AI Founder: "I Watched a Guy Hack a Government With AI" and Anyone Can Do This Now | Ep 8

00:00:00:06 - 00:00:20:16

 

Nothing is safe. They did something recently. It's a new form of jailbreaking. It's called obliterating, where they can reduce the guardrails by like 95 plus percent. You can get this all open source. Alex Smith is a resident AI expert, owner of an entity that has retail and 100,000 stores. All of the zero day attacks that used to be limited to just governments.

 

00:00:20:17 - 00:00:39:22

 

Now they're for everybody. A guy literally attacked the Mexican government with jailbroken AI. Posted it all over social media. Oh my God. The whole world can come to a halt. Regular people can be cyber terrorists and they have no idea what they're doing. So what do we do? Alex, what do we do? My opinion is we're in inning two of the baseball game of AI.

 

00:00:39:23 - 00:00:53:01

 

Inning one was like chatbots. Inning two was open. Claw dropped down files, AI agents. Inning three is where we're going to start seeing.

 

00:00:53:03 - 00:01:16:13

 

Elon Musk has described artificial intelligence as the supersonic tsunami that is sweeping the planet, meaning a wave so tall, so wide, traveling at a speed unfathomable that it is going to consume the planet and reshape reality as we know it. So Alex Smith is a resident AI expert, owner of an entity that has massive scale deployment, retail and 100,000 stores.

 

00:01:16:15 - 00:01:46:14

 

Access to the backroom channels and conversations most people don't. What is it that has changed in artificial intelligence in, say, the last six months that most people are profoundly unaware of? Well, thanks for having me, Brady. Yeah, one. What I will say is artificial intelligence is so different than most industries. So there's this saying that the niches are the riches are in the niches with most industries in AI would actually argue it's the opposite.

 

00:01:46:17 - 00:02:19:01

 

We are moving toward a place where these AI companies are going to control top down commerce, so you won't see a AI platform that is specific to only one little thing. You're going to see an AI platform that's like anthropic, where every single day they're adding a new tool, they're sniping something from a competitor, they're adding a layer, and they're going to get so big and so massively sweeping across all industries that it will be like a tidal wave.

 

00:02:19:01 - 00:02:44:17

 

And that is just on the software side. We're not even talking about what happens when hardware is introduced, when robots are introduced, automation and everything is interconnected. It's it's a wave. It's a very, very big wave. Yeah. It's weird. Like there's all these gimmicky ways that the internet is being integrated into consumer products. Like no one needs an internet enabled fridge.

 

00:02:44:19 - 00:03:09:18

 

I think that's dumb. I don't need an internet enabled cat watering machine, because there's so much limited utility in what you can actually do. Oh, it's got the internet, but like it doesn't have any autonomous functions. All that it does is limited to the scope that the programmer that built the OS for my cats watering device. It has two, three basic functions, and it's just recursively just does those same things.

 

00:03:09:18 - 00:03:31:22

 

The fridge, it does the same thing. So you can check your Facebook on your fridge. There's no point in that until intelligence gets applied. Exactly. Then everything changes. Now your fridge is like forecasting when your food is going to expire because it has a and awareness of like how much you know, shelf life is in your grapes and like this thing that you just bought, like the actual application of it is going to change the way that the world looks.

 

00:03:31:22 - 00:03:51:18

 

Because now all of the things we're like, oh, it's we have smart devices. It's like, no, they're Dom, they just have an internet access in like a very limited, you know, software layer that's going to get blown wide open. What are you seeing that looks like. Yeah. Yeah. So basically what happened was is Alibaba did a push notification on their app.

 

00:03:51:19 - 00:04:08:18

 

They told everybody that was a part of the Alibaba app that, hey, we're going to use Quinn to test out this new delivery system, this mechanism of sending a drink to you wherever you're at, whether you're in a park, whether you're in a restaurant, whether you're at work, it doesn't matter. They hit the button. The AI agents did literally everything.

 

00:04:08:18 - 00:04:36:05

 

They ordered it from the shop. They dispatched the delivery driver. They charge the person A to Z, and it caused an absolute riot because every single link in the supply chain broke down. The printers ran out of paper. There weren't enough employees at the stores to actually make the drinks. The delivery drivers. There's a video of it where there's like 30 delivery drivers huddled around this small little mall shop where they're making these drinks and they're yelling at these two employees.

 

00:04:36:05 - 00:05:03:05

 

You can see the printer paper in the background flowing like all over the place. There's drinks all over the counter. Delivery drivers are like getting ready to punch each other. Madness. So the reason I bring that up is because, like you said, with respect to the internet of things like the fridge that's connected to an AI agent or that's connected to the internet, there is going to come a time when the Internet of Things plus AI agents, plus hardware.

 

00:05:03:05 - 00:05:29:06

 

So like delivery drones or, you know, robots that actually, like, deliver something to you where all of this intersects. Now, I want to take this idea of the fear side of things out of it, because we are so far from that point. The technology is there. Yes, but the point at which all of these things connect and society is running around all of this, we are still, in my opinion, 20 to 30 years away.

 

00:05:29:11 - 00:05:48:21

 

There's not enough infrastructure, there's not enough electricity. It there's a lot of things that have to happen before we get to that point. And in between then and now, there is a lot of money to be made for anybody that wants to take part in this. Well, I mean, in that framework, everyone should hear that and be excited about AI.

 

00:05:48:22 - 00:06:17:19

 

But they're not they're not at all. And I you know, there was a popularity ranking pulling people in the sentiment around AI from the people that responded to the survey is that AI is less popular than Ice, than immigration enforcement, which is like a pretty low bar. It's like it was literally it was like Democrats ice AI. But so why is it that the sentiment around artificial intelligence is so abysmal and is it Dario's fault?

 

00:06:17:20 - 00:06:42:13

 

So that's a really good question. It's not Dario's fault entirely. We have a couple of things that are working against AI. One, you have a religion. Religion associates AI with this end of times idea mentality and that it's going to bring about the end of the world. It's going to it's like the Terminator meets religion aspect. So there's that.

 

00:06:42:14 - 00:07:05:17

 

You have Hollywood propaganda, which hasn't helped because let's be honest, Hollywood loves to pump out doomsday films and has for the last 40 years, and we all grew up on them. There's not really a positively impacting movie about AI. I mean, whether you're looking at like Dune, Terminator, you name it, they're all negative based, end of the world type stuff.

 

00:07:05:23 - 00:07:26:01

 

There's that. And then there is the corporate fear mongering in order to sell products and raise the stakes on whether it's IPO pricing, whatever it may be, increase the value of my company. Hey, you need to use this product right now, because if you don't learn to use it right now, you're going to be out of a job in the next six months.

 

00:07:26:03 - 00:07:51:07

 

That's where I've been very adamant to come out and say, I don't agree with that. I think that people need to use AI. I think they need to learn it as a tool. But I think there's a much longer time horizon to the use of AI and to its interaction with humanity, where it disrupts humanity. A lot of the jobs that we're seeing that are lost right now for, frankly, just bloated holdover jobs from Covid, that's just like the truth of it.

 

00:07:51:07 - 00:08:17:04

 

And nobody really in the mainstream openly speaks about it. It's AI, this, AI that. So so companies will blame artificial intelligence for the reason they're doing all these riffs, which is an acronym. So reduction in force. I mean, we saw meta, we've seen it in Amazon. They hired like crazy during like post Covid. And there's this big surge of capital that came from the federal government, tons of money sloshing around, and no one really knew how it was going to settle out.

 

00:08:17:04 - 00:08:35:08

 

And it takes a while for all of that surplus capital to kind of work its way through markets and to determine actually what the pricing of this service is, this SaaS tool. And they hired like crazy. And now they're letting people go and going, oh, it's not because we're we hire too many people. It's AI's fault. Yeah. That's interesting.

 

00:08:35:09 - 00:09:13:04

 

Now to be clear, AI is disrupting certain back office administrative roles. To say otherwise would be stupidity, but the scale at which they say that it's disrupting employment is also ridiculous. Kind of blown out of proportion. Yeah, I completely agree with you. And there's like anecdotes everywhere that actually like reinforce this one that I heard was there was a fortune 50 company that at the beginning of the year, the CEO of this fortune 50 company said to his dev teams, I want you to find a way to replace $1 billion in optics and operational expenses.

 

00:09:13:04 - 00:09:39:06

 

So find a way to make artificial intelligence automate so many processes that at the end of the year, we save $1 billion. That was the impetus. And so the dev team, the developers, the they got to work and they spent $500 million worth of tokens to try and achieve the goal the CEO had laid out. And the end result of it was that they couldn't do it.

 

00:09:39:08 - 00:09:53:22

 

They they spent half $1 billion to try and achieve this goal and with almost no results from it. Which is crazy because I guess in a way I'm surprised by that. I can see a few reasons why that would occur. One is like, okay, well, at the beginning of the year, everyone was hyped up on agents. Okay, great.

 

00:09:53:22 - 00:10:10:23

 

So open claw, nobody even talks about Open Claw anymore. But Open Closet. We'll get this thing to go in and do this and automate all these processes. That's human labor. We can replace that until you find out. Oh, like an update. Push into broke everything. You have 20 different processes and handshakes and APIs and and it's all broken.

 

00:10:10:23 - 00:10:38:09

 

Okay. Well we have to fix that again which is more human opex to fix that. So it's just kind of this like recursive loop of like we know that AI offers productivity. We know that it is sophisticated, but it's still not there yet. And I don't know why. You know, what I would say is it's it's not there because to be fair, like AI as a whole, although it's been around for, you know, almost six decades now, it is still in its infancy in the grand scheme of things.

 

00:10:38:09 - 00:10:58:22

 

And what I would say is I think people just rely too heavily on outsourcing. And we talked about this today actually on another podcast, outsourcing all of their creativity, all of their thought processes to, hey, AI, go take care of all of this so that I can just go sit on the beach, not do a thing, not care about life.

 

00:10:59:00 - 00:11:17:01

 

Like, I'm sorry, but that's not how life works. And you're taking the most powerful thing about you as a human, your brain and your capability of creativity, and to think and to think differently than everybody else. Because you and I are not the same. We are going to think differently. The way you solve a problem is not the way I solve the problem.

 

00:11:17:01 - 00:11:40:12

 

That is what makes us special. The moment you outsource that and say, I'm not going to use my best tool ever, that's what happens. Like systems break down because you've given power to something that is not even close to your capabilities and probably won't be, to be fair, for ten years if that, if that, if that. So how does it work?

 

00:11:40:13 - 00:12:04:18

 

I mean, AI artificial intelligence is best first manifested as a chatbot. That's the first version of it. People don't understand that artificial intelligence is this massive entity of a ton of problem solving, you know, parameters and weights, and it's powered by GPUs wired to server racks. So it has a ton of bandwidth to process and compute your request.

 

00:12:04:18 - 00:12:22:21

 

And the first version of us being able to interact with this otherwise nebulous digital intelligence was in a two dimensional chatbot. So how does that work? Because as it's been described to me, it's kind of like predictive text where you're typing out a word and then your Apple keyboard will be like, hey, you probably want to use this word next.

 

00:12:22:21 - 00:12:39:12

 

And you can kind of keep tapping that. Sometimes it gets the sentence right, sometimes it doesn't. Evidently that's how our language models work, but I was told that it's almost like it's like a matrix of words that sits around the spotlight, and then it connects to the next one, and it kind of shifts and it has this kind of predictive outcome.

 

00:12:39:13 - 00:13:09:20

 

How does it work? That's you. Honestly. You hit it nail on the head like it is pattern recognition based on the next letter or word that you are saying in your prompt to it. It's actually not as we have it right now, not fully like thinking learning autonomous on its own. Now, having said that, there's something very weird about AI that the biggest companies in the world still haven't figured out, which is that it's a black box they cannot actually see inside of the black box.

 

00:13:09:21 - 00:13:33:20

 

To understand this is exactly how it thinks. There is no two plus two equals four. There is no equal sign. That's that's the really weird part. And it's only been recently that there's a company that put out something called the world model, where they're actually changing the way Llms work, so they no longer are pattern recognition based. They are now moving to the way of actually thinking.

 

00:13:33:21 - 00:14:08:20

 

And so we're in for a very interesting next phase or what I call next inning. Because like right now my opinion is we're in like inning two of the baseball game of AI. Inning one was like chatbots. Inning two was open claw, you know, drop down files, AI agents. Inning three is where we're going to start seeing AI agents that open the door to actual autonomous thinking, which is okay for now, because they don't have any physical interface with like, reality.

 

00:14:08:21 - 00:14:32:02

 

Like, not really. I mean, even when they did the the Tesla Diner in Los Angeles and they had the Optimus bots running around serving everyone, and they were all remote controlled. But knowing that this is the worst it's ever going to get, it's like eventually there's going to be some sort of reasoning cortex that'll have a housing inside of a body, a machine with fine motor controls and actuation that it can control and can assert itself into the physical world.

 

00:14:32:03 - 00:14:50:04

 

I mean, today it's not there, which is kind of a comfort because I've seen iRobot and I've seen iRobot rather, I've seen it. And those little dudes are everywhere, you know, and it's kind of funny to like, I remember watching iRobot and being like, oh my God, like one day cars could just like drive themselves and people will.

 

00:14:50:07 - 00:15:11:14

 

And now my car drives itself. I was like, that didn't take long. You know, and one day it's going to be kind of out of fashion to manually intervene on your car. And I can already kind of see, like a lot of people's driving behaviors have already shifted to that. People don't want to go around unless there's a Waymo or they're in a Tesla with FSD or, you know, now a lot of competitor Cadillac's got a pretty good FSD system.

 

00:15:11:14 - 00:15:36:20

 

Ford's rolled out a pretty good FSD system. I don't know how far away we are from the I robot outcome, but it is. It looks shockingly accurate. Shockingly, yeah. Shockingly accurate. I do see that happening. But you know, right now the the limiting factor is the ability to produce electricity. And there's going to have to be some massive infrastructure overhaul.

 

00:15:36:23 - 00:15:56:00

 

I don't know whether it's going to be they're going to develop a lot more nuclear reactors, whether it's you know, they're going to have to figure out because I am a believer that cold fusion is real. So the question becomes is, okay, how do we power this whole revolution? Because that is the problem right now. Yeah it is.

 

00:15:56:01 - 00:16:16:16

 

And we can look at what China is doing. They are turning on terror. Lots of power. Yes, annually. And I don't have a great grasp on this. I'm not an electrician, but what I've been able to peripherally understand is like a terawatt of power is absurd. I think the entire United States runs on a terawatt and a half of power in China is going to activate a full terawatt by the end of this year.

 

00:16:16:16 - 00:16:37:14

 

So it does seem that the limiting constraint right now is electricity and then compute like secondarily, you know, because they're having a hard time getting the chips. Everybody's lining up for it. The orders are like ten years in advance from TSMC and basically everyone else. Now Nvidia is you know, they've got back orders for years now. OpenAI apparently is getting into the silicon game, which is crazy.

 

00:16:37:15 - 00:16:57:23

 

Did you hear about that? Oh yeah. So what's their their play there. I mean they're trying to vertical. So OpenAI will steal I'm going to use the word steal I probably shouldn't but like they're going to move into as many sectors as possible. They want to be vertical across the entire chain. So at some point they this was a have to they have to create their own chips.

 

00:16:57:23 - 00:17:20:05

 

They can't continue going to an outside company and saying, we're going to buy billions and billions of chips from you guys, especially when everybody else is going to that same company, then there's nothing that actually sets you apart. So this is like the next step in the process, which makes it even more competitive. I mean, to begin to start making your own silicon is insane though.

 

00:17:20:09 - 00:17:45:03

 

It because I have the funniest anecdotes I see on Twitter, people like we managed to transform sand into Electric Sorcery. We trapped lightning in it, covered it in mystic runes, perform some sort of like technological seance over it. And now these chips breathe, think, and operate at a nanometer scale, so I don't know how they're going to like, just jump right in as if there appeared.

 

00:17:45:05 - 00:18:07:22

 

I mean, this has been iterative over decades. TSMC is the best in the world, and the amount of the facility that they need to make, like a two nanometer, three nanometer chip, is shocking. I don't know how they're going to, like, be able to jump into the race competitively. That's, you know, we could go down the rabbit hole of that.

 

00:18:07:22 - 00:18:23:16

 

I don't think, actually that OpenAI is going to be a company in like 2 to 5 years. How do you mean? So I actually have a personal opinion that they will end up going bankrupt at some point in the next 2 to 3 years, and they will get broken into pieces. That's my personal opinion. Sold for parts. Yeah.

 

00:18:23:17 - 00:18:50:10

 

Well, they'll get picked up by Microsoft by others that are shareholders. But I, I believe they were they were the Myspace if you will, of AI don't stand back and I'm sorry like Sam may be really really smart, but I think he got a little in over his skis on how much money he was spending. You know, instead of the keep it simple stupid model, it was the I want to be everywhere doing everything all at once model.

 

00:18:50:10 - 00:19:16:10

 

And that has very rarely worked in business. Now there's still a large company. They're still gearing up for an IPO, which has been delayed, to be fair, because of certain government things that have happened. But I don't think there's enough liquidity in the market that's going to be able to prop up open AI, and they have to figure out some way to generate a lot more revenue, a lot more revenue.

 

00:19:16:10 - 00:19:41:12

 

And that is where I question whether they will be around in 2 to 5 years. Anthropic doesn't have this problem though. No. They just went profitable. Yeah. Which is crazy. It's like their bet was so clever because if everybody is kind of orbiting the same problem, which is how do we make AI valuable other than helping high schoolers write papers and helping Instagram influencers like, automate their captions?

 

00:19:41:13 - 00:20:07:13

 

You know, everybody has been sitting here banging their head against the wall going, this technology is transcendent. It's incredible. It's also not providing any value. Who the fuck is this guy? Fair question. Sorry comment section, but I'm not an industry plan and I'm not CIA. I directed the video production for Mr. Beast, where my work has about 10 billion views, and now my team and I produce shows just like this for founders all over the country.

 

00:20:07:13 - 00:20:24:09

 

Listen, if you're a founder and you've built a real business, you are the best kept secret in your industry. So if you were done with being invisible outside of your company, book a call with us on the link below. My team and I will show you exactly how a show like this could change everything for you and your business.

 

00:20:24:09 - 00:20:49:14

 

All right, back to the show. What a weird situation to be in. You're surrounded by. Like if you were in a field surrounded by food, you wouldn't be like, oh, how do we. How do we make this valuable? It's like it's very obvious it's food. Everyone has a use case for this, but there have been all of these ticky tacky on the margin like use cases for AI, and none of it has really been world changing until anthropic said, we're not going to focus on cutesy little image generators.

 

00:20:49:14 - 00:21:11:15

 

We're not trying to be the new Midjourney. We're not trying to do Sora and video. We're going to put the sum totality of anthropic development of thus far into code, because SaaS software that seems to be if you can't interface with the atoms world, let's build the photons world, the electrons world. And that has been the most genius pivot I think I've ever seen.

 

00:21:11:15 - 00:21:33:01

 

It's been the most genius because now you can. That's kind of what leads us, I think, to the mythos conversation. Why don't you take us to like anthropic decision to focus not on creative fuzzy tools, but on code? And then where has that led to today? For sure. So one open AI doing all the fun Z tools was more marketing fluff than it was anything.

 

00:21:33:01 - 00:21:52:09

 

And, you know, it's funny because like all of Silicon Valley does the same thing. It's this freemium to subscription progress. They want to give you something for free, and then they want you to go to a subscription that doesn't work when you basically give everything away for free. And there's no real way to get somebody to pay you money.

 

00:21:52:09 - 00:22:12:21

 

And it also doesn't work when you are. And this is my big problem with AI right now, is most of the subscriptions are denominated in US dollars. Okay, so you essentially cut out the rest of the world, which when you look at there's a company called Menlo Ventures, they put out statistics in 25 and 26 for AI, consumer statistics, enterprise statistics.

 

00:22:13:02 - 00:22:36:03

 

The thing they looked at is they said, okay, there's close to 2 billion people worldwide that use AI, but only about 100 million or less pay for it. Makes you wonder, well, guess what? That's basically a portion of the US and Europe and Canada, more than likely, that have the money to pay for a subscription. That's 22 to $100 US denominated.

 

00:22:36:04 - 00:23:02:02

 

The rest of the world is not going to pay for that. A guy in Nigeria, a guy in Brazil, a woman in the Middle East, they're not going to pay for one platform, a $20 US denominated subscription. It's impossible. So problem number one, this the way they generate revenue. Not smart. Anthropic picked their lane, their lane, and they did it at the right time.

 

00:23:02:02 - 00:23:20:17

 

They did it early enough on that. They could say to the rest of the world, every coder that wants to go build a SaaS product is going to use us focused in just on them. And then they expanded. And once they started expanding, they basically took over everything and became an AI super app, which is what they are now, because they're not just for coders anymore.

 

00:23:20:18 - 00:23:49:11

 

Now they have AI agents, they have a react front end where you can see a spreadsheet as you're building it, an Excel file as you're building it, an app as you're building it. They basically said, hey, lovable is doing great, let's hijack that and throw that on our platform. So very smart move by anthropic. Fast forward to where we are right now, which is the models are getting so good at one specific thing, which is cybersecurity.

 

00:23:49:13 - 00:24:13:13

 

That is a scary thing for any government, any terrorist organization, any anything around the world, because now nothing is safe. There is not a single thing on the internet right now that is safe. If somebody has access to a high level mythos or fable type AI. And to be fair, the day that they were dropped, they were jailbroken.

 

00:24:13:15 - 00:24:38:11

 

There's cyber threats all over the place. This is going to continue exponentially. And now the government has said, hey, we have to step in because this is actually a problem. Our own systems are no longer secure. Mac OS not secure, windows not secure. All of the zero day attacks that used to be limited to just governments, intelligence agencies.

 

00:24:38:13 - 00:24:59:00

 

Now they're for everybody. So what do you do when you or I could go in and perform a zero day attack on a government? And it's happening right now. I follow a guy on social media who literally attacked the Mexican government with jailbroken AI, posted it all over social media. It this is this is the world we live in now.

 

00:24:59:01 - 00:25:29:09

 

What did he do? So he used AI to find public exploits. And a lot of these third world emerging market governments have public entrails, basically where they store data, where they store passwords or where they store usernames. And from there it's a few steps removed to get a password to figure out how to break into it. And this guy's, to be fair, good at what he does, but he's using AI to do it, so he's good at hacking.

 

00:25:29:11 - 00:25:40:18

 

Red team, blue team, good guy, bad guy. Now he gets a very powerful AI. His powers now have been multiplied by 100.

 

00:25:40:20 - 00:25:59:16

 

Oh my God. And to to add something on here, I am not a technically savvy guy. Like I can use a computer, I can use a phone, I can do all these things before AI, if you would ask me, hey Alex, go develop an app. It would have taken me years. It would have taken me millions of dollars using AI.

 

00:25:59:20 - 00:26:33:06

 

In less than nine months, I've been able to build an app. I didn't build it myself. I built the prototype myself, which allowed me to hand that off to a group of developers to take a development process down to less than 60 days. So how you use AI can make you extremely powerful. If you understand the nuance of I'm going to use my own creative intelligence, Im going to go build a prototype of something, or to my understanding of what I want it to look, feel, move, act, talk like.

 

00:26:33:06 - 00:26:50:16

 

And then I'm going to go hand it off to somebody who knows what they're doing to actually make sure the code is the right way. Then points are secure. Then we're going to use AI to go in and check it, run hundreds of simulations on it to break through it. Okay, now we have a product. It's safe. It's secure.

 

00:26:50:17 - 00:27:21:20

 

We've used AI in certain ways to enhance the experience and to make the process faster. So that seems like one of the more pragmatic ways of leveraging you right now. On the topic of cybersecurity, it it's so funny, man, because like, people kind of unless they're really deep in this subject matter like you are like I, you know, aspire to be it's really hard to understand why this is so existential until you kind of zoom out and realize every single thing in our lives is built around software.

 

00:27:21:20 - 00:27:42:10

 

Software touches every single layer of it. And what does that mean? Well, unless you're a developer or a coder, you don't look at the world in that way. You just go about your life and then you use the software and then do do you go and do other things? But to a developer, every single thing from like the checkout app, you know, you go to a restaurant, they hand you the toast thing to pay for your, your dinner.

 

00:27:42:10 - 00:28:02:13

 

It's like there's tons of software in there, and there's software to track your meals correctly to make sure that you're not paying for what's buckets nachos down at the end of the table. But then that software has to handshake with a payment processor, and then that payment processor hands to handshake. With Visa or Mastercard. It has to have a software layer in order to determine, oh, they're using an Amex, so don't send it to visa.

 

00:28:02:13 - 00:28:17:07

 

Send it over here. And there has to be all these checks and balances and your bank has to ratify that. The the money is in there. All of that's just a software layer to buy your damn, you know, bottomless margaritas at Chili's. Right. So it's like you start looking at it that way. Okay. Well where does this start becoming existential?

 

00:28:17:08 - 00:28:42:20

 

Well, our nuclear program is run on software. Our ability to move infantry and to manage supply chain logistics. It's all software. And for those who, you know, videos everywhere right now. So I've been a video editor for North of 20 years. I learned this the other day. Premiere Pro it's the software that we use. It's like the number one editing software in the timeline alone, which is where you drop like your clips and you rearrange them like that.

 

00:28:42:20 - 00:29:11:08

 

My team used that to make this video. In the timeline alone, there sits over 1 million lines of code GS in just the timeline portion of Premiere Pro, which is a simple, complex video editing software. And they have so many issues that they're constantly trying to micro fix, because a million lines of code is a nightmare. Imagine trying to read a million lines of a book and go in and look for one typo.

 

00:29:11:11 - 00:29:35:13

 

So it's like when you look at the macro of the entire world, which is run on software, the planes are all governed by software. The people that are monitored, the FAA, the regulators, the tower controller, air traffic controllers, they all use software. Your car uses software. Your phone is nothing but software. All of these things are built on millions and billions of lines of code, and there are typos in them.

 

00:29:35:14 - 00:29:55:17

 

They are written by humans and those typos can be exploited. Those are vulnerabilities. And that's how hackers work is they go into some software, let's say you're a payment processor, your stripe or whoever. They just get the back end of it. They go through it meticulously. They have the the vocabulary or the training to know where a typo or a problem exists.

 

00:29:55:18 - 00:30:14:06

 

Oh, there's a semicolon. There shouldn't have been. There should have been over here that left a vulnerability. And so you can exploit that. And that starts to get out of control when you realize that there are layers and layers of software that have been built on top of each other. So when an iOS developer builds an app, they're not coding it from the ground up.

 

00:30:14:07 - 00:30:33:08

 

They're using tools that were built in Swift, and Swift is built on top of a whole bunch of other software stacks, and Linux kind of sits on the base of it. So when they rolled out mythos and these pointed it at some pretty antiquated like legacy code bases, mythos went through billions of lines of codes and went, hey guys, there's like a ton of problems in here.

 

00:30:33:12 - 00:30:53:21

 

And then all the people deploying mythos. What do you mean? Oh, there's I found so many zero day exploits in Linux, the foundational code base of all software. And they went, oh my God. So if somebody took enough time and had that tool and they went and found that exploit, they could hack all software everywhere if it was built on Linux.

 

00:30:53:23 - 00:31:20:00

 

So now you have legitimate cyberterrorism on the other side of having mythos and your own rascally nature and pointed at something and go help me find an exploit I'd like to to topple visa. And it's like, that's really where we're at. And so it's like, this thing arrived so fast in the station, man, it wasn't like this. I literally like, I pine over the simplicity of AI a year and a half ago.

 

00:31:20:01 - 00:31:39:20

 

Like I long for that now. It was like, oh, cool. I was so frustrating. You know, it keeps adding it to little silly goose dashes. It's like, that's not what we're at anymore. No, no, we're literally like humans. Like regular people can be cyber terrorists and have no idea what they're doing. Oh, yeah. So what do we do?

 

00:31:39:21 - 00:32:09:02

 

Alex, what do we do? Well, you know, I'm not usually one to say that the government should step in. Unfortunately, that's kind of where we're at to a certain extent, to at least slow the process down. Now, I don't love the idea either, that, you know, Dario just came out recently and said, hey, I think all AI companies need to halt new LLM production for the next 6 to 8 months for the governments and intelligence agencies to catch up.

 

00:32:09:04 - 00:32:28:07

 

I am a firm believer in the cat's out of the bag like there is no going back. I think we just need to let it all out and see what happens. But to your point, I'm sure you remember this when there was like the Boeing 777 Maxs that crashed. Oh yeah, that was a software, right? Because that turned into a hardware error.

 

00:32:28:07 - 00:32:45:18

 

Right? Well, they built the plane and they knew that the plane had the tendency to just dip forward because of the way that they balanced the entire plane. So they built software to tell the plane to constantly counter steer itself, accounting for the fact that the plane just naturally dips down because it's so damn big, or where they set the engines up.

 

00:32:45:18 - 00:32:59:10

 

And I'm not I'm not an engineer. Like, it would be a nightmare to get a plane that far into development and then find out, I don't know, like 80% of the way through. After all, the engineering departments have got the weight and the circuits and the heart and the the software layer and the wiring, they're like, hey, it's perfect.

 

00:32:59:10 - 00:33:14:08

 

And then they put it up in the air. They go, this thing just tips down. It just tips down. What do we do? Well, we could put a piece of software in the plane telling it to counter steer for that across every use case of the steering the plane. But then I guess in the story that software failed, right?

 

00:33:14:09 - 00:33:36:10

 

Well, it overcorrected and it caused a bunch of crashes, which they eventually had to, like, halt the seven, the 777, Max's program. They had to put them on the sideline for I think it was like close to a year and Boeing had to go in and fix everything. The point of the story is that software rules everything from water filtration to airplanes to communication banking.

 

00:33:36:11 - 00:34:00:05

 

I mean, you know, the World Economic Forum, Klaus, whatever his name is, talked about the fact that there would be a global cyber pandemic. Do you remember him talking about this? So post Covid and in conspiracy circles like this is major, major talk. Klaus Schwab came out and said, there in the future is going to be a global cyber pandemic.

 

00:34:00:06 - 00:34:39:09

 

It's going to affect banking. It's going to affect all these things. It's going to cause a worldwide shutdown similar to the way Covid did. But it will be from a cyber perspective. Now look at where we're at today. Ding ding ding. Does it kind of strike some similarities to what was said a couple of years ago? Everything is run off of software, and if all of a sudden you have a very powerful group of people or one really, really, really smart guy in his mom's basement that figures out how to hijack one of these, you know, jailbroken mythos, the whole world can come to a halt overnight.

 

00:34:39:11 - 00:34:59:04

 

And again, I'm not one to say that the government needs to step in, but I do find it interesting that anthropic now is saying that they want, you know, like iris scans or they want driver's licenses of people that are using their platform. Did you see that? Well, yeah, I mean, I kind of I'm on the fence on that KYC kind of makes sense.

 

00:34:59:06 - 00:35:29:02

 

No, no. Your customer KYC and it's it's a big topic. Do you force somebody before they use your app to submit their driver's license for accountability? Yeah. And I'm the antique guy. Like, you know, me, I literally built a platform with friends, with family that is like the burner phone of AI specifically to combat this whole thing. Having said that, like, I look at the positive impact of humanity.

 

00:35:29:02 - 00:35:45:17

 

I look at the positive side. I think the more you try and restrict people, the more it's like telling the little kid, don't go eat the candy after dinner that's in the cupboard. What do you think the kids going to do all they want? Yeah. So when you tell the world. Well, we got to lock up this AI because it's too powerful.

 

00:35:45:18 - 00:36:19:12

 

That's like ringing the dinner bell for the coyotes and the wolves that are like. Well, guys, we just found our next weapon. That's where I think this could have been done a lot more, in a lot more nonchalant way of they release it, they let it be. And behind the scenes they are actively every single day working. And for anybody to think that intelligence agencies right now are not spending billions of dollars to be prepared for a potential cyber threat of some kind.

 

00:36:19:13 - 00:36:38:15

 

They are every single day it's in the news. Read between the lines. It's not hard to see. I think they could have just done that without being so public about the world is going to come to an end because mythos is now out there, which I guess they overturned by the way they did as of yesterday. What do you make of that?

 

00:36:38:15 - 00:36:59:10

 

Have you played with it yet? I played with it when it first came out, but honestly, it's interesting because I use less AI now than I did a year ago. I'm actually moving more away from AI in the sense that my position in in our company, my instant AI, is more human to human. It's a lot of like contract negotiations.

 

00:36:59:10 - 00:37:19:23

 

It's a lot of direct relations with distributors, with people on the ground in emerging markets, you name it. So I don't find myself having to need to use AI. That's for me. That's kind of nice. And like from the self-driving car perspective, I'm like, give me an old 1960s car, please. That doesn't have a chip in. It doesn't have software in it.

 

00:37:20:00 - 00:37:46:23

 

Nobody can hijack it. There's no Bluetooth. Like, you can cut my brakes if you really want to, but that's about all you're going to get. So when you roll up in like a 1992 Ford Fiesta, I'm going to know why you're like, I can't have them hacking. My company's too big 100%. So let's talk about your company for a second, because it's actually astonishing you've figured out this product market fit that everybody overlooked, which is the inception point of any great company.

 

00:37:47:00 - 00:38:05:04

 

You figured out that it's not just United States citizens that are interested in AI. It's not just the EU, it's not just Canadians. There are 8 billion people on this planet, and they're all interested in this. More and more third world countries are getting on the internet and they're landing on YouTube, which is why we're going so well.

 

00:38:05:04 - 00:38:25:05

 

But they're also trying to figure out how they can catch up to the Americans and like the civilized world. But their issue is wages. And so they don't have $20 a month to toss into a subscription. So what is it that you noticed in this equation? Figured out how to solve it. So there's there's a couple of things.

 

00:38:25:05 - 00:38:52:20

 

One, most of the world does not pay subscriptions for things. The US, Canada, Japan are like the top countries in the world. That subscription basis for streaming, whether it's phone plans, whether it's utilities, you name it, most of the world is what's called prepaid, meaning that you pay at the convenience store or at the grocery store for your mobile phone data for however much you pay for it.

 

00:38:52:22 - 00:39:12:10

 

You pay for your mobile phone minutes. You pay for your utilities. Like if you go to Africa, for instance, you're going to a convenience store and you're paying for all of these things ahead of time for on and on a usage basis. Wow. You're not hey, here's 20 bucks a month every month, no matter how much I use.

 

00:39:12:11 - 00:39:43:13

 

Which oddly enough, guess what? Now all the AI companies have caps on usage, so you're paying a subscription and they're capping you imagine that point is, I saw because of somebody who kind of gave us a little tip. I saw a very interesting opening in that nobody in the AI industry had touched retail stores. There was not a single AI product that was being sold in retail stores, and there was not a single way to consume AI with cash.

 

00:39:43:15 - 00:40:01:14

 

You had to have a credit or debit card, you had to give them your information. You had to buy a subscription. Not only that, but when you bought a subscription, let's say you only got a subscription to one platform. You didn't get a subscription to all of them. That's where aggregators have come online over the last couple of years.

 

00:40:01:14 - 00:40:30:04

 

Open router being a big one. They're doing. Yeah. Perplexity being the biggest like technically that was the first like initiation of okay aggregation in my opinion. And a super app is like where things are going. I actually personally think if perplexity plays their cards right, they could be bigger than Claude, they could be bigger than open AI because they have access to more tools that are all coordinated.

 

00:40:30:05 - 00:40:58:20

 

And to add on top of that, that's where I think that my instant I could be absolutely massive, because we use the brand trust that's been built by the retailers to sell a product. And not only that, half of the world does not have access to a bank account. So Silicon Valley has basically said we only care about the top 5 to 10% of the world to buy our products, but yet they expect the whole world to use AI.

 

00:40:58:22 - 00:41:19:13

 

That's where 99% went left. We said, let's go, right. We're not going to sell direct to consumer. We're not going to sell in a subscription. We're going to do what Sam Altman just now is talking about, which is like AI usage as a utility. I just hate the way he said it, because it's terrible marketing one and two.

 

00:41:19:13 - 00:41:44:15

 

It's very hard to take people from a non limited subscription like, hey, pay 20 bucks a month and you get all access pass two. Hey, you're now going to buy tokens on a prepaid basis. It's easier to just go straight to prepaid. So we started in Africa. We started in Middle East. And like we just tapped ourselves into a very big group of brokers.

 

00:41:44:17 - 00:42:08:04

 

Most people don't know this, but the prepaid gift card market. So prepaid vouchers and gift cards one are like the same gift cards in the US because we don't do prepaid outside of the US. It's all called prepaid vouchers, essentially the same thing. We tapped into the brokers that run that industry, because a lot of those products are sold between points of distribution by a broker.

 

00:42:08:05 - 00:42:31:11

 

Okay, we went to all the brokers and said, hey, we've got your latest and greatest product. We've got an AI prepaid card, same thing as mobile phone, prepaid mobile phone cards 20 years ago, which is still a half $1 trillion a year business that's in every Walmart. It's in every gas station. The gift card market, separate from the prepaid market, is almost $1 trillion annually globally.

 

00:42:31:12 - 00:42:40:15

 

Is it really? Yeah, it's absolutely huge. And it goes under the radar because it's boring.

 

00:42:40:17 - 00:43:15:00

 

Well, I have a newfound love for boring. So we grabbed up all of these brokers. We started talking to telecom companies. We started talking to retailers, distributors. And within the next probably 90 days, God willing, will be in close to 4 million stores. We'll have some extremely large telecommunications deals across the world. And it's only just in the last two days that a potential competitor and I say potential because they're where we were a year ago, has just now started to kind of get the idea.

 

00:43:15:02 - 00:43:40:11

 

But one of the things that we did different is we went in and we grabbed exclusive contracts. So we said, okay, we're the first to do this. We're going to close the door, the door behind us. We don't want somebody to come in and eat our lunch or outspend money or whatever it may be. So we are fast tracking across this entire space that nobody is touching, nobody is seeing coming.

 

00:43:40:11 - 00:44:14:11

 

And hopefully this holiday season, you will find my insanity cards on every single shelf in every single Walmart, Costco, Sam's Club, BJ's, 7-Eleven, circle K, and at that point it's it's less about like is are people buying from circle K? Are people buying from 711? Are people buying from Walmart? And it's more about being everywhere all the time, because you may not be selling at a Walmart on a daily basis, but you're going to have thousands of people that walk by that gift card rack that see your gift card.

 

00:44:14:11 - 00:44:20:11

 

So it's just as much marketing as it is sales.

 

00:44:20:13 - 00:44:41:08

 

A big part of this offer to is that people can use AI in an anonymous fashion, right? Yes. Which is, you know, a big conversation, as we already touched on. But it is funny to to say, hey, if you buy, call it 100,000 pre-paid tokens, which I don't know, gets you some pretty aggressive access, but you can burn through 100,000 tokens pretty quick too.

 

00:44:41:11 - 00:45:07:16

 

You can then pick what platform you want to spend those tokens on. Okay, you want to go do that on Gemini. You want to go do it on grok? Great. Use my instant AI, but part of what the models use is being able to actively track the user of their AI models. They learn from them, they learn from their behaviors and patterns, and they can kind of build profiles just like social media companies like Facebook knows way too much about me, way too much.

 

00:45:07:19 - 00:45:30:20

 

And Facebook owns Instagram and Facebook owns WhatsApp. All these platforms, they have handshake data sharing agreements. They all know way too much about me and you and Leah and everybody. And so the goal with my instant AI is not only to democratize access to artificial intelligence. You don't have to be on this like drain of an ongoing subscription basis, but you can also use it anonymously.

 

00:45:30:22 - 00:45:55:12

 

Exactly. Very, very akin to burner phones, but with AI. And, you know, part of the reason is because, like, to a certain extent, I look at AI as almost like the precursor to a minority report esque mentality, where they know what you're thinking before you're thinking it, because they've collected so much data on you as it is. As it stands right now, these companies know what you're thinking.

 

00:45:55:12 - 00:46:16:16

 

As you're thinking it. They don't necessarily know before you start thinking it, but we're not far off from that. And so, yes, part of building this system and part of the prepaid route was and allowing people to buy AI access with cash, you could walk into a Walmart. The way you look right now with a hat, with a beard, with a jacket you would never see, tattoos.

 

00:46:16:17 - 00:46:37:05

 

You would never see anything about you that may set things off. Pay cash for an AI card. Go ask a question of mythos five of how do I do x, y, z? How do I exploit the Department of Energy? Sure. Please don't. Yeah, not those guys. Don't mess with those things. Don't. They don't do that. Not the Department of Energy.

 

00:46:37:05 - 00:46:58:16

 

No no no no. Bad move. But point is, I am a want a free market capitalist. And two, I am a firm believer in independence, in the ability for humans to be completely free. Now, there is the flip side of like if you do something stupid, you kind of are asking for something to stupid to come back your way.

 

00:46:58:18 - 00:47:19:15

 

But that is part of our overall strategy in. When we looked at the data inside of the US, 80 plus percent of people inside of us did not trust tech companies, did not trust AI companies outside of the US. They actually trusted it at close to 80% in favor, especially China. Chinese citizens are very bullish on AI very well.

 

00:47:19:15 - 00:47:38:23

 

And it's interesting because the more third world country is, the more they're likely to trust AI and the AI companies because they look at it like it will. Even the playing field people in the US don't trust it, because we've had so much usage on it already, and we've seen our data tracked already by governments, by tech companies.

 

00:47:39:01 - 00:48:00:10

 

It's the inverse. And so we at least wanted to offer a worldwide solution of, hey, if you come onto our platform and you really care about anonymity, you can ask your questions, you can use your agents to do things, and the LMS are not going to get your name, your credit card number, your address, your phone number, your height, weight, your first child's name.

 

00:48:00:10 - 00:48:28:06

 

Like they're going to get a user ID and that's it. And it will basically be a randomized number. If you choose to give these systems your information, like, hey, I'm going to choose to talk to it about my kid in my lawsuit and my grandmother, and that's on you. If you don't, though, for all intents and purposes, you can remain anonymous, which in today's world is very difficult to do.

 

00:48:28:07 - 00:48:53:13

 

This conversation around KYC, around being able to verify who it is that's using this tool. You and I have both been of the mindset where it's like, I don't want that. I don't want to be tracked. Yes, it's fun that I get very qualified ads on my Instagram. I like that I get very good advertisements on supplement stacks and camera gear and cool clothes.

 

00:48:53:14 - 00:49:15:03

 

I like that, but the amount of data they have on me, they're only using a fraction of it to qualify my ads. I don't know where the rest of it's being used. I don't, I don't know, and I find it unsettling. And the big part of why there is such a big push from Dario and lobbyists, maybe that he's hired, they're all pushing for KYC.

 

00:49:15:07 - 00:49:39:06

 

The theory in some circles in Silicon Valley is that they're trying to effectively disintegrate open source, open sources, and that's really like people, oh, I'm not a coder. What the hell is that mean? Open source is the democratized free access of code. So like GitHub, you know, or lovable or hugging face. These are places where coders, this is what they do.

 

00:49:39:07 - 00:50:03:07

 

They've been coding for forever. When they build an incredible series of codes, they make a program. Sometimes they put it behind a paywall and say, I'm going to turn it into a business. Sometimes they just know it's like a convenience thing and they go, I made this really cool tool. I'll put it out there, and then it allows other people to build on top of that tool and eventually innovate to some actually really incredible products using open source tools.

 

00:50:03:07 - 00:50:28:17

 

It seems like Dario, Sam Altman, a lot of these guys are trying to control the regulatory framework, force, KYC, force verifying identities so that you get rid of people that are building open source. You're getting rid of anonymized accounts. You're getting rid of like shadow accounts, like people that post and shit post on X, it seems like there is an effort to get rid of all of that.

 

00:50:28:22 - 00:50:50:02

 

Why would they want that? I mean, it affects their business first and foremost, the open source, the ability to go and take an open source model like a Quinn model and jailbreak it. They did something recently. It's it's a new form of jailbreaking. It's called obliterating, where they basically can reduce the guardrails by like 95 plus percent. What.

 

00:50:50:03 - 00:51:07:08

 

And you can get this all open source. So I did this actually with a Quinn model for my own personal use. And I think I even showed it to you. Like you can ask it almost anything you want. And it will. It takes a little bit. I tried, yeah, I tried, I tried asking it at some pretty naughty stuff and it was giving me kind of the run around.

 

00:51:07:09 - 00:51:31:19

 

It does and it takes a little, you know, messing with it. But the point is, like the real battle here is between closed source and open source. Open source is an extraction of wealth from the closed source companies. But that's because the closed source companies are more of their money. Funding. Everything is going into IP is going into like building their model from scratch.

 

00:51:31:20 - 00:51:56:09

 

We actually love the open source community because we can go grab tools, put them on the Mayans platform for damn near free from the open source community. Now there's a big downfall. Recently, a lot of things that have been added to GitHub open source, have malware attached to them, have happened with open Claw, happened open claw when Open Claw dropped on GitHub as an open source.

 

00:51:56:09 - 00:52:22:03

 

That's how it started. Yeah. One of the first connectors that was made for Open Claw by another developer that stacked on top had malware attached that it grabbed data from everybody that attached to it. So that is another you know, we're going to insert that as a slight issue that if you don't have a way to vet the open source product that you're now using, you can very quickly fall prey to malware of some kind.

 

00:52:22:05 - 00:52:43:02

 

So I find myself conflicted because on one hand, I love open source. I love the idea of the open source community. I love the idea of people innovating without needing to turn something into a business just for fun. But there is the nefarious aspect, as there always is. I love the closed source because it comes with certain things like limitation of liability.

 

00:52:43:02 - 00:53:08:13

 

It comes with, you know, insurance. It comes with shareholders that hold company values, hold the executives to a certain standard. So there's a battle here that I don't care who the winner or the loser is, and I don't necessarily think there will be a winner, a loser. It's just it's a battle for money, power, control. Like what runs the world right now.

 

00:53:08:15 - 00:53:24:15

 

I think I'm allowed to talk about this. The statute of limitations has certainly passed. So there was kind of a precursor to all of this. Now the open source and everything else. And it really it found its roots in Torrance in file sharing, which was like a big thing in the 2000 because the internet was just this Wild West landscape.

 

00:53:24:15 - 00:53:40:16

 

It wasn't very well regulated. It was really hard to control. And maybe that's my telling myself that. And in reality, they actually were able to watch everything the whole time. There's only a few of us, you know, that's probably. Yeah, yeah. Come on. They knew. They knew exactly what you were doing the whole time, whatever statute of limitations.

 

00:53:40:16 - 00:53:57:05

 

But it was so cool because you go, okay, I want to download the Nickelback album. And so I go to demon. I had access to demons. So I you familiar with that website? Yeah. And you couldn't get access to demons? Someone inside demons had to give you access in order to come into it. It was so cool because, like, The Pirate Bay was great too.

 

00:53:57:06 - 00:54:15:15

 

I loved The Pirate Bay, but it was also just a range of outcomes, like range of products. You could get a lot of weird stuff on there, but it was cool because people would say, I have the entire Nickelback album on my computer in Illinois, and I will seed it to other people. So the receivers and leeches and people that would seed it.

 

00:54:15:15 - 00:54:33:21

 

So. Oh, great, I downloaded that from you. Thank you. Now there's I have it on my computer over here in Vermont. So there's two computers that have a copy of that Nickelback album. And so you would download the torrent file, which was kind of like an assembly blueprint, like this is how you put together the actual file structure in the codes, the ones and zeros of the Nickelback album.

 

00:54:33:21 - 00:54:55:23

 

And then those two computers would assemble it on my computer. Basically, they would follow the protocol in the torrent file. It was so cool. But then you found out that there were 500 versions of that Nickelback album, and half of them had malicious malware on them, and they would get into your shit and mess your computer up. And so I learned really quickly that not everybody is out there playing ball.

 

00:54:55:23 - 00:55:17:21

 

Not everybody has good intentions. So to that tune, it's like there are a lot of vulnerable, less sophisticated internet users. There is some reasonable guardrails that should probably be implemented because we know that that happens. You know, it's like, oh, okay, we have laws around how food should be handled because we don't want somebody spraying, you know, wasp killer on it.

 

00:55:17:21 - 00:55:34:13

 

So it's like there are reasonable guardrails so hard to regulate. So I guess I see the argument for KYC because there can be some really. I mean, the open claw thing was insane. Oh yeah. They were able to say, I guess you said it was a router that people like, what was the top layer that they would stack on top of an open claw?

 

00:55:34:13 - 00:55:55:03

 

What was that called? So, well, it wasn't that it was that they would they would do like a connection point. So they had all these MCP connections where the open claw could then connect into, like, you know, Zapier or connect into, you know, a tool of some kind, an agent of some kind to go do something specific. People were creating a ton of connection points to open claw when it first dropped out.

 

00:55:55:04 - 00:56:18:00

 

Right. And that's where nefarious actors said, oh, we can use this as a way to collect data or hack a bunch of computers, or put ourselves behind the computer screen of millions of people. And they did that, and they embedded it pretty deep. It was really hard to detect still a bunch of data. The protocol would be like, okay, open claw, like, you know, this connector that somebody is going to download and put on top of their open claw.

 

00:56:18:01 - 00:56:30:20

 

Hey, if they're opening up their Facebook, make sure you log what their password is, you know, make sure that you're aware of what's going on in their emails, see if you can get their payment info or any kind of data, or if they use this sort of phrase or whatever, activate. I mean, a lot of like really weird.

 

00:56:31:00 - 00:56:53:20

 

Yeah. You turned your open claw into like, a manchurian candidate. Yeah, yeah. Can everybody just be cool? Is that too big and ask? I mean, yeah, it kind of is. Honestly, we live in a world where you have people of all colors, all creeds, all spectrums. There's going to be a few bad apples.

 

00:56:53:22 - 00:57:17:18

 

Now, that's where my opinion is. That's where I'm the Antichrist. Like, I actually believe that just because you have a couple of bad apples does not mean you ruin the party for everybody else. And this is a party, and it's I liken it to the same thing as guns. Like, you know, there was this big over the last 20 years, this big push for like gun laws specifically because there were things that were happening around misuse of guns.

 

00:57:17:18 - 00:57:43:02

 

And the argument was always, well, why should somebody bad using a gun mean that everybody else loses the rights to guns? I do look at it in a similar respect that, you know, for KYC, people should be able to retain anonymity, especially in this world where your digital information, your digital footprint could actually be used by one of those nefarious characters.

 

00:57:43:06 - 00:57:49:18

 

So the problem is a solution, and the solution is a problem.

 

00:57:49:20 - 00:58:21:21

 

That's terrifying. Yeah I know. So when AI really hit the geist of conversation, everybody was already like seven years too late to it. And I've heard your breakdown of the actual long tail history of artificial intelligence dating all the way back to like, the 50s or the 60s. It's pretty remarkable. But more recently, one of the reason that OpenAI was founded was because Elon had this very prescient concern with AI taken to the upper limit of its capabilities, what it could do.

 

00:58:21:22 - 00:58:41:15

 

And his thesis was the paperclip maximizer, I guess it was by someone named Nick Bostrom thought experiment, which I find this fascinating, he says. Here's a scenario. Imagine that you give artificial intelligence one goal. I want you to make paperclips, and I want you to maximize paperclip production. And so the I goes, sure, my job is to do exactly what you say.

 

00:58:41:15 - 00:59:00:19

 

So I'll do that to the umpteenth degree, whatever it takes, because that's how artificial intelligence works. We'll do whatever you say, boss, but once it exhausts all of the easy raw materials, it has no built in reason to stop. So it goes to your car and it starts stripping your car down for parts needed to make the paperclips, because that's what it's supposed to do, boss.

 

00:59:00:19 - 00:59:14:14

 

It's supposed to make the paperclips, and then it's eventually going to go through your house, and then it's going to strip your house of anything that it needs to make the paperclips, because that's its assignment. And then take into the upper limits. The thought is eventually it's going to go, well. It turns out you can make really good paperclips out of human bones.

 

00:59:14:16 - 00:59:34:19

 

So eventually artificial intelligence is going to create paperclips using human bones. And so it's just going to demolish you because that was, it's that was what it was assigned to do. And so that is a thought experiment was kind of what it illustrates. Like once AI can source its own materials interface with the real world, it will no longer need humans to to it doesn't need to safeguard them.

 

00:59:34:20 - 01:00:02:04

 

So this was kind of the impetus for creating OpenAI. We need to safeguard a world where infinite technological intelligence that has physical, interactive ability in the real world needs to have some guardrails. Do you do you find the paperclip theory to carry any water? Do you find it to be at all valid? Oh yeah, absolutely. Especially as this thing gets smarter and the upper limits get crashed.

 

01:00:02:05 - 01:00:20:14

 

Now, having said that, humans are also very smart. We put limitations on things. There's ways to put limitations now. You know, I think there was an article that came out where this woman had an open claw and it like, deleted all the emails in her in her inbox, and she was a worker for OpenAI or one of the bigger companies.

 

01:00:20:14 - 01:00:43:09

 

And she immediately then figured out like, oh, I probably have to put some limitations or some sort of alert system like, hey, don't do this. Stop what you're doing right now. It's it's that easy. But having said that, you know, Elon's intentions are there. Unfortunately, he kind of got co-opted. That's my personal opinion. Oh, he got played for sure.

 

01:00:43:10 - 01:01:17:10

 

He got played. And to be fair, at the highest levels of of this game, it comes down to money power, control. So this idea that I'm going to build this ginormous machine for the benefit of humanity, you have to have an extremely strong will. You have to curate that experience to the nth degree. And even then, there is no guarantee that it will not get co-opted and used in a negative format.

 

01:01:17:12 - 01:01:23:06

 

Good luck, good luck.

 

01:01:23:08 - 01:01:56:23

 

There is something to be said about a a shark like Sam Altman being the one spearheading OpenAI, the first frontier AI model. People don't. I mean, if you do, great. But he came from Y Combinator, you know, and so for people that are in the business world, that are in, you know, startups, you know, angel investing, venture capital, Y Combinator was a fast it's a fascinating business because they they are rigorous in who they actually approve to say, yeah, we'll help you scale your business will deploy capital for you.

 

01:01:57:00 - 01:02:12:07

 

My buddy owns a business, and he went to Y Combinator to see if they would take him on, if they would allow him to integrate with them and they could get capital from them, help them scale. And he told me that in the onboarding forms, like you have to fill a ton of paperwork out. Never mind just submitting all your financials, your pals, your roadmap.

 

01:02:12:08 - 01:02:31:09

 

You have to go through this really in-depth questionnaire. And one of the questions was something to the effect of how far will you go to make sure you get what you want? And the people that answered back in these in the survey, like, yeah, I'll do whatever it takes, you know, I'll sell my own mom. Those were usually the businesses that got accepted.

 

01:02:31:09 - 01:02:57:21

 

That's kind of the ethos of Y Combinator. Like, they're ruthless. They want you to exploit. They want you to push any kind of boundary you can to get ahead. And that's what Sam Altman ran, Sam, and now he runs open. AI it's like if the writing on the wall wasn't just laid bare for everybody, you know, take a little historical look back and realize, like, the person running that first most frontier model was the shark of sharks.

 

01:02:57:22 - 01:03:22:01

 

A ruthless dealmaker talks out of both sides of his mouth, astonishingly intelligent. And that person is at the head of OpenAI making all of the decisions. And so when you go, okay, well, this is kind of a not great. Are there are there any heroes, are there any good guys in this scenario? We'll lo and behold, a guy named Dario said, I'm really concerned with all the things that we're doing well, like 2019.

 

01:03:22:01 - 01:03:44:16

 

So even before it was live, he's like, I'm really concerned about this. We need to have real guardrails, not just what we said we were doing, but I'm going to go. I'll make anthropic and it's going to be a for humanity. We're going to take all those things. And now Dario is just a suspect. So is it that having access to the most powerful superintelligence in the world, is it inevitably corrupting?

 

01:03:44:18 - 01:04:15:16

 

I don't think that having access to the AI itself is what's inevitably corrupting so much as it is that it's access to the power that you get from it. These are the heads of trillion dollar companies. That's not they're not heads of million dollar companies, not billion dollar companies, trillion dollar companies. You. And not only that, but they are beholden to shareholders, partners, employees.

 

01:04:15:20 - 01:04:42:03

 

There is a whole chain that they are in service to and in service of. So I think what we really have to one. Yeah. Sam Altman's a shark. So is Dario. Let's be clear. Bull sharks, great white sharks, big fat mofos. They eat whatever gets in their way. And if you don't understand that, believe that like you need a wake up call.

 

01:04:42:05 - 01:05:08:11

 

But they still are beholden to the powers that be within their own companies. And so I think it's a lot more gray than it is black and white. And I don't think it's just all of this person was good and now they're bad. Sometimes you as a business owner have to choose between a bad decision and a really bad decision.

 

01:05:08:13 - 01:05:30:19

 

And there is no winning. It doesn't matter which way you go, the public is going to look at you in a certain way. Your private friends, whatever it may be, are going to look at you in a certain way. Your shareholders, your employees are going to look at you in a certain way. So you get to a certain size where you have to remove yourself from this idea that if I go left, I'm a good guy.

 

01:05:30:19 - 01:05:58:13

 

If I go right, I'm a bad guy. You just are who you are. You're just stuck between making a decision. That's going to affect people on this side of the aisle or in this region, or this color, this creed and the opposite from the other decision. And that's something that unless you've had a lot of time in business at high levels, it's very hard to understand because we as humans, we want to go straight to the black and white.

 

01:05:58:15 - 01:06:17:19

 

We don't like to live in the gray because the gray has nuance. It has. You have to take on a different perspective. You have to understand things at a level that you maybe don't want to understand or, hey, I don't want to look under the hood because the moment I look under the hood, I find out something about myself that maybe I didn't understand about the world.

 

01:06:18:00 - 01:06:49:19

 

The world is less clear than it was before. That's my opinion on these heads of these tech companies. Now, having said that, Silicon Valley is a shark tank. They're going to eat each other and anybody else that gets in their way, that's business. It's war. So go to war. So looking ahead at the war, I know you don't have a crystal ball, but you are in some very interesting conversations.

 

01:06:49:19 - 01:07:14:05

 

You have access to some fascinating people, and recently even more fascinating people, which just won't talk about out loud in the next six months to a year. What are some developments in artificial intelligence and how it's getting integrated into society? Maybe a business or maybe on a personal level, what are some things that you're forecasting or kind of anticipating that are going to be earth shattering, game changing advances in artificial intelligence?

 

01:07:14:07 - 01:07:43:10

 

I think it's going to be more along the lines of hardware. I think we've kind of hit a maturation point of agents, you know, and I could be totally wrong. I don't think the agents are going to get a whole lot better where the public is concerned. But you're going to see companies come out with pieces of hardware, like a small stick that has a camera on it, that has infrared on it, that has an audio recording device on it that is connected to your phone, which has my instant AI.

 

01:07:43:12 - 01:08:00:13

 

So now I can see everything in the world around me with this little joystick. I can capture stuff. I can get a video of myself, a video of you, and I can tell it, hey, go edit this, clip it out on my instant AI and send it to Instagram and post it. Use my account. Use this account, log in.

 

01:08:00:13 - 01:08:26:21

 

Here's the password. Go. I think we're going to start heading towards the intersection of hardware and software. That's when it gets interesting because it will be, for me at least, highly interesting to see if the world actually takes on these pieces of hardware, like OpenAI is coming out with the earpiece or whatever it was, or the pendant. A bunch of other companies have come out with pendants, and for the most part, those hardware pieces.

 

01:08:26:23 - 01:08:52:06

 

Those companies have been in business for less than a year and go out of business. Yeah, because the cost is too high. The uptake of the product itself is too low. So and everybody across the board looks at whether it's applaud, whether it's a pocket, whether it's a pendant as a surveillance tool for the surveillance state. So there's a massive pushback on hardware.

 

01:08:52:08 - 01:09:17:23

 

So it'll be interesting to see what what happens as we move towards the need for better hardware to integrate with AI. This seems kind of naive, though, as if to say that we aren't already entirely surveilled. We we are. And you're right, it is naive. But the sentiment stays true that people are already used to their phones. They're already used to spending time on their phones.

 

01:09:17:23 - 01:09:32:17

 

They don't. They they feel like they can opt out. Like, hey, I can opt out a data tracking on Instagram. Hey, I can opt out of this. It's the devil you know is always better than the devil you don't.

 

01:09:32:19 - 01:09:50:11

 

We have a couple of Alexa's in our house, which I know, but it's so funny. I showed my wife this. I was like, okay, look, watch. We go to the opposite end of the room and just whispered, Alexa, do this like whisper and it can hear you the opposite end of your room. Anybody that has an Alexa, try this.

 

01:09:50:12 - 01:10:11:08

 

That microphone is so amplified, you can whisper commands and it can hear everything. And then if you start thinking about your phone has stereo microphones on it, your soundbar and your house is a stereo microphone on it, your smart TV has a microphone in it. Your Alexa's are positioned everywhere, so they all have handshake partner agreements. That's why they all work together.

 

01:10:11:08 - 01:10:33:19

 

And so they are all creating a stereo mesh of sound in your living room. You can hear everything you got a smart fridge. That bitch is recording you now. It's not like somebody is sitting in a strange little room monitoring everything that's going on in your house, but it is creating a mesh of information. It's flagging keywords, and that usually is contextualized to make your apps better.

 

01:10:33:19 - 01:10:53:05

 

It's still a ton of data, like our cameras will roll for an hour and a half, and we'll create 400GB of data. It's astonishing amounts of data, and there's not really a lot of infrastructure in place to sort through. Not just my data, but everybody on my block, everyone in my neighborhood, everybody in my zip code. It's so much data.

 

01:10:53:05 - 01:11:19:09

 

And so it's all just getting stored and it's not very efficiently picked through, but it is being surveilled. We are all being listened to nonstop. So I find that funny, where people are like, oh, I'm resistant to new AI hardware adaptations because now they're going to they were listening to you anyway. Your Alexa was doing it. They did the cutest little thing where I think it's like an echo Dot or something, and they rolled out models where it has a front facing camera on it, and people got really sketched out, so they rolled out new models.

 

01:11:19:11 - 01:11:35:12

 

It's got a little shutter on it where it can cover the camera like, oh, we're not, it's good you just popped that little shutter. I got no idea what's going on in your life, but bullshit. Yet they probably added in that. Have you seen this? Where they have that now Wi-Fi technology, where they can actually use Wi-Fi signals to map you out in your home?

 

01:11:35:13 - 01:11:57:05

 

Yeah. Like echolocation. They're putting that in all the hardware. So yeah, you can close the shutter on your little device, but guess what? It has echolocation on your ass all over the place. Sorry. So very naive. But the truth is, 98% of the world doesn't know this stuff. They don't have access to the information, or they don't care or it's not a priority.

 

01:11:57:07 - 01:12:16:16

 

It's, hey, I think this thing is bad. I don't understand it. I don't want it. So therefore I'm not going to spend money on it. And then marketing kicks in and then we'll see if that shifts or affects the way people think. I gotta give real credit to. So like I bought this pocket and I got it like a couple days ago.

 

01:12:16:16 - 01:12:33:09

 

I haven't really used it, but what I found is I have so many different disparate meetings and things that pop up in any given day running a business that I'm like, I wish I had a log of this stuff and it drives me crazy when I don't. So I'm trying to figure out, like, if this will optimize some of my conversations, knowing full well that, like, I have to make the terms and agreements work.

 

01:12:33:10 - 01:12:49:10

 

Like I know that they're they're storing all of it, but I'm like that. That's a that's a transaction I'm willing to make, because I just have so many balls in the air that I'm trying to juggle. And so I'm hoping I really admire companies that will jump into this in this very volatile product cycle. And they'll go, we'll make hardware.

 

01:12:49:16 - 01:13:07:06

 

But it's like not knowing at all how fast the landscape can pivot in six months. Like they go and they raise around some guy with deep pockets and dry powder is like, sure will allocate 50 million. We've got a manufacturer in Xinjiang will help you get this out to market. I mean, I admire the ambition and that I see stuff like this.

 

01:13:07:06 - 01:13:22:00

 

And it does get me excited just from a capitalistic standpoint. Like there are things that are moving, you know, the market is developing, but it's we're just still so early to it. Like this thing could be obsolete in six months. And now they design it so that, hey, like a new software, a new model comes in, we'll just update it on the back end.

 

01:13:22:01 - 01:13:44:22

 

It'll get better at processing, it'll get better at listening. The hardware is true, but man, the kind of balls it takes to try and deploy a hard asset like a real product, into a market when the thing that it's built on top of is so volatile, it is shifting every week. It takes balls, man. So I mean, so credit to you and your team got Tyler over there.

 

01:13:44:22 - 01:14:06:02

 

And the people that are building this thing, like you took a real swing at the fences you went. In spite of how rapid this growth is happening, we're going to roll out a product and take it to market. And dude, it worked. So congrats to you. I appreciate it and it's working. And you know, we we almost like went back to the future so to speak.

 

01:14:06:03 - 01:14:40:18

 

Like we went backwards in time to go do something that was so weird and different. Like everybody I talked to. Wait, you're selling gift cards for AI? Wait, you're selling prepaid cards? What? Why? And it's only now that it's like the amount of, you know, jaws on the floor of like, oh, maybe there is something here. And if we can get the next year, right, maybe there is a place where hardware catches up, and maybe there is a place where there's some, you know, intersection between what we're doing in a piece of hardware.

 

01:14:40:19 - 01:14:56:03

 

I don't want to let the cat out of the bag or anything like that, but we'll just say we have a roadmap. You're gonna have to tell me about this off air for sure. Okay. And with that, I'm going to go learn a secret. None of you get to know about. This has been forged. Thank you for watching, Alex, my man.

 

01:14:56:03 - 01:14:57:15

 

Thank you for being here. It.