Subscribe
Share
Share
Embed
Explore your curiosity - Interesting people with fascinating stories.
Life consists of three things:
-The stories we tell others
-The stories others tell us
-The stories we tell ourselves.
Matt, welcome aboard to the show. Thank you so much for joining us today.
Speaker 2:Hey, Ari. What's going on, man?
Speaker 1:So so, Matt, let's talk about Italy. At the very early days
Speaker 2:Yep.
Speaker 1:Of GPT, they outlawed ChatGPT. They did. They've turned around since, so I think that outlawing basically lasted three weeks or something like that, can't remember. But now they're building infrastructure and they're part of this whole kind of axis of AI that's happening through Italy. Tell us the story.
Speaker 1:What's happening here? Why is it happening this way? And and maybe we dabble a little bit into why they said, nope. No. Thank you initially and changed their mind.
Speaker 2:Sure. Yeah. No problem. So what's going on with AI in the European Union, right? Because that's really the larger entity that's kind of controlling things right now.
Speaker 2:Artificial intelligence is a society changing technology. On a personal note, I believe it is the primary enabler. I think that humanity can leverage to get to a tycoon civilization. Like it is that important. It is that impactful.
Speaker 2:And as such, nation states realize the potential of what is possible with artificial intelligence. And so it is now a battle of nation states, right? Or that healthy competition, right? That reshuffling of the world order in today's political climate, right? If you will.
Speaker 2:So, the primary nation state, currently in the lead as of right now is The United States Of America. Second being China. Third being the European Union. Very distant, distant third. So the EU got together like, oh, no.
Speaker 2:We gotta do something. So everyone is in catch up mode to The United States Of America where where it's at right now. And the European Union has drafted a massive, massive, federal stimulus. Right? Right around 200,000,000,000 Euro.
Speaker 2:Right? To literally proliferate, artificial intelligence, Gigafact, they call them like Gigafactories, all around metro areas in Europe, public education, developing their own large language models, like really building that talent pipeline. I mean, it is it is a highly, highly impressive kind of like whole cloth solution to, you know, really pulling up, 500,000,000 people and the world's largest economy into the age of AI. So it's really cool stuff.
Speaker 1:So let's play a game of the good, the bad, and the ugly. Let's do it. So, so let's start from the maybe the bad or the ugly. Is the, are these mega factories, whether in the EU or in Texas, are they going to be the birthplace of the Terminator?
Speaker 2:Really? We don't know. No one knows. No idea. Like, I mean, so you can say that like, is it possible for some, you know, like rogue artificial intelligence to, you know, like build robots and then take over the world?
Speaker 2:Yeah. I mean, who, who knows? Right? There might be a teapot orbiting Jupiter as well. Right?
Speaker 2:I mean, prove that one. So but
Speaker 1:Do do we need to build robots, though, or is it enough? And do we need general artificial intelligence, or is it pretty damn dangerous right now?
Speaker 2:Well, right now it's impossible. Right? From a just a physical standpoint. I mean, the world does not have enough compute. The world does not have enough data, like storage to be able to actually hit the AGI.
Speaker 2:Right? But they're rapidly iterating towards it. Now, this work of course is happening within, you know, dedicated, you know, facilities, very high-tech, very closely guarded, very highly paid, very smart people. Much smarter than I am. And so as they develop, you know, this cutting edge technology and it iterates, you know, yes, there is always the potential for misuse, right, with any new technology.
Speaker 2:Because again, I think that people misconstrue some kind of like esoteric threat from AI. Right, is is like science fiction, you know, that has really put into our brains and kind of inoculated us to what AI is. But we have to remember that artificial intelligence at this point in time is a tool. All right? I mean, fundamentally no different than a hammer, a wheel and axle, a screw, a ramp, like it is a tool.
Speaker 2:Right? It is a very, very powerful tool. And it's really how you leverage artificial intelligence within your processes and workflows is what the average out is what the the outcome is. Right? So we're using artificial intelligence for protein synthesis and protein folding to discover new drugs and all that stuff.
Speaker 2:That's amazing.
Speaker 1:Right? So we're gonna we're gonna get to the good in a second. Let's stay in the ugly.
Speaker 2:I'm sorry. So yes, it can be used for bad. Yeah.
Speaker 1:Okay. It can be used for bad. I agree with you. We don't have AGI yet, artificial general, the GE being important intelligence. Here's my argument though.
Speaker 1:And let's get your perspective since we're in the bad and the ugly right now. Organized crime is a thing. The United States has enemies, right? You used to, you know, work in adjacent area to that. Should we be concerned about the weaponization of AI, which is basically a form of a nuclear weapon that can't be detected that can right now can only live on really, really strong computers that cost 20 to $200,000 But second computer, you know, quantum computing comes around.
Speaker 1:I don't know if that's soon or not. You could have really dangerous AI in its current form in your pocket, on your watch. Correct. And even, and even without that, you know, being at mobile, even if it's on a server somewhere in Russia or China or anywhere, it's still really dangerous, not because it's going to go and do whatever it wants. It's not going to start building Terminators, but it could hack into places.
Speaker 1:And in fact, organized crime has unguarded or unrailed GPTs and other modules, which can do some pretty dangerous stuff, which are trained on exploitations. This gets real ugly real quick. If you think about abusing the general population through hacking and through scams, if you talk about targeting basically infrastructure, right, does the electric power grid really have the level of security that it needs to be secure against AI hacking where it's not humans, but it's Yeah. Like, this gets really scary really fast. And What do we do?
Speaker 1:Like, if you don't completely disagree with the premise No. No. No. No.
Speaker 2:I I I see what you're getting. Hey. It's like, hey. There's lot of stuff that can go wrong. What can we do as humanity to Yeah.
Speaker 2:I mean, we need our arms around this problem and all
Speaker 1:that humanity. United States. How does United States protect itself from people that don't necessarily have an agenda for United States to be here tomorrow?
Speaker 2:Understood. So, yes, in in that respect, like, as you're framing it, yes, we do need the legal frameworks in The United States Of America and internationally. Right? Because as you've correctly pointed out, yes, there are threat actors to The United States Of America. Like, just like there are threat actors, right, to every other nation state.
Speaker 2:Right? Everyone's got an enemy out there.
Speaker 1:Being one of them.
Speaker 2:Yeah. For for for some reason. Right? For some reason. Right?
Speaker 2:We just have to be mean to each other. But irrespective of that, this is our reality. Right? So anyways, yes. And so we need those legal frameworks just like we have with nuclear weapons, biological weapons, even the general laws of war and the Geneva Conventions.
Speaker 2:Yes. There needs to be international Let's
Speaker 1:let's break
Speaker 2:it down.
Speaker 1:This is one of probably three or four verticals that we'll talk about the legal Let's international just talk about this for a moment and then we'll jump to the other frameworks.
Speaker 2:Okay.
Speaker 1:Why do we have the Geneva Convention and what would the Geneva Convention look like for AI?
Speaker 2:We have the Geneva Convention to govern how we treat each other as nation states. Right? How nation states are supposed to interact. Because if we did not have a common set of rules like everything else, our society would collapse. Right?
Speaker 2:It would be chaos. It be anarchy. So that's why we have the Geneva Convention. Why do we need a Geneva Convention for artificial intelligence? Because it is such an impactful technology as you've previously mentioned, like with like nuclear weapons and biological weapons.
Speaker 2:Right? Like, I mean, you already see where artificial intelligence is being used for profit. Right? I mean, just what the other day, Elon Musk, you know, told his engineers because he sure didn't do it to flip a couple switches and then all of a sudden, I am Mecca Hitler and blah blah blah blah because of what one of the system prompts said to check Elon's own tweets. Right?
Speaker 2:And then ground the response in Elon's own tweets and it started pooping out Nazi stuff. I mean, you know, yes. We absolutely need a regulatory framework around artificial intelligence in The United States Of America. The key point of regulation, again, because this is a tool. Remember, AI is a tool.
Speaker 2:So, like, we cannot regulate an automobile. Right? I mean, we can rip you know, what somebody does with an automobile, we can regulate the consequences. Right? But we can't say, alright.
Speaker 2:You can only drive your automobile on this one type of street or in this part. Right? You see what I mean? So like that, we have to have consequences for AI misuse. We have to have consequences and legal frameworks for data protection and personal privacy that aren't there yet.
Speaker 2:I mean, and then on So let's break this down
Speaker 1:because it's hard. Right? If if if China launches sarin gas into The United States, it's kind of clear what happened, right? Like that's a violation of international courts. Yeah.
Speaker 1:How do you know that, China or anybody else, right? We're just, we're just using China as an example.
Speaker 2:Yeah, yeah, understand it.
Speaker 1:It could be Russia, anybody. I don't want to pick on any one state. How do you even know that they're using a technology in a way that would violate these new accords that haven't really been written yet? Yes. Because at the end of the day, you just have a bunch of noise, right?
Speaker 1:There's weird posts that could be people, could be not like, is this something we can really identify and know that it happened?
Speaker 2:Well, so again, and I am not into AI forensics at that level where I can tell you these are the exact techniques that we use to identify these specific use cases by blah blah blah. Right? Disclaimer. Disclaimer. Disclaimer.
Speaker 2:But what I can say, it's like any other threat detection. You know? It's like any other aspect of cybersecurity because it's not like artificial intelligence is sitting here in my pen. You know? No.
Speaker 2:Right? We know the threat. We know the incoming threat vectors. It's your mobile phone. Right?
Speaker 2:It's an Internet connection. So chances are if you're standing in the middle of a baseball field with a glove in your hand, you're pretty good and you're pretty immune from the threat of AI at this point in time. Okay? So it's it's just like anything else. I mean, and then then you have to look at the, really, the liability, right, of where the AI is injected.
Speaker 2:Okay? So if it's AI propaganda, what was the effect of propaganda? Like how do we punish propaganda now? You know, is there some sort of international propaganda punishment mechanism that we can leverage? And if so, then that would then funnel through current legal frameworks.
Speaker 2:And then that's how we would handle this issue, that issue in my mind. Also again, as not, not as a barred lawyer commenting on this either. So, you know, and so it would really just be not so much of, oh no, we have to account for all this new threat space. It's more of, oh no, how do we apply these specific incidents to the current legal frameworks and then map them to this new dynamic? Right.
Speaker 2:So that's, and then, but because we have to do that, right, because we have to have these legal linkages, whether by actual written law, you know, by Congress, federal, state, local level, or by, precedents set by the judiciary, you know, because we don't have that developed yet, you know, yes, people are going to, you know, make the foddles and foibles and mistakes as they would in incorporating any new technology into society. Right? We didn't have drinking and driving laws until we had a problem with drinking and driving, you know? And so, and that's the way that society changes and matures. And so This that's the
Speaker 1:really important because, but let me make it simpler by separating it into two things. One, there's all our free letter agencies that are, and they separate into two types of agencies, inbound and outbound, right? So on the outbound side, meaning protecting from external threats, really what they're doing is what they've done forever, which is map out the risk and the threat environment. And then you have the buckets of threats and then you kind of put protections for each level of buckets. And then you have a way to identify it.
Speaker 1:You have the forensics and you have your countermeasures, which are going be very real countermeasures. We're going have security people who will be working on this with AI people. Correct. Then we'll have the criminals who are also working on it. So we have all this, this, this outbound work.
Speaker 1:Now you started to talk about the inbound work. Well, what happens if somebody in The United States uses AI in a way that is dangerous, but they're a citizen or they're, you know, plagiarizing? Government's going to need to catch up and have laws for that. Yes. So basically you're painting this really nice picture which is very, very practical.
Speaker 1:It's a picture of look, there's work that we need to do, it's a lot of work, but the way that we've done this work historically, like we know that there's a threat of you know a Chinese agent or a Russian agent or whatever coming into our soil with a bunch of dangerous chemicals and dumping them in some kind of water purification plant and send poison to houses, for example. We know that's a threat. We've put together mechanisms to prevent that, to detect that. You know, that's one of the reasons why TSA, you know, checks our socks and our liquids and all that Yes. Know, like it or not, thank you to nineeleven.
Speaker 1:We've put mechanisms in place to prevent that.
Speaker 2:Correct.
Speaker 1:And diplomatic operations, etcetera. The same thing needs to happen for AI. That's your point, basically.
Speaker 2:Yes. Yes. And so I think and, again, it's just like in incorporating any other new technology into our society. I think what really trips a lot of people up with artificial intelligence is is why it's so destabilizing, and that's the velocity of the application. Yes.
Speaker 2:Okay? That's why it's destabilizing. So for every new major technological advancement in human history, okay, we, we being humanity, have had to wait for that those technological enablers to Right? Catch Yeah. So like with the automobile example.
Speaker 2:Okay. Great. Thank you so much. Henry Ford, you got the assembly line. Oh, wait a minute.
Speaker 2:Now we need the roads. We need to grow the rubber trees. We need to train the mechanic. All this infrastructure needs to be built out. Well, guess what?
Speaker 2:With AI, the infrastructure's already there. Right? So it's
Speaker 1:not scary because it happened in a few years. It took more than thirty years for the microwave to come around. The microwave, you can buy it in for $200, 150, you know, at, at Target. That first microwave was the size of a room. It was massive.
Speaker 1:Right. And that technology kind of matured over thirty years. We kind of figured it out. It slowly went out to the market. It became cheaper and cheaper and cheaper.
Speaker 1:And now everybody has a microwave in the room, but it took thirty years. AI kind of appeared two or three years ago when somebody let the cat out of the bag. The corporations had it for ten years or versions of it. Somebody let it out of the bag and now there's an arms race. The problem is, which you're describing now, that this arms race is moving faster than any technology in the history of mankind.
Speaker 2:That's exactly it. Succinctly.
Speaker 1:And and the problem is that the technology itself is making the technology move faster.
Speaker 2:Correct. So it's it's iterative. Right? Like, we and and that's why people say AGI is inevitable. Right?
Speaker 2:Because AI, again, is a is essentially a process improvement. You know? It is Yes. The Von Neumann machine of, of, of process collapse, of stack collapse.
Speaker 1:Yeah.
Speaker 2:And there it is. Yeah.
Speaker 1:I have made a lot of predictions over the last twenty years. My predictions are usually right in two verticals. One on what's gonna happen more or less. Two on the timeline. About three or four years ago, one aspect of my prediction started to fail consistently.
Speaker 1:It's the timeline.
Speaker 2:Okay.
Speaker 1:I said, oh, this one happened for three or four years. It happened three months later. So, so this, this concept of collapsing and kind of self, self improvement of the technology has completely made us as human beings unable to predict when the next big thing is gonna arrive.
Speaker 2:Incorrect.
Speaker 1:We just we just incorrect. Why do you agree? Why do you think that's true?
Speaker 2:Well, because what you're looking at when you look at AI when you really sit back and you study, AI incorporation into our daily lives, right, what you're really looking at is you're just looking at a rapid, essentially, of stacks. Right? And what that does is that allows you to then predict, to mathematically predict when your business process stacks will collapse due to a Genetic AI implementation. That's, that's the five That thousand dollars
Speaker 1:is true in the world where you have a set technology, that set technology is impacting the corporate world, which you focused on. And then you kind of see you can kind of create a predictive analytics to say, okay, this is gonna continue to percolate out to the environment at this rate with the set technology. My point is different. That set technology could have a leapfrog moment.
Speaker 2:Yes. Okay. You're talking about that iterative improvement in capability. Yes. Yes.
Speaker 2:Okay. Never mind. I'm sorry. I completely So misunderstood no.
Speaker 1:That was a great point. I love your point. I love That's your about the percolation of AI into the public. Great point. We should come back to that.
Speaker 1:Sure. But that next leapfrog in AI, nobody knows when it's gonna happen. When it happens, it's gonna hit all of us by a storm. Yes. And that's kind of scary because suddenly chat GPT, whatever, will be able to do something new that we could only imagine and we thought it would take five, six, seven years.
Speaker 1:It might be two months, might be three months. And we see this happening again and again and again. We just had ChatGPT murder a Blovable, right? So Claude, for example, can basically do everything Blovable can do. So they just released a version and they murdered a billion dollar company.
Speaker 1:Now Lovable is gonna survive I think, they're gonna go into the corporate world, they're gonna stay away from Chattypiti, but the thing is that it had this kind of leapfrog moment. And that's the thing that's becoming incredibly difficult to predict. Here's my, and with this we'll finish the bad and the ugly. If these leapfrog moments are happening all the time, almost once or twice a year, if not faster than that. Do we need to put some guardrails?
Speaker 1:I mean AI as it is right now, we're not too scared, but should we be concerned of these leapfrogs? Should there be a regulator regulation where we're like, on, before you release this new nuclear version of AI, let's take a look at what it can do because we don't want to arm, you know, our enemies. We don't wanna arm the public to do dangerous And
Speaker 2:I okay. So I think you're looking at it in a so you're you're right, but let me go ahead and answer that from a different angle. Right? So you're looking so you concluded that statement or that question by saying if we armed the enemies, are we like, you know, utilizing AI as a threat. Okay?
Speaker 2:That's not where the threat to our society really lies. Okay? It's it's not in in we do a new AI thing and then we give a new AI capability to a competitor. Like, that that's not the threat to our society. The threat to our society, again, reverts back to the velocity of the adoption of AI into our society, which like you just said, is the leapfrog moment, the one that we can't predict.
Speaker 2:Right? And when leapfrog moments happen, they normally happen sector wide. Right? So we're talking about finance. Okay.
Speaker 2:So once Claude, whatever, whatever, finally figures out finance bro ease. Right? You know, Wall Street's gonna have a very, very bad day. Right? You know what I mean?
Speaker 2:Like, in all those, you know, those young college kids and they're, it's not gonna be a good day for them. So it's gonna have some market sector wide implications. If you reframe that, right, and you throw in some, I think, Nassim Taleb, right, if I'm saying his name correctly, you essentially have sector wide black swan events for like one or two a year for the next five to ten years. That's what you're looking at for the American economy and for the global economy, right, as artificial intelligence proliferates and, like you said, hits those leapfrog moments. So this is what we need to do.
Speaker 2:The guardrails that you're talking about are really with the systems, the social support systems of our society. Okay? Because we can handle layoffs. Right? Again, that's not the issue.
Speaker 2:We have job retraining programs. We have unemployment insurance, all that stuff. But we cannot handle the volume that's going to happen when entire sectors get decimated and that transit.
Speaker 1:And that's what is destabilized. Start happening at the same time. Right? Correct. Might be one, two, three sectors, but then it might be five altogether.
Speaker 2:Look, dude. Alright. So here's the thing. Like, is math. Alright?
Speaker 2:And it doesn't matter if you are a coder in Silicon Valley or if you are a super awesome quant on Wall Street. You are doing the same fundamental thing. So as Silicon Valley is developing these coder agents that are self iterative, and their encoding is just math in and of itself.
Speaker 1:That's true.
Speaker 2:Like, as soon as it hits software development, again, finance, stack of cards. Right? And so and then you look at the other, you know, like mathematically, numerative fields. Hey, accounting. That's gone.
Speaker 2:But wait a minute. Look at the adjacent fields. I don't know. Are you a, a law firm that specializes in accounting fraud? Well, guess what?
Speaker 2:No. You're not. You're not a firm anymore. Right? So, and that's the ancillary cascade effects as well.
Speaker 2:So again, sector wide black swans and then the secondary and tertiary effects.
Speaker 1:I agree with your predictions. We don't know when this is gonna start rolling out, but we're seeing hints of it right now. Engineers that studied computer science, just come out of college, can't find a job. They go working in Walmart. That is just a small signal of what is going to come.
Speaker 1:Is there a, let's do some scenario analysis. Is there a version, we talked about the kind of horrific, terrifying version. What's the positive version of this? And let me give you a few economic arguments. The biggest force that has created equity worldwide was not Martin Luther King or any advocate ever.
Speaker 1:It's actually been technology. So if you look at the distribution of equity, it's been the wheel, it's been the written word, it's been the printing press, it's been technological advancements like the internet, potentially AI. So now you take, for example, a gardener, but this gardener is brilliant. They're basically Einstein, but in today's world they're a gardener and they didn't have the life circumstances to, you know, go and learn how to be engineers. In the world of AI, they go to ChatGPT, their brilliance of ideas of innovation can turn into code within today it's many, many hours, but tomorrow maybe within an hour.
Speaker 1:Now this gardener with their brilliance can build a business, become multi millionaires. So it's creating a potential equalizer and the ability to redistribute equity in a way that only technology can do. Is this, is there potentially a utopia that could happen due to this powerful equalizer?
Speaker 2:Absolutely. Yeah. Like, that's, that's the whole point. That's again, that's one the things that I do what I do. I mean, like, want to get humanity to Star Trek.
Speaker 2:Right? Like, you talk about the what you're because what you're really referring to in this scenario, that I hear is really the democratization of intelligence, right? The, that, and you know, the opening or I'm sorry, the collapsing of those barriers, right. That would otherwise inhibit that entrepreneurial spirit. Just startup costs in general.
Speaker 2:Right? Well, I completely agree. I've actually, I've got a, I've partnered with a program right now in another endeavor of my life, that we just ran a pilot of an agentic AI platform for underserved communities, right? To serve as the backend for their business. Right?
Speaker 2:So stuff like that is happening.
Speaker 1:Us more about this. This is this is beautiful.
Speaker 2:Yeah. Sure. So it was just a, it's AI for ROI, you know, because it needs a catchy slogan. And and, and just finished the pilot in Africa, and it's an agentic AI platform to handle, again, back end business processes that would otherwise be prohibitive from you or anyone else starting a business, such as human resources. Right?
Speaker 2:How do I hire? How do I manage my payroll? Like accounting. You know, how do we do our taxes? All that stuff.
Speaker 2:And really reducing those those barriers to entry for a lot of underserved and un economically disadvantaged communities. So that's the goal.
Speaker 1:So there's a there's a there's a beautiful opportunity here.
Speaker 2:Oh, yeah.
Speaker 1:And I think we we need to grab the the bull by its horn, so to speak. Mhmm. Here's my analysis of it. Tell me if you agree.
Speaker 2:Shoot.
Speaker 1:And always feel free to tell me that I'm stupid or dumb or saying the
Speaker 2:wrong No. No. No. There's we don't need to be rude.
Speaker 1:It's the only way I learn. Okay. I'm just exaggerating. I'm going to argue that the way that we as The United States, in kind of a narrow frame for a second, the way that our population doesn't go into serious economic devastation because of the introduction of AI will be predominantly through early adoption. Meaning that I work with firefighters, PD, every day.
Speaker 1:My what I want to start, and this is a program that I'm working on, is AI training for firefighters and police department. So, you know, how can they use AI? Because, you know, if they can use AI, if your gardener can use AI, if the knowledge of how to leverage this to kind of empower oneself and build the business of their dreams, if you know kind of anybody can do it, then when AI starts to devastate industries, we're going to have the power to create new industries, to create new innovation, to create new. So I would argue that we as the, let's say Department of Education, really need to rethink our syllabuses. We really need to think what our universities are doing right now Because if our universities only function is to make people productive and they're coming out of the universities and they can't find a job, something's going really wrong here.
Speaker 2:So I think there's one one thing I'll agree with and one thing I'll disagree with. The thing that I wanna disagree with is I wanna disagree with the premise that our universities are simply for making you more productive. There is so much value in a liberal arts education, and it will be so much more valuable in the future. Right? Because in the future, no one's gonna go get their master's in accounting.
Speaker 2:Okay? But everyone but you're gonna have a still a lot of people going to get their master's in philosophy. Does that make sense?
Speaker 1:I agree with that 100 Yeah.
Speaker 2:And then, the second part, to that was yes. Right? So again, as in and and you're looking at that transition phase again, and that's that, what we just talked about, those regulatory guardrails and those, you know, social support structures to, to, to take, you know, to, to process, right, these mass amounts of people. But yes. So, okay.
Speaker 2:The industry collapses. All right. Say retail collapses or whatever. But Walmart, all of a sudden we got a million or 2,300,000 people just out of work just sitting there. Right?
Speaker 2:So yes. Why can we not equip these people with artificial intelligence and boom, you're solopreneur. Go for it. Live your dream. Do it.
Speaker 2:You know? Like, why can't we do that? Yes. We absolutely can. And that will be a viable path forward for a certain percentage of them.
Speaker 2:And so we should absolutely, as a society, open that pathway up. Because what are we gonna When?
Speaker 1:When? When the industry collapses or today? Prior.
Speaker 2:Now. Now. Now. Now. Now.
Speaker 2:Now. So now you were talking about, Department of Education, which by the way, is set, thanks to the big beautiful bill to go away, right? To be done with. I mean, but so whatever, if, if in the future we still had a Department of Education and if I could design a national AI education and upskilling strategy, what it would look like is it would look like the establishment of those and those AI upskillings for the current workforce plus access to AI. We need a federalized LLM.
Speaker 2:Right? Period. We need a public LLM flat out. Right? And provide at no cost to our citizens.
Speaker 2:I mean
Speaker 1:So hold on. Hold on. You said something I think amazing. I haven't heard anybody else say that. You're arguing that LLMs or AI should be like water.
Speaker 1:Absolutely.
Speaker 2:Is such a Dude service. Bro. Alright. Okay. Look at this.
Speaker 2:You wanna talk about a public service? Okay. So, in a certain percentage of the population, you're always gonna have free riders. Okay. You will.
Speaker 2:No matter what the system is.
Speaker 1:Hold on. Hold on. For the for not everyone is an economist. What does a free rider mean?
Speaker 2:A free rider is somebody that does not contribute, you know, to the overall society. You know? You know, they're, they're commonly given, disparaging names, you know, welfare queens, stuff like that. Right. But in many respects, it's a lot of times somebody downed their luck, but there's a certain percentage that like purposefully, no, I'm just going to skate my way through.
Speaker 2:We know that.
Speaker 1:So so so what you said is that some of them are abusing the system.
Speaker 2:Correct. Some of
Speaker 1:them are
Speaker 2:always gonna have that.
Speaker 1:A certain percentage is abusing, and you can't you kinda can't deal with that. Mhmm. But you can do incentive design and try and lessen it. Yes. But for yes.
Speaker 1:Let's not move this into supply demand and economics and incentive design, but for the people that that don't want that, that they want to be productive members of society, those are the people that we're really focused on.
Speaker 2:Absolutely. Yes. Because you know what they're gonna use AI for? They're gonna use it the same thing I used AI for. I'm an internationally certified AI implementation specialist or whatever they call me or whatever the title is.
Speaker 2:And I did that all through AI. Like, I literally just back and forth with AI. Study buddy. Let's go. Let's get smart.
Speaker 2:And I've been doing that nearly every day for two years. Right? I have my own AI subscriptions. I have my own instances. I build my own agents.
Speaker 2:If it's too complex for me, I got guys that I went to hackathons with that I just call them up and be like, Hey dude, we need to get a rocket on something.
Speaker 1:So how do we
Speaker 2:It's amazing. Get Yeah, sorry.
Speaker 1:That you're doing is a privilege that you can do. How do we, how do we basically create a shock absorbent for The United States at large to be prepared when the industries start, let's not say collapsing, let's say morphing. So how do we create that stock absorbent? So we're talking about education, we're talking about having free LLMs and access. What is this program that, you know, politicians at some stage, we hope, will come up with?
Speaker 1:What does this program look like?
Speaker 2:In my personal opinion, it is, so where there is destruction, there's always the potential for renewal. I'll just leave it at that, when we look at the current state of the United States government. So, and however you feel about this current political environment, what you need to understand is this too shall pass. Right? The sun will rise.
Speaker 2:The earth will turn. This too shall pass. And then when this does pass, right, we as a society are going to need to rapidly, and I mean rapidly, rebuild the supporting infrastructure that our society is gonna need not just to stabilize itself, but also to continue to advance and stop regressing. That's what needs to happen. The only way that we as a society can do this, right, in the resource constrained future, right, because national debt, climate change, resource depletion, etcetera, okay, is through essentially the entire rearchitecture of our federal processes with artificial intelligence by incorporating AI into them.
Speaker 2:And the only way that we can do that is to ensure that we have the proper regulatory measures from congress laws on the books regulating the artificial intelligence development and implementation process. That's where we have to regulate. So that's where we focus the regulators on that's on on that narrow band. And then what that does is that leads to the ability of the nation state to properly rearchitect the functions it's gonna need to survive.
Speaker 1:So what that means is basically you're you're painting a a top bottom picture. You're saying Congress is going to have to do what it does, which is create laws. Those laws will have to look at all of the processes in a wide spectrum from the Department of Education all the way to basically money that we hand out to re educate people if they lose their jobs to how, you know, we train the next generation. So basically every vertical of how our society builds and operates because It's an operating system. Right?
Speaker 1:Society is an operating system. Yes.
Speaker 2:Our laws are nothing but an instruction manual for an operating system.
Speaker 1:That is
Speaker 2:what our laws are.
Speaker 1:We we basically the constitution, we build That's right. So with the constitution, I hope won't change, but the laws that support the constitution as well.
Speaker 2:We can get to that one too. So everything that you see right now as far as this social regression, right, this attack on individual rights, Right? Whether it be gay marriage, whether it be body autonomy if you're a woman, whether it be voting rights, gerrymandering, and all these societal ills that have kind of prepped in over the past about sixty to seventy years, those are the weak points. Right? That's what needs to be buttressed.
Speaker 2:Those are the cracks in the foundation that we need to fill in order to be able to build for the future. Right? So what are constitutional amendments am I talking about? I'm talking about getting money out of politics. Boom.
Speaker 2:I'm talking about term limits. Right? We all know what the ills are. Right? I mean, can point to, you know, like the the again, money out of politics, term limits.
Speaker 2:Those are my two big ones. Enshrining body autonomy into law, you know, like gay, lesbian, transgender protections, all of those. Bake that into the document. We need to update the kernel so we can rearchitect the I'll
Speaker 1:agree with I'll agree with corruption. I think everybody that's that's nonpartisan.
Speaker 2:Yeah.
Speaker 1:We don't go into politics in
Speaker 2:this Yeah. No. I got it. I'm yeah. Understood.
Speaker 1:But I'll say that there is a there's half of The United States that would not agree with some of the statements that you made. Mhmm. And so we we don't go into politics with this show on the show, but I think everybody, you know, cross party will agree on the corruption. Yeah. That's not a thing that anybody would disagree to, but there's other aspects that quite honestly, think there are ethical aspects.
Speaker 1:And I think that depending on your ethical framework, then you could believe that in one thing and in different ethical frameworks you can believe in another thing. And I'll tell you a short story. I was in China, somebody lied to me, I called it out. I was wrong because my ethical framework was the truth matters. But what I did in China, I actually highlighted somebody who didn't do what they're supposed to do in front of other people.
Speaker 1:So I basically shamed them. So I broke saving face. So my argument is that you can have the different ethical constructs that lead to different decisions. What you did right now is you showcased your own ethical construct and I have no criticism for anybody's ethical constructs. Yeah, of course.
Speaker 1:But there's other people in The United States who have different ethical constructs and at least on my show, I don't I don't disparage anybody's ethical construct. Yeah. So I won't say that I agree with with, you know, certain aspects of, you know, the pro life versus the the woman's rights. That that's ethical constructs. I I don't go there.
Speaker 2:Ari, that's fine. But what you're talking about is resolved through the constitutional convention process.
Speaker 1:Okay. So here's my argument. Yeah. Ethical constructs will never change in the constitution because there's gonna be too much disagreement between different ethical positions within The United States. That's something that's never going to change.
Speaker 1:So I don't think that's going to change. If I had to make a bet, I'll put money on that. However, corruption, absolutely. I hope that the corruption, that is cross partisan. We see corruption on the left and the right.
Speaker 1:That's definitely something we should improve. I am all for that. I don't think anybody's against it. But my question is different. In the context of AI, do you think there'll be needs to change the constitution ignoring ethical constructs?
Speaker 1:Will there need to be changes to the constitution in the context of AI?
Speaker 2:Yes. I can definitely see that, and trying the right of privacy into the constitution.
Speaker 1:Perfect. Okay. Let's talk about that. Okay. Why the right of privacy?
Speaker 1:What does that mean? What does it look like today? What do we need to change?
Speaker 2:Okay. It looks like today, it looks like absolutely nothing. Right? So, I mean, think about how many apps you've clicked through. Did you read all this did you really read the terms and conditions?
Speaker 2:No. You didn't. So it's getting to the point, right, where artificial intelligence, you know, I mean, is essentially, you're right. You're being trained on the entire Internet. Right?
Speaker 2:You've heard that before.
Speaker 1:Yes.
Speaker 2:Okay. So what does that mean? That means a sufficiently powerful artificial intelligence can essentially run facial recognition on every picture stored in a database that's accessible to it. Right? Okay.
Speaker 2:So have you ever taken a picture or video with someone that you don't want to go be seen by anyone else? And are you sure it was ever deleted? Right?
Speaker 1:Or has somebody else took a picture of that by mistake because you were in the background?
Speaker 2:Exactly. Exactly. So and keep in mind, this is everything. Like, this is traffic cameras. This is driver's licenses.
Speaker 2:So what does occur? Right? So when you look at what's happening across the world, this is what I see. Right? Data right now is currently being structured.
Speaker 2:And I mean all the data, right? The databases are getting connected. The APIs are going out. The AI is accessing the information. And once we get to the proper level of compute, well, I'm going be able to pull and, and, know, you turn that temperature down, of course you get the weights and I can get anything on you that I want.
Speaker 2:Yeah. Right? Because it's out there. So
Speaker 1:believe So hold on. Let me try, let me translate that to, to kind of simple English. This MCP server that you're talking about right now, what it really is is the ability for the AI engine to connect to any data in the world. We're doing this slowly, but it's gonna accelerate significantly. The temperature that you talked about is basically mechanisms to reduce hallucinations, to be less creative, to give factual data.
Speaker 1:There's additional methods to reduce hallucinations. We can talk through that. It gets too geeky and mathy but we can talk about separation of concerns. We can talk about all that but there's a lot of methods that we're currently and you know multiple parallelization operation of AI agents and then selection of correct results versus hallucinations. So there's so many things that are happening right now that are going to solve the hallucination issue to a 99.999%, Right?
Speaker 1:We're going to get to Six Sigma. And so it's going happen slowly, and then all at once. But your point is that if all information is correct and AI can actually do the job,
Speaker 2:what happens to privacy? Exactly. That's exactly it. What happens to your privacy? Because, I mean, once you combine artificial intelligence, excuse me, with quantum computing, cybersecurity's gone.
Speaker 2:Right? And our cybersecurity is let me
Speaker 1:let me explain that. So cybersecurity today is based on the fact that you have encryptions. Encryptions basically translate into the fact that it takes a certain amount of time to break the encryption. That amount of time can be hundreds of years if the password is long enough. Quantum computing comes along, thousand years turns into three seconds.
Speaker 1:Security's out the window. So these, like quantum computing, another existential threat to privacy, right? I mean, that's one way to look at it, but also will probably solve cancer. That's another way to look at it, right? Everything has the good, the bad and the ugly.
Speaker 1:So, you know, it's very important to be able to do scenario analysis and basically understand what the spectrum looks like. You know, from that perspective, you know, quantum computer comes along, encryptions fall. We have to have quantum safe encryptions basically. That's a thing we could develop. And then we have all the data connected, so privacy falls.
Speaker 1:So we need these laws about privacy. What does that look like?
Speaker 2:A good question. Not a privacy lawyer. I mean, it's like when you get into the nitinoid of the actual legal, that's that's a great question. In general, it's going to be essentially sovereignty over one's I think a lot like what's happening in the EU. Right?
Speaker 2:Sovereignty over one's likeness, sovereignty over one's intellectual property, sovereignty over one's voice, you know, and, and just really, you know, maintaining your person and really codifying the fact that essentially you have a digital twin is really what we're doing. Right? And so And you
Speaker 1:own that digital twin.
Speaker 2:Exactly. And it's a part of you. I mean, because you know, right now with all of your, I mean Ari, all of your personal information is thrown into marketing algorithms and you know, Google AdSense makes money off of it. Right? So in the future, you know, are all of our children just gonna be little monetizations?
Speaker 2:Is that what we're really, really going for? Like, where where everything about their life is is just exploitable at the drop of a hat? We talked earlier. Right? I think we let off the conversation with threat of propaganda and artificial intelligence into the population.
Speaker 2:Like, it feels right into that. Because again, if I can there is no right to your if you have no right to your own likeness and it's manipulation, then, you know, you you already see the deep fakes of, you know, children that have been kidnapped and, you know, that scam that goes on. Many more like it. But yeah, and that's right. Exactly what I think the flavor is going be.
Speaker 1:Perfect. Absolutely. I think that one sentence you said is that you own your digital twin and that's going to be the starting point for all of this. And I've had this argument with many people. Many say, no, it's legal.
Speaker 1:Sure, it's legal today. There's a really It no should be legal. That's right. But I've had this argument like, oh, if you write a prompt, create a book on this topic in the voice of somebody else, that's gonna be illegal in the future, in my opinion.
Speaker 2:Mhmm. Well, and also when we look at when we talk about illegal. Right? So that's essentially having a copyright over your own voice. Yes.
Speaker 2:Okay. Well, then now we have to relook at copyright law because I mean, what happens? Hey, I can tell my LLM, give me a children's story starring Mickey Mouse and Bugs Bunny, and it'll do it. Yep. Just fine.
Speaker 2:Well, okay. Well, now I've turned around and I've generated a, you know, essentially a product that I can then publish and sell. And if I publish and sell, I will, you know, get hit with copyright. You know, does the, does the actual GPT, right? Does the LLM does, who has the liability for that?
Speaker 2:That those copyrighted characters were essentially used without paying that, that fee to the publisher or to the rights holder.
Speaker 1:We've never charged the photocopy machine the copyright fees. Exactly. However, you know, copyright law as it is today, if I do it for myself, nobody's going to care. If I start selling that on the internet, whole different story, then the lawyers come after me. This is why my son is studying to be a lawyer, because I don't think lawyers are going anywhere.
Speaker 1:I think they're going to be a big part of the future.
Speaker 2:Well, Ari, but if I can create my own story with Mickey Mouse, why do I ever need to buy Disney's?
Speaker 1:Well, the second you start selling that, then you're in trouble. Now if it's for self consumption
Speaker 2:And everyone's self consuming? Because why would I go out and spend $50 for a Mickey hat I'm sorry, for a Mickey video when I can just have the LLM create it for my kids? Hey. And I can I can rig I can rig the prompt? I want it to reflect these values.
Speaker 2:I want it to I want the story to be like this. I mean, why? Why would I ever purchase from Disney? Why would I ever go see a Marvel movie?
Speaker 1:You got me there. You got me there. Got me there. I didn't even think about that. Yeah.
Speaker 1:That's brilliant. That's absolutely brilliant. Matthew, no better way to end this show with a huge compliment. What a wonderful Thank you. Clearly, you have a brilliant mind.
Speaker 2:I appreciate I'm
Speaker 1:an absolute delight to go head to head with you.
Speaker 2:Man, I appreciate it because I am running on fumes, brother. Been trying to like, yeah, this, I think this concept, so yes, gotta do this again when I'm fresher.
Speaker 1:Ladies and gentlemen, this is there you have it. Matthew Fumes is still brilliant. There's one question that we ask every single one of our audiences and that is if you had to go back to the most difficult part of your life, maybe your twenties, maybe later on,
Speaker 2:Mhmm.
Speaker 1:And you had to give yourself advice. What would that advice be?
Speaker 2:This too shall pass.
Speaker 1:Matthew, thank you so much for joining the show today.
Speaker 2:God bless.