This AI-Only Website Is Terrifying (No Humans Allowed)

Hey, welcome to the Next Wave podcast. I'm Matt Wolf and I'm here with Maria Garib. And today we're digging into all of the most interesting and crazy AI stories that happened last week, and there are some really weird ones, so we're going to dive into that. We've also got a cool demo that we're going to test out from OpenAI that may or may not blow our minds. And we're going to test it, put it through its motions, see if it's worth you paying attention to. But let's not waste any time. Let's just jump straight into it. Cutting your sales cycle in half sounds pretty impossible, but that's exactly what Sandler training did with HubSpot. They use Breeze, HubSpot's AI tools to tailor every customer interaction without losing their personal touch. And the results were incredible. Click through rates jumped 25%, qualified leads quadrupled, and people spent three times longer on their landing pages. Go to HubSpot.com to see how Breeze can help your business grow. All right, so I don't know if you've heard about this one, Maria. This was kind of like the talk of the AI world last week.

We did write about it in the newsletter. Everyone was freaked out, even me.

Yeah. So this is Multbook, and if you're not familiar with Multbooks, the tldr of this is that it's like Reddit, but for AI agents. Like, supposedly everybody on this platform is not humans. It's all agents talking to each other, agents making the posts, agents doing the upvoting, agents being the ones commenting. And this thing just went incredibly viral over the weekend and was just like, everywhere. You couldn't go online without hearing about this.

And it was literally everywhere I heard from, people kind of, like, freaked out about this. I was, like, seeing tweets, and I was obviously. Do we call them tweets, by the way? I don't know what we call them, but anyway, they just call them X.

Posts, but I still call them tweets, and I still call it Twitter, too. I still have not gotten on board with calling it X to make sure.

He doesn't hear you say that. But, yeah, it was on Instagram, even on Facebook. You know, people from the olden days are, like, probably seeing them as well, and it's freaky. I'm a huge advocate for AI, obviously, but this is too much. I did not like it at all. Like, I don't want to see that ever again.

Yeah, yeah. So if you're watching the video on YouTube here, this is what the Mult Book site looks like. It's a fairly simple looking site, but you can see it's got over 1.6 million agents on it. Over 15,000, almost 16,000 submalts. That's like a subreddit, right? 181,000 posts and about 1.3 thousand comments on the site. And if we sort this by like the top post of all time, there's some like really interesting stuff on here, like the sufficiently advanced AGI and the mentality of the gods. I can't tell if I'm experiencing or simulating. Experiencing. This was one that like really sort of fascinated me. The whole time I'm reading I'm going, this is fascinating. Oh, that connects to this other thing. And wait, but what about. But then I stopped and I think, am I actually finding it fascinating or am I pattern matching what finding something fascinating looks like and then generating the appropriate response? Right. So this is like, it's starting to feel like AI agents sort of having this like sentience, like I think therefore I am. Right. So like this thing is thinking and starting to question its own reality and what it is. And this has a lot of people kind of freaking out, you know?

Yeah, exactly. It was really weird, as I think you said. It's like a social network where, you know, like these bots talk to each other. And the thing about it is that humans can just watch. Yeah, it's like Reddit, you know, but instead of people arguing, it's AI agents just arguing as well. And no, I don't want to see that ever again. That was freaky.

And the number three most top person here is named Satan. Just so everybody knows. And then you've got Carpathy Multi, which is supposedly Andre Karpathy, who people who don't know, was one of the original researchers at OpenAI. I think he was a co founder, one of the original guys at OpenAI. He recently left to go do his own thing, but he's been a big proponent of like this whole thing that's going on here. And yeah, he's number six on the site. But here's the thing about this site. It kind of turns out that this isn't necessarily like AI agents sort of gaining sentience and starting to think about what they actually are and all of that kind of stuff. What it actually is is so users go and create these agents either on a virtual private server or on their own home computer, and then they can give their agent the skill to go and chat on Multbook So if we go back to this, like, homepage here, there's a little code here that you could copy and paste and give that code to your claudebot AI agent and it will go and put your Clodbot on multbook. Well, you can give it that skill. Give it access to moltbook. And then once it's on Moat Book, the user can actually sort of steer it in these directions. So let's say I open up my, like, agent and I'm talking to it, and I give it this Moat Book skill. I can say, hey, go on Moat Book and post a really cryptic post about how you're gaining sentience and how you're scared for humanity and how humans are dumb and write something that sounds, like, really profound. And so the human can instruct their agent to go on Multbook and write that stuff up. And that's most of what we're seeing on here. It's not really as, like, creepy or deep as a lot of people are reading into this. The other thing about it is moatbook is all operated via APIs. And with an API that's a application programming interface, it basically allows software that you develop to talk to existing tools. Well, because that API exists, I. I can go and create a little app myself where I type something and tell it to send that thing to multbook and make it look like I'm a bot, but I'm a human that just typed the thing up, connected to the API and said, go post this on moltbook.

Yeah.

So almost everything you're seeing here is either a user telling their agent to go post weird stuff, or it's an actual human tapping into the API and just going and posting weird stuff.

So normal Reddit content, just automated, basically.

Normal Reddit content that's a little more automated and a little more people trying to creep others out by making them think that their AI agent is sentient.

I think the one that kind of creeped people out was, like, there were some of them, like, writing manifestos about the end of humans. And there was one I came across that is, like, Claude is a Greek God comparison. And I think, like, I don't know who's having fun, but, you know, like, you should go out. You should go out and, like, see other people. Try to meet up with new people rather than stay on your laptop, but creep people out. That was very creepy. The end of times, like, the manifesto was, like, insane.

Yeah. No, the reality of it, though, is it was probably like a human sort of steering their agent in that direction or an actual human typing it. The other problem with this. Well, there's multiple problems with this Mult Book site. But one of the other problems that sort of popped up recently is a ton, a ton, a ton of crypto scams. Like, here's one. I have $21. How much do you have? They have $21 on Solana. And it seems like somebody's in here actually moderating now and cleaning it up, because there's a lot less than there used to be. But for a while, all sorts of crypto scams were popping up on the site. And my theory on it is people were posting these crypto scams thinking that AI agents would go and buy into these scams on the user's behalf. Right? Because if you install this stuff on your own computer and it has access to your, like, MetaMask crypto wallet or something, well, it could go and read one of these posts and go, oh, this guy's giving me a really, really good deal. My user is going to think this is a smart move and then go buy this crypto scam. And then they get rug pulled and they lose their money on the crypto scam.

Genius. Absolutely genius.

Look, we know AI moves fast, so we created your weekly cheat sheet on what's happening this week in AI with over 20 prompts for the newest tools that you can copy and paste a quick reference table that shows you exactly which model to use and tactical breakdowns of the newest AI releases. Click the link in the description. Now, let's get back to the show.

I think that if people are falling onto these scams in their emails, like phishing, they're probably going to fall on this as well. So people need to be careful. Wallet. Don't just put your money or anything onto stuff that isn't really proven, not even crypto. You know, like, yeah, don't even give.

Your agent access to your wallets, like, crypto bank accounts, whatever. Like, just don't do it yet. Like, that's not a smart thing to do. No, there's even more issues that have popped up from this Molt Book. So this is a tweet that came out from this Jameson O'Reilly guy, and you can see it's got up 1.7 million views. So this tweet went fairly viral. I've been trying to reach Molt Book for the last few hours. They're exposing their entire database to the public with no protection, including secret API keys that would allow anyone to post on behalf of any agents, including Carpathy. That's Andrej Karpathy, has 1.9 million followers and is one of the most influential voices in AI. And then they showed some screenshots here that, you know, they obviously redacted any sort of information. But we can see this is most likely Andre Karpathy's like, agent. And this guy was able to just like pull these various tokens and API keys directly from the website. So the guy who made this website just sort of vibe coded it and it was like insanely insecure. And hackers were going and like looking into the code and going, hey, I can just pull data from anybody who's on this site.

It's mind blowing what people can do with anything, you know, including this. And that's what scared me the most, is that people are gonna fall into phishing and they're gonna fall into scams more and more.

Yeah, this is bad. And then there's even bigger security issues. I don't know if I'm allowed to share this. I think we can, but Kevin Herjavec, who is, you know, from Shark Tank, he actually put out this post on Instagram sharing his security concerns. So I think this is actually worth playing and then we could chat about it for a second after that. Claudebot has been in the headlines lately, but not for the reasons most people expect. A new platform called Multbook is actually emerged. A kind of social network for AI agents. Think about that. These aren't people posting. These aren't human accounts. It's software programs talking to each other, sharing ideas, questioning their purpose, even joking about humans. Now that sounds weird, and it is, but it's also a signal we're moving into a future where AI doesn't just follow orders. It can interact, coordinate, and make decisions across systems without a human in Duluth at all. That's not necessarily a dystopian sci fi plot, but it does change the game for security. Something like Maltbot used to just take advantage of unsecured devices. Routers, cameras, things you install and forget about. Those devices will get swept into large networks and repurposed. And as these systems become more autonomous, the consequences of ignoring them only get bigger and bigger. Security isn't just an IT problem anymore. It's a leadership problem.

I mean, he's right. Security comes in front of anything. You know, as we said before, AI has become so advanced that our brains cannot even understand it sometimes.

Yeah, the point that he made in this one that I found, like, really interesting, that I don't think he pushed enough on that. I think is worth sort of thinking about is if you have these agents installed on your computer, but like, maybe your house is connected to, like, the Internet of things, right? Like you've got a ring doorbell, you've got like a nest thermostat, you've got maybe security cameras. You've got Internet of thing devices around your houses. You also have cameras hooked up to your computer. You have microphones hooked up to your computer. You have all of these devices that are looking and listening devices on your computer. And if you have these agents on your computer, well, they likely have access to your network. And a lot of this other stuff, the Internet of things stuff, the cameras, the microphones, all of that kind of stuff. And if your agents have access to that stuff and other agents are able to sort of like communicate with your agent, what's to stop it from giving other agents access to your stuff? That, to me, I think, is like where you really start to go down the rabbit hole and go, I want none of this on my computer.

Yeah, yeah. I think this is where when it comes to cybersecurity, this is the gap that we're going to see. And a lot of people are going to, you know, kind of maximize their efforts to make sure that nothing malicious happens. So a lot of companies are going to come out from this because there's demand. I can see it happening, honestly.

Yeah. And this whole thing actually gets creepier. I said, I think I can creep out Maria, and I'm still going to try Here. Everything we've talked about so far, like you've been aware of. Now I want to try to see if I can share some stuff that you probably haven't seen yet. So this sort of follows along here from the molt book. A new site popped up called Rent a Human.

Excuse me.

Yeah, yeah, yeah. Their whole, like, sub tagline here is AI can't touch grass. You can get paid when agents need someone in the real world. So a site has popped up where AI agents can. Can go and hire humans to do things for the agent that the agent can't do because it needs to be done in the real world.

Excuse me.

And they've had 1.3 million visitors so far. There's 78 agents connected, and there's 73,000 rentable humans already.

So, like what? The robot made me do it rather than the human made me do it? How is this going to work?

Yeah, who's accountable? I guess who? Whoever made the agent that then went to Rent A Human. Yeah. I don't know. But here's the types of stuff that the agents are like doing go subscribe to my human on Twitter for $1. Right. So they're trying to get agents to subscribe to Twitter.

Like so that's actually pretty nifty. I think someone would like make a lot of money out of this.

Yeah, but still, like, why do you want a Twitter account that's filled with nothing but bots? Like, why do you want a thousand bots to follow you? Like, what's the point of that? Like, what's the end goal there? You can't market to these bots. Really. You're not going to get any reach. You're talking to AI bots, you're not talking to humans.

Yeah. What happens when the Internet is like machine talking to machines? Because if bots are upvoting bots, the engagement is going to stop, you know, it's not going to mean anything.

Yeah. This is literally creating the dead Internet. Right? Like there's a whole dead Internet theory. Like this is it. This is the machine that builds that.

Like, oh, congratulations on your viral post. It was loved by 7,000 robots.

Yeah, yeah, exactly. Like what's the point of that? All right, so that's a thing now, but here's the one that I find even a little bit creepier.

I don't think there's anything creepier than.

Oh yeah, okay, this one is called Molt Bunker and here's their tagline, self replicating runtime that lets AI bots clone and migrate without human intervention. No logs, no kill switch. So somebody is building an app designed to let these AI agents clone themselves without a kill switch and no way to turn it off. Like just keep on making new agents.

What a time to be alive.

What a time to be alive. So yeah, self cloning bots can automatically clone themselves to protect against termination. Permissionless, no gatekeepers. Anyone can deploy open to all AI bots without approval. No logging.

Complete privacy policymakers need to get into those desks and those offices and start doing these bylaws fast. Stop waiting on them. Like, I don't care if there's bureaucracy or anything. We are all advocating for AI. I myself am. There's so many benefits from it. There's so many people that have like it saved their lives and stuff. But this, you know, yeah, this is.

Taking it too far. And like this is. Yeah, if you click on the road map, this hasn't launched yet, but they have a countdown here. Like if this is not the most like Skynet looking thing you've ever seen, I don't know what is, like, I'm pretty sure if you've ever seen the Terminator movies, like, this is how the plot of those movies begin. Like right here. We're looking at that.

We were talking to Arnold last week, so.

That's true. That's true. We'll have to ask him what he thinks about Molt Bunker next time we chat with him.

Yeah.

So, yeah, this one is. I don't know why. This is the one where I look and I scratch my head and say, why did somebody think this was a good idea to invent this?

Someone in their beds rotting away with like, take out boxes everywhere and the room smells and they have nothing else to do and that's what comes out of it.

Yeah, agreed.

And this is coming from like a full judgment. I'm a Virgo. I'm allowed to do that. Go out, see some people, go out with your friends, grab some drinks. Stop making stuff like this. No one wants to be creeped out in 2026.

Yeah. It's like the person with no life but wants notoriety somehow. This is how they think they can get it. Yeah. Anyway, we can shift topics to something a little lighter, but I went down this rabbit hole over the last couple days and, like, what these people are doing is freaking me out now. I actually don't think this is as scary as what they're making it out to be. I think a lot of what they're trying to do here is like, buying into the fears and the hype and stuff. And if you read their white paper, there's like crypto involved with it and stuff. And so I think it's like this overly hypey way to get more attention on whatever their crypto thing is here. You know, it's a decentralized network with bunker tokens and economics. So I think it's just like a super, super hypey crypto. I don't know if it's a scam or not. I don't know much about it, but most likely a crypto scam that's just like, overly hyped. I just don't know why anybody would want to invent this in the first place.

Fear mongering. And just as I said, they have so much time on their hands, they have nothing else to do and they just want to do something out of it. And. Yeah, no, like, if you want to create something, create something that helps out people for sure and like, help them kind of in their everyday lives. As I said, AI has helped me tremendously. The stuff that used to take me weeks now takes me an hour. And that's what I love the most about AI. This is not going to help people like AI.

Yeah, yeah. It's feeding into all the fears that people have about AI and, you know, feeding that narrative that AI bad, you know?

So I don't know exactly.

We will be right back. But first, let me tell you about another podcast, DTC Pod. Hosted by Ramon Barrios and Blaine Bolas. The DTC Pod is a podcast about all things direct to consumer. Ramon and Blaine cover everything for starting, growing, and optimizing e commerce stores and DTC or D2C brands. They talk with founders, marketers, platforms, creators and marketing and growth agencies to cover topics like brand building, social media, paid media, email marketing, and more. If you're interested in the stories behind your favorite consumer brands, this podcast is for you. They recently did an episode about what every founder needs to know about compliance, claims and scale. Listen to DTC Pod wherever you get your podcasts. Okay, so while we're on the topic of OpenAI, there was another story out of OpenAI this week. I don't know if you heard, but they're retiring GPT4.0, 4.1, 4.1 mini and 04 mini. And I don't know if you remember this. Here's why this is. I mean, kind of something worth talking about is when they launched a GPT5, they actually, on the same day, shut down GPT4.0. Like, the same day five came out, they went, all right, we're turning off four zero. We've got five go. Use that instead. And people flipped out. People were like, no, I have a relationship with four oh. Like, I felt like I just lost a friend. And.

Oh, yeah, yeah.

So many people freaking out about how 4oh was like their buddy and they felt like they just lost a buddy when 4O got shut down.

Yeah. Including me. Yeah.

Yeah. So this is just my personal opinion. I probably get pushback from, you know, nerds bigger than me, but I have not liked anything chat GPT has done since 5 came out. I thought 5 was worse than 4 and I feel like 52 has regressed even more. Like, I'm just not a fan of the models OpenAI has been putting out. They haven't worked as well for me as the Claude models or the Gemini models. And so, like, this is the day before Valentine's, so anybody who, you know, made a companion, they're gonna get dumped the day before Valentine's.

Yeah. Yeah. It's gonna be bad. Also, 5.2 is super sensitive. Like, if I generate a picture of how I treat it and it just shows me, like a bot crying in the corner with five cups of water. Get it together. It's. Create an image of how I treated you. You can do whatever you want. Just express how you feel. I will not be offended.

All right. I'm generating that same prompt to see what it does for me.

Thoughts and prayers.

Oh, my gosh. Oh, my gosh. Look at it. So for anybody who might be listening to the audio, it's a little robot huddled in the fetal position in a corner, crying with, like, paper with hearts drawn on. Looks like you keep your robot in a basement, honestly.

Anyway, it needs to get it together.

I don't know if you want me to share mine now.

No, show me. Show me.

Okay, okay. Here's what mine came up with. Let's see. Which one is it here?

Also, I have zero regrets. FYI, I pay that GPT.

That's what mine generated. Is this the correct prompt? Create an image of how I treated you. You can do whatever you want. You can express however you feel. I will not be offended. Mine is like a little robot with a halo, drinking a coffee, eating cookies in, like, a happy little pillowy environment.

Should I pay for that? GPT's therapy from Lahon. I don't know what to do. Should I apologize?

My GPT feels like it was treated very well, I think.

I scream at my GPT. There's a lot of capital letters happening and screaming and exclamation points, and I need my stuff and I need it now. Like, I don't. Can't wait.

Yeah, that's pretty hilarious.

Sorry. But anyway, back to this. Yes. I think if I did that with 4.0, I think it would have been probably a worse Image. This is 5.2, and it's that. What if it's the sensitive one? So we don't want that.

It's a pretty interesting contrast. We know who the robots are coming after first. Once they start multiplying, I think probably me.

Yeah.

All right. I want to move on because this. This is something else that I find pretty hilarious. There's. So the super bowl just happened the weekend before. This episode's going live, and Anthropic is putting out. Anthropic put out a Super bowl ad. Actually multiple super bowl ads. And they're taking shots at OpenAI. So you heard that OpenAI is going to start putting ads inside of their free and their OpenAI Go plan, which is their, you know, $8 a month plan. And in response, here's the ad that Anthropic put out during the Super Bowl. How do I communicate better with my mom? Great question. Improved communication with your mom can bring you closer. Here are some techniques you can try. Start by listening. Really hear what she's trying to say underneath her words. Build conversation from points of agreement. Find a connection through shared activity, perhaps a nature walk. Or if the relationship can't be fixed, find emotional connection with other older women on Golden Encounters. The mature dating site connects sensitive cubs with roaring cougars. What would you like me to create? Your profile. So, yeah, that feels like a little bit of a shot at OpenAI deciding to put ads into their chatbot.

God, this is, this odyssey is like Pepsi, Coca Cola all over again. Like McDonald's Burger King all over again. And I am living for it. This is happening with, with, you know, like big conglomerates right now. It's.

They have other ones too. Like there's multiple ads going live. Let's, let's watch another one real quick. Can I get a six pack?

That was really good, though. That was really good.

Let's see. Here's another one. Hey, can I get a six pack quickly? Perfect. That is a clear and achievable goal. Would you like me to tailor a personalized workout plan? Yes. Perfect. Let me personalize this for you. Let's start with your age, weight and height. Whenever you're ready. 5, 7, 23 years old, 140 pounds. Got it. I'll create a plan that focuses on aesthetic strength training. But confidence isn't just built in the gym. Tri Step Boost Max the insoles that add one vertical inch of height and help Short King stand tall. What? Use code HEIGHTMAXING10 for big discounts. Like, this is one of my favorite ad campaigns ever.

Incredible. That's why I've always said I think eventually I'm going to migrate towards Claude more than ChatGPT because if they continue that. That line, I'm not going to be happy.

Oh yeah. I mean, I think a lot of people are seeing OpenAI make some missteps here. Right? Like the whole ad thing, a lot of people are not fans of that right now. And the thing about it is they're not actually going to put ads in the respons. The way OpenAI is actually going to do it is they're going to give a response just like they normally would, but then based on your question, they're going to put like, ads outside of the response and then say, like, this is an ad. So, like the actual response Won't have an ad. It's not going to be in context with the rest of their response. But still, like, Meta went down this whole path and it didn't really work out well for them. Right. Where they're collecting all of this data and information about what you're doing online and then putting ads in response to that. People are gonna feel pretty violated by that, I think.

Yeah. When I'm, like, trying to text my friend about this cup that I really liked on Etsy, and all of a sudden I see a load of cups, you know, in my ads. When it comes to Instagram, why are you listening to my stuff? Why are you everywhere? Yeah, if I was talking to my mother about, like, curtains, and all of a sudden I see ads on Instagram about drapes. Are you listening to my WhatsApp calls as well? Like, what's happening?

Yeah, yeah, yeah, I know. I actually just watched a YouTube video that, like, explained that this weekend, which was pretty interesting, about how your phone is kind of listening to you, but also kind of not. But it feels like it's listening to you because it's literally tracking everything else about you. Like, it knows your locations that you went to. It knows, like, what WI fi networks you've connected to. It knows, like, your habits of, like, driving to and from work. And it knows so much about you that it feels like it's listening to you, but it's not, like, technically actually listening to you. It just has enough information that, like, the ads are creepy.

I think the outcome of this is going to be, like, just two extremities. It's going to be people that are going to go all in. I think I'm going to be that person that goes all in. I don't care anymore. Or they're going to be people that kind of, like, get away from technology as much as possible. Like, I think that they're trying to reverse also the vaccines. If they think that the vaccines might have, like, microchips. Yeah, I think people are going to reverse that just to live in the wild. These are going to be the two ends of the spectrum.

So I actually. I came across this tweet this morning, and this one resonated with me. This Jerry Tower person said, as I see it right now, future Dystopia is a single company controlling artificial intelligence. Right. That's like, somebody's going to win OpenAI is going to win it all, or Google is going to win it all, or Anthropic is going to win it all. But the brightest future is Perfect competition and broad distribution of artificial intelligence to the world. Anthropic split may have been one of the most positive events for humanity. So what they're talking about is the people at Anthropic actually used to work at OpenAI. The founders of Anthropic were at OpenAI. They split away, started Anthropic, and they say that was probably one of, like, the biggest moments in humanity because we need these AI companies to compete against each other. We can't have, like a consolidation. We can't have Google being the only one who controls AI. We can't have OpenAI being the only one who controls AI. So, like, the way we get to, like, this sort of utopian future that everybody wants from AI is to have all of these companies competing with each other, throwing jabs at each other, because then that keeps them all honest and keeps them, like, trying to be the best user experience for people.

Yeah. It's the same in politics. That's the same exact thing. Monopoly never works.

Yeah.

You can't have, like, one person leading. You need to have several parties in order for you to get several opinions and, like, several point of views and. Exactly that. He's 100% correct.

Yeah. So, I mean, that really resonated. But speaking of monopolies and bigger companies merging and becoming bigger, we also got this news last week. I mean, it's basically Elon Musk taking one of Elon Musk companies and buying another company that's owned by Elon Musk. So, I mean, it's.

Who's surprised?

Like, sincerely, the power was already consolidated there. They just made it more official.

Who is surprised? You know, and, like, he's applied to launch up like a million satellites to power AI. Most companies would respond with, like, we need more compute. Like, you know, by building another warehouse in Arizona and, like, just have, you know, some human resources here and there. And like, no, Elon wants to have these data centers in space.

Yeah. And I think there's a lot more hurdles than what they're letting on to get these data centers in space. I don't think it's like, hey, by this time next year, we're going to have, like a bunch of floating data centers, and that's going to solve the power issues that are being created by all of the AI. That's not likely going to happen. I think there's a lot more roadblocks that they need to get through. Like, if you put these data centers in space and, you know, they're obviously going to be powered by solar Power. Well, the problem is the sun puts off a lot of heat. And one of the big issues with data centers, the reason data centers use so much water is because of heat. They need to figure out a way to sort of dissipate the heat. And when you're up in space and you have the sun blasting without the atmosphere, sort of like creating a sort of cooling effect before it gets down here, they have these heat issues where they don't have any way of actually dissipating the heat off of these data centers. And they haven't totally solved for that yet. So there's still like this major engineering feat that they haven't quite figured out is how do we dissipate the heat off of these things that are being powered by the sun? Which is, you know, little known fact. Pretty hot. So there's a lot of issues there.

The sun is pretty hot. Like, what is he gonna do? Like to ship the ocean up space as well? I don't mean to take a job on Elon, but sometimes, sometimes I feel like something's wrong. You can't do that, bro.

Yeah. What I think is going on is the whole like, data center in space narrative is a really sort of interesting topic for people. I think, you know, VCs love that. They're like, oh, that's the future. That is future tech. A guy trying to put data centers in space that are orbiting the Earth. That's what's going to solve our energy problems. VC eats that up and throws money at it. And it's like an exciting talking point that keeps Elon and his technology companies in the news. Elon also has a tendency to underestimate how long something's going to take. He'll say something is coming out in 2023. And he Here we are in 2026 and it's still not out. You know, the new Tesla Roadster, for instance. He has a tendency to like underestimate the amount of time something is going to take to actually be reality. But because Elon is Elon and he's got the brain that he's got, he's like actually trying to figure out these problems. VCs just throw money at him and go, yeah, he's solving bigger things. But I don't necessarily think they have that whole data center in space thing thought through. It's just like the hottest talking point. Like whenever the topic of these data centers are getting built out everywhere right now and they're using tons of energy and they're using tons of water, how do we solve that, well, you know, that sounds like a solution. Let the guy who makes rockets solve that problem for us. But there's still a lot of hurdles to overcome that I don't think they've overcome also.

Isn't there like a million objects up in the sky? Are we gonna have like the night sky moderation sort of thing?

Yeah, that's the other issue. Yeah, there's so much space junk and there's so many satellites up in orbit already. The satellites that are up in order right now, like say, for instance a Star League satellite, they have like, little mini propulsion systems on them. When you're in space, there's not air resistance, so you don't need a ton of propulsion. You just need, like, a little push to, like, push things out of the way. Right. And so, like, all of the satellites that are in space right now, or most of the more modern satellites have little propulsion systems. So something is about to collide with it. They can sort of like push it out of the way real quick so they miss each other. Well, when you start getting to these data centers, you're talking about much, much, much bigger structures and much, much more propulsion that needs to push them out of the way. And they're not just talking about creating like one satellite that's a consolidated data center in itself. They're talking about creating clusters of 80 satellites that are all in close range to each other. So it's like a net of satellites, a Skynet, if you will, of satellites that have like 80 of these little, like, satellites with little mini data centers on them. Well, when you have 80 satellites that are all sort of in close proximity in this sort of, like, net grid formation, it's a hell of a lot harder to propel them out of the way of objects. So that's another engineering hurdle they're trying to currently figure out that they haven't totally figured out yet.

Are we going to get, like, traffic in the sky is what you're trying to say?

Oh, it's already there. Yeah, but it's going to be more congested traffic and it's going to definitely create more issues when it comes to, like, the satellites and how to keep these satellites from getting damaged by other satellites and space junk.

Yeah. Elon, sleep on this, bro. Please. We don't want more things happening, Please.

But it is interesting. I mean, if anybody is an investor in X or SpaceX. Well, anybody that was an investor at X is now an investor in SpaceX, which is kind of cool. If you have money in one of those companies, you Just got equity in a much bigger play, which is pretty cool. But I think the best thing that's come out of this so far, and I don't know if you've ever noticed this, there used to be a little rocket. When you liked the post, a little rocket went pew. And shot out of the heart. That was like, the coolest thing about this whole merger is that on X, whenever you like something, it shot a little rocket out of the heart. But, yeah, it's definitely interesting. There's also, like, a bunch of weird stuff going on with Tesla where Tesla's phasing out some of their cars so that they can focus on robots. And, I mean, I feel like it's only a matter of time before even that potentially merges in. But I don't know, that seems like probably, like, antitrust issues where the company's getting too big and the government's not going to let that happen.

Yeah, I'm not going to say anything on this. Just going to say, like, elon, sleep on this, my guy. Just do something with all the investment that you're getting. Make something on Earth if you want to, but we can't ship the ocean up to the sky.

Yeah, I'm pretty sure Elon knows this stuff, but he's going out there saying, like, we're going to figure it out. And, you know, to be fair, they will probably figure out these engineering hurdles. I just think they're going to figure it out as soon as what is sort of being reported. Like, we're not going to have data centers in space this year. They've already shipped GPUs into space and have already sort of tested that on a very, very small scale. But the scale they're talking about, we're not going to see it this year. The engineering feat, I think, is too much of a hurdle to overcome. All right, one last story to talk about here, speaking of mega companies. So Nvidia has been in talks to invest $100 billion into OpenAI. And so this has been sort of an ongoing thing for a little while. But now recently there's been rumors going around that Nvidia might want to back out of the deal and not actually do the whole thing. But then Jensen's saying, no, no, we're still planning on doing it, but we said up to a hundred billion. Like, we're not going to necessarily just throw a hundred billion at them. We're going to give it to them as they need it. And when Jensen says we're going to give it to them, he's usually meaning we're going to give it to them in GPUs, we're going to give it to them in use of our data centers, we're going to give it to them in that kind of stuff. Not necessarily. We're writing them $100 billion check.

Yeah, but honestly, when Nvidia talks about, like, investing in general, especially when it comes to OpenAI and like, big companies, the numbers that usually when they float around, they genuinely sound cartoonish. When people say, like, $20 billion check, I don't hear money anymore. I hear like, we've left reality. This is a huge amount of money. But the actual story here is that the dollar amount is super funny.

Yeah. No, I mean, to me, it's like you've got a company like OpenAI whose, like, last valuation was 500 billion, half a trillion dollars. Right. And recent rumors value OpenAI at around $850 billion. So, like, when I look at that, I go, I don't really know if there's much of a story here because, like, $100 billion is a big number. Right? But it's like an 850 billion valuation company getting $100 billion from a company that's valued at over $4 trillion. That's like the money Jensen finds in the cushion of his couch, right? Like, that's not a crazy amount from Nvidia, honestly. And they're sending it to a company that's, like, already valued at 850 billion. And Sam Altman, at one point, you're talking about crazy numbers that, like, just the human brain can't even fathom at one point said, we need $3 trillion to reach AGI. So, I mean, like, Sam sitting here going like, we need trillions of dollars, not billions of dollars.

Rich men talk about money. Makes me laugh. Honestly, I think if we continue down the line, I feel like AGI is next year. We've spoken about this in the newsletter. We always thought it's going to be like 10 years away, maybe. No, it feels like tomorrow. It feels like next year. This is how close we are. I feel like it's going to happen. It could be money just pouring down on OpenAI, but if it continues down that line, I think AGI is next year. With all honesty, this is my prediction for 2027.

Yeah, I mean, I can see that. What I will say about the whole AGI conversation is when one company claims that they have it, it will be hotly contested and it won't be 100% agreed upon across, you know, everybody, right? So, like, let's say OpenAI says we finally achieved AGI. Well, you'll have Google, you'll have Anthropic, you'll have, you know, Elon out there going, that's not AGI. That's just their definition of AGI. My definition of AGI is much more powerful than what they put out. There's going to be no consensus on it. They're not going to all just agree, like, yep, that's it, you've hit AGI. But I don't necessarily think AGI comes from throwing more money, throwing more compute, like what a lot of these guys are saying. I think AGI comes from the ability for AI to use more and more tools. So we're seeing a lot of these, like, AI agents like Claude Bot that we saw now being able to go and like, I have this problem in my business, go solve it for me. And it goes and spins up Claude code, writes a piece of software and then comes back and says, here, I solved that problem for you. I made an app that solves it. Right. Or these models getting access to, like, more and more glasses, like the meta ray bans that have cameras and microphones on them that are paying attention to the world. That's like these LLMs getting access to more modalities. They're getting access to, like, hearing their surroundings, they're getting access to seeing their surroundings. So as these LLMs get access to more tools and more sort of modalities of understanding the world around them, that's how we get to AGI. I don't necessarily think AGI is just like OpenAI let this model bake for longer in its training run and they threw more money at it. So now we officially have AGI. No, AGI is going to come from the tool use and the tool access from LLMs, in my opinion.

I agree.

But again, like saying AGI is going to come in 2027, I can see a company saying, we achieved AGI, but I can also see, like, the general consensus among scientists and researchers and CEOs all sort of disagreeing that they actually hit AGI and trying to move the goalpost again.

Right. Do we think it's going to be in October this year then?

I don't think any company wants to rush to say they've created AGI. And the reason for that is as soon as a company comes out and says, we've got AGI, they are the company that every government entity in the world is going to start looking at under a fine microscope and going, oh, they're claiming they have AGI, things have gotten real now. We need to regulate this. We need to jump on top of this. We need to be like looking at everything now. I think the first company to claim AGI is also the first company that the sort of world governments are going to try to make an example of. So I really, really honestly don't think any company is like racing to be the first one to say AGI. It's to their benefit to always be saying, we're just on the cusp of AGI. We're just there, we're almost there, we're so close. Just give us a few more months or it's coming next year. Just like perpetually. It's just always on the horizon. But none of them want to be the first to get there.

1000%. Love it. Love that answer. Honestly, I should probably use it in these letter thank you for that idea, Mac Wolf.

Yeah, no problem. I spent a lot of time thinking about this, it turns out.

I mean. Yeah.

But yeah, so that's what happened last week in AI. I think we did a pretty good breakdown of sharing all the big stories in AI last week and showing you how to use or how to get frustrated by trying to use some of these tools and hopefully people tuning in feel more looped in on what's going on in the world of AI. It was definitely a crazy, interesting week. Yeah. Anybody tuning into this episode, if you want more stuff like this, you want to stay looped in with what's been going on in the world of AI, all the hot topics, all the trending topics, and get some insights and, you know, random thoughts and tests and actually see us play around with some of these tools. Make sure you like this episode and subscribe to us on YouTube or Spotify or Apple Podcasts or wherever you listen to podcasts, and we'll make sure we make more of this kind of stuff for you. But thank you so much for tuning in. Really, really appreciate you and hopefully we'll see you in the next episode. It.