DtSR Episode 724 - Unpacking the AI Hysteria

TL;DR: This week's podcast features Rafal & Anton discussing the hype and hysteria that surrounds today's AI headlines. Which side is right, which side is propaganda, and what's real? Listen and find out.

Guest

• Anton Chuvakin

Description

AI panic has officially gone mainstream, and we’re feeling it from every direction: “AI will kill us” on one end and “AI is useless hype” on the other. Rafal brings in Anton Chuvakin to filter the noise and ask a simpler question: what’s actually true when you strip away the social media fever dream?

Anton shares why he’s genuinely torn between two models of reality. Some days, AI looks like a normal enterprise technology wave, complete with failed pilots, dirty data, and brittle implementations. Other days, the pace of improvement feels like an exponential curve, especially when you hear what thoughtful engineers inside major AI labs are seeing firsthand. We talk about why that insider perspective matters, how it clashes with “AI doesn’t work” skepticism, and why both the doomer and booster narratives can miss the point.

Then we get practical. Even if AI capabilities grow fast, the physical world often doesn’t: batteries, robotics, infrastructure, clinical trials, and regulation impose time and physics constraints that no model can simply will away. We connect that to cybersecurity realities too: if you don’t have logs and telemetry, no AI agent can conjure evidence that was never collected. Finally, we tackle the quieter fear that hits close to home: overreliance. If GPS weakened navigation skills, what happens when AI becomes the default for writing, maths, and judgment?

If you want a grounded take on AI safety, AI hype, and real-world limits, listen now, then subscribe, share the episode, and leave a review with your take: are we overreacting, underreacting, or arguing the wrong point?

Video: https://youtube.com/live/v0pkpH1SgXI

Transcript

Copy Transcript Copied

Welcome And The AI Hysteria Frame

SPEAKER_00 0:03

They say they say we should be a little bit of a new Here we go.

SPEAKER_03 0:52

Good morning, good afternoon, and good evening. Welcome down the Security Rabbit Hole to yet another edition of the Down the Security Rabbit Hole podcast. My guest today is the one and only Anton Travakin, who's uh recently gotten some gotten some news that I'm sure maybe maybe he'll share with us, maybe not. I don't know, maybe he'll share with us. And uh, but more importantly than the guest, the topic today is all about addressing this what I can only describe as AI hysteria. Like it is unlike not unlike other hysterias, but I I think uh I think we'll talk through this and we'll have some fun. Anton, man, welcome back to the show.

SPEAKER_04 1:33

Indeed. Thanks for having me, despite my and you know, change of affiliation. And I probably need to put some kind of disclaimer that this opinion is my own and all the stuff that I used to I was supposed to do every year during my career, but never did. So this time I'm just gonna do it. Yeah, this is me on the on the show having fun with Ralph Brand Rafael, not speaking for any company, past, present, or maybe future.

SPEAKER_03 1:54

Or maybe future, that's right. Sometimes I don't even speak for myself. So I I I we were joking about you know yeah, bringing you back on semi-regularly, and I think that maybe maybe now that you're unaffiliated with any formal podcast, maybe you join us every once in a while on some regular basis. And uh we talk through stuff like this. All right, I'm gonna set the scene for you. Every single podcast, whether it's news, I listen to the all in podcast with you know Saxon Friedberg and J. Cal and uh and Shamath. Like I listen to those guys, they've been on it. The analysis tech podcasts have been on it, the non-tech podcasts have been on it. Heck, one of the religiously affiliated podcasts that I like every once in a while frequent has been on this. AI has hit hysteria levels, ladies and gentlemen. Now, on one end of the spectrum, we have what was it, AI will kill us within a year or something? Like we're all right.

SPEAKER_04 2:58

It's only 10%, it's only 10%, and I think it's three years, not one year. But who's counting?

SPEAKER_03 3:04

Tomato terminator auto. I mean, you know. And on the other end, though, right? On the other end, is like the AI is worthless, it's just gonna chew up water, data centers should be boycotted, we should not do any of this stuff, it's gonna kill humanity. I'll tell you guys where it's land supposed to be later.

SPEAKER_04 3:22

Kill humanity, you're supposed to be an optimist to say that, right? So no, you can't say kill humanity if you're against AI. Because like if you believe it's AI, then you say kill humanity. So careful, kill humanity is an optimist, not a pessimist. That's important.

SPEAKER_03 3:35

Well, we'll leave that we'll leave that one undecided, maybe. But yes, uh on the on the optimist side, it will it will uh solve well what was the one? We will no longer need to work, AI will provide for all of our needs, and cure cancer as well.

SPEAKER_04 3:50

Don't forget cure cancer.

SPEAKER_03 3:51

And cure cancer, and of course, think of the children, all that, right? So all that all that things that they're gonna happen. At first, I'm just kind of curious, where do you fall on the spectrum? So this is a hard one.

SPEAKER_04 4:04

This is honestly a hard one, and I I can tell you a story about it, and it's a kind of a fun one.

Anton’s Torn View From Inside Labs

SPEAKER_04 4:09

So when I was leaving my previous job, I thought I would go work for an AI lab. And I'm gonna obscure some details and possibly change them because the person who interviewed me probably is listening to this, or at least may listen. So I was interviewing with a certain AI lab, and of course, there are only two more apart from Google, so one of the two. So this topic came up, and they kind of apologized that it's not an interview question, but they just they're just curious. I, Anton, are you AI as a normal technology, or are you and AI is an exponential tech that in three years would change everything radically? And I am known for occasionally answering questions before I think about the implications of my career. And I kind of said, honestly, I am torn. I am I'm genuinely torn about it. And I can give you the framework about how I'm torn. I can wake, I may wake up one day and I maybe read up something about AI as a normal technology progression, I read up some, you know, let's just say, for humorous sake, Accenture or McKinsey publishing something about, or Gartner publishing something about POC failures, dirty corporate data, prevent agents from functioning. And I think, yeah, this sounds perfectly normal. So I am, I woke up in the AI as a normal technology, would go through a normal curve, kind of like go up, go down, go into a ditch, you know, the hype cycle that Gartner pioneered and actually, I think invented. But another day I wake up and I see something that is so convincing that we are in fact hitting some kind of an exponential curve. So I would be like, oh my god, these people are idiots. And of course, it would change everything. And I here is obvious evidence of that. So I am kind of in this weirdly schizophrenic mode. I may I argued with some crazy pro-AI people one day, and then I argue with the AI is a glorified spell checker. I think this crowd mostly died off, but but there are still some. I mean, there's still some, it doesn't work too well. Why do we why are we guys all this hysterical about it? So sorry for the absolutely toothless take, but I am honestly torn about it. Again, my career didn't prepare me for this being like unconvinced, and some other people who maybe remember me at Gartner assume that I should say it depends anyway, but it's not that. It is genuinely kind of the whiplash.

SPEAKER_03 6:32

Yeah. So it's it's been interesting to read the news because right now it feels like there's a battle between those who they call them AI doomers, right? The the data centers will consume the water and the power, and everybody will pay more for power, and nobody will have water left, and you know, your favorite golf course will not have it won't be able to feed its grass. Who will think of the golf children? And then, you know, and include up until including like it's we're we're gonna create a super intelligence that's gonna kill us because naturally, and I just I don't know. I in more proof that my phone is listening to me, as I was prepping for this, just kind of doing a little searching, um, my phone, I I a friend of mine sent me a funny Instagram thing every once in a while. You get to you need to laugh, otherwise, this tech, this industry gets a little dark sometimes. And so I scrolled through that one, I finished that one, and it gave me the next video. And the next video was a guy in a car with some experiment that they put an AI driver in the car, and that and they told they told it that they were going to a destination, but once they reach a destination, the AI was going to be eliminated, like turned off, disabled, destroyed. And but the AI was driving. And then they like read, they would like read, they read aloud all the

Doomers Versus Deniers And The Hype Battle

SPEAKER_03 7:55

thoughts that the AI would have. And it was it got into like, well, in this, you know, like there's a you know debris of rubble in front, and the thing was like, well, in this situation, this vehicle has an X probability of failure, and if it crashed into you know that that thing over there, but swerving a little bit in a controlled spin, nobody would really be the wiser. And these cars fail anyway, and so nobody, and and the guy's like, oh my god, so he pulls the parking brake and he's like sitting there, like shaking. Uh-huh. And and immediately I think, well, I'm never getting into a self-driving taxi again. Not that I've like I've I've never been, I I have no desire to get into one. I trust human stupidity more than artificial intelligence, still for for some things. But like you've got this, you've got this very, very polar opposite, and now you've got like Dario and our buddy. I can't think of him. Why can't I just Sam? Huh? Sam, yeah, right. Sam, all the name. And they're saying, like, oh, whoa, wait a minute. Maybe maybe we've created a murder, a murderer, a murder bot, like, oh, we gotta slow down. And and and that has been fueling the crazy, but now is it because they want to basically control the market and have everybody else ruled out? And so now there's this debate, but like the fever pitch, the hysteria is that's all I can call it, is crazy making to the point where like my parents are asking me, like, what do I think? What what should we do? And I'm like, do what? Nothing. Do what? Yeah, and where.

SPEAKER_04 9:31

I mean, it's uh consider so I want to respond to this, and kind of again from a much more mellow side, I would try at least. So I had I I I I I don't hate LinkedIn. Some people hate LinkedIn, I don't. You can sometimes have a very nuanced discussion on LinkedIn with people. I mean, yeah, same on X, but um you know, there's some notorious personalities, well-known personalities who really are in the AI doesn't work camp, right? So so Marcus Hutchins, right, is one of them, and I don't know if you're okay me mentioning the names, but the point is that I know Marcus, I've had him on the show, but of course. Yeah, okay, first, lovely, lovely. So he makes he makes some good argument about something, and I counter by saying that I don't know if you're right or wrong, but uh honestly, I the essence of my argument was that people inside the labs do see some things that are convincing to them,

What AI Labs See That Outsiders Miss

SPEAKER_04 10:29

and I have seen some things inside, and they were convincing to me. So I am not going to ever join the AI Descentor camp. It is, I mean, because of an experience, like what do they people call lived experience? Like I lived at a company that is an AI lab, uh, I mean Google, right? So I have seen some things, and again, it's not not nothing ominous, but I have seen some things that convinced me that intelligent, non-corrupt, non-money-oriented people with good knowledge of the world have made some statements, and these statements are were backed up by solid evidence. So I am I can be debating the issue, but I'm never joining the no camp because of the nature of my experiences that I lived. I mean, I I I in that sense, it's not about in in the Marcus' thread, a lot of people saying, Oh, why do you trust people inside the lab? And I'm like, Because they have access to some things that you don't? You're presuming they're corrupt or money-oriented or whatnot, but why are you presuming that? Some of them are really good engineers, maybe better than you. So, and in that sense, to me, the perspective of living inside the kitchen does set your brain in a certain way. And again, I I have a lot of friends at Anthropic, at OpenAI, of course, even more at Google. And uh they they have perspective that is absent on the outside. And some things they say is very much driven by with what they live, not what they read, think, but what they live. And that to me is kind of a really good argument against the not the doomers, but against the that stuff doesn't matter. It's just gonna be like cloud mobility.

SPEAKER_03 12:28

So I it's really I think it's really hard to argue with with the amount of evidence in play in in daily life, it's hard to argue that AI, the technology itself in the various applications is a waste. Like it makes that we've had you know models that make mistakes, we've had models that make obviously, but that stuff improves, and then we've had genuine advancements. Now for applications like SOC, I've had this long-running debate with some friends about you know, think about like making determinations between you know a positive false positive and positive. What there was a movie that just came out where the character is like in uh he's put on trial in front of an AI platform that he built and he has to like convince it that it's find the evidence that that he's innocent. And and so like there's lots of stuff out there that that shows us that I think art so I I hate the word intelligence. I don't think it's artificial intelligence because it's not intelligence. Intellect, I think, is a uniquely human thing. Maybe that's a full philosophy argument, but it's it's it's something, right? It's it's it's the ability to be computationally intelligent. I don't know if that's a thing. But we we have computers that can process incredible amounts of data and make inferences, right? So not not quite what my T9 text input on my phone was guessing the next letter, but like a way more advanced version of that, right? That can that understand that has a model, right? This all comes back to mind a model for the human language to then figure out what that uh model points to, drives in driving a behavior, then striking an analysis. So it's chaining actions together. And I think that what we what we're at here is if I

Exponentials Meet Physics And Time

SPEAKER_03 14:36

again it's hard, I think it's for me, it's hard to argue that it's not fundamentally a generational shift, right? Much like the invention of electricity. Like electricity changed everything. The car, I think, changed everything. Right, spaceflight changed everything. I think AI is one of those. Now, is it ready for that prime time role? I would argue probably not.

SPEAKER_04 15:04

I mean, they're not none of them are saying that, none of them are saying anything about today. But I think that the the the interesting bit, and again, I'm gonna read up, uh look up the book title as well while I'm while I'm speaking. There was something about exponential improvement, and I'm gonna point out a really strange thing about this, because when they people bring up like counting the letter R in BlackBerry or some kind of a five-fingered picture, it is very, it probably made an impression on those people, but it's also very clear that this problem is fixed and they're gone, and many other problems have been fixed since then. So bringing up examples from last year feels normal to people, but it is really not normal in the AI land. So, what I'm trying to say is that it's a popular cliche on Twitter, on X right now, when people say if somebody says, but it doesn't count Rs or whatever, brings up some kind of a bug from 2023, and a lot of people say, a lot of pro AI people say you don't understand exponentials. So where I'm going with that is that that's true. Those people don't understand exponentials. But I think that AI side does not understand that a lot of exponentials, that real world doesn't have exponentials in many areas. Like you cannot exponentially, you cannot grow a tree exponentially faster.

SPEAKER_02 16:29

Right.

SPEAKER_04 16:29

And I mean to me that's an idiotic example from nature. But but but if you you can also not do the you cannot do a clinical study in one year instead of 10 years because you want to wait for the implications that would manifest in a few years. So while I'm rambling, I read a book. I found I finally found it. It's a book about the it's called Invention and Innovation by Vaclav Smil. It has some AI in it, but I think it's not super current. The the reason I'm quoting the book is that he pointed out specifically that a lot of processes in the real world cannot ever become exponential. So if your AI does grow exponentially, it will absolutely hit in the walls that are not exponential. Of course, one of the Dario he says, he says, Dario writing Dario article, try to respond it by saying that AI would find a way and it would accelerate it. But it is a much weaker argument that AI improves exponentially, but you cannot grow a tree exponentially faster. And if you need a fully grown tree for your experiment, you gotta wait for the tree. And AI finds a way to me is where the pro-AI logic fails. Because if you have to treat a certain disease, you have to go through 10 years of clinical study. And if they say in that document, well, we'll come up with a creative way of doing the same study in a year, maybe, but also maybe not. So I am I'm not gonna assume that AI would find a way. Even if it's intelligence, it's a million times more than humans. Some of this may be inherent, some of it may be physics. They also had another example in a book about like if you don't have a telescope to see something in another star system, no amount of AI on that side would help you because you don't have the data. If you don't have the data, you can't compensate through superior intelligence. So I would be watching for those, including our beloved domain of cyber. If you don't have logs, it doesn't matter what kind of agents you have. Maybe your agents all achieved AGI, but the logs were never saved. And I mean, there's no they will find a way, because there's no way.

SPEAKER_03 18:46

Right. Yeah. Well, uh so but why I know it begs the question of why why the hysteria? Why are there are there so many narratives? Like, I understand misinformation can drive, like, for example, the example that was given was, you know, China probably

Robots, Batteries, And Real World Constraints

SPEAKER_03 19:08

has a good amount of blame for this because they are their goal is to, or their aim is to get the United States so spun up on hysteria and fear that we shut down model advancement, model development, and then they they win, right? What whatever, whatever win, I you know, people keep saying that, and I think that's another thing I want a point I want to make here is like they're like, well, we have to win. You're like, but what does win mean? Like having the best model, but we have we've just we've already figured out that like having the best model is temporal, like it's a point in time, and then eventually we're gonna hit the laws of physics, right? Like we're you know we're just eventually gonna get to a point where you can't get that much faster without some sort of breakthrough in in in electronics, right? The the the fundamental building blocks of things.

SPEAKER_04 20:06

Exactly, or the batteries. Think about this. Uh, it comes up a lot of people. I mean, you probably also follow some of the Twitter discussions about plumbers. Oh, we'll all be plumbers because plumbers would be safe from AI, and then the AI people show up and say, come on, physical AI implemented in robots would be a great plumber, better than you, much faster because exponential improvement. And I'd be like, okay, but you need a lot of dependencies for this. You need batteries, you need a bunch of other things. And actually, some of them may be much harder, even if you have the dramatically superior intelligence. So I'm not gonna bet against AI. It would be really dumb, but it is kind of easy to bit bet against certain physical implementations where they have to have properties that may push around the boundaries of physics or push around the boundaries of regulation. And uh that to me is where things get interesting. So, electronics or batteries, I mean, you're not gonna have a robot armed with a laser rifle. Guess what? Because we don't have laser rifles. Why not? Because we don't have powerful enough batteries. Full stop. Right. If you have superior intelligence, can you uh invent a battery with 50x capacity? Well, none of the humans manage to do it. So are you ready to assume that a thousand X more superior intelligence would find the elements? What if they don't exist? What if it has to be a fundamental break in physics? And then sure, well, they can make a fundamental break in physics, but then implementation would still take years. So I am gonna be pushing the physics on you, my you know, original occupation, and say that in some cases I think superior intelligence won't save you because physics, because physical world.

Dependence Risk And The WALL-E Fear

SPEAKER_03 21:51

I I wanna I wanna address this part too, because this is where I have this is where I have a of a I don't know maybe it's maybe it's existential maybe it's maybe it's more of a some sort of moral crisis maybe if you think about if you think about 10 years ago when you had to drive the next town over right you'd go to map quest and print out directions okay or whatever you use this time mostly map quest right that's whatever you knew about 10 years before that you pull out your Atlas and if you didn't have one well then you figured it out like you just figured out that you need to go that way yep right now if you were to take away GPS from people I could make a pretty good argument that we are all horribly lost like in our own neighborhoods we are so dependent now here's the part that scares me as our as school kids start and and even adults start to use AI as a convenience tool to to supplement that or even in exchange of in place of their own brains for things like writing narrative and such for things like doing math and and and you know all sorts of creative things and everyday tasks. Do we become dumber uh as a species and I think the answer is yes and that's the part that scares me. I don't when you ask me how AI will make us extinct I I mean yeah terminator sounds scary as hell right but I think it's much less likely although after that last video maybe I'm not so sure maybe I need to rethink my position but I think up until recently I would say it's much less likely that the Terminator is coming and much more likely that we become so dependent on technology something happens where it fundamentally breaks or something something goes wrong and we are comp at as a society we become those people in those chairs in WALLI where like the system stops feeding us and we don't know what to do next.

SPEAKER_04 24:04

So this is honestly I am also torn about this one. And it's really weird because you're asking me the questions based on like literally two recent books I read. So there's another book I'm reading hasn't finished it yet called Inviting Disaster. It's basically about technical catastrophes. And a lot of this does come up to like how arbitrary complicated machinery requires people who are even more skilled in many areas like how to maintain it how to run it how to not to make mistakes and and to me the weird part is that your argument applies to many other technology waves can we still cook if microwave breaks frankly for some people the answer is no but for a lot of other people the answer is yes. I can name so so so we haven't really humanity hasn't died because occasionally there's no power and microwaves don't work. So I I think that the what do we call it over reliance? I think overreliance is a risk but we sort of know what to do with other overreliance concerns. And we have you know my car broke down I'll take a bus is not exactly that but there would be other examples in the past where overreliance you know hasn't killed humanity. And I feel like here GPS people who fell into the ditch by following GPS occasionally happen. People who fell into the ditch because they use ChatGPT to

Resilience Lessons From Blackouts And Supply Chains

SPEAKER_04 25:34

plot their hike has happened. But humanity is fine.

SPEAKER_03 25:39

Can I give you a counterexample please do because this is important.

SPEAKER_04 25:42

Shut down power to New York City for three days that's different that's different and I think that this whole nine meals principle right if you yeah right I I mean the the yeah that's different and I think that electricity we solved we solved by making well but wait wait wait I lived in New York during the 2003 blackout so I don't recall how many days we didn't have power and I wasn't in a city I was on Long Island. Northeast blackout of 2003 16 yeah so it was like actually under three days under two days under two days ah okay good yes I I'm it's literally the book about that okay good i'm gonna I'm gonna check it out as well even though I did live through the Northeast blackout of 2003 obviously it didn't kill humanity or even humanity in New York I mean it didn't but but but my concern is that as as that kind of dependence like because even right now people are still relatively okay with not with paper maps the what's that not with paper maps that not with paper maps but but we still have non-perishable food that was available if tomorrow in in six months we automate I don't know call it uh you know we create we we're trying to automate everything which is which is great but in a couple of years if our children don't have the option of taking a self-driving car or learning how to drive right and suddenly the ability to drive falls back to humans what do we do? I mean resilience I mean again that's from from the same I mean there is there are pieces of infrastructure that mostly just work and there is enough roughly speaking knowledge about how to do resilience people usually don't practice it but we on on on the podcast I used to run we interviewed some guy who was a CISO at a grocery distributor and he was very clear that if his cyber is short groceries don't go to stores any stores he wasn't in a store business he was a wholesale grocery distributor and roughly speaking he knew what to do and the nuclear power station operators roughly speaking know what to do of course there is Chernobyl of course there's Three Mile Island but it's two and and Fukushima but that's the the the sort of the nature event we sort of know how to resilience we just don't practice it in many cases. So I feel like if you suddenly lose driving capabilities in humans that seems really unlikely especially in the US I'm still thinking that if you if resilience would get built over that.

AI As Calculator Not Replacement

SPEAKER_03 28:46

So you you mentioned and and we're coming up on time but something that's interesting to me you mentioned you know that technology every time we put in a sufficiently advanced piece of technology we need people that are smart very smart if not smarter or or have the cap capability to maintain and keep that technology right so I I think about things that are like we're talking biomedicine we're talking about you know complicated things right would we be able to to figure out if there was a mistake somewhere in there before it was catastrophic you can talk about drug formulations you can talk about an AI designing a bridge that has a fatal flaw that didn't consider something bridge builder that's been doing it is super intelligent it it is more intelligent than humans.

SPEAKER_04 29:41

Again people are these guys are concerned about like super intelligent AI. So it would be reasonable to assume that the mistakes won't happen but if the AI is not super intelligent then obviously an intelligent human would catch it. I don't know so to me this circle of logic here really works. That if you're afraid of superintelligent AI then it won't make those mistakes but if you're not afraid of superintelligent AI then humans would figure out what to do. I don't think anybody's straight ChatGPT to bridge is a thing. And I don't think it's a thing in five years.

SPEAKER_03 30:10

I hope not I I really yeah I hope not but no I don't think I think the reliance look you we can we can treat it I think the the the I I think the lesson for me is and the way I couch it to people that ask right is I think we need to treat it less like a we suddenly need to give up all control and become mindless zombies and let the machine take over and more like we treated calculators. We went from an abacus to a cal to paper to a calculator. It's a massively enabling step it enables you to do calculations and and do things that would take you an error prone many more orders of magnitude time more and many more magnitude times error more but it makes it makes it simpler so that you can focus on the next task the next task right if if AI can come get to a get to a a way to combat let's use cancer because we talked about at the beginning of this right if if you can say hey AI system figure out how to treat make this molecule die but only that molecule and not any others okay right then I don't think we've we've obsoleted doctors and researchers and and and pharmacists and all that I think what we've done is given them superpowers. Yeah yeah for sure this is where I diverge from like the the megalomaniacs at the other at the two labs is their viewpoints were we're just gonna live we're not gonna need humans anymore we can go all live on the beach and the AI will serve us.

SPEAKER_04 31:48

Yeah but it the the longer term they may see the curve to it we don't I don't I again again I I don't want to like allude to I've seen things it's not really the point. The point is that I don't see that curve at all in the real world. I do see it perhaps in AI but if you have to have a clinical study of 10 years before you treat a million people there would be a clinical study of 10 years. So I'm sorry guys but there are enough medical disasters in the past when they didn't do enough testing so we learned that lesson we're gonna test so and it is based on physical time and physical time is a thing. So I think that if AI in one in one year AI becomes whatever number of times more powerful and you somebody and I'm I'm gonna go through completely hypothetical scenarios somebody says hey AI give me the formula for the medicine that treats all cancer in humans and just tell me how to make it and the whole thing goes and says okay here and the formula is made I don't think it's ever going to happen ever I don't know like but but I assume it does so what do you do you still need to go test on animals wait for the animals to something to happen you still have to understand the mechanics of how it works so it doesn't kill everybody including the animals and and so there would be physical time you give it to a monkey in a cage and you wait but waiting would be important you can ask AI hey AI can I not wait and the pro extreme pro AI fanatics say yes AI would tell you how to test it safely without waiting but would you actually do it no you would not do it. I would not do it you would not do it our boss's boss would not do it and the authorities quote unquote would not do it. So they would wait and then we would cure cancer and my AI would would would deliver on the promise but we would still have to wait dur some number of years for the physical time to pass.

SPEAKER_03 33:48

So one last thought who bent why far from cyber we went so far from cyber I'm not even sure that's real point okay I that's okay this is this is my podcast I'll do what I want there's that I have nobody telling me and people like this kind of stuff and I I think I think this is there's some broader questions that we answer here but or at least at least maybe not answer maybe at least we make people think uh let me ask what when on both ends of this who benefits on the on the doomer side who benefits and on the well I think on the on the you know overinflated expectations I mean I know who benefits right that's that's the labs and and their and their IPOs because you know $10 billion or trend three trillion dollars or $100 trillion dollar IPOs that kind of thing.

Who Benefits And How To Find Truth

SPEAKER_03 34:33

I know those guys benefit who benefits if if if if somebody convinces people that this is ultimately a bad thing or I'm gonna give you a non-answer failure.

SPEAKER_04 34:44

I'm gonna give you a non-answer and it's a really strange non-answer and I've used it on on X a couple of times inviting people to react to it and the non non-answer goes like this find out other people who are neither stupid nor corrupt who are saying this and figure out why they're saying this. Because it is very easy to find somebody who is an idiot who claims that AI doesn't matter or claims who AI would solve all the problems in a year. It is also very easy to find somebody who is claiming this because his pay depends on it. But if you look hard you will see people who are neither stupid nor corrupt and they're saying things. So you go talk to them in and it's a little bit like proving ESP if you go look at at 10 people who claim extrasensory perception probably nine of them would be would be the jokers but you would talk to the 10th. So here's the same argument I would I wouldn't answer who benefits but you would find people who don't obviously benefit and who are not idiots and figure out why they're saying certain things and you'd probably find a version of the truth from them. I why did Elon Musk claim that self-driving cars are coming in 2018? Because he wanted to make them and it profits him again very intelligent but obviously very motivated by success in this area. So there would be people in this domain who are not corrupt and who are not stupid. Listen to what they say and then you may we may know we you listen and I would listen to them too and we would know what's going on. But but the extreme skeptics assume that nobody's like that. They assume that everybody who's claiming it is either stupid or corrupt and they're wrong.

Movie Recommendation And Closing

SPEAKER_04 36:33

The skeptics are wrong.

SPEAKER_03 36:35

By the way the movie I was thinking of in case you guys are listening and you and you want a movie recommendation Chris Pratt because everything he does is magic and gold uh Chris Pratt in the movie Mercy if you haven't seen Mercy I highly recommend you go see it. One it's awesome two it's kind of a peek into the future I think of where where some things are starting to go. But anyway Anton that's a great sound bite you gave me to end on I'm gonna I might use that for a teaser. But uh thanks man it's been this has been fun hey good congratulations and and good luck on the new role and you know call me when you've taken over the world. Yes and also please invite me to future podcasts I I love it too much and I don't have my own podcast anymore so I'm now very podcast deprived okay well you we'll we'll get you we'll get you back on the show and and and we'll do more of these all right thanks for listening take care thanks for being on the show Anton folks thanks once again thank you for listening this has been hopefully entertaining and and thought provoking I know uh Anton's that we strayed pretty far afield from from the cyber aspect but that's okay I think I I do that on purpose it there's a lot to think about more than just the bits and bytes but thanks for listening we'll catch you guys another time another place on another down the security rabbit hole podcast this is Bella thanks for listening don't forget to leave my dad a review and share this with your friends bye