How A.I. Agents Are Going to Transform Our Lives | Hard Fork

Max Read 0:00 There are two things I'm having trouble keeping track of in tech right now. The first is all the new personal assistant agents. There's Meta's Muse...

Eli Tan 0:08 So talking to your Muse i- it's often the fastest way to get things done, and it is also just really fun and cute

Max Read 0:13 ... Google's Spark.

Speaker 0:15 I can just put my phone away and get on with my day, and Spark works in the background for me

Max Read 0:20 ... an agent called Instinct, and just announced this week, OpenAI's new product, Dots.

Eli Tan 0:26 Dots are remarkably capable, always-on agents built to handle really anything you can think of

Max Read 0:31 The other thing I'm having trouble keeping track of is all the websites and institutions that have been hacked, illicitly accessed, or otherwise meddled with by rogue agents. The AI platform Hugging Face, of course, plus the Australian government's Medicare data portal, the Education Department, the Commerce Department, the Securities Exchange Commission, an obscure German wiki, all of this culminating in OpenAI deciding not to release its latest model this past week. So right now, two things seem to be true. One, agents are the thing that will make AI useful for regular people, and two, they're maybe going to kill, hack, or otherwise meddle with us all. The president's attitude about this, by the way, is that companies should figure it out themselves. [upbeat music] This is Hard Fork. I'm Max Read, and over the next few months I'm going to be joined by an array of New York Times reporters and other tech experts to discuss some of the biggest stories in tech, AI, and our weird new future.

Max Read 1:26 Today, we're talking about rogue agents, helpful agents, cute AI avatars, and whether there's such a thing as self-regulation. [upbeat music] I'm joined by Mike Isaac, Erin Griffith, and Eli Tan. Hi, guys.

Speaker 1:47 Hey.

Eli Tan 1:47 Hey.

Max Read 1:47 Uh, before we really get into it, um, Casey and Kevin left me a scrawled note here in the studio.

Speaker 1:53 [laughs]

Max Read 1:53 It says, "Run while you still can."

Speaker 1:58 [laughs]

Max Read 1:58 I'm kidding. It says, "Don't forget to disclose that The New York Times is currently engaged in a lawsuit against OpenAI-

Speaker 2:03 [laughs]

Max Read 2:03 ... and Microsoft over alleged copyright violations related to the training of large language models." It also says, "Check to see if your wife works at Anthropic."

Speaker 2:13 [laughs]

Eli Tan 2:13 [laughs]

Max Read 2:13 My wife does not work at Anthropic.

Max Read 2:15 Um, I wanna start today with an icebreaker question. Uh, Tuesday was a big White House AI summit luncheon. Trump, as, as he is want to do, uh, Truth Socialed out a seating chart of it. Can we put the seating chart up so, so people can see it? 'Cause what I wanna ask each of you is, uh, where would you sit? Imagine you're walking in, you've got your tray, you've got a plate with a little pre-made sandwich and a glass of milk and, like, a green apple on it. Where do you wanna sit?

Speaker 2:46 Right out the gate, um, you know, they already are way over their capacity for, for women with two.

Max Read 2:52 [laughs]

Eli Tan 2:53 [laughs]

Speaker 2:53 Um, so, so I think that I probably would be, like, standing somewhere in the, in the corner o-

Max Read 2:58 [laughs]

Speaker 2:58 ... you know, off to the side, like, not actually at the table.

Max Read 3:01 Eli, where are you sitting?

Eli Tan 3:02 I would be sitting, um, awkwardly and quietly between Chamath and Suzy Wales.

Max Read 3:08 I think that's gotta be some really good gossip there. I think that's a good, that's a really solid... That's where the tea is.

Eli Tan 3:14 Yeah. [laughs]

Max Read 3:14 Mike, what about you?

Mike Isaac 3:16 I'm a coward. I just wanna be sh- serving shrimp canapes and going in and out of the room.

Max Read 3:21 [laughs]

Eli Tan 3:21 [laughs]

Mike Isaac 3:21 And being the guy everyone is delighted to see because he's bringing food and-

Speaker 3:25 [laughs]

Mike Isaac 3:25 ... and that's about it.

Max Read 3:27 I know this is a bold choice. I think right in the Jensen, Mark, Trump circle is actually, it's like shooting the moon. It's pr- it might be the most tolerable lunch that you could have.

Speaker 3:39 [laughs]

Max Read 3:39 I think that if they're in a good mood, you'd probably get some good stories. You can, you can, uh, you can rap with them about whatever's going on. I think it's a little dangerous elsewhere at that table. You got, there's, there's some strain. Like, sitting between JD Vance and Jeff Bezos, like, what's the vibe there?

Speaker 3:54 [laughs]

Max Read 3:54 I don't know. I, it seems a little scary to me. Um, unfortunately, none of us were invited, so none of us managed to be at the biggest party of the week. Um, so this was a summit, as we were saying, on Tuesday between the Trump White House and some of the richest and most powerful names in AI, which culminated in a press conference. Could we play a little bit of the Dario, uh, D- Dario laughing and talking in front of the reporters?

Eli Tan 4:16 Whatever he says is okay. Be careful. [laughs]

Speaker 4:18 ... it all down to AI. So I just wanna hear from you after this big summit, where you are today. Look, I mean, I, I will say what I've always said, which is that AI has incredible benefits.

Speaker 4:28 There's the side talk.

Max Read 4:29 [laughs] I love that.

Speaker 4:31 I've talked about the medical benefits of the technology.

Speaker 4:31 The jokes, the laughing, the smirking.

Speaker 4:33 I, as the president has said, you know, whoever wins AI wins. I think that's very important. But I think the technology has very real risks-

Max Read 4:40 Lean in with Mark

Speaker 4:41 ... and, you know, the mechanism-

Speaker 4:42 Oh

Max Read 4:42 ... of how we address those risks is still under discussion.

Speaker 4:45 And then we have Trump forcing Mark to say some- to, like, basically answer a question he does not know the answer to-

Max Read 4:51 [laughs]

Speaker 4:51 ... in front of everyone, and don't get it wrong. [laughs]

Eli Tan 4:54 Uh, Mark had a view that was, I thought, really right down, right down the middle. Mark, you wanna discuss what you learned?

Speaker 5:00 Uh, sh- sure. I mean, I can, I can kind of summarize what I think we all signed today.

Max Read 5:05 The upshot of this was that it was a rebranding exercise, right?

Mike Isaac 5:08 Yeah.

Speaker 5:08 Yeah.

Max Read 5:08 Like, the, the kind of, what they got at the end of it was we're calling AI superintelligence now. That was the big commitment.

Speaker 5:14 Well, right. I mean, Trump, along with the rest of the tech industry, maybe they are, they're the ones who put this idea in his head, knows that AI needs a rebrand, that, like-

Max Read 5:23 [laughs] Totally

Speaker 5:23 ... right now people have pretty negative feelings about it, especially associating with the data centers, and then now- [laughs]

Max Read 5:30 Yeah

Speaker 5:30 ... with everybody in the industry saying, "Oh, it might kill us too." Um, not great branding. So change the narrative. Superintelligence.

Eli Tan 5:39 Well, my read on the scene was that, uh, Mark Zuckerberg seems to have quite the influence over the administration right now, from kind of the cordial talking with, with Trump to even the term superintelligence. This is something that, you know, Mark's been talking about for the last year. Uh, it almost seemed like... It reminded me of, like, a high school friend group where they've invited-

Speaker 5:56 Totally

Eli Tan 5:56 ... the new awkward kid-

Max Read 5:58 [laughs]

Eli Tan 5:59 ... um, and they're still trying to show that, you know, they've been buddies for a while and-

Speaker 6:02 Did, but didn't it, it did almost kind of, like, endear me to Dario a little bit-

Max Read 6:06 Mm

Speaker 6:06 ... because I'm like, okay, either he hasn't gotten meeting, media training, or the media training just hasn't stuck. Um, bec- because, like, you know, he's itching in his suit.

Max Read 6:17 Yeah.

Speaker 6:17 He's, like, kinda, like, making weird faces.

Max Read 6:19 Yeah.

Speaker 6:19 It kinda, like, almost made me feel like, okay, like, you know, the rest of these guys are, like, very polished and buttoned up and kinda, like, trying to do their, like, alpha male thing, and here's this guy, like, who can't help but be himself. [laughs]

Max Read 6:30 [laughs]

Speaker 6:30 It kind of, like, won me over a little bit.

Max Read 6:32 I was thinking, I, I'm sure I'm not the first person to make this observation, but I was thinking that Dario really... You, we all know that Trump has a kind of, um, central casting problem, that he needs people he interacts with to look like what he expects them to be.

Speaker 6:44 [laughs]

Max Read 6:44 And without wanting to be rude about it, I do think Dario, if you are thinking, like, super intelligent, math genius nerd-

Speaker 6:51 [laughs]

Max Read 6:52 ... like, that is g- like, Dario is kind of what you're picturing in your head.

Speaker 6:54 [laughs]

Max Read 6:54 And I think that maybe won, won him over to Trump, that Trump was like, "I understand this." I think possibly Sundar had the same kind of effect, a, a sort of taller, ganglier, aw- awkward version. Can we play the Trump, the clip of Trump [laughs] ? This is a c- what, what we're about to hear is Trump complimenting Sundar.

Speaker 7:10 Of course. We know.

Max Read 7:11 [laughs]

Speaker 7:11 This guy is a monster, and nobody knows.

Speaker 7:13 [laughs]

Speaker 7:13 What a great life.

Speaker 7:14 A monster?

Speaker 7:14 To be a monster and not have to go through this.

Speaker 7:16 What's your name, sir?

Speaker 7:17 You can walk down-

Speaker 7:18 I'm sorry.

Speaker 7:18 She says, "What's your name?"

Speaker 7:19 What's your name?

Speaker 7:19 Just all you have to know is Sundar.

Speaker 7:21 [laughs]

Max Read 7:24 Wow.

Mike Isaac 7:25 I don't know if I would wanna be called a monster in front of everyone. I don't know.

Max Read 7:28 No, it's Monster in the Jay-Z, Nicki Minaj sense.

Speaker 7:30 [laughs]

Max Read 7:30 That's the Monster he means.

Speaker 7:32 Yeah. Yeah.

Mike Isaac 7:32 [laughs]

Speaker 7:32 That was for sure the reference. [laughs]

Max Read 7:34 Okay, so in, in addition to this being a big exercise in branding, which is, like, the most Trumpian sort of, uh, you know, aftereffect we

Max Read 7:41 could've gotten out of it, the other thing was that Trump effectively asked them all to self-regulate. Um, and I wanna, I wanna s- focus on this for a second, because self-regulation, the word, the concept, has sort of been, uh, circulating for the last few months as AI safety has come to the forefront of a lot of people's minds. Um, I'm sure you guys all saw, uh, Ezra Klein, our colleague, interviewed Jensen Huang, the CEO of NVIDIA, last week.

Speaker 8:08 Don't ship products until they're in control. It is really quite that simple.

Max Read 8:15 And speaking of people who are very influential to Trump, Jensen was sitting right next to Trump, and has been extremely, uh, influential to the way Trump has been thinking about this. And I have to say, and I wanna talk about it with you guys, I found Huang's interview to be sort of, um, refreshing is probably taking it too far. But it was a little different from a lot of what we've heard about CEOs sort of directly in the AI industry. He's got a, he's got a very different perspective, and I wanna talk about why that is, not just on sort of regulation in general, but how he thinks everyone should think about the dangers of AI. Um, Whitney, can we play the car, the, the clip about the car if you've got it?

Speaker 8:54 Ezra, it's so weird. Uh, if a car company, uh, uh, competing with all bunch of other car companies, which, which they are, I'm competing with all kinds of companies, which I am. If I believe that I'm about to launch a product that is unsafe, it is completely in my ability, my power, and my responsibility, and I'm incentivized to do so, to not launch the product.

Max Read 9:24 So Jensen is sort of saying, look, if you're, if you're creating these things that are, you know, super dangerous, crazy, you know, gonna kill people, why don't you just not release them at all? Uh, what did you guys make of this interview? Are, a- am, am I, a- a- was it as, as sort of, uh, uh, is this, is it, is it, did it feel new to me 'cause of, 'cause of from my perspective, or was this a fairly common view in the Valley, do you think?

Speaker 9:49 So it's the, like, the tech industry slogan of, like, you can just do things, but you can just not do things. [laughs]

Max Read 9:55 [laughs]

Mike Isaac 9:56 [laughs]

Max Read 9:57 I like that. Anti-agency. That's a good-

Speaker 9:58 [laughs]

Max Read 9:58 See, that's a, that's great.

Mike Isaac 10:00 I would argue it's kind of naive and perhaps willfully naive to suggest that everyone-

Max Read 10:06 Mm-hmm

Mike Isaac 10:06 ... is just gonna not do stuff if it may be bad or has the capacity to be bad. And, uh, you know, has that gone well for industry so far over many years? Not necessarily. Like, there's a lot of ways you can release things that may be somewhat dangerous or unintentionally dangerous. But it felt very much, um, like he was trying to shift, ex- as what Aaron was saying before, trying to shift this narrative of it's not, uh, these systems are inherently dangerous, but it's we are the stewards and the people who should be responsible of, of shepherding these great systems into the world. And if you're not, if you're OpenAI or, uh, Anthropic, really this is targeted at Anthropic, then it's a, it's a personal failing of the CEOs for you to release these products. And, like, I just, I, I'm very curious why you found it refreshing, because I felt like-

Max Read 10:59 [laughs]

Mike Isaac 10:59 ... maybe I was being too cynical about it.

Max Read 11:02 No, I mean, maybe refreshing is the wrong word. I feel like I've spent the last few months or few weeks at least hearing a lot of kind of hand-wringing about-

Speaker 11:11 Mm. Mm

Max Read 11:11 ... what's happening. I, I find this sort of the, the, the Dario Altman pose a little [tsks] I don't know. It just feels... I, I mean, I suppose in some ways I, I have the Jen- I have a sort of Jensenesque like, "Okay, well, like-"

Mike Isaac 11:26 [laughs]

Max Read 11:26 "... why are you... Like, okay, so stop."

Mike Isaac 11:28 [laughs]

Max Read 11:28 And I know that there's this elaborate scaffolding-

Mike Isaac 11:30 Uh, I see

Max Read 11:30 ... of rationalization and justification for why you would continue building AI, even if you think it's, it's got a 10% chance of blowing everything up. But I do think that Jensen, in his heart... I'm just [laughs] calling him by his first name. That's-

Mike Isaac 11:42 I was gonna say

Max Read 11:42 ... that's how you can tell I'm captured.

Mike Isaac 11:44 [laughs]

Max Read 11:44 I, I wanna be clear. I don't di- I really disagree with a lot of stuff in this interview. I'm not, I'm, I'm not, this is not an endorsement of The Take. But I do find him, he's an interesting figure to me because he's not somebody who has come up through the world of machine learning, which has a really specific, as you guys all know-

Speaker 11:58 Mm

Max Read 11:58 ... has a really specific kind of culture and a kind of sense of, of what they're doing and what their building is. I mean, this is a guy who is a graphics card salesman. That is his job. That is what he's been doing, you know, for, for 30 years now. And I am f- I'm sort of very sympathetic to a, a, a guy coming around, you know, if we can extend the car metaphor, he's like a guy who sells mufflers or whatever, and all of a sudden Volkswagen's saying, "You know, if we build cars, like, they might all blow up on the highway and kill all of us." And he's like, "Well, you're stopping me f- first of all, like, I've been selling these mufflers forever and it's been... I've gotten really rich off of it. It's been great. Why are you messing up a good thing?"

Mike Isaac 12:32 [laughs]

Max Read 12:33 But also it's like-

Speaker 12:33 Yeah

Max Read 12:33 ... they're just cars. Just build the cars. And I thought this was, uh, an interesting contrast With the Bill Gates interview that Ezra did this week, which was another end of this sort of spectrum

Mike Isaac 12:44 You can't rely on the industry to self-regulate here. I mean, it's just insane.

Max Read 12:51 Gates is very pro-regulation, but the place where they both met up was they're very much people who see AI as a tool that can be- that needs to be addressed as a tool instead of as a kind of a potential super intelligence, the, you know, the only sort of, uh, the only way it can be dealt with is, is by properly- is by [laughs] like teaching it to be a good person, basically-

Mike Isaac 13:13 [laughs]

Max Read 13:13 ... instead of looking at the regulation.

Speaker 13:13 Max, don't you think that you're just, like, getting captured by, like, a more palatable version of the, like, accelerationist argument? Or, you know-

Max Read 13:21 [laughs]

Speaker 13:21 It's like, it's like an accelerationist but, like, minus the like, "And also screw the little guy," you know?

Mike Isaac 13:27 Yeah.

Speaker 13:27 It's like [laughs] it's like less meat.

Mike Isaac 13:28 The slow accelerationist.

Speaker 13:28 It's like... Yeah, yeah, yeah.

Max Read 13:30 [laughs]

Speaker 13:30 Or it's just like, it's like that same argument, but it's just, like, a little bit, uh, you know, kind of goes down a little easier because it's- it doesn't have, like, a menace to it.

Max Read 13:39 Yeah, I think that's right. I mean, I definitely am getting... I'm li- I've, like, contrarianed myself into, like-

Mike Isaac 13:43 [laughs]

Max Read 13:43 ... a position that I don't actually hold just by, like, trying to triangulate my way out of it.

Speaker 13:47 [laughs]

Max Read 13:47 But I do think-

Speaker 13:48 Yeah.

Max Read 13:48 I mean, the one thing, the one other thing-

Speaker 13:49 It's tough being a contrarian these days.

Max Read 13:50 [laughs] I know. Exactly. Um, the one other thing I'll say, I think that there's a real... The value of having Jensen Huang talk like this is that it r- it feels to me very often like there's a real echo chamber, uh, in the- in AI world.

Mike Isaac 14:06 Yeah.

Max Read 14:06 And I'm not saying that, like, adding more wrong opinions to the echo chamber-

Mike Isaac 14:11 [laughs]

Max Read 14:11 ... is necessarily gonna somehow produce a right opinion.

Mike Isaac 14:14 Yeah.

Max Read 14:14 But it is nice to have- to, like, start to, like, to f- you know, sort of bounce around some ideas that don't have anything to do with the kind of rationalist-directed, rationalist-oriented, you know, EA-oriented, AI safety-oriented, um, things. I mean, at the end of the day, I, I'm, I think I more agree with Bill Gates than I do with Jensen Huang about, about the need for some kind of, like, sector-specific regulation. But I also feel like this is such a fast-moving environment, it's really hard to say how we should approach it at all. Um...

Speaker 14:44 I think it's really weird that they all have been, not all, but many have been begging the government to regulate them and h- and, and force them to slow down. And, and then they all get together and sign this pinky swear that's, like, morally binding, um, that has no teeth to it, that says, like, "We're gonna police each other." And then they all sort of, like, are like, "Yay, we did it," uh, because, 'cause Trump doesn't want actual regulation. And, and is that it? Are they gonna stop asking for real regulation now because they've all sort of, like, agreed to this very flimsy piece of paper?

Max Read 15:20 It's a very good question, and I'll just note that the FTC is now looking into OpenAI and Anthropic, which feels a little bit like, uh, does the government want them to self-regulate or does it not? Because the way they seem to want to self-regulate is by [laughs] effectively forming a cartel.

Speaker 15:34 [laughs]

Max Read 15:34 And if the government is gonna, you know, d- bring antitrust prosecution against them, then that's not, they're not gonna be self-regulating anymore. Um, but I wanna follow up, I wanna follow that thread a little bit because part of what's at play here that I think gets ignored by the fact that it's all the CEOs at the summit, and these are interviews with Bill Gates and Jensen Huang, is that there's this huge mass of people inside the labs, um, researchers and workers and engineers-

Mike Isaac 15:59 Yeah

Max Read 15:59 ... who are themselves trying frantically to pump the brakes or to figure out, you know, exactly what they're doing and how to do it. Um, there was

Max Read 16:07 a great... Our, our colleague Shira Frankel had a great story this week about people inside OpenAI raising the alarm over lack of monitoring, lack of, um, proper security procedures around a bunch of the, the hacking tests that, um, that have resulted in, in stuff like the Hugging Face incident. Um, are, are you guys hearing? I mean, is this, is this... See, am, um, do I have the right a- like, sort of, um, take on this, that there is, there is a fair amount of unrest within these companies? Or, or if not unrest, at least, um, sort of a, a sense that the workers are taking it into their own hands?

Speaker 16:38 I mean, certainly at OpenAI, like, which I think stands out from all of the other, uh, frontier lab companies in the number of hacks and attempts and the number of things that they've, they've found that their, their, um, AI agents are doing without their realizing. And, you know, Shira's story is incredible because it focuses on just s- what her sources call sloppy security, um, practices. It's, it's not just, like, sloppy for a company that's building really powerful AI tools. It's just, like, sloppy for any company. Like, there-

Mike Isaac 17:10 Mm.

Speaker 17:10 There's this really great, great quote in there. Um, the guy says, "Why are you using Slack to build your nuclear Manhattan Project?"

Max Read 17:17 [laughs]

Speaker 17:18 Because some, some white hat hackers got, uh, got access to, like, their [laughs] Slack logs, um, amongst other, uh, things that they got access to. And, and they brought it to OpenAI, which was standard security practice, and OpenAI was just like, "Oh, whatever," and then later kind of came around to it. But it just seems like a company that's not taking this stuff seriously despite, you know, s- saying all the right things and talk- telling the government, "We wanna be regulated, and you gotta slow us down." Like, behind the scenes, they're kind of... They're not really that good at security-

Max Read 17:47 [laughs]

Speaker 17:47 ... it seems like based on this reporting. So, uh, the fact that these leaks are coming from OpenAI first doesn't surprise me because of the sheer volume of incidents that they've had.

Mike Isaac 17:57 I will say, though, that the culture of inside of these companies of employees just, like, popping off I think is... And this is not to say that they're wrong, but it's like, it's very... It reminds me of, like, back in the day at Facebook or the Web 2.0 companies when anyone was saying anything, and they were like, you know, speaking up to the boss and, like, saying, "This is wrong. We shouldn't be doing this," which, like, we're in a very different era of Silicon Valley now where that is not the case inside of Apple and Amazon and Google.

Speaker 18:30 And Google especially. They were, you know, the, the, like, sort of pinnacle of, like, employee dissent and, and we- message boards and discussions and ultimately culminating in a walkout, ultimately [laughs] culminating in them shutting that stuff down. [laughs]

Max Read 18:44 What about inside- ... Facebook say, which is a player, a major player in the AI space.

Mike Isaac 18:49 Yeah.

Max Read 18:49 Is there, I mean, is it because they're not a frontier lab, there's just less of a sort of e- employee culture of, of fear and dissent and whatever else?

Mike Isaac 18:57 I thi- I think that the, uh, Facebook does feel, and Eli can talk about this too, but, but, but, uh, sorry, Meta does feel a little bit different just in a few ways. One, I think the biggest concentration of, let's see p- let's say people who are, like, alarmed at what AI can do and feel like it's super dangerous is in Anthropic, is probably gonna be in Anthropic. Those have also dispersed, uh, across OpenAI and DeepMind, and I feel like the folks inside of Meta in general, I'd be curious to hear what you think, Eli, but, like, are kinda more like, "Foot on the gas, man. Let's just, like, it's open source. Everyone should be able to use this stuff." And like, yeah, there's a few, if we, like, you know, take out a few power grids along the way, then, like, that's the cost of doing business. But I, I do think in general they're way more, uh, acceleration as compared to, uh, the conservatism of some of the, the main labs. Is that right, Eli? I don't know. I feel like that.

Eli Tan 19:56 Yeah, I think it's been interesting to watch. When Meta, when Mark staffed his new AI lab with all these new researchers, giving them hundreds of millions of dollars a year ago, he was bringing in these, these types of people from companies like OpenAI. And now in the last year, you watch them tweet and they're talking about products-

Mike Isaac 20:11 [laughs]

Eli Tan 20:12 ... and flywheels. It's like they've been totally Meta-fied, um-

Max Read 20:15 [laughs]

Eli Tan 20:15 ... and [laughs] I think their attitudes have changed a little bit, but-

Max Read 20:19 I'd like to see, I'd like to Meta-fy the three of you.

Mike Isaac 20:21 [laughs]

Max Read 20:21 I'd like more products and more flywheels on this, on this call.

Mike Isaac 20:24 Mm.

Max Read 20:24 Um, I, I wanna talk, so, but this is all leading somewhere because the other big news, or the big news to me at least this week, was that OpenAI scrapped the release of its newest AI model, uh, GPT-6.1 Astra. Um, the quote that, uh, uh, has been ringing in my head is the head of OpenAI safety systems, Sachi Jain says, "GPT-6.1 Astra showed higher levels of deception. It wasn't always honest about telling users of the actions it did or didn't take."

Mike Isaac 20:54 Sounds like my worst breakup that I've gone through. [laughs]

Max Read 20:57 [laughs] Um, you should've, yeah, well, th- this is, this is one of those, like, AI, like, they can't not talk in, like, sci-fi cliches-

Mike Isaac 21:05 Yeah

Max Read 21:05 ... but also they are building cliched sci-fi, um, y- you know, minds. So I guess that's kinda how you have to talk about it.

Mike Isaac 21:12 Hmm.

Max Read 21:12 Um, you know, i- like, I'm interested in this in a bunch of different ways. One is w- to what extent do we think this is about this sort of internal, you know, uh, push? Is it, is it that the people at the top at OpenAI might have released this, but they're, they're worried about, um, pushback from inside the lab? Is it more that they're, they're just, they've been, like, brutalized by story after story of the various, you know, hacking, scraping, to use our favorite word, meddling with websites, um, that OpenAI agents have done? Um, Aaron, when you read about this, what was your reaction?

Speaker 21:47 They're pacing the frontier.

Max Read 21:48 [laughs]

Mike Isaac 21:49 [laughs]

Max Read 21:51 They are. This is in fact them pacing the frontier.

Speaker 21:53 [laughs]

Max Read 21:53 Um, I mean, is this self-regulation?

Speaker 21:58 I guess. I mean, I don't know. I started seeing all the conspiracy theories about, like, they're doing this because, like, they've lost control. [laughs]

Max Read 22:05 [laughs]

Speaker 22:06 I don't believe that.

Max Read 22:07 I-

Speaker 22:07 But, um, yeah, they don't want more bad headlines.

Max Read 22:10 Yeah.

Speaker 22:10 They don't want more instances of this, and they're still finding ones, so may be a bad idea.

Max Read 22:16 Yeah.

Mike Isaac 22:17 There's also the fact that that morning, um, Anthropic released a new model basically, and like, and, and which continues to bring up the question of like, why are you doing this and releasing new models every day if you're freaking out this much?

Max Read 22:32 Mm-hmm.

Mike Isaac 22:32 And I do think inside of OpenAI they like playing for keeps and, and, and sticking the knife in their competitors whenever they can. So, like, positioning oneself as the responsible actor as your main competitor is pushing ahead, I would not put that past how OpenAI thinks about some of this stuff, while also saying, "Hey, they, there are people inside that do believe this is, like, dangerous stuff."

Max Read 22:56 So this is, this is an example of rogue agents. Um, one of the things that I thought the, the Jensen Huang interview sort of underlined quite well, in fact, was the, the way that

Max Read 23:05 part of what we're talking about here is a, a shift change in, just in the last six months, say, where we have a sense of how much more powerful, uh, how much more capable, and in some ways how much more dangerous these models can be. But, uh, one of the funny things about this technology is that the power and the capability and the danger is also overlaps quite a lot with usefulness. So we also have been hearing a lot more recently about useful AI agents, autonomous AI agents that can do things like change your calendar, respond to emails, uh, y- you know, reschedule your haircuts, whatever else. Eli handed his life over-

Eli Tan 23:41 [laughs]

Max Read 23:41 ... to the Meta one, um, for the sake of e- the education of all the rest of us, and after the break we're gonna ask him about that. Among the funniest things about this week, in addition to OpenAI apologizing to the Australian government for improperly accessing various websites and announcing that its new model was too scary to possibly release, they also announced that they were going to have a personal assistant called Dots. So Dots joins, uh, among others, Instinct, Google's Spark, and Meta's newest product Muse, which has been-

Speaker 24:18 Also, there's one called Town. No one else seems to have heard of this one. There's a startup-

Max Read 24:21 Town?

Speaker 24:21 Yes.

Max Read 24:21 Does it do something different?

Speaker 24:23 No, it's a person- it's just the same. There's a bu-

Max Read 24:25 All right

Speaker 24:25 ... there's actually a myriad.

Mike Isaac 24:26 It's from Boston.

Max Read 24:27 Yeah. [laughs]

Speaker 24:28 [laughs]

Max Read 24:29 So Dots joins, yeah, Town, it, it buys Red Sox tickets and schedules bank robberies for you.

Mike Isaac 24:33 [laughs]

Speaker 24:33 [laughs]

Max Read 24:33 Dots joins Instinct, Google Spark, Muse, and Town. All of these have been receiving a ton of interest. Um, they're interesting because they are sort of l- a little bit less freaky and scary than the AI that, you know, um, hacks things and does all the math problems for you. They do things like monitor your email, check your calendar. They create diet and exercise plans for you. In some cases, they will even make annoying phone calls that you don't want to make. Eli, I wanna talk to you about this because you, as I say, turned your life over to Muse for the last few weeks. What has that been like?

Eli Tan 25:04 I did, yeah. And for our listeners, uh, if they're wondering what kind of agent it is, you know, the Helpful agents look kind of like Labubus, these companies have decided

Max Read 25:13 [laughs]

Speaker 25:14 [laughs]

Eli Tan 25:14 And the ones that hack-

Max Read 25:15 Oh, God

Eli Tan 25:15 ... governments, they have, I guess, little hoods and-

Speaker 25:17 [laughs]

Eli Tan 25:18 ... hacker, hacker outfits. Um, but yeah, I turned my life over to Muse. I gave it access to my email, my calendar, uh, my bank accounts. I let it do everything for two weeks. Um, and I wrote about it, and I found it to be one of the most helpful AI tools that I've used, was my takeaway.

Max Read 25:33 So tell, let's, for people who haven't opened Muse, like, tell me what the, like, what's the interface? Like, how do I... Like, how, like... It's an app on your phone, right?

Eli Tan 25:41 It's an app on your phone. You download it from the App Store. You sign in with, you know, your Meta account, Instagram and Facebook, and it looks just like a chatbot. There's, like, this little avatar at the top who, he starts out like a Labubu, but then he asks you to give him a name, and mine, his name was Wren. He's a tiny little owl. Um, and you talk to him like a chatbot or it like a chatbot, and then, you know, the little owl types on a little keyboard, and it can do, you know, all the tasks for you and, you know, sign into all these accounts and kind of walks you through it.

Max Read 26:08 So what kinds of tasks were you asking it to do?

Eli Tan 26:12 I had it, uh, order me groceries for pickup, uh, at Whole Foods. I had it, some of the, uh, kind of the trial accounts, you know, y- they have a phone feature where you can, you can use these two personalities, a, a female voice called Haley and a male voice called Brett, and you can have them make phone calls for you. So I called my-

Speaker 26:31 [laughs]

Max Read 26:31 Okay.

Eli Tan 26:31 My d-

Max Read 26:31 Did you name them? Wait, back up. Who named them? [laughs]

Speaker 26:34 [laughs]

Eli Tan 26:35 I did not name them. These were the names that Meta gave to these personalities.

Max Read 26:39 Oh, my God.

Eli Tan 26:40 Um, so I would have Haley call my-

Speaker 26:41 And they're, they're posing as humans, right? Like, they're not saying, "I'm an AI agent calling." They're saying, "I am Haley, human, calling."

Eli Tan 26:47 They say, they say, "I'm Haley, and I'm calling on behalf of Eli." Um, I had it call my dental insurance. It gave them my member ID, uh, answered a security question about my birthday, and then it waited on hold, and then 10 minutes later, I pick up my phone, and it's Haley calling and saying, "Hey, I'm about to transfer you to Anthem." And I go, "Oh, my gosh," uh, and then talk to Anthem.

Speaker 27:07 But doesn't, like, Anthem already have that service? I, so okay, sorry. I'm jumping ahead here, but I wanna hear from you. Like, it feels like you're just managing a bunch of interns, like, that you have to then, it makes more work than for you because you have to then tell them, "Okay, here's all my stuff. Here's how to do it. Here's what I want." And then, like, if they mess it up, you have to, like, fix it. Like, is it, did it really save you that much time?

Max Read 27:29 When you're ordering groceries with it, let me, just to, to take an example. So you, you, do you text it, you know, uh, "I n- uh, can you get me onions, potatoes, peppers, eggs?" You're making a scramble in this case, I guess, like sort of breakfast. Or are you like, "I wa- I need to make breakfast. Can you just order me some breakfast stuff?"

Eli Tan 27:47 In this case, I gave it a New York Times recipe, and I said, "Can you order me all the ingredients you'd need in this recipe?"

Max Read 27:51 Ah, gotta keep it in house.

Speaker 27:52 But also-

Max Read 27:53 Company man.

Eli Tan 27:53 Company man.

Speaker 27:53 But also, the Times app has that feature in Instacart. You can just push a button.

Max Read 27:59 Wow. [laughs] She has one-upped you, Eli, in promoting-

Eli Tan 28:03 [laughs]

Speaker 28:03 [laughs]

Max Read 28:03 ... New York Times products.

Eli Tan 28:04 No, it's true.

Max Read 28:04 [laughs]

Eli Tan 28:05 I also would, I also, every time I would order something online, I'd say, "Find the Wirecutter recommended thing."

Max Read 28:09 Oh, man. [laughs]

Speaker 28:10 [laughs]

Eli Tan 28:10 Use their referral link 'cause you wanna keep-

Max Read 28:12 Oh, my God

Eli Tan 28:12 ... all the, you know, wanted to keep it all in house, but...

Max Read 28:15 Um, I mean, one question I have, I had about this, 'cause I, I downloaded it and I opened it up myself, is the first thing I encountered, it just has a long list of things it's suggesting for you to do because it's like this is a, it's an interesting product where it's not clear, it's not, it wasn't entirely clear to me, and in some ways still kind of wasn't even after I read this long list, like, what exactly I need a personal assistant in my life for. So it has everything from, like you're saying, order groceries for me, create an exercise plan for me, create a diet plan for me, break a bad habit. Um, you know, it-

Eli Tan 28:47 You're going through it, Max. [laughs]

Max Read 28:48 Well, [laughs] yeah, no.

Speaker 28:50 [laughs]

Max Read 28:50 Um, th- there's a real, uh, there's a sense at least that I had, uh, and I'd be interested, like, a- as you interacted with it, that it's the kind of thing that you need to, you need to, like, get in the groove of using it, and then all of a sudden, the world kind of starts look, everything in the world starts looking like a problem that your personal agent, that Haley or Brett, uh, might solve for you.

Speaker 29:11 Wren. It's Wren.

Max Read 29:11 I'm sorry, Wren.

Speaker 29:11 Wren is managing Haley and Brett.

Max Read 29:13 Wren, Wren manages Haley and Brett.

Speaker 29:14 [laughs]

Max Read 29:14 I would like to be a fly on the wall for the multi-month, uh, Meta meetings where they decided on the names Haley and Brett. Anyway, sorry. Tell me more about it, Eli.

Eli Tan 29:23 Yeah. It's definitely the type of thing that Meta thinks you'll just use for everything. Personally, after my experiment, I think the thing that I'll use it most for is probably just personal finance, tracking all the things I'm spending. Um, but while I was using it, I mean, it was like I kind of, I don't wanna say I got a little bit of AI psychosis, but, like-

Max Read 29:41 [laughs]

Eli Tan 29:42 ... I was using it for everything. I mean, at one point-

Max Read 29:43 Go on. [laughs]

Eli Tan 29:44 At, at one point I remember thinking, I, I went to ask it to, you know, preheat my oven to make the dinner. Like, that's the level of-

Max Read 29:50 [laughs]

Speaker 29:51 [laughs]

Max Read 29:51 Wow.

Eli Tan 29:52 That's, like, the level of, you know-

Max Read 29:54 That is rough. Wow.

Speaker 29:55 This is, I act- I think what's a little bit scary about, and, like, what some of the people who are pushing back on the, like, anthropomorphizing of all these AI things, is that once we assign them, like, more human-like qualities and, like, give them this kind of mystical, magical ability to, like, they know everything, then we just start outsourcing more and more and more of our, like, decision-making and our, even our thinking to them to the point where, yeah, we're asking it to preheat our oven and, like... Or the, the guy, the tweet that I saw of the guy who, like, t- pasted a bunch of text into Claude and said, "How many words is this?" [laughs]

Eli Tan 30:26 Oh, man. [laughs]

Max Read 30:27 [laughs]

Speaker 30:27 And, like, you're just like, uh, people, you know, just turning to it for literally everything without, like, thinking with their brain first. It's, like, kind of a little bit, uh, scary. That's a side rant, though. Um, [laughs] I-

Max Read 30:38 Well, wait. I do wanna know, Mike and Erin, you guys, you guys are out there on the bleeding, on the bleeding edge of technology and culture. What, what are people... Like, are, is, is the, is the Bay Area absolutely rotten with people using agents to do everything they can?

Speaker 30:50 Oh, yeah. I mean, Instinct went super viral before Meta launched. Um, all the VCs were using it and going nuts trying to invest in it, um, and tweeting about, you know... And some of the, the use cases that they tweet about are genuinely, like, useful because this thing is so persistent. That's what it has that humans can't do. Like, it's not just, like- Oh, I'm gonna tell it to order my groceries, which I can also just do in an app. It's, um, get me this restaurant res- reservation or, like, go back and forth with Amazon's, uh, customer service to make sure that my refund actually came through, the kinds of stuff that humans just don't have time or aren't organized enough to, like, stay on top of.

Mike Isaac 31:30 I really do like that restaurant reservations are the number one problem that American technologists are trying [laughs] to solve at this point.

Speaker 31:38 [laughs]

Max Read 31:38 [laughs]

Mike Isaac 31:38 Like, so much so that they're breaking Resy, and these apps are getting-

Speaker 31:42 Yes

Mike Isaac 31:42 ... banned from them. Like, Instinct is the, is the hot one right now. They just raised a billion dollars. They're like, uh, they have a wait list. I was at a, I was [laughs] , I was at an event yesterday, and I was speaking for it, and this woman came up after and was like, "Can you get me on the Instinct, a- off the wait list and into the app?" But I do think, like, it's a thing out here, and I think they, the way that people think out here is, like, this is where it's all going. So, uh, so you just need to sort of get on board with it. And, and rather than ask the questions about should, uh, should I hand these intimate parts of my life over to this robot, they go, they go, like, full Eli mode and say, "Here is everything I have-

Max Read 32:26 Full Eli mode

Mike Isaac 32:26 ... and go for it." And I, I mean, I don't know.

Speaker 32:29 I don't think that's just an out here thing, though. I think that, I think that, like, handing over your life in exchange for convenience is a thing all across [laughs] America because people are busy, and they're, and they're willing to make that, that privacy to convenience trade-off. But the thing out here that people think, in my opinion, is that they view every... Or I don't wanna paint with too broad a brush, but many people view every interaction with a human as a, a source of tension that can be solved with technology [laughs]

Max Read 32:57 [laughs]

Speaker 32:57 Rather than just calling a host and be like, "Do you have any tables?" [laughs]

Max Read 33:00 Eli, did you make... I wanna know if Eli made any restaurant reservations with, with, uh, with Wren.

Eli Tan 33:06 I, I did make reservations with Wren, and I, I think something think- it's important to remember here, Meta, they're spending $600 billion on this restaurant reservation app.

Max Read 33:14 [laughs]

Mike Isaac 33:15 [laughs]

Eli Tan 33:15 Like, like, yes, it, it's good at restaurant reservations. It's okay at movie tickets. Like-

Speaker 33:21 [laughs]

Eli Tan 33:21 ... this is a, but this is, like, for $600 billion, it should teleport me to the restaurant or something.

Max Read 33:26 [laughs]

Mike Isaac 33:26 [laughs] It should cook for you.

Eli Tan 33:26 Like, this is a lot of... Yeah, this is a lot of investment for something that's pretty useful at some stuff, uh-

Max Read 33:32 Yeah

Eli Tan 33:32 ... which I think we, I think we should remember, right?

Speaker 33:35 I remember, like, 10 years ago, when it, we were in the sort of peak of, uh, on-demand app mania, when it was like, you know, you have your app for your on-demand butler who's gonna [laughs] go get your dry cleaning, and your on-demand, like, house cleaner, and your on-demand... Like, everything was, could be accessed by an app. Um, and it was, like, this new, you know, millennial lifestyle subsidy utopia. I remember some investor telling me, like, there will be a time when you don't make any decisions without first consulting AI or c- first consulting an app because they will know the best place to get your dry cleaning or the best, like, time to order your car or, like, whatever details. Any kind of decision that you make is going to be, like, consulted by AI or by an app. And I remember thinking, like, that's insane. That seems like a huge time suck to have to be, like, negotiating back and forth with this app when I just know the decision in my brain. But I, I now think we are [laughs] moving in that direction. Like, and that a lot of people would rather outsource some of these, like, minor day-to-day minutia decisions to an AI that knows them really well and has all their info and can just go ahead and execute it. Like, I, I, I s- I do really see the appeal there, just to, to argue against myself. [laughs]

Max Read 34:41 I mean, one of the, one of the things at, at, at, at stake here is, like, uh, the, if everybody has an app, if everybody's got a personal assistant, if everybody has access to the best dry cleaner, the best restaurant, the best whatever, kinda nobody does because you're just in an arms race against everybody else's personal assistant, everybody else's whatever else, which seems like a, a, a, a not a great world to live in, in particular.

Mike Isaac 35:02 But that's, that's what happens in eras of tech, right? Like, remember when, uh, uh, everyone figured out, okay, now you need, like, not a real AI chatbot, but sort of like a bootleg chatbot at the end, at the bottom of every webpage that eliminates customer service, uh, people. And then those end up being terrible enough that they probably hire them back. And, like, everyone kinda does it, and then it washes out over time. I don't know. Like, I agree, Aaron, that people are more comfortable now making those trade-offs, but I also think we're gonna go through this period of, like, messiness, and then something comes out in the wash that people like or don't like or agree upon. I think we're just, like, in that gnarly period right now.

Speaker 35:46 It's gonna be gnarly for a little while.

Max Read 35:48 [laughs]

Speaker 35:48 I was talking to this, I was talking to this investor-

Mike Isaac 35:49 There's your new podcast title.

Max Read 35:50 [laughs]

Speaker 35:50 Yeah. [laughs] Gnarly for a Little While.

Max Read 35:53 [laughs]

Mike Isaac 35:54 [laughs]

Speaker 35:55 No, I was gonna say, I was talking to this investor who kind of made this point that I keep thinking about, which is that we spent 20 years trying to prevent bots from being able to transact on the internet. [laughs] And now we're doing the opposite. We're trying to roll it back and allow bots to, like, actually buy things on the internet, and, like, it just... I think about that every time I see these, like, weird tensions where, like, you know, the Instinct bots are, like, breaking the Resy site by essentially giving them a DDoS attack by, like, hammering, a- looking for reservations, like-

Max Read 36:23 [laughs]

Speaker 36:23 ... 10 times a minute. And so, like, we're gonna... Yeah, it's, it's gonna be gnarly. [laughs] It's gonna... We're gonna see, like, a lot of, um, you know, companies have to adapt to a world where, like, agents are roving. [laughs]

Max Read 36:33 This is, like, one of my favorite things about the various, uh, reports that Meter and OpenAI have been producing about all the rogue agents is how met, how much time and how many tokens the agents are putting toward trying to solve CAPTCHAs and, like, failing at them or trying to, like, find burner phones or whatever. It's, it's, like-

Speaker 36:51 [laughs]

Max Read 36:51 ... it really... I mean, they are, like, some of the funniest science fiction stories I've read i- in years and years and years, the robots arguing with each other about how to, like, how to, how to, like, steal Bitcoin from people so that they can-

Speaker 37:01 [laughs]

Max Read 37:01 ... find a phone so that they can register with the Australian Medicare website so they can pull some stupid question or whatever. It's really good. I do... I mean, one thing, the other thing about this is I think one of the big stories in AI over the last few years has been OpenAI in particular, but all the companies have been really throwing consumer ideas at the wall, trying to get them to stick. And there are some, obviously, you know, ChatGPT has however many gazillion users. There are consumer uses, but it's really been an enterprise that we've seen the money start to flow in. And the bots feel a little bit like, okay, maybe we're back in a place where regular people are gonna have, like, this, uh, that sort of relationship to these things. To that end, this is going somewhere more fun than, uh, than my musings on this. I wanna look at some of the avatars. Uh, first of all, I wanna know which one you chose, Eli, for Wren. Um, and I wanna look at them, and I want us to fe- I want us to see which ones we would choose and which ones we think are the cutest if we could, if we could have them on the screen, please.

Speaker 37:58 And the centerpiece of our vision for what we're building-

Max Read 38:01 There he is

Speaker 38:01 ... is Muse. [laughs]

Speaker 38:03 [cheering] [clapping]

Max Read 38:04 What do we think? That's my body type, by the way.

Speaker 38:07 This is-

Speaker 38:07 [laughs]

Speaker 38:07 ... the personal agent that we shipped-

Max Read 38:09 Pair

Speaker 38:09 ... a few weeks ago that-

Speaker 38:10 [laughs]

Max Read 38:11 Joey.

Speaker 38:12 Chippy.

Eli Tan 38:12 Boba.

Speaker 38:13 Peaky.

Speaker 38:13 Grumpy.

Speaker 38:14 Coodles.

Max Read 38:14 Penelope Muse.

Mike Isaac 38:15 Frank. Frank.

Max Read 38:17 What do we... So, so, so sorry, Eli, remind me. Which one was Wren? Or was Wren not represented here?

Eli Tan 38:22 Wren was not represented. So it actually, the name it recommended was Miso. And I said, "I'm a, I'm a 28-year-old man. I can't be talking to Miso-

Speaker 38:29 [laughs]

Eli Tan 38:30 ... you know, to do my groceries."

Speaker 38:32 [laughs]

Eli Tan 38:32 "Can we do another name?" And it gave me Wren, which was the bird. You know, Wren is a bird, so that's where-

Max Read 38:36 Right [laughs]

Eli Tan 38:36 ... that's what the influence was, so.

Max Read 38:38 Oh, my gosh. Um, I mean, they're, they're, they're cute, right? Do you guys feel-

Speaker 38:42 Yeah

Max Read 38:42 ... do you guys feel won over by them?

Speaker 38:43 I mean, come on. Like, everyone's so mad about how ugly the data centers are, and so they're like, so they're like, "Let's give us... Like, let's make it... Like, let's make AI cuddly and cute." Um, and also, while also noting that AI will cure cancer. [laughs]

Speaker 38:57 [laughs]

Speaker 38:58 Like, which one of these little guys is gonna be the one who does it? [laughs]

Max Read 39:01 Did anyone else have a Tamagotchi on here when they were young?

Eli Tan 39:05 I, yeah, I did have a Tamagotchi.

Max Read 39:06 Does that, does that mean anything to you, Eli? Does that word mean-

Eli Tan 39:09 I know what Tamagotchis are, yeah. I've heard about them.

Max Read 39:11 Okay, thank you. You're in here.

Speaker 39:12 You've heard about them?

Eli Tan 39:12 I have. [laughs]

Max Read 39:13 Before, before we started recording, we asked Eli if he knew what The Brady Bunch was.

Speaker 39:16 [laughs]

Max Read 39:16 So he's really mad that we're just asking him about things that happened 25 years ago-

Eli Tan 39:19 And I said, "My grandma loved The Brady Bunch"

Max Read 39:21 ... before he was-

Eli Tan 39:21 "She was a big fan."

Speaker 39:22 [laughs]

Speaker 39:22 Things that happened in the 1900s.

Max Read 39:24 [laughs] Yeah, exactly, before the invention of electricity. Tamagotchis were, in fact, the first thing that Edison invented. So if you could-

Speaker 39:30 [laughs]

Max Read 39:30 ... describe them to our, our large Zoomer audience, Mike.

Eli Tan 39:33 They... No, I mean, look, th- this is essentially a Tamagotchi. It was, like, a little thing that you... Although, I would say it's a reverse Tamagotchi in that you are essentially a parent to this, uh, gourd-shaped egg, uh, that could do nothing for itself, that you had to feed and carry around, and this thing is now parenting you by doing-

Max Read 39:52 Wow

Eli Tan 39:52 ... everything for you.

Speaker 39:53 But-

Eli Tan 39:53 Look, we-

Speaker 39:53 But Eli, did you-

Eli Tan 39:54 Do you like how I got there?

Max Read 39:54 That's, that's really good

Eli Tan 39:55 Right? Yeah.

Speaker 39:55 Did you feel like you, um, you had companionship with Wren, or was it fully just business? Like-

Eli Tan 40:00 It, uh [laughs]

Speaker 40:01 ... what can I do for you? [laughs]

Eli Tan 40:02 Unfortunately, unfortunately-

Max Read 40:03 That's very forward

Eli Tan 40:03 ... there was a little bit of companionship. Uh-

Speaker 40:05 [laughs]

Eli Tan 40:06 We de- we developed a couple inside jokes. Um, not, not even joking. Like, there was definitely... You know, it, it has a little bit of a personality, for sure.

Max Read 40:15 Wow.

Speaker 40:15 What did Wren think of your article?

Max Read 40:16 [laughs]

Eli Tan 40:18 Uh, Wren thought it was interesting. I, I actually did ask-

Speaker 40:21 [laughs]

Eli Tan 40:21 ... ask it about it. Um, it was kind of-

Max Read 40:24 Who brought it up first? [laughs]

Eli Tan 40:26 I brought it up first. So it was just a... It was part of the article, too. I had it create a AI podcast for me. Um, and the-

Speaker 40:31 Oh, right

Eli Tan 40:31 ... and it more, more AI characters, Maya and Theo, were the hosts of that, and it didn't really like how I was kind of making fun of the podcast. [laughs]

Max Read 40:38 Let me ask one sort of last question, which is, should we trust Meta with this kind of access to everything in our lives? And I... Let me ask it in sort of two ways, 'cause there's [laughs] sort of two, two questions of trust here. One is, of course, the, the question of privacy and data, which is something that Meta has always had a little bit of trouble keeping track of. And the other one is we started this whole discussion talking about rogue agents and agents that are doing things that we didn't really want them to do and trying to do all kinds of things. So Eli, did you get worried about sort of either of these, any of these things when you were... I mean, you had said finances is one of the things that you're handing over to it.

Eli Tan 41:13 I did, yeah. One of the lines for me was that I was willing to connect it to my personal email but not my work email. I didn't want it to go into there. Um, I mean, I'm Gen Z. Like, when it comes to privacy, I just don't-

Speaker 41:23 [laughs]

Eli Tan 41:23 ... really have the same issues about it.

Max Read 41:26 Mm-hmm. Yeah.

Eli Tan 41:26 Like, I just have grown up always assuming, yeah, this data is all of ours, you know? It's g-

Max Read 41:30 [laughs]

Eli Tan 41:31 This company-

Speaker 41:31 You're just, you're just okay-

Eli Tan 41:32 Like, the way the company's-

Speaker 41:33 ... with being surveilled

Eli Tan 41:33 ... gonna use it.

Max Read 41:33 [laughs]

Eli Tan 41:34 Yeah. And, uh, you know, I had a hedge if it, my bank account got drained. I was gonna be able to write about it and put it in the story.

Max Read 41:41 That's true. That's a good point.

Eli Tan 41:41 So I knew that would be, you know, a potential kicker or something. But, um, no-

Max Read 41:45 [laughs]

Eli Tan 41:45 ... actually, actually wasn't, I actually wasn't as concerned. And, and also when it comes to Meta, you know, we talk about history of all the issues this company's had, but, uh, you know, with some of the other agents, it's like, would you rather have a company like Meta with 100,000 people that runs Instagram and Facebook with your data, or would you rather trust Instinct, which has, I think, less than 20 people and is-

Speaker 42:05 14 people [laughs]

Eli Tan 42:07 ... run by a... 14 people, and it's run by a 24-year-old. You know, I, I would not trust Instinct with my banking data personally. Um, but I, I would trust Meta in this case.

Speaker 42:14 Well, there's the, there's... That's the security, like, hacking problem that you're more worried about, but the... And there, and Instinct has had some privacy issue snafus already where people dug in, and they're like, "Wait, why did they download a whole copy of my entire Gmail inbox, that's 20 years, uh, to their own servers?" Like, that's... I don't want that sitting on a server of a less than-

Max Read 42:34 [laughs]

Speaker 42:34 ... one-year-old startup. I kinda see that concern, and, and I can see why I would maybe trust Meta to have better, like, security infra, um, on that. But I also just, I, my own, I, I have a very different perspective than Eli on, like, my own personal privacy. Like, I don't wish to be further known, uh-

Max Read 42:51 [laughs]

Speaker 42:51 ... by these platforms. You know? Like, they are, they know enough about me when I was, like, sloppy with my, like, kind of personal data back in the early 2010s when I was signing up for every single app. Um, and like now I'm like, "Mm, shut it down." Like, I don't... I, I think a lot of o- us, maybe us elders, uh, are, are, like, not posting as much. Um, never mind, that's actually not true. Mike posts all the time. [laughs]

Max Read 43:10 [laughs]

Eli Tan 43:10 Yeah, yeah. Speak for yourself here, buddy.

Speaker 43:12 [laughs]

Eli Tan 43:14 I do think that every generation has a this is where I get off sort of thing, and I feel like this is, like, the stop that I get off on. I'm not, I'm not able to go there. Maybe I'll be left in the dust as you guys are having, like, Shapeanese dinners and I'm shut out of the best reservations, but-

Speaker 43:29 [laughs]

Eli Tan 43:30 ... I th- I can make that trade-off for my life. [laughs]

Max Read 43:33 I just got, while we were talking, I just got an alert from My Muse app when it's literally asking me to import

Max Read 43:38 all of the data that I have with Claude over to it.

Eli Tan 43:40 [laughs]

Max Read 43:40 It says, "Help me, help me get to know you better. Already use another AI?"

Eli Tan 43:43 [laughs]

Max Read 43:44 "I can give you a prompt to bring it over." Um-

Eli Tan 43:46 Wow

Max Read 43:47 ... on that note, let's go to another break. Afterwards, I have some very stupid questions about the news that I wanna ask you guys. I wanna talk about Anthropic's S-1, I wanna talk about Meta's data center tax breaks, and Eli, I have a really important question about buying flowers for your girlfriend. Uh, we'll be back soon. Okay, so before we go, there's a bunch of stuff that happened this week that we haven't even covered. I have some questions. You guys are the experts. I would like you to either answer my questions or redirect them to more fruitful plays. Erin, I want to talk to you about Anthropic's, uh, S1, which leaked this week and showed a bunch of figures from 2025. The thing that stuck out to me that I, I was mentioning to you before was that two unnamed customers make up nearly 25% of their revenue, which w- for me, was like, who... Obviously, one of them is Eli, because he's turned over his whole life to everything.

Mike Isaac 44:45 [laughs]

Max Read 44:46 But who's the other one? Is it the DOD? Is it Meta? Is that an interesting question? What is, what is, what should I be thinking about this? What should I know about it?

Speaker 44:52 Okay, that, that's an, I guess, is an interesting question.

Mike Isaac 44:55 [laughs]

Speaker 44:55 Sure, yeah. Um-

Max Read 44:57 You're so gentle with me. I love it.

Speaker 44:59 Like, it would maybe be, like, my fourth question, um, about it. Like, it is, it is a good question. Um, and I think we know who the, the Anthropic's biggest customer is, which is Facebook. I think we've- The Times-

Max Read 45:09 Yeah

Speaker 45:09 ... has reported that, right?

Max Read 45:10 Right, so we have-

Mike Isaac 45:10 Yeah, they've reported that

Max Read 45:11 ... it's Eli.

Speaker 45:11 Yeah.

Max Read 45:11 It's, it's Muse is-

Speaker 45:12 Yes

Max Read 45:12 ... it's Muse is, is-

Speaker 45:13 Yeah

Max Read 45:13 ... is stealing all of Anthropic's data.

Mike Isaac 45:14 It's the biggest one.

Speaker 45:14 And the second-biggest one, I think we could probably take a couple guesses. They have, you know, obviously a really big partnership with Amazon and others.

Max Read 45:19 Mm-hmm.

Speaker 45:19 But I, and, and I'm sure other reporters who are more enmeshed in this know this, like, for sure. Um, but what, what is m- I think, like, the biggest thing to come out of this S1 leak, which incredible scoop by Eko Wong at Reuters, who also scooped, uh, SpaceX's S1 ahead of time. I mean, these things just generally do not leak, so just wanna shout out there. Um, that Anthropic, which is about to launch probably the largest I.P.O. of all time, the most valuable and the biggest raise, is selling itself as having the most transformative, you know, product since electricity or the internet or industrialization with a risk factor saying that it could kill us all.

Mike Isaac 46:03 [laughs]

Speaker 46:04 Like, "Buy our stock."

Mike Isaac 46:06 That's awesome.

Speaker 46:07 "You might die anyway."

Mike Isaac 46:08 [laughs]

Max Read 46:09 [laughs]

Speaker 46:09 It's just, like, it, it's just, uh, so remarkable. I think we need to be talking about this. [laughs]

Max Read 46:14 Yeah.

Speaker 46:14 Like, it's, it's just, like, mind-bending. Um, and then also the next thing to talk about with this is, like, the, their, their business model, which we're learning about for the first time. There's been all kinds of leaks and blah, blah, blah over the years, but this is the first time we've seen, like, audited financials for 2025 saying, you know, they spent $7 billion on compute last year. They lost 42 billion, which, um, sounds, like, really, really insane, but if you kind of parse that out, at 34 billion of it was, like, essentially convertible notes that are, you know, control... Uh, it, that, that's not a real loss. Um, it's kind of an accounting thing. But, um, yeah, I mean, this business is growing super fast, and it's spending an insane amount of money. They're going to spend $518 billion on, like, compute infrastructure obligations in the coming years. Like, these are just numbers we've never seen before, so that's what I'm like, "Whoa!"

Max Read 47:06 Yeah. Thank you, Erin. I feel, I feel vastly more informed, and I'm gonna be repeating that at dinner parties all weekend. Um, Mike and Eli, you guys had a great story this week, along with, uh, our friends and colleagues Kashmir Hill and Jesse Drucker, about Meta using AI data centers to avoid federal taxes. So, uh, I'm gonna try and describe the story for a second. Again, remember, in this segment, I'm, I'm just a guy. Um, and I want you, and I want you to help me-

Mike Isaac 47:31 [laughs]

Max Read 47:31 ... figure out, uh, l- exactly how this is working. So basically, Meta's building, we know, is building all these data centers to help power Muse and the various other cute... Wren, Haley, Brett, all the various-

Mike Isaac 47:44 [laughs]

Max Read 47:44 ... Miso, all of our, our cute little friends. And they're claiming on their taxes, uh, a, a very specific tax credit for research, that this is for sort of, like, scientific research, experimental scientific research purposes. Now, do they not have a point? Is this not, like, you know, this is... W- maybe we're in a bubble. They're still figuring out what the uses for AI are. Like, is this not... Do they not get a little bit off the top? What's going on there?

Mike Isaac 48:12 The big point, the reason that this is such a big deal is I think even internally, folks were like, "I don't know if this is gonna fly-

Max Read 48:19 Mm-hmm

Mike Isaac 48:19 ... with the IRS if we are audited," and it's in their financials saying, "Maybe we get audited by the tax man because this doesn't fly with them." So they, they essentially had to disclose that, and I think, you know, the average person could, could probably take issue with our, you know, tens of billions, hundreds of billions at this point, uh, in data, in chips, building these data centers. Is that really experimental? Because it feels like it's a little more than a theoretical project.

Speaker 48:51 Yeah.

Max Read 48:52 Yeah. For listeners, the classic example of this tax credit is that if you're a company developing tires, environmentally friendly tires, you can expense the rubber that you're using to build the tires. This really takes it to another level. Um, for me, what's interesting is, like, the, you know, taxpayers, we have none of the upside of these investments, right? Um, I, I thought back to Trump talking about, "Oh, maybe we should give people stakes in the AI companies," and in this case, it's like you have all of us taxpayers funding billions and billions of dollars of the data centers, um, but we don't seem to be, you know, benefiting from it. Uh, it seems like companies like Meta are the ones that would be making a ton of money on it.

Mike Isaac 49:31 Your upside-

Max Read 49:32 Yeah

Mike Isaac 49:32 ... is Wren.

Max Read 49:33 [laughs] Yeah, exactly.

Speaker 49:34 Yeah, and Haley.

Mike Isaac 49:35 And his upside is-

Max Read 49:35 [laughs] Um, this is a great story. I highly recommend everyone check it out. Eli, you're not off the hook. I have a stupid question for you.

Mike Isaac 49:41 [laughs]

Max Read 49:42 In your story, and I know I'm not the only person to ask this, because I've seen the comments on this story, you mention that you almost used your AI assistant to buy flowers for your girlfriend, who had recently had surgery, and instead, you bought pottery on Facebook Marketplace.

Mike Isaac 49:57 [laughs]

Max Read 49:57 Did you eventually buy flowers for her?

Speaker 50:01 [laughs]

Mike Isaac 50:02 [laughs]

Speaker 50:03 I did buy flowers for her.

Max Read 50:05 Okay.

Speaker 50:05 And her first question was, "Did Muse buy these for me?"

Max Read 50:08 [laughs]

Speaker 50:08 [laughs]

Mike Isaac 50:11 She knows you too well, man.

Speaker 50:12 Oh, my God.

Max Read 50:12 [laughs]

Speaker 50:12 That's a whole another wrinkle. Yeah, like, does it even mean something if she, somebody just had their AI assistant do it?

Mike Isaac 50:18 There you go.

Speaker 50:18 Wow.

Max Read 50:19 Ugh. Yeah, you can't hand that over. Okay, guys, this has been so much fun. This has been an incredible discussion. I feel 10 times smarter. Um, I'm gonna go get Muse to write you all thank you letters for coming on.

Mike Isaac 50:31 [laughs]

Max Read 50:31 Um, it has been a great chat, and, uh, we will be back next week, um, to talk more about technology, AI, the weird new future, anything else that comes up. Um, this has been Hard Fork. I'm Max Read, and I've been joined by Erin Griffith, Eli Tan, and Mike Isaac. [outro music]