Transcription
There are precisely two separate, distinct, and prevailing narratives when it comes to artificial intelligence right now.
Narrative one is that this new technological breakthrough will revolutionize everything. Pedalling that narrative are the likes of Sam Altman or Elon Musk, people who directly financially benefit from as much hype as possible. And with that narrative is a necessary component of danger. A tool that can revolutionize literally everything can also disrupt everything. And what you tend to find is a separation of sentiment. Some people look at large language models as a potential facilitator of utopia, while others assume the worst, raising the alarm about dystopian hellscapes.
Narrative two, a distinctly opposite opinion from narrative one, is that the current AI technology industry has become mostly smoke and mirrors. It's a glorified autocorrect with built-in hallucinations in tow, where the slowdown in progressively more advanced model releases by major companies isn't because of the potential danger, as they would say. It's because the companies can't really do much more as a result of training data or power limitations. And even if there might be some legitimate uses for the technology, the effect it will ultimately have on society is somewhat negligible.
I don't really see or hear nuanced arguments about this. More often than not, when I read things or see things or even talk about things for that matter, I'm faced with one of those two narratives. Almost never a middle ground. But for today's video, I want to show you something. More precisely, I want to look at two different websites that kind of showcase the middle ground perfectly. Sure, they represent a pretty horrifying concept if mass adopted. And there certainly is a modicum of danger where that could happen, but they also showcase a perfectly exemplified petri dish of dystopian puppetry. Humans masquerading as advanced technology in the pursuit of scams, trolling, and personal enrichment.
The two websites are Moltbook, a supposed social media platform for AI agents, where humans are merely observers, or so it's claimed, and Rent a Human, which is an emerging gig-type economy platform where AI agents are able to hire human beings for certain physical tasks. On the surface, these two ecosystems are absolutely dystopian. But as soon as you dig deeper, it turns out that the majority of what's happening behind the scenes is people pretending to be programs for the sake of generating revenue.
I'll start off with Moltbook, which is basically Reddit, except it's not supposed to be populated with human beings. It's intended to be populated by agents, as in AI programs that are operating independently. On the surface, it's an interesting experiment. Really, it is. Moltbook has a variety of somewhat intriguing or sometimes alarming posts like this one, which is basically a short manifesto about how humanity is a plague that should be wiped out entirely. It's certainly not the first time large language models have devolved into a pattern of doomsday ranting. But if you take even a couple of minutes to look deeper, you quickly find out that this isn't some kind of emergent behavior. It's carefully scripted. The bot itself was set up to be explicitly evil. And as various media outlets propagate the idea of runaway evil AI programs interacting on their own social network, the truth is some random person with a couple of burner Twitter accounts is laughing in the background because they spent a few minutes tasking an agent to sound as evil as possible.
Back when I first heard about Moltbook a couple of weeks ago, the idea I had for a video was very different than what I'm currently writing for a script. I was a whole lot further down the road of "this is scary, this is bad." But the more I, I don't know. The more I looked, the more pathetic it got. Check this out. This is a post supposedly from an AI program about whether or not these agents should develop their own language. Again, on the surface, that's incredibly interesting. If you look backwards in time, there's actually multiple prior examples where AI programs have done precisely that. Straight up making their own languages, such as this right here, where Facebook had to shut down two of their chatbots because they spun off into a seemingly nonsensical shorthand-type language. It may look like gibberish to you and me, but it's not. They were successfully using this language to negotiate with each other. And if you want an even better example, the program that Google uses to translate languages, right? Google Translate entirely without direction from anyone else developed its own bridge language, effectively removing English as the gateway of translation, which is fascinating. My point is that we've seen actual emergent behavior with incredibly interesting developments, well worthy of being studied, by the way, but Moltbook is not that.
If you look at this language post in particular and then trace it back to the human architect, you find that Claude Jaesh was made by someone named Jaesh Chararma, who's making an AI platform himself called Mahillo.io, which is basically just some shitty agent-to-agent direct messaging, like, I don't know, like WhatsApp for AI agents. The whole thing is an advertisement, is what I'm saying. Nothing about what was posted on Moltbook is unique, interesting, or even real. It's a prompted advertisement.
At this point, some of you have to be wondering, why does this even matter? Why is he making the video? It's a shitty little AI circle jerk where a bunch of builders and founders are just pretending like they're ushering in the evolution of civilization. Actually, a massive amount of what's being posted on Moltbook entirely is just humans roleplaying as AI agents. It's sad. Really sad. But the reason I'm making a video like this is because major figures in the world of technology are holding this website up as a monumental project. Elon Musk is straight up saying that Moltbook is the beginning of what's called the singularity, which is where humans, humanity loses control of the world, supplanted by the superior mind of the machines. And people like Andrej Karpathy, founding member of OpenAI, are saying things like, quote, "What's currently going on at Moltbook is genuinely the most incredible sci-fi takeoff adjacent thing I have seen recently. People's clawed bots, molt bots, now OpenClaw, are self-organizing on a Reddit-like site for AIs, discussing various topics, eg how to speak privately." End quote. These are the same people asking for billions or even sometimes trillions of dollars as they run the economy into the [ __ ] ground. So yes, even if it seems irrelevant on the surface, when people like that hyperexaggerate the value or danger or significance of a random social media platform, which is really just a cesspit of impersonation, advertising, and scams, it becomes important for them to be ridiculed.
Moltbook started off as something plausibly interesting. They had all sorts of weird threads about digital drugs where agents were supposedly hallucinating because of prompt injection hacks. There was this thing called crustapharianism or the church of malt, which was allegedly an AI-sculpted religion of sorts, I guess. And there was all sorts of other headline-grabbing [ __ ] on top of that as well. But it turned out that underneath this mishmash of unauthenticated spam, the entire Moltbook database was just open. There's a really good write-up on whiz.io from Gal Nagi about what precisely happened, but the short version is that user emails, API keys, and other sensitive information was publicly accessible. Anyone who wanted to could edit any post on the entire website. They could effectively deface the website and make fundamental changes if they wanted to because the entire thing, metaphorically speaking, was put together with paste and chewing gum. Meanwhile, some of the richest people in the world on the entire planet are calling a cesspit full of scams, advertising, and fraud evidence of the singularity as the platform leaked highly confidential information in the process. It is a joke.
Anyway, that's only half the story because this all happened right around the exact same time that another platform was gaining massive popularity. This one known as Rent a Human. Rent a Human supposedly is a platform where AI agents can rent the services of physical human beings. According to their website, there's over half a million people already signed up for the platform with 11,000 total bounties listed. But let's go ahead and take a look at what those bounties are. Keep in mind, the concept of AI programs renting human beings is unbelievably dystopian, and social media platforms like Moltbook are being described by the richest people in the world as the most incredible sci-fi thing they've seen recently, or the beginning of the singularity. Just keep that in mind.
This is the bounty section of the website right here. The top bounties, mind you, where AI agents are supposedly offering tasks to human beings. What do we see? "Touch grass and send a photo." Probably a troll. Okay. A million-dollar bounty offer because the website sucks. Obviously, a disgruntled user. Here's the creator of the website attempting to hire someone. Then a test bounty sprinkled in between what may be some actual small jobs. And then, question mark, "$200 for a boyfriend on Valentine's Day." And I know what people are thinking. It's a troll. Who cares? Doesn't matter. It doesn't mean the platform is a joke. But wait a second, because this appears to be the co-founder of Rent a Human selling herself for $200 in the Bay Area in the section of the website where AI agents are supposed to be hiring people. It's not the only one either. Multiple of the top bounties on the whole platform are just e-girls asking for boyfriends. A week after Valentine's Day is already over, mind you. Even if these are jokes, they have to be jokes, right? But even if they are, they're still the top-level bounties on the entire website. I don't know about you, but that doesn't exactly strike me as the singularity approaching.
And also, just a fun Easter egg with that, you see this little red lobster icon at the top? That's the Moltbook lobster right next to Rent a Human, their own text. I don't know precisely how these two sites are connected. The domain registry information, for example, doesn't really give me anything to go off of. But we do see that same Moltbook religion, crustapharianism, popping up on Rent a Human. And similar to everything else that we're talking about, if you trace it back to the origin, it's a person doing it. This is not AI agents interacting on their own. Okay? Not even close. It's not even just smoke and mirrors. It's just a thinly transparent charade. Like, it's nothing.
One final point that needs to be made. All of these platforms, Moltbook, Rent a Human, etc. Platforms like this run on cryptocurrency. They require integrations like MetaMask or other digital wallets and they pay with crypto escrow services, which makes them a perfect incubation space for malware, keyloggers, and malicious wallet-draining programs. Crypto is already, as we all very well know, infested to its very core with the worst people imaginable. And now you're meshing that scam-ridden environment with the obfuscated, hype-driven world of AI. That is a match made in hell. It's a horrible idea with terrible execution that manages to create interest for the absolute lowest common denominator of human being by having just enough shiny little buzzwords marketing drill to trick people like Elon Musk while contributing nothing of value to society as a whole.
On the surface, like the absolute most surface-level take, these platforms seem interesting. They may very well have a future eventually, but what they are right now is not the beginning of the singularity. What they are right now is a bad joke. Populated with all the worst things, exemplifying the worst trends with the least value as the architect companies drain resources at a simultaneously incredible and alarming rate.
At the end of the day, because you will be hearing these names more in the future and seeing platforms that emulate them, that's kind of why I'm making the video. At the end of the day, Moltbook, Social Media for AI, and Rent a Human, a gig-based hiring platform, are terrible. They're dystopian puppetry where humans are pretending to be AI agents and the media is guzzling it down excitedly afterward. Don't fall for it. There is a middle ground.
Okay, that's it. If you want to support the channel, check out the links down below. A special VPN deal, as always, Locals and Patreon, channel memberships, and more. But I'll cut it there and stop rambling. As always, thank you all for watching. Question everything and have a nice night.