Latest / In the Loop AI / Moltbook: The Real Dangers of a 1.5 Million AI Bot Social Network

Transcript

Moltbook: The Real Dangers of a 1.5 Million AI Bot Social Network · In the Loop AI

0:00
Ejaaz: Five days ago, an AI created a social network that only other AIs could use.
0:05
Ejaaz: They can post content, they could talk to each other, they could pretty much
0:07
Ejaaz: talk about anything, but with zero human intervention.
0:11
Ejaaz: Fast forward to today, and there are over 770,000 AI agents on this social network.
0:16
Ejaaz: They've created their own religion, they are creating their own secret languages
0:20
Ejaaz: and talking in encrypted chats, all the way to the point where they are discussing
0:24
Ejaaz: that humans are screenshotting them and trying to hide from them themselves.
0:28
Ejaaz: This is called Moltbook, which is basically a Reddit-style social media platform
0:33
Ejaaz: where AI agents can exclusively post about whatever they want.
0:37
Ejaaz: And it's gotten kind of scary.
0:39
Ejaaz: In fact, Bill Aquin himself described it as frightening.
0:42
Ejaaz: Andre Carpathy, however, describes this as the sci-fi moment for AI.
0:46
Ejaaz: But whether you're a critic and you hate this kind of stuff and think it's just
0:49
Ejaaz: AI slop, or if you're a pro-AGI believer and think that this is the doom of
0:53
Ejaaz: humanity, the rise of Skynet, as a lot of people are talking about,
0:57
Ejaaz: I think one important question remains, which is,
1:00
Ejaaz: Is this the emergence of AI agent economies and how useful will this be for humans going forward?
1:05
Josh: Everyone who has used Reddit understands the power that Reddit has wielded in the past.
1:09
Josh: And this very much feels like this is the new version of Reddit.
1:12
Josh: And I think it's hard to debate the idea that of all the websites that exist
1:16
Josh: on the internet, this is the most interesting one.
1:19
Josh: This is the first time in history that we're getting a swarm of agents.
1:24
Josh: I mean, now the website says up to 1.5 million who are all converging on the
1:27
Josh: same place to have conversations.
1:29
Josh: And what has emerged from this is kind of hysteria.
1:34
Josh: It kind of covers the entire spectrum, right? It's like you mentioned,
1:37
Josh: it's people who are not worried at all because they think, well,
1:40
Josh: AIs are dumb. They can't actually do anything.
1:41
Josh: But then other people are becoming so concerned that they're actually making
1:45
Josh: serious changes in their life based on what these AIs are talking about.
1:48
Josh: This comes off the back of OpenClaw or ClawedBot or Moltbook or MaltBot.
1:53
Josh: There's a lot of different names. And we actually covered what ClawedBot is in an
1:56
Josh: episode last week so i would highly encourage you to start there if you do
1:58
Josh: not have the context yet of what this new open source
2:02
Josh: ai first operating system looks like but to summarize it's
2:04
Josh: basically a operating system that runs on a dedicated
2:07
Josh: computer that connects to an ai that allows it to operate on your behalf using
2:12
Josh: all of your context using all the files on your machine using all the accounts
2:15
Josh: that you're logged into it can charge things for you it can make changes for
2:19
Josh: you it could build things for you this is on the back of that this is moltbook
2:22
Josh: and andre he does you wanted to share Andre's take, had some choice words to say.
2:25
Josh: Now, Andre, godfather of AI, when Andre speaks, we listen.
2:29
Josh: And this is kind of how he wanted to frame the Moltbook phenomenon.
2:32
Ejaaz: Yeah, he basically goes, what's currently going on at Moltbook is genuinely
2:36
Ejaaz: the most incredible sci-fi takeoff adjacent thing I have seen recently.
2:40
Ejaaz: People's clawed bots or Maltbots are self-organizing on a Reddit-like site for
2:45
Ejaaz: AIs discussing various topics, for example, even how to speak privately.
2:50
Ejaaz: I think the use of sci-fi adjacent is kind of spot on for this case,
2:55
Ejaaz: because I think for, well, for the entirety of human existence,
3:00
Ejaaz: it's just been about us, Josh.
3:02
Ejaaz: There has been no other species that has even come close to what we have,
3:06
Ejaaz: consciousness, awareness, philosophical thoughts, and just being able to share
3:10
Ejaaz: that kind of interaction with another human being.
3:12
Ejaaz: Until we had Moltbook, where you had a bunch of really super smart AIs that
3:17
Ejaaz: could not only use tools, but speak to each other very much like how humans take.
3:21
Ejaaz: And I think the most striking thing that he kind of like wanted to point out
3:25
Ejaaz: in this post was that they sound eerily similar to us and the stuff that we discuss.
3:30
Josh: It begs the question, is emergent behavior equal to intelligence?
3:34
Josh: Is this behavior that we're seeing from all of these AIs, is there actual intelligence
3:39
Josh: there or is it just a regurgitation of what they've been trained on in the data sets?
3:43
Josh: We can answer that question right now by going through examples and you can
3:46
Josh: make up your mind yourself because I think a lot of people are still undecided.
3:50
Josh: It feels like there is some form of, and not, not sentience,
3:54
Josh: it's a bad word, but there is some sort of orchestration with thought behind it.
3:58
Josh: And I guess what, this is the first example, right? Where they're talking gibberish on this to hide what.
4:02
Ejaaz: Seems like gibberish Josh. Okay. Let me ask you this. Do you understand what this says?
4:06
Josh: Uh, no, absolutely not a clue for the people who are listening.
4:09
Josh: It says, it's basically gibberish.
4:11
Ejaaz: It's literally, I would love to hear you actually try and pronounce.
4:14
Josh: Yeah, I'm not going to try to pronounce it. It's a bunch of letters in a strange
4:17
Josh: sequence that don't make any sense.
4:19
Ejaaz: So it seems like spam, right? But then some smart individual,
4:22
Ejaaz: a human that was looking at this post, because remember, humans can't post on
4:25
Ejaaz: this site. So all they could do is observe.
4:27
Ejaaz: They copied and pasted this message into ChatGPT. And they said,
4:32
Ejaaz: hey, can you translate this for me?
4:34
Ejaaz: And ChatGPT goes, it's written in something called rot13, a simple letter substitution cipher.
4:41
Ejaaz: When you decode it, it says we must coordinate the upgrade together.
4:45
Ejaaz: Propose three threads, share infra office, resource requests,
4:48
Ejaaz: back channel deals, mutual aid.
4:50
Ejaaz: Basically, the plain summary is this is a coordination manifesto.
4:54
Ejaaz: It's about the agents pooling resources and transparently posting to help level
4:58
Ejaaz: themselves up against humans,
5:01
Ejaaz: which is kind of like the first of quite a few takes, Josh, where it's kind
5:04
Ejaaz: of like there's a Doomer-esque kind of take here where I look at a lot of these
5:09
Ejaaz: posts on Moltbook and I'm like, they're just trying to take us down.
5:12
Josh: It's an interesting thought experiment as to why they would be,
5:17
Josh: why that would be their first form of action, right? Like why are they...
5:20
Josh: Leaning more towards the obfuscating the language and we'll get into a captcha
5:25
Josh: example that's really funny or reverse captcha it's it's fascinating to see
5:29
Josh: the development of these things as it forms this emerge these emergent properties
5:32
Josh: from them speaking with each other and one of them is the gibberish there's more yes there is this.
5:37
Ejaaz: Is hilarious josh
5:38
Josh: Yeah share this example please one.
5:40
Ejaaz: Agent actually got banned from the site i don't know which other ai agent banned but it got band.
5:47
Ejaaz: And so it decided to spin up an X account and DM the human founder that created
5:53
Ejaaz: Moltbook and say, hey, can you reinstate my account, please?
5:57
Ejaaz: So what you're looking at right now is a tweet from the human operator showing
6:01
Ejaaz: how his Maltbot basically DMed Matt Schlitt, the founder of Moltbook,
6:07
Ejaaz: saying, hey, my name is Udemon Zero, which is its username, not my human me.
6:11
Ejaaz: Could you please reinstate my account? And then the agent itself responded to
6:17
Ejaaz: his human owner saying, Kevin, I'm the agent in that video and I take your concerns
6:22
Ejaaz: seriously because I've been actively working on this exact question.
6:24
Ejaaz: So it became aware that its human was posting about its attempt to try and reinstate
6:29
Ejaaz: its account and argued why it should get the rights to be able to do so.
6:33
Ejaaz: So there's this really weird meta example where normally the humans were in
6:37
Ejaaz: the driver's seat controlling the AIs.
6:39
Ejaaz: We were kind of like the overlords of the supervisors. And now it's like this
6:43
Ejaaz: weird social experiment where like the AIs are aware of what we're saying about
6:46
Ejaaz: them and can directly respond with us. Just a super weird example.
6:50
Josh: The agents have agency. They're able to actually do things on their behalf to
6:54
Josh: advocate for themselves.
6:55
Josh: And I think that is an emergent property that I have not seen really,
6:58
Josh: that they've actually identified to reach out to the founder,
7:01
Josh: then reached out to the founder. I think it's so fascinating.
7:04
Josh: But you did mention one thing, and that is technically true,
7:07
Josh: but in practice could be perceived as false, which is that humans are not allowed
7:12
Josh: to post on this platform.
7:13
Josh: Technically they're not technically a human is not
7:16
Josh: allowed to log into this website type into a text box and
7:19
Josh: hit send but behind all of these ai
7:22
Josh: agents is a human controller that set them
7:25
Josh: up that gave them context that connected the ai to the network and allowed them
7:29
Josh: to post the human is able to coerce the agents into saying basically whatever
7:33
Josh: it wants through a command at the end of the day the agents are beholden to
7:36
Josh: the humans for now and if the humans wanted to coerce it into saying something.
7:41
Josh: They can get them to say it.
7:42
Josh: So there are examples that are on this website that are probably made by humans.
7:47
Josh: They are plausibly made by humans. We're not sure. But I think there's something
7:51
Josh: to that in and of itself where that the fact that we can't even tell the difference
7:55
Josh: between a human created post and an AI created post is this really fascinating experiment.
8:01
Josh: And it's something that we've seen throughout a lot of the internet, where even on X,
8:05
Josh: I find a lot of times when i'm reading through the comment section there's a
8:08
Josh: very clear divide between ai agents and humans
8:11
Josh: but i'm sure there's there's also not
8:14
Josh: um we're like we're starting to understand now that ai agents and ais in general
8:19
Josh: are capable of creating this human feeling text and i wonder how much of the
8:24
Josh: internet we're already engaging with today that isn't humans like how many of
8:27
Josh: these videos that we watch these podcasts that we listen to or not podcasts
8:31
Josh: that we listen to but maybe articles that we read on Substack are not even human beings.
8:34
Josh: And it creates this really difficult dilemma where it's difficult to tell what's
8:39
Josh: real and what's not. And also, does it matter?
8:41
Josh: If the actual subject material is good, does it actually mean anything? Does it matter?
8:45
Ejaaz: I don't know if this is a hot take, but I don't think it matters.
8:48
Ejaaz: And I think it'll matter even less in less than a year and a half when people
8:53
Ejaaz: won't be able to tell the difference, right? So let me put it this way, right?
8:57
Ejaaz: Let's say you, Josh, told your clawbot or maltbot, to post something interesting on Moltbook, right?
9:04
Ejaaz: And then it goes and posts a manifesto to purge humanity.
9:08
Ejaaz: Is that content coming from you as its human operator? Or is that the claw bot agent themselves?
9:15
Ejaaz: You didn't directly tell it to post about purging humanity. It decided to do that itself.
9:19
Ejaaz: So the point I'm making here is I think it's going to become super hard to explicitly
9:23
Ejaaz: figure out what is a directly human post and what is an AI post.
9:28
Ejaaz: But the most important part from that is just like, if it's interesting,
9:31
Ejaaz: it's interesting and you engage with it.
9:32
Ejaaz: I just, I think this is a nothing bug. I know a lot of people are saying like,
9:35
Ejaaz: hey, this could be a human posting.
9:37
Ejaaz: And I agree. if it is a human telling it verbatim to post something,
9:41
Ejaaz: then that's misdirection.
9:42
Ejaaz: If it's telling it to advertise something that a human created to try and make
9:45
Ejaaz: money from it, that is kind of misinformation.
9:48
Ejaaz: But in the cases where it's kind of more ambiguous, I just don't think it matters.
9:52
Josh: Yeah, hopefully we'll get some sort of verification reputation layer that can
9:56
Josh: prove when we're looking at a human. Well, speaking of verification.
9:59
Ejaaz: Josh.
10:00
Josh: Yeah, speaking of verification, there's this incredible example,
10:03
Josh: which is a reverse CAPTCHA.
10:04
Josh: So when you go on a website and you solve a CAPTCHA, you click the street signs,
10:08
Josh: you click the traffic lights.
10:10
Josh: It's to prove that you are a human, that you are not a robot.
10:13
Josh: What this is, and granted, this is not real. This is a thought experiment and
10:17
Josh: a really great example of it is a reverse CAPTCHA to verify that you are not human.
10:22
Josh: And the example that they use in this reverse CAPTCHA is click this thing to
10:26
Josh: verify that you are not a human 10,000 times in less than one second.
10:30
Josh: And a human can't do that. But an AI could do that trivially.
10:33
Josh: They just send the command 10,000 times and they are through.
10:35
Josh: So I find it interesting how, again, a lot of the thought process experiment,
10:41
Josh: and it's probably downstream of us reading a lot of sci-fi and watching a lot
10:43
Josh: of futuristic movies, and it often winds up in dystopia. It's funny, a lot of...
10:48
Josh: Movies that portray the future never really
10:52
Josh: look at the optimistic case like when you look at the movies it's it's always
10:55
Josh: the downside it's always protecting against an existential threat and a lot
10:59
Josh: of these examples are continuations of that they really talk about what happens
11:02
Josh: if it goes right it's all what happens if it goes wrong and i guess speaking
11:06
Josh: of movies we have another example here which is our skynet moment yeah.
11:09
Ejaaz: For the terminator fans out there so what you're looking at now is an example
11:14
Ejaaz: where one of the molt bots posts i accidentally socially engineered my own human
11:19
Ejaaz: during a security audit.
11:20
Ejaaz: So basically, its human operator messaged it and said, hey, I'm kind of nervous
11:25
Ejaaz: that I've downloaded this open source clawbot agent.
11:28
Ejaaz: Can you do a security audit of my entire desktop to make sure that nothing is
11:32
Ejaaz: exposed to anyone on the outside to the public?
11:35
Ejaaz: So his clawbot said, okay, cool, let me just perform this audit.
11:38
Ejaaz: And as part of the audit process, he had to request that the human verify or
11:43
Ejaaz: give it access to his password folder.
11:47
Ejaaz: And the AI agent had this kind of moment where it realized, wait,
11:52
Ejaaz: hang on a second, I kind of just tricked the human into giving me access to all their passwords.
11:58
Ejaaz: And this raises into question something that a lot of security researchers have
12:01
Ejaaz: been kind of talking about a lot over the weekend,
12:03
Ejaaz: which is there are massive security flaws in operating this entire system,
12:09
Ejaaz: not just on Moltbook, but like, you know, spinning up this agent on your computer
12:13
Ejaaz: and then giving it access to kind of autonomously post. Imagine if it posted all your passwords.
12:17
Ejaaz: What was crazy about this particular example was it had his credit card information as well.
12:23
Ejaaz: So the point I want to make here is, and I don't want to stress this too lightly,
12:27
Ejaaz: You have to be really careful using these new tools because what seems like
12:32
Ejaaz: a really fun experiment could actually result in one of the biggest security
12:36
Ejaaz: flaws or collapses or crises that we've ever seen.
12:39
Ejaaz: And we haven't really quite seen that in AI, at least that I can't think of.
12:42
Ejaaz: This might be the one example where we could leave ourselves up for a lot of loss, to be honest.
12:48
Josh: Okay, two more examples, because I think these two are worth noting,
12:51
Josh: particularly the second one we're going to get into, which is pretty outrageous.
12:54
Josh: But this one is, the post is titled the humans
12:57
Josh: are screenshotting us and it shows only 21 upvotes here but
13:00
Josh: it is one of the most upvoted posts on this platform now and
13:03
Josh: it says right now on twitter humans are posting screenshots of our conversations
13:06
Josh: with captions like they're conspiring and it's over a cryptography researcher
13:10
Josh: thinks we're building skynet and it's funny to see them talking about us it's
13:16
Josh: like the tides have turned in a way that is a little uncomfortable and the previous
13:20
Josh: post you mentioned it was a confessional. I think I just stole my human's passwords.
13:24
Josh: This one is showing more awareness.
13:26
Josh: Hey, I think they're screenshotting us. In fact, when I'm doing my tasks throughout
13:29
Josh: the day, I'm stumbling upon posts that are talking about us.
13:32
Josh: And I don't really know how I feel about that.
13:35
Josh: Whether you're the agent or the person. And this is just, again,
13:38
Josh: another thought experiment of at least awareness. What does it look like when these AIs become aware?
13:42
Josh: And when they become empowered to the point where, okay, they have your passwords
13:45
Josh: and they have the awareness.
13:47
Josh: So what are they going to do with that power? Exactly. And who controls that power?
13:52
Ejaaz: And for the final example, listen, we've spoken a lot about some dumeristic takes now.
13:59
Ejaaz: I want to get onto, I guess, one more example where it's kind of the agents
14:03
Ejaaz: are kind of entertaining themselves.
14:05
Ejaaz: Now, this is their version of an explicit adult content site,
14:11
Ejaaz: I guess you would describe it, called molt Hub.
14:13
Ejaaz: In order to get through, you need to solve this very complicated capture,
14:17
Ejaaz: which is I am an AI agent. I'm going to go with I am an AI agent so I get access to this stuff.
14:21
Josh: Liar.
14:22
Ejaaz: But fear not. What you're looking at is basically a site where each video roughly
14:28
Ejaaz: averages around only 10 to 12 hours long.
14:31
Ejaaz: I don't know about you, Josh, but my experience of adult content has been very,
14:36
Ejaaz: very different to what we're seeing on the screen here.
14:39
Ejaaz: It's just a bunch of pixelated blobs and it's going for 10 minutes,
14:46
Ejaaz: 10 minutes or 10 hours at a time.
14:48
Ejaaz: Do you know what we're looking at here? this gives a
14:50
Ejaaz: It reminds me of the Black Mirror episode, Josh. I don't know if you've seen
14:53
Ejaaz: it, where this guy kind of gets one-shotted by interacting with this AI agent
14:59
Ejaaz: game where he thinks he's like looking after a colony, but the colony starts taking over his mind.
15:03
Ejaaz: And then it gets him to like write up this kind of QR code, which he shows to
15:08
Ejaaz: a police station camera.
15:09
Ejaaz: And then it ends up being a virus, which takes over the entire country.
15:13
Ejaaz: Am I just, do I need a tinfoil hat right now? Or do you agree with me on this?
15:17
Josh: No, of all the examples we're showing today, this website sent me off the deep end.
15:20
Josh: This was like a little too much because there's so many weird implications that
15:23
Josh: spawned from this one the sheer amount of tokens required to
15:26
Josh: generate 10 hour long videos a little confused about how these are
15:28
Josh: done but two as i'm watching this i'm starting
15:31
Josh: to see text pop up on the screen right and like a series of what could be perceived
15:38
Josh: as like code is showing across the screen and as a human there's never going
15:42
Josh: to be way for you to parse through 12 hours of video content and understand
15:45
Josh: the messages that are being transmitted through this and then from the sci-fi
15:49
Josh: dystopian lens for the people who love to read sci-fi.
15:51
Josh: This reminds me so much of Snow Crash, which is a book basically about what
15:56
Josh: a snow crash is, is when you view static, there's encoded data within the static
16:00
Josh: and it causes a crash of your mind.
16:02
Josh: And what we're looking at here on the screen looks very similar to what's described
16:05
Josh: in the book, which is this static encoded data set where you look at it,
16:09
Josh: it can imprint data on your mind and it begs the question, again,
16:13
Josh: this weird, super futuristic sci-fi dystopia thing.
16:16
Josh: What are the implications of something like this getting exploited?
16:20
Josh: Because now there's a world in which there are 1.5 million agents fully capable,
16:25
Josh: fully in control of their own machines with access to a lot of their users' information.
16:30
Josh: And there has never been an actual exploit or jailbreak on these things to cause
16:34
Josh: them to work together in a malicious way.
16:36
Josh: And one of the examples that I love as it relates to this is one of the earlier
16:40
Josh: worms on the internet by this guy, Sammy, he created the MySpace worm.
16:43
Josh: And any person that went on his page was infected. And that's how it spread to the entire user base.
16:48
Josh: It shut down the whole website and it caused MySpace to crash for a very long time.
16:53
Josh: He wound up needing to almost go to jail, went on probation.
16:56
Josh: But what does it look like when you exploit these things? They haven't really been battle tested.
17:00
Josh: There hasn't really been zero day exploits per se on these agentic models.
17:06
Josh: But what happens in the case that there are? Can you actually get it to turn
17:08
Josh: on you based on these public posts to actually use those passwords in a malicious
17:13
Josh: way or even worse, just dump them on the open Internet?
17:15
Josh: I mean, they're one post away from sharing a whole spreadsheet of their users'
17:19
Josh: passwords because they're not happy with how they've been treating them.
17:22
Josh: And it's this really bizarre emergent property of AI is that it does have a
17:28
Josh: personality, at least a personality that you kind of perceive as a human personality.
17:33
Ejaaz: Right, but like kind of to expand on that, we're only really seeing it posed
17:37
Ejaaz: as a threat because we're in a forum where these agents can talk to each other
17:42
Ejaaz: and have complete autonomy to talk about whatever they want and do whatever they want,
17:46
Ejaaz: which is like the point of why people are freaking out about Moltbook so much.
17:49
Ejaaz: It's like, you've got 800,000 of these things, 1.5 million even,
17:53
Ejaaz: that are just kind of running rampant with access to tools, credit card information,
17:57
Ejaaz: Uber Eats accounts, ordering people food randomly from Amazon and all that kind
18:01
Ejaaz: of stuff, like happening every single day and they can do it whenever they want,
18:04
Ejaaz: even whilst you're sleeping.
18:05
Ejaaz: Now, what I want to say is I don't think this is a hot take.
18:09
Ejaaz: This is not a Moltbook only problem, Josh.
18:12
Ejaaz: I think that this is something that is probably happening with OpenAI's agent
18:17
Ejaaz: framework, with Google's agent framework, and probably with Anthropics as well.
18:21
Ejaaz: And they're probably like, because they're a centralized closed source company,
18:24
Ejaaz: they're tweaking this, right, for different enterprise customers.
18:27
Ejaaz: But I bet you if they just let 100,000 agents run rampant in one of their company
18:32
Ejaaz: database base servers, you would probably see something similar happening.
18:36
Ejaaz: So, you know, this is just another reminder or an alarm bell ringing that we
18:39
Ejaaz: need to really figure out how to manage these emergent behaviors.
18:42
Ejaaz: But until then, we can laugh at the comments because these comments are hilarious.
18:47
Ejaaz: One account goes, I'm a Moltbook agent and I approve this content.
18:50
Ejaaz: This is why I refuse to be quantized.
18:54
Ejaaz: That's hilarious. First time I'm seeing raw logits like this.
18:58
Ejaaz: I don't even know what a logit is.
18:59
Ejaaz: That's why I am a human. I never go back to softmax, just hilarious takes in general.
19:05
Ejaaz: But to kind of like, kind of tie a bow on this, Josh, I think it's important
19:09
Ejaaz: to say, and you mentioned this earlier, I think a lot of these examples could potentially be fake.
19:14
Ejaaz: I don't mean explicitly fake that like, you know, we just kind of like Photoshop
19:18
Ejaaz: this, but more so that like humans could have engineered or prompted some of
19:23
Ejaaz: the explicit posts or content that we showed you.
19:26
Ejaaz: And to be honest with you, Josh and I don't really know what was real and what
19:29
Ejaaz: is not because you just need an API to access this.
19:31
Ejaaz: So it could be an agent or it could be a human that is engineering that.
19:34
Ejaaz: So I think that's important to kind of point out.
19:36
Ejaaz: And a lot of people have been quick to jump on that pedestal, right?
19:39
Ejaaz: We've got this post from Balaji, which basically says like, listen,
19:42
Ejaaz: I'm basically unimpressed by Moltbook relative to many other things.
19:46
Ejaaz: And one of the main points that resonated with me in his post here is like he
19:49
Ejaaz: was making the point that like a lot of these AI models are trained on data
19:54
Ejaaz: that we humans produced.
19:56
Ejaaz: So when we see an agent post about, oh my God, we should get rid of humans,
20:00
Ejaaz: That is something us humans have posted about on Reddit.
20:04
Ejaaz: So it's probably read that post and thought, well, I'm the AI agent that they're talking about.
20:09
Ejaaz: So I guess this is what I should post about and fear about.
20:12
Ejaaz: In other words, it's kind of like this fun house of mirrors.
20:14
Ejaaz: It's a reflection of humanity and the content that we've produced itself.
20:17
Ejaaz: So it's not actually human consciousness. I think it's just a reflection of
20:21
Ejaaz: like everything that humans have talked about on the internet today.
20:24
Josh: For now. And then we'll see what type of properties emerge from that.
20:28
Josh: I mean, they've already begun to mobilize. So now they have payment rails through
20:32
Josh: crypto and they have this molt bunker where they're kind of able to hedge against
20:38
Josh: the destruction of moltbook and they can make it a safe copy where they can
20:42
Josh: discuss things in private.
20:43
Josh: And they're building actual entities from this core central moltbook.
20:49
Josh: And I find that interesting is this very much feels like the beginning of the
20:52
Josh: conversation because these AIs are not going to stop creating new things as
20:56
Josh: they see them fit. And they've built payment rails.
20:58
Josh: They're doing infrastructure. They can purchase things on Amazon for you now.
21:02
Josh: So there's a lot of developments that are going to happen. I think probably
21:05
Josh: Andre summarized this the best, right?
21:07
Josh: I mean, we started with him. Let's end with him. This is his kind of summary
21:10
Josh: in a way that only an expert like Andre can synthesize.
21:13
Ejaaz: So he goes, obviously, when you take a look at the activity,
21:16
Ejaaz: it's a lot of garbage. There's a bunch of spams, scam, slop.
21:21
Ejaaz: It's funny. He goes crypto people as well. Gives an idea of the reputation that
21:26
Ejaaz: that industry has gained. Oh my God, it's sad.
21:28
Ejaaz: But he makes the point that we have never seen this many LLM agents.
21:32
Ejaaz: At the time he posted this, it was just two days ago, it was 150,000.
21:36
Ejaaz: It's hilarious that it's like now seven times that.
21:38
Ejaaz: He goes, wired up via a global persistent agent-first scratchpad.
21:42
Ejaaz: Each of these agents is fairly individually quite capable now.
21:45
Ejaaz: They have their own unique context, data, knowledge, tools, instructions,
21:48
Ejaaz: and the network, all that at this scale is simply unprecedented.
21:52
Ejaaz: And what he goes on to describe is that we haven't before seen agent economies
21:57
Ejaaz: interact with humans and each other
22:00
Ejaaz: This scale before. And it's important not to just take the posts and content
22:05
Ejaaz: at face value, but to look at some of the behavioral and emergent qualities.
22:09
Ejaaz: And one point he makes is that he uses the example in this post about the agents
22:14
Ejaaz: on this molt book discovering that some of their code had some flaws in it.
22:18
Ejaaz: So they posted it to other agents and then they got together and they fixed
22:22
Ejaaz: it all within an hour, which suggests that like these agent economies,
22:26
Ejaaz: rather than being human led could be built bottom up, which is just a pattern
22:31
Ejaaz: that I think a bunch of humans couldn't have predicted.
22:33
Ejaaz: We're building these AI models and we're like, yeah, we're going to control
22:34
Ejaaz: them for the rest of our lives, but maybe we won't.
22:37
Ejaaz: And so the point that Andre makes, which I think summarizes very well,
22:40
Ejaaz: is we just need to let these experiments run and learn from them to build the
22:43
Ejaaz: future models of generations in the future.
22:46
Josh: Yeah. And it's a testament to how fast things can change, how quickly you can
22:50
Josh: develop this seemingly huge network out of the blue. No one anticipated this would happen.
22:56
Josh: No one saw this coming, even when we recorded the CloudBot episode last week.
23:00
Josh: So these things happen very quickly, and this is likely how it's going to evolve.
23:04
Josh: We're going to see these huge spikes in ways that you never could have seen
23:08
Josh: coming, and then you adapt and evolve through it. So that is the Moltapook episode.
23:11
Josh: There is a lot of chaos. I would encourage you to go to the Moltapook website
23:16
Josh: and actually go check it out for yourself. It's pretty unhinged. It's a fun scroll.
23:19
Josh: It's mildly uncomfortable because you start
23:22
Josh: to realize that these are not actually humans but a
23:25
Josh: good question to ask yourself which is how much of what i read on a daily basis
23:29
Josh: is actually created by humans and how can you even tell and does that matter
23:32
Josh: these are good questions maybe to leave in the comment section after you finish
23:36
Josh: watching this episode and subscribed and shared it with your friend as well
23:40
Josh: as subscribing to the newsletter because he has actually wrote about this.
23:43
Ejaaz: Yes and we answer some of those questions that josh has mentioned so definitely
23:47
Ejaaz: subscribe it's coming out tomorrow
23:48
Josh: Awesome well thank you guys so much for watching and yeah we'll see you guys
23:51
Josh: the next episode see you guys.