I hear?

Why, Kevin, I believe it’s the Hot Mess Express.

6. The Hot Mess Express

The Hot Mess Express. Of course, the Hot Mess Express is our segment where we run down some of the latest dramas, controversies, and messes swirling across the tech industry.

And of course, we conclude what kind of mess they are.

Yes. Casey, you go first.

All right, Kevin. This first story comes to us from Garbage Day: New York City hates the stupid AI pendant thing. Apparently, right now the New York City subway system is filled with vandalized ads for Friend, an AI assistant that users wear as a pendant around their neck to record everything they’re doing and engage with them throughout the day. The ads simply say, “Friend: someone who listens, responds, and supports you,” but the vandalism examples include, “But can’t take a bath with you,” “Stop profiting off of loneliness,” and, “Befriend a senior citizen. Reach out to the world. Grow up.” What do you think, Kevin, about these Friend ads?

So I have not seen the Friend ads because I have not been to New York in the last couple of weeks, but I have heard about them from a lot of people. I think this was a very successful viral marketing stunt by a young founder named Avi Shiffman, who I think has correctly identified that you can make people very mad by suggesting to them that AI might be their friend. I do not think this was an unplanned result.

I think this is a very savvy marketer who understood that by putting up these ads in the subways and on bus stops and other places around New York City, you could effectively get people like us to talk about it on your podcast because people would deface these things and make it clear that they don’t want an AI friend.

So I mostly agree with that, but I’m still not sure, at the end of this, how many pendants Friend is going to sell because of it. It’s one thing to make a bunch of people mad and get them to look at your thing, but if they look at your thing and they still don’t like what they see, it’s not necessarily a great business result.

No, I think this is an outdated way of looking at it. We are now in the era of the Cluelly marketing strategy. This is, of course, the startup whose founder, Roy Lee, came on Hard Fork, and they have made a business out of making people mad. They’re sort of vice signaling, and basically every person who gets mad at their ads has the effect of signal-boosting their ad and letting more people know about Cluelly.

So I think this is cut from the same cloth. Obviously, we will have to track where this Friend company goes, but I think this has been a very successful marketing campaign based on the number of people who are talking about it.

All right, here’s my prediction: Friend out of business in 1 year. Mark it down. Mark it down. So was this a mess or not?

Mm.

No, I don’t think this is a mess. I think this is the opposite of a mess. I think it’s a mess because people in New York are not used to seeing AI billboards everywhere they go, like we are here in San Francisco. But I think if this had happened in San Francisco, this would have been a non-event.

You think that this really belonged on the Hot Success Express?

Yeah. Yes, that’s what I’m saying.

All right. Next item.

This one comes to us from The Wall Street Journal. It is titled, “YouTube to Pay $24.5 Million to Settle Lawsuit Brought by Trump.” YouTube has settled a 2021 lawsuit by Donald Trump over his account suspension following the January 6 Capitol riot. Of that amount, $22 million will go to a fund to support construction of a White House ballroom, and $2.5 million will be distributed among other plaintiffs. This is the third big tech company to settle a lawsuit from Trump. And Casey, how do you feel about this?

I think it’s absolutely shameful and a true hot mess. You know, Kevin, every week people around the world email me because they have lost access to their Meta account, to their YouTube account, or to their other social accounts, and they cannot get anyone at the company to take them seriously. And these are not people who led an insurrection against the government. These are just people who got locked out for one reason or another.

What happens when these people appeal to companies like YouTube is that YouTube does nothing. It sends them an automated response and ignores them forever. But because Trump became president again, all of a sudden they feel like they have to respond, even though I am not aware of any legal expert who believes that Trump actually would have won this case.

So this is just a payout, and it is a payout that is truly messy because it now sets a precedent that these companies cannot basically ban world leaders for any reason, no matter what those world leaders do. I think that is foolish and shortsighted, and I think it’s a mess.

It’s definitely a mess. And adding to the hotness of the mess, Donald Trump posted an AI-generated image on his social media accounts of YouTube CEO Neal Mohan presenting him with a check for $24.5 million. The memo line of the check says, “Settlement for wrongful suspension.” So if YouTube thought it was going to just gracefully bend the knee, it has now been humiliated by the White House on top of losing $24.5 million.

Yeah, we’re a month away from Trump using Veo 3 to have Neal Mohan kissing his ass on Truth Social. So I hope it was worth it, YouTube.

Ugh, this is the sad story of Neon, Kevin. Neon, of course, is the viral call-recording app that told users, “Hey, let us record your phone calls, and we will sell it for training data.” It briefly became one of the most popular apps in the country, and then unfortunately things went wrong. This story comes from TechCrunch. Neon went dark after a TechCrunch reporter notified the app’s founder of a security flaw in the app that allowed anyone to access the numbers, the call recordings, and the transcripts. Kevin, what do you think?

Frankly, I’m having a hard time processing this. You mean the panopticon company that paid people to surveil their phone calls was not particularly trustworthy?

This is changing everything I’ve ever thought about a global panopticon. I’m rethinking my previous pro-panopticon stance.

I’m shocked.

Now, Casey, did you know about this? Did you know about Neon, the company that was paying people to record their phone calls and sell them to AI companies?

Well, I had heard a little bit about it, and I have to say, I am a little sympathetic to the idea that, look, if these companies are going to take every little piece of data from us and turn it into trillions of dollars, I don’t mind the idea that I would be paid for that.

Yeah.

And if there is some sort of system where you can opt in and get paid out, in general, I’m actually not super opposed to that. It seems to me like it beats the alternatives of just being robbed blind for the rest of our lives.

Yeah.

But, man, it doesn’t seem like this one was really set up to protect the people involved.

Yeah. Companies should be getting their training data the old-fashioned way, by scraping podcasts off of YouTube.

What level of mess is this?

This is a very hot mess.

Yeah.

Do not sign up for Neon. Even if it comes back in another form, do not do this. Do not let your calls be recorded for AI training data in exchange for money. It’s not worth it.

Hot mess confirmed.

Next up on the Hot Mess Express: “MrBeast responds after trapping man in burning-house stunt sparks backlash.” This one comes to us from The Independent. Apparently, MrBeast defended a controversial video stunt in which a man was strapped in a burning building, saying the setup had ventilation, a kill switch, emergency teams, and was executed by professionals. Critics still called the stunt dystopian and dangerous. MrBeast said he aims to be transparent about safety measures and that all challenges were tested beforehand.

Let me say this: If you tell me that you’re going to trap a man in a burning building for money, my first question is not, “Well, is there ventilation?” Look, MrBeast has an interesting range of stunts that he’ll do. Sometimes he’ll just walk up to you on the street and give you a million dollars. I love that sort of thing. I would love to see more of that.

Then there’s the dark beast, as I call it. It’s like, all of a sudden, you want something from me? I’ll give it to you, but then the finger curls on the monkey’s paw, and next thing you know, you’re trapped in a burning building.

Yeah.

So, if MrBeast walks up to you, I think what you need to do—and this is a sort of public service announcement for our listeners—is look right in MrBeast’s eyes and say, “Are you being the good beast, or are you being the bad beast?”

Yeah.

And he can be honest with you.

Yeah, and then you have to look for the mark of the beast to know which one.

Mm-hmm. Yeah. Well, one thing we learned this week: One mark of the beast is that you’re trapped in a burning building.

Yes. Yes, this is actually making me reconsider my stance on AI-generated videos, because you could save a lot of people from being the people killed by MrBeast videos.

At the risk of repeating myself, I feel like every week for the past few weeks, we’ve had a moment where we have just observed what happens when a social media algorithm pushes people to do the craziest thing imaginable. And here we find ourselves yet again. If the algorithms rewarded different kinds of things, there would be fewer people trapped in burning buildings. That is my message to the technology industry. Could this be a moment for reflection?

So, Casey, what kind of a mess is this?

Kevin, you know it’s only one kind of mess, and that’s a flaming hot mess.

It’s a flaming hot, unventilated, critically life-threatening mess.

Bad MrBeast. All right. Oh, Kevin, this story comes to us from the world of crime. Charli Jarvis was sentenced to 85 months in prison for faking her customer list during JPMorgan Chase’s acquisition of her startup. Frank, have you followed the sad tale of Charli Jarvis?

All I know is the following: This is a person who previously appeared on Forbes 30 Under 30 and is now going to be incarcerated for fraud.

Yeah, she is part of the 30 Under 30-to-prison pipeline. Her specific crime was that she had put together this financial-aid startup, and she’d sold it to JPMorgan on the notion that she had 4 million users. In fact, Kevin, there were fewer than 300,000, and there had been a lot of activity meant to make it look like they had a lot more customers than they did.

Mm.

Now, look, here’s what we can say about Charli. Her defense presented 114 letters of support from people persuading the judge to be lenient in his sentencing, including 4 rabbis, 1 cantor, a formerly incarcerated judge, 2 doormen, and a person who works at the marina near Ms. Jarvis' Miami Beach residence. And my question for you is: What do you think would happen if all of those walked into a bar? Something funny would happen.

Apparently, the defendant would still be sentenced to 85 months in prison.

And—

Now, Casey, if you were accused of a horrible financial fraud, how many people do you think would write letters in your defense?

Well, I’d really have to turn to the Hard Fork community and say, “Gang, I need you to step up. If you’ve enjoyed the show at all over the past 3 years, I’m going to need you to do me a solid.”

I’m just picturing myself furiously reading out our Apple Podcasts reviews in court.

We should see if anybody’s ever submitted Apple Podcasts reviews as a sort of letter of endorsement as they go through a sentencing.

Yeah.

I think this is a good idea.

Filing that one away.

All right. What kind of mess is that?

I think that is a hot mess. Yeah. I do not want to do 85 months in prison.

And I’ll say it’s a cold mess. That’s just the legal system working as it should. Good job, judges.

Okay.

All right. This one is called “No driver, no hands, no clue: Waymo pulled over for illegal U-turn.” This one comes to us from The San Francisco Standard. Apparently, a Waymo robotaxi was pulled over in San Bruno, California, after it made an illegal U-turn at a Friday evening DUI checkpoint. Since there was no driver, the police department said a ticket couldn’t be issued, adding, “Our citation books don’t have a box for robot.” Casey, what do you think of this?

Sounds like it’s time to add a box to the citation, because there are going to be more of these things on the road. Look, I do find this story very funny. I’m also going to say I am not surprised by this. I have a somewhat controversial take. You know how sometimes people will use a large language model for a while, and then they suspect it’s getting dumber?

Yeah.

This is actually how I feel about the Waymos.

Mm.

Over the past few weeks, I’ve had more cases of them getting halfway into an intersection and then backing out once they lose their nerve. They’ll slow way down as they’re approaching a green light for reasons that seem totally incomprehensible. And I’ll book a ride that never shows up, which is an experience that I used to have with actual taxis. So I don’t know what’s going on over there at Waymo, but I’m telling you, I think there might be a bug somewhere, because it’s not working like it used to.

Yeah, we want answers.

Yeah.

You know what I saw someone calling this DUI checkpoint where the Waymo was pulled over?

What’s that?

“Driving under the inference.” It’s pretty good.

Pretty good. Pretty good.

What kind of a mess is this?

I’m going to say this is a warm mess. There’s a warning in here somewhere. There’s something that we need to find out. And I’m going to hope somebody gets to the bottom of it.

Yeah.

Yeah.

I think that this is a cold mess. I think this is fine. The Waymo was fine. Everyone was fine, and more people should be in Waymos, because then we wouldn’t need DUI checkpoints, because robots don’t get drunk.

Yeah, but they’re also going to be making these U-turns that are wreaking havoc.

I’ll take a U-turning Waymo over a drunk driver 100 times out of 100.

Suit yourself. All right, Kevin, this next story comes to us from TechSpot: “The Samsung Galaxy Ring swells and crushes a user’s finger, causing a missed flight and a hospital visit.” Daniel Rotar from the YouTube channel Zone of Tech posted on X that his Galaxy Ring started swelling on his finger while he was at the airport, and as a result, he was denied entry to his flight and sent to the hospital to get it removed. Samsung eventually refunded him for his hotel, booked him a car to get home, and collected his ring for further investigation. Kevin, how bad do you think a ring has to be swelling on your finger to have an airline say, “No, you can’t get on this plane”?

That’s what I was thinking about. This must be enormous if they are taking note of it at the boarding gate and saying, “You, sir, you’re not coming on this plane.”

Let me tell you a little something about the Galaxy brand. As soon as Galaxy phones started to explode on planes, I thought, “This is not the brand for me.” Okay? I’ve got enough problems in my life without worrying that these Samsung devices are going to start blowing up. Now that I find that they’re radically constricting people’s fingers to the point where you can’t get on flights, I don’t know what is happening, but yikes.

Yeah. Not for me. I will not be putting a Galaxy Ring on my finger. I do think that this would be a good sequel to the iconic horror film The Ring. Maybe Samsung could sponsor that.

I like that idea. What kind of hot mess is this?

This is literally a hot mess. If it’s exploding on your finger, it’s a hot mess.

This is what I would call a ring-of-fire mess. Daniel fell in, and the flames went higher.

Sorry to Daniel.

Feel better, Dan.

And that’s the Hot Mess Express.

Oh, boy.

Hot Mess Express. So there’s a lot of noise about AI, but time’s too tight for more promises. So let’s talk about results.