OpenAI’s $38B AWS Deal and Cybersecurity’s New Reality | TSG Ep. 961

Hey, everyone, I'll tell you that Open AI sure does. Get around, huh? Uh, you're watching Textron Gang.

Hi everyone. Happy Thursday, and welcome to our Thursday edition of Textron Gang. I'm Alan Shiel, and it's great to have you on here.

As usual, we got some interesting topics to explore with some very interesting people to explore them with. Let me introduce you to our interesting people today. Uh, our gang for today is Jeff Rech of the IDSA.

Jeff, welcome. Thank you. Our, our DevOps extraordinaire, ambassador Garima, Bob Powell.

Garima, good to see you joining us, I think for the second day in a row. Haley from Hudson, Ohio. Steven Foskett.

Steven, it's good to have you back on today as always. And of course, I would win every day if I could, but you don't, don't threaten me. I shouldn't say that, Steven.

Don't threaten me. I, we'd love to have you. And, and of course, speaking of being on every day, he is our every day chief content officer.

And, and Kos here with us. Mike Azar, gang member. Welcome.

It's great to have you on here. Um, so Mike is, I, I teased in the opening boy, this open AI is gaining some reputation for getting around, huh? Well, so the deal here is $38 billion partnership with AWS, and I guess OpenAI continues to need as much compute infrastructure as it could, but I feel like, you know, this is the latest in a long series of these announcements, and they seem to have an insatiable appetite for computes.

And now my, you know, and I don't mean to be overly snarky, but Alan, I'm kind of like, well, who doesn't have an open AI contract these days? And if you don't, what does that say about you? I, I agree with you.

So, I, I got two, I've got two things I, I want, you know, I, I spent a lot of time as a, a biz dev corp dev person, chief strategy officer, but you know, it was biz dev and corp dev. And, you know, we used to do back in the day, what we used to call a lot of Barney Barney deals, right? You know, everybody remembers Barney your favorite purple dinosaur.

Um, well, we get me. That's what the Barney thing's about. Yeah, right?

I mean, we would do a press release. I love you, you love me. We're a big happy family and, and all said, and good, but not a lot came out of it.

I'm beginning to get that Barney feeling with some of these open AI deals, because quite frankly, as much money as they're able to raise, and I, I, I've heard, you know, they were an open source, but they changed their kind of corporate structure now, and they are preparing for a, a, an IPO. Perhaps they can't do a for-profit thing. Um, and the IPO may value the company at upwards of a trillion dollars, they said.

But that's speculation when you look at the amount of money they've pledged on these compute deals, right? It, it, it's, it's, it's, you know, it's far exceeds their ability to pay as of today. Now, in a lot of these deals, they pledge, we're going to use you for compute this way for X dollars, and then you are gonna pledge x plus dollars back to us for using our services.

So it becomes a net, net zero, you know, uh, deal. And, and maybe that's the key to it. I, I don't know.

I, I look, I, I applaud them. It must be great to work in the biz strategic relationship department there because maybe second only to Nvidia. Um, but I applaud them for going out and, and making all these deals.

I, I just question whether they can actually be done. And, and from AWS's point of view, again, kudos to them. Look, they've got a nice stake in anthropic, but Google has a nice stake too.

And Anthropic just did a deal with Google, AWS hedging their bets here, bringing, bringing open AI into it. And of course, open AI has that whole Oracle relationship as well. You need a scorecard here.

These are like, these are like high school students pre-AIDS, you know, the way they're, they're swapping around here, like, free love brother. I don't know. What do you guys think?

I will pitch in and then maybe I'll let, uh, Steven and Jeff, uh, talk about other, uh, regulatory pressure and the risks which they foresee. But let's start from the basics. Uh, 38, uh, billion dollars deal with Amazon and OpenAI.

I happened to be, um, at a dinner table with a few folks from, uh, MAWS, and we were pretty excited to see the momentum, uh, because particularly, uh, because of the multi-cloud approach, which, uh, is in making here. But of course, uh, uh, digging into details about this multi-year agreement, which, uh, enables OpenAI to access hundreds and thousands of high performance Nvidia GPUs, uh, and 10 millions of CPUs hosted on a AWS cloud platform. What it means to the ecosystem is primarily, uh, for the UX US tech hub ecosystem.

I think most of the investment probably will foresee in some specific hubs. Uh, and, uh, this is again, you know, how much geographical distribution or widening regional economic disparity happens through this is, uh, we have to watch out for that. Um, the massive scaling opportunity, which OpenAI would get from this.

Um, again, uh, the scale of compute and the CPUs, which, uh, are in the promise, uh, and is in making. It's also like, uh, um, ensuring that we have next generation AI models, uh, and also ensuring performance enhancements and, uh, reach, reach to a certain extent. Um, multi-cloud AI ecosystem.

I mentioned that. Operational advantages specific to AWS customers, because I, of course, uh, people who are riding already riding the AWS wave will could benefit out of this. And validating that, uh, you know, I think AWS was lagging behind a little bit in the AI ecosystem.

So with this, they are solidifying their position, but there are more cons to this. And again, um, I, every, like, pros are very obvious, but cons, I think we'll have to call out cons as well. I feel that there will be a certain amount of competitive tensions between the cloud providers, uh, AWS Microsoft, uh, in particular, how complicated this relationship can become.

We will see in the next, uh, few months ongoing regulatory scrutiny on the big tech partnerships. And I will call out some of the recent past examples, which are very evident for everyone to notice. I mean, Google, um, and, um, Google, Google search man, monopoly is no new news to anyone here.

So that is something which properly also signifies that how, uh, like big tech regulatory scrutiny is kind of, uh, speeding up. But again, could we act, uh, in time with these kind of mega big deals happening? And, you know, we don't act within a certain timeframe.

And then the, we have gone past that stage, for example, what happens with Facebook, Instagram, uh, you know, all these kind of platform merging a single platform. I have to say that you, EU has done a great job because eu uh, regulatory and scrutiny has, uh, significant F presence. And when this, uh, acquisition of Ed o and Figma, um, was supposed to happen, I think they had a substantial regulatory pressure and they scrutinized it.

Similarly, I think, uh, I can give you some examples of, uh, acquisitions which happened in, uh, Canadian telecom ecosystem, which was also scrutinized to a certain extent. Why I'm, uh, advocating and I'm bringing, uh, these examples is that, you know, this big, big tech advantage, right? I mean, as you Ellen mentioned this multi, uh, year agreements.

I mean, it's not substantiated how, and how, how it is supposed to be rolled out, of course, which regions will benefiting out of it and all that. But to a certain extent, it also shows that, you know, we have policy making and regulatory gaps, and we have to scrutinize these mega deals in a more sensitive way so that we don't provide, uh, uh, adva like disadvantage, uh, from a small, medium business enterprise perspective. But I mean, I'll stop here and I'll probably, uh, take support from Jeff and Stephan for their views as well.

Well, Steven, let me ask you this question. To Garima's point, and maybe come to the heart of the matter, if OpenAI and all those big hyperscalers are basically sucking up all the GPU capacity, do I have any hope of getting access to GPUs? Or is, are we basically locking up the distribution of the GPUs and that's gonna become an issue?

Well, I, I think, uh, Garima is a hundred percent accurate in her assessment here. Um, this is all about locking up and, and monopolizing this market. And, and in fact, I would say that this whole announcement with open AI and AWS, um, remember the Stargate announcement with Oracle as well, and the fact that they've already been using Microsoft Azure as their primary hardware, um, this is a financial engineering move and market dominating move more than anything else.

It's all about them essentially staking not just a claim, but staking control of all of the GPU resources out there. Um, ultimately though, the question isn't, will open AI dominate the, uh, GPU hardware resources, the next gen resources in the, in the coming year? The answer is yes, uh, both through a combination of finance, uh, or, you know, financial engineering as well as engineering, engineering and, and of course partner work.

Um, but what will the result of that be? And I think maybe that goes to your point, Mike, that you know, okay, OpenAI ends up with essentially a monopoly on global GPU resources and a trillion dollars of, uh, fantastic IPO money. Then what?

Um, I personally believe that the era of ever am beginning, uh, models is at an end. And I think it's time for companies to focus on actually useful, um, applications of AI productive applications that, that generate revenue and, you know, deliver business results instead of continually trying to, you know, slurp up more data, build bigger supercomputer size, you know, not even supercomputer data center, multi-data center supercomputers to, to build an even bigger model. Um, I don't think it's gonna get to artificial general intelligence with the current, uh, transformer technology that we have.

And so I feel that this is entirely pointless. And ultimately what's gonna happen is we're gonna see practical applications that use smaller models, more distilled models that can run on much smaller hardware. And the whole, um, you know, supersize me approach that OpenAI is taking is I think a dead end.

Alan, you're a biz dev guy, and I always wanna know, like these kind of things, but where did this $38 billion number come from exactly? Did somebody stick their finger out the window and basically, you know, magically decide what the cost? Well, it, it is, it is probably tied into a certain amount of, of Rackspace compute, and it very well may be tied into what open AI is getting back from, from AWS as well.

But guys, let me, let me be clear here. First of all, when I, um, you know, I used to help my friend Brad Feld in the VC world, and one of the things they look at when, like, when you send in your pitch to a vc, they don't just say no automatically in spite of what you may think. They, they actually do look at it things.

They look at size of market, is this a big enough playground, right? Another thing they look at, do you have any proprietary technology? Well, this before open source was such a big thing, but do you have any proprietary technology?

A third thing they look at, though, barrier to entry. com revolution, everybody in the garage who could do a PowerPoint slide could be the next Al Vista or the next Netscape, or, you know, that's what Google was, right? And, and so it was, it was in every man revolution here in the AI era.

It's the big boys who are putting up barriers to entry that are just now, I don't think our normal monopoly, you know, antitrust measures figure in here. 'cause this isn't about mergers and acquisitions. This reminds me more, and, and forgive me, I'm gonna date myself, but the Hunt brothers tried to, to corral this world silver market, right?

And they came very close. It was, it was kind of out of that movie only in America where the Dukes tried to get the orange juice futures market, right? The hunt brothers in real life nearly cornered the worldwide silver market one time.

And it, as a result, there were all kinds of regulations that got put into place. So that doesn't happen again. But it's the same thing here.

You've got, in essence, these seven or eight companies, and we don't need to name 'em all, but we know who they are, right? That are, you know, represent 80% of the growth of the s and p 500 that are really locking up cornering the GPU market and the compute market. And Steven, in your point, I take a page out of Jurassic Park, right?

Nature will find a way, You know, Alan, I, I think nature will find a way, and I think there's another analogy to look at. There's a lot of analogy we could use. I believe this is a nuclear arms race in ai, in that the superpowers are going to, you know, take a look at the current nuclear situation here.

The world has enough nuclear weapons to destroy the world at least 80 times over, right? I believe that's a situation we have here. And the big superpowers are going to get even more simply as a deterrent, whether that's in a valid word here or not.

Uh, but I think that's what's gonna happen. And to Garima's point in other, uh, jurisdictions, it could be more effective. But I don't think you could point any example to me where a US regulation got in front of a problem ever.

So I I, I don't think that's gonna happen here. And, and the last one I'll use, 'cause I live in the world of analogies, is I believe the AI bubble, which we're in, um, I'm, prove me wrong, please, um, has a lot of hot air balloons and, and OpenAI is the biggest hot air balloon, and everyone wants to grab onto the hot air balloon as it goes up. If enough people join, it's gonna come down.

I'm not saying OpenAI is gonna crash and it's gonna go away, but, um, there's no way this can be sustained financially or technology. I think, uh, this will also, uh, invite some risk management into ai, which is a good news because, you know, people are also overseeing this, like strategic leadership is overseeing this, that the, this can all, like, this can be seen as soft acquisition. Come on.

You know, we, we already know that these kind of mega deals, um, you know, it's also how to go, uh, around the policymakers and regulatories, uh, to kind of ensure that they don't come in their way. But, uh, I can, uh, see that there will be a lot of risk management, uh, associated with a, like onboarding to cloud as well as, uh, using open AI models. I mean, again, for all sort of goodness what they have done to the market.

They are still training, uh, their models, uh, from open source, open data, right? And, you know, trust me or not, I think this will also bite back. Mm-hmm.

It might be time to reissue that classic book. What color is your parachute? 'cause when that bubble pops, man, you gotta figure out what you're gonna do.

Well, if I can add one more thing into this conversation, uh, just this morning as we're taping this, the news came out that, uh, the hedge fund investor who inspired the film, the Big Short, has just put a billion dollar short against Nvidia and Palantir betting that the AI bubble is going to burst. Uh, we are not a, a financial podcast. Uh, we're, you know, I'm not a good source of financial news, but I'm looking at that thinking, boy, I wonder if that guy knows something.

Maybe let me inforce though, we're not making to giving you any investment advice. This is, yeah, we have enough problems just talking about the technology, but, uh, it, it is, But there is research being done that, you know, and it's years away from any fruition, and it may never happen. But it's basically arguing that the entire, uh, model, the way we build them is, uh, based on FORTRAN and all those core technologies.

And there's another way of thinking about this that's gonna be a little more reliable and a little more declarative using a different set of math algorithms and models. So people are doing research in this space. So, you know, the current AI models may not be the end all be all here, but you know, we're, I think that's where we are for the next five years at least.

No. So here, here's where it is. And it goes back to what I said about Jurassic Park and nature finding a way, right?

Right now, we, what we have is what I call Cadillac ai, but we may not need a Cadillac for every use of ai. We may need a Chevy and Ford and a Pontiac for those who remember those great Pontiac cars, right? And, and there will be Pontiac brothers out there who develop a, a Pontiac AI and afford ai, and someone's gonna make a better assembly line to assemble it.

And so Cadillac won't be the only brand of AI out there. And it may be just, as you know, during the seventies and the gas crisis, the rest of the world look at, looked at those big boats that Cadillac was producing, the 76 El Dorado convertible, one of my favorite cars of all time. Um, and looked at them as just insane.

Given the world of, of how it had flipped. We may see the same thing here. We may see the same thing here.

Very true. All right, let's take a break. We're gonna come back and, and let's go on to our next Block, which is driving under the AI influence.

I didn't realize it was a crime. You're watching text Young, You've earned it. The spotlight, the responsibility, the weight of teams, companies, and entire industries fall on your shoulders.

Lives depend on your decisions. Your home life included that work. You are protected physically and digitally.

Nothing gets through your team without a fight. But in a globally connected world, everyone sees you, including those who mean to cause you and your organization harm. And now home your sanctuary attackers see an opportunity.

Your digital front door is wide open. And what compromises your home can breach your boardroom. Because the devil's greatest trick isn't targeting your workplace firewall.

It's convincing you that your personal life isn't at risk. Black clerk, digital executive protection, defending the new attack surface your personal life. Hey folks, we're back and talking about a little bit, uh, more ai but in a different context.

Uh, the folks over at Tesla are talking about putting a, a new GPU inside the cars that's a little more powerful. And that got everybody talking about, well, just what are the safety concerns around these issues? And Steven, I know you follow it more closely than I have been, but what is your take here?

What's going on? 'cause we have been putting GPUs in these cars, but this next generation is a whole other thing. And well, I don't know, do we really wanna hallucinate while we're driving?

Uh, certainly not. Um, so I, I first off, uh, you may not believe this in five minutes, but I am a Tesla owner and I think it's a pretty good car. Um, setting that aside, this is really, really dumb story.

Um, so let's dive in so many. Um, but technically speaking, um, Tesla has, uh, hardware in it to do the self-driving features. Um, and you should put self-driving in scare quotes because of course it's not really self-driving.

It's, uh, driver assistance. And, um, you know, and for, for, from the very first car, they've, they've had now, uh, four generations of self-driving hardware in the cars. They call it AI five.

Now that's the new one that they've introduced. But before that it was called Hardware one, hardware two, et cetera. Uh, the current cars have hardware four, which came out in 2023.

Um, in that time, Tesla has ramped up the amount of processing power along with the sensors that are used to do any kind of driver assistance features. So, for example, uh, the original ones relied on Intel's mobile I platform, uh, which used, um, I believe, uh, Intel Atom CPU course. Uh, then they moved to an Nvidia AI computer.

Uh, then they moved to, uh, a Tesla designed, sort of designed chip that uses arm, uh, a 72 cores, I believe. Uh, now they've got a new platform, you know, that uses, um, arm cores and is another Tesla design. Uh, and the new one they're claiming is going to be much, much more powerful.

Uh, one of the things that they're saying is that it's gonna have, um, uh, I don't remember the number, like 40 times as much memory, which wouldn't be a big surprise or nine times more memory because the current one only has 16 gigs of memory. And as I said, it's basically a 72 cores with 16 gigs of memory. It takes 160 watts of power consumption.

This is not a powerhouse chip. This is a decent, strong chip for automotive applications. The new one they're saying is gonna be like an 800 watt chip with 144 gigs of memory and, you know, all sorts of performance and everything.

And the problem with that is, who cares? The hardware four actually does a pretty good job of doing driver assistance features, despite what some people might say. And despite the crazy man in charge of the company, um, it honestly is a pretty good car with pretty good features.

The problem is that even at 160 watts of power consumption, it drains your battery. Uh, so if you leave the car locked in a airport parking garage for a while, uh, you're gonna go from 60% to 40% battery pretty quickly. And if it's wintertime, you might find yourself at 20% battery and the thing will automatically shut itself off.

Um, that's because frankly, running a little computer, even 160 watt computer, um, and that's maximum TDPI doubt it uses quite that much, but that's gonna take some power. So, Elon, uh, being Elon is gonna make some crazy statements here to justify what they're doing. And the crazy statements around the AI five platform, uh, which again, is just hardware five, that's just the next generation chip that they're gonna use in the Teslas, is that these things are gonna be all sorts of useful.

Uh, he's saying that these chips are gonna be used outside of Tesla cars that maybe they'll, uh, you know, be used in robots or flying cars. I know, uh, that sounds crazy. It's crazy.

Um, also, uh, that he suggested maybe they would use them in, uh, satellites. 'cause starlink, you know, they're sending satellites up there. Now, we talked on the Tuesday episode about how crazy it is to try to dissipate heat in space and have an AI data center in space.

Well, uh, it's not any less crazy if it comes from Elon's mouth. Um, so all of this is to say that the promises they're making are really not likely to happen. But the most crazy slash stupid promise is one of the things that he announced, which is that these chips are so powerful that they will allow Tesla to use, uh, idle parked cars as an AI supercomputer, a distributed AI supercomputer.

Um, that makes no sense whatsoever. So let's think about it. So do you want your car to be sucking down 800 watts of power from your a hundred kilowatt hour battery pack to do AI processing with a slow network connection?

And 144 gigs is more memory, but not crazy amounts of memory and not a lot of storage. I mean, there's no practical world in which it makes sense to use Tesla, you know, of 50, 60, 70 plus thousand dollars cars as, uh, AI supercomputing nodes when you could do exactly the same thing at a thousand dollars a pop in a data center somewhere that has a, a stable supply of power and a stable network connection. This is just another promise, like Elon Musk saying in 2019, that Tesla buyers would have an appreciating asset because full self-driving would allow them to lease their car out as a robo taxii when they're not using it.

That was pretty crazy. But the idea of using the fleet of Tesla cars as an AI supercomputer, it makes no sense. Technically, it makes no sense financially, it makes no sense for owners.

And this is the dumbest thing I've heard in at least 36 hours. And that's saying a lot for 2025. So wait a minute.

Let's say I did lease the GPU in my car back to them. What am I getting? Like, you know, is this gonna be like, you know, 2 cents a minute, or how does that kind of pricing, what, what would I actually get for my trouble?

Well, see, that's the problem. Exactly. What would you get for your trouble?

So even if these chips are nine times faster than the fine, but not really impressive hardware for chips that you've got in your car already, that's nothing compared to a modern, you know, H 100. I mean, you would need literally, I I, I don't have the numbers here, but you would need, let's say a whole bunch of cars to theoretically even the, the, the, the theoretically comp computational power of a single, you know, Nvidia, uh, device. And yet you'd also need bandwidth for data and storage.

And none of that would make any kind of financial sense. The best possible scenario would be that you'd get pennies. But the truth is you'd get negative pennies because the other thing that's happening here is you're wearing out your battery and you're wearing out your hardware already.

We've seen that, um, you know, GPUs used for AI processing have an actual useful lifespan, more like 18 months than the, you know, six, seven years that you might get out of a conventional computer. And that's because they're run hard and hot all the time. Well, that's, do you want that to happen to your car?

Do you want your car's computer to have an 18 month lifespan? Uh, do you want your battery to have an 18 month lifespan? None of this makes sense, and there's no financial model that would make it make sense for owners.

Mm-hmm. So, I don't know, agreement, is it me or does it feel like sometimes Elon Elon's just making these grandiose statements and then maybe some engineers will attempt to execute it, but by the time, you know, it's supposed to come to fruition, nobody remembers what he said five years ago and nobody calls him on it. So I think, uh, there is, uh, you know, pros and cons of everything.

And again, uh, with Aon, I, there are some things which I think it's also good because of course, uh, Steve, you are killing the story here, but I wanna give some credit to the disruptions in the market he has, uh, uh, done so far. I mean, starlink, for example, is, uh, a great initiative of a non-terrestrial kind of, you know, network which we can create for countries like Canada. You know, there is also like substantial amount of change, which he brought in through like the work which she's doing with the Tesla software, for example, particularly the user plane control plane, uh, you know, ecosystem.

I would agree to what Steve has been saying is that thermal, uh, efficiency and energy efficiency is the one which we all watch out for because, uh, I live in Canada. I live in a colder climate, so I would like to have, uh, you know, a car which is also reliable in that climatic ecosystem, the edge, uh, you know, cases which we have, like the urban complexity, for example, or the climate complexity, which we face in Canada. And if you go to Asian demography, it is even more complex because roads are subject to any kind of disruption, right?

So those are the kind of things which I would like to see if, uh, you know, from a user brain control plane perspective, we can make more enhancements through these kind of efficient, uh, chips, which is this, um, a, I think it's a AI five, which, uh, is the newest one. Maybe that's something which I would watch out for. And I'm, I am positively looking forward for the software updates in this.

Now, uh, he has been claiming also that the, uh, this will be used by Optimus, which is this robotic arm, and he will integrate, uh, a ai, uh, with cars and robots. I think it'll have some impact on some kind of a business model generated out of it, but we probably will, we'll have to take, uh, Steve's word here and see and be carefully watching out for this, uh, change. Steven, are you also staying in, correct me if I'm wrong, but you know, one of the reasons I stay away from electric cars is, you know, I feel like you drive for two hours and you're charge for two hours, and now maybe we're gonna drive for half an hour and charge for three hours.

I don't know. Well, let's just say that that's not an accurate impression of electric cars. Um, I'll start by that.

Uh, as someone who owns only electric cars, uh, one Tesla and one BMW electric car, I'll say that, um, uh, they're great and you don't drive for two hours and charge for two hours, so, we'll, we'll, we'll start with that. Um, But in Elon's crazy world where you're sucking down 800 wat watts of power continually and continually transform transferring potentially gigabytes or even terabytes of data over the cellular network in order to do, to handle that AI processing. Yeah, yeah.

You're looking at the 50 mile range on your brand new model. S you know, it, it's, it that, again, this is why this doesn't make sense. Um, you know, the last thing you wanna do is attach, um, you know, try to run things off of batteries.

Remember Seti at home, and, um, you know, I loved Seti at home. It was so cool. It was, you know, you could do like, you know, and then they did protein folding at home.

And, uh, so when you weren't at your laptop or your desktop or whatever, it would do, you know, computations, distributed calculations and computations. And it was super fun. Um, it was not financially viable for you, the user.

It was gratifying you were contributing to science contributing something that you cared about. Um, that's the only scenario I can see where something like this would make even a lick of sense. So Steven, so both Seti and the protein folding was about using that utilizing unused capacity or underutilized capacity to, to tackle these real world problems.

The fact of the matter, that was also the model behind Uber, right? Most people who owned a car didn't use it all the time, for most of the day, it's sad, idle. And if we could put those cars into use, people could actually do, make money from the better utilization of their vehicle.

Yep. And remember Elon promised that in 2019, he said, and I quote, Teslas will no longer be a depreciating asset. They will be an appreciating asset because you'll be able to use them as part of the robax fleet using full self-driving.

I'm sorry, that will not happen to, you know, uh, Garima's point, uh, no one who knows anything about this stuff will say that there's a car that can drive a hundred percent autonomous in a hundred percent of conditions at a hundred percent of locations. It just isn't possible. And so Elon's promise then was, was nonsense.

And this suggestion now is just complete nonsense as well. I'll also add one more thing to this, uh, to substantiate what Steve is saying, that the business models are weaker, uh, not only in this case, but also if you see non-terrestrial network or satellite network, this is like, uh, adding cost to the business, adding cost to the user, I think it makes sense for some particular use cases, right? And it's the same thing with self-driving cars and, you know, inference engines and the cars from control plane, plane user, plane management, plane level.

But when it comes to, let's say, non-terrestrial network, it also has a high cost to the business. And it's not non-substantive how this would play around in the longer run. Uh, if, if, I may agree, but first of all, I have to say, you said known remembers what Elon said five years ago.

Steven does. I Was that obviously? Yes.

Yeah. Um, uh, but, and you may remember I've said in the past, I have a physics background. Whenever you use a lot of energy, often it, it manifests itself as heat.

We don't know the conditions that cars are gonna be in or the environmental conditions the cars are gonna be in when they're utilized this way. And we have, if you haven't seen yet, a car fire for an electric car is horrendous. It's very destructive and hard to put out.

And we haven't even talked about what the extra high level utilization it's gonna create with this new heat output, especially if a car car's in an ENC closed garage with no ventilation. I mean, there's a lot of environmental issues that you cannot control when you do this autonomous, let's just suck up some computing power. Yeah, none of it makes sense, does it?

Mm-hmm. And by the way, most the Uber drivers that I know are not using their family vehicles to drive around town. They have actually become professionals and they have a car specifically for that purpose because the family car would just can't take that kind of wear and tear.

How many Uber drivers do you know, Mike? Oh, well, I seem to count. The same guys seem to drive around my town.

So it's about 10 of them, and it's usually one of those 10 guys. But look, everything you all said is you like the way I used you all, you'll gotta love it when, when Alan Hummel throws a Y out there. Everything you all said is absolutely true, but I'll say this, our children will not, or grandchildren as the case may be, will not know a time when people drove CARSs.

I firmly believe that. I agree. This may not be the pack.

They live in a time where somebody else is driving their car. I I do think we will lick this problem. It may not be Elon who solves it, and it, it won't be the Tesla Gen five or whatever he calls it.

Um, but I, I do firmly believe we are destined for a future of autonomous vehicles, both trucks, cars, vans, trains, planes and automobiles. Alan, I agree. It's gonna take infrastructure to get there.

Not cars only. Where's that, where's that damn flying car you promised me 10 years ago too, right? Go.

We get astro on it. Anyway. Hey, let's take a break.

We're going to come back here and get on our sea block. We'll get, and we'll get back to something a little more mundane like ransomware. You're Watching Textron Gang, Discover Textron Group, the epicenter of tech innovation.

We are your go-to for reaching it, leaders and practitioners worldwide. Our secret impactful content that sparks awareness, engagement, and top quality leads with us. You'll access editorial websites, streaming videos, virtual events, custom content analyst research, and more.

Join our satisfied clients. Let's revolutionize your tech journey. Contact us today and tell your story to the world in the most powerful way with Techron Group.

Hey folks, we're back in Alan's point. We're talking about an article up on Security Boulevard that says that there's been an uptick in ransomware incidents. And this comes after several months of maybe the industry collectively patting itself on the back saying we've been driving these number of ransomware incidents down.

But Jeff seems like, you know, every time we think we whack this mold, the tactics and techniques change and we're back to square one. Is that what's going on? Um, I think it's exactly what's going on, Mike.

I think you have a good grasp of the situation. You know, Terry wrote a couple of good articles on this on text wrong, and the challenge that we have is we're trying to solve ransomware with technology. And first of all, the weak point for ransomware has always been people always, it always takes someone clicking something from ransomware to work.

It rarely happens completely unattended. I know there are cases, but that's rare. So if you take a look at what have we upgraded in humans, it's nothing, uh, yeah, we've made them aware.

We, we put out PSAs, we, we have conferences saying, by the way, here's what you should do. Never click a link. But people still do it.

It's human nature. And that's something we haven't changed. I'm not sure we're going to necessarily add that to the fact that ransomware is a very profitable business when you're successful at it for two reasons.

One, you can get a lot of money without having any physical, um, harm, um, risk to yourself. I, I'm not encouraging people to, to be ransomware extras, by the way. Um, that's the first one.

And the second one is that, um, you're always gonna find someone that can do it. And now, because of those two conditions, the business of, I can sell you any parts of ransomware, uh, toolkits that you need at a relatively low cost on the dark web, it's not difficult, especially if it's state sponsored. But it's not difficult for really anyone, um, to be able to say, I wanna go buy a, a list of, um, IP addresses.

I wanna buy a list of email addresses to Target. I wanna buy a list of, um, or I wanna buy a a, an engine that can run when people do click so that I can get all of their cryptocurrency. It's not hard to do that.

I will offer, as Terry Robinson said, ransomware never really went away. I think two things have occurred that make us think we've done better. One other things have come to the front, so there's a little less reporting of ransomware that may be hard to believe, but I, I've seen instances where it's true, where I've talked to organizations that said we had it, we dealt with it, it's over o Okay, um, we'll see if it's over.

So that's one, uh, that other thing to overtaking it. And the lower reporting, I think, um, contributes. But the main is that there are dark web based businesses that offer these tools that the bad guys are gonna buy at a relatively low cost and they can shotgun it.

If they hit a hundred thousand targets, if they go for a hundred thousand targets and three hit, they win. So I really think that's where we are. So, so you're Saying that, you know, there are people who are getting mugged out there and because there's not gonna be any recourse, they don't bother reporting it to the police.

Gee, that never happens. Nope. Never.

Even when there is recourse, they don't necessarily, for a whole bunch of reasons, they don't want the attention, Especially if you're an immigrant today in America or something like that. But you know, this, this whole thing here, and Terry did do a nice job on these, this reminds me a little bit of the MIT versus Wharton report, dueling reports on the, uh, success of ai, right? How can they both be, right?

But yet they are, it's kind of caught them. Um, I, I think with ransomware, we're, we're seeing a similar thing. I think the amount of ransomware attacks continues to tick up, continues at a good steady clip, if not increasing.

I think where we're seeing a decrease is the effectiveness of these ransomware attacks. I think, you know, and it was a report we did here just a couple days ago where we saw the, the amount of ransomware attacks that resulted in a payment of ransom went from 28% to 23%, which is a sizable decrease. Right?

Um, and why aren't people paying? Couple things. Number one, I think companies are wise now that they've gotta have an impeccable backup solution that if they do get ransoms, they can restore without having to pay the, the, the ransom money.

So they, it kind of renders them a little bit of mute. Secondly, advent of cyber insurance and other ransomware professional negotiators, if you will have, um, really had a huge impact on how much we pay when we pay, who do we pay it to? Right?

It, you know, we, we've seen the government in years past come out and say, it's illegal to pay the ransom. Don't pay the ransom. It's negotiating with terrorists, right?

Then we sort of go the other way. Well, maybe you should pay the ransom sometimes. But, you know, we've developed sort of a professional cadre now of our level of people who specialize and they'll say, okay, what kinda ransom malware are we talking about?

Who's the gang behind it? Are they honorable thieves? You know, and, and stuff like this?

And, and so we, we have gotten better at dealing with ransomware and, and you know, and we have gotten better at spotting phishing and stuff like that. But what we also have to remember is what are the vectors of where ransomware gets in? It's still predominantly a phishing attack.

And when we talk about phishing attack, a good friend, it'll rear its ugly head here again, AI has had a huge effect on the quality of phishing, spear phishing and just general cast a wide net phishing that we get in our email boxes today. We've all seen it, right? You gotta really look at these emails there.

I think though, when I talk to people, they can recover their data, but it's not instantaneous by any shot. So basically they now do the math and says, it's gonna take us two days, maybe three days to get this back going again. And then the CFO looks at it and says, well, three days worth of revenue is still lower than the ransoms that we're paying the ransom.

So Steven, I don't know what you're seeing, but Yeah, I, I, I definitely am hearing, I I'm not seeing it because I don't have direct involvement with people paying ransom to ransomware attacks, but I, I'm definitely hearing about people paying the ransom and I'm hearing incredibly about ransomware gangs actually living up to their side of that bargain and unencrypting data, which is just absolutely nuts. Uh, the other thing that's happening too is that there's a lot of, not just ransomware, but, um, for encryption, but ransomware for release, you know, data release, essentially, I will release this thing. Extortion.

Blackmail, right? Yeah. But one thing that I'm definitely seeing, as you mentioned, is that the, the sophistication of these systems has gone way up.

So we talk in the article here about as a service, you know, you can literally just say, Hey, I wish I could be a ransomware guy or gal, and guess what? You too can do ransomware. All you have to do is sign up for one of these services and you can ransomware all you like, which is wild.

Uh, the other thing, as you mentioned is ai. I am the recipient as a president of a company. I'm the recipient of a lot of ransom or a lot of, uh, phishing email, and I've gotten pretty good at it.

Um, but there's a, it's getting so much better and so much more clever. It used to be that you could just spot 'em instantly. To the extent that I was like, why doesn't Google just delete this garbage?

It's obvious that I'm not, and our, you know, Nigerian prince doesn't have a, a bunch of money for me, but nowadays I'm getting a lot more, uh, carefully targeted things, um, things that, uh, are supposed to emulate the voice of other people in the company. Um, you know, in fact, the funny thing is, I actually get some email for other addresses within my own company, and often that email comes from me and it is intact. And somebody emulating my voice in talking to HR or it and saying, you know, Hey, this is Steven, I want you to do this or this or that, which is hilarious 'cause it comes to me.

But the point is, that's the sort of thing that we're facing now when it comes to phishing. We're facing very clever things the other day. No kidding.

I got a voicemail in my own voice asking me to change, uh, payment for a client. Um, that's wild. And that's the sort of, you know, to, to Jeff's point, you know, the people are always the, the door that these things go through, whether it's ransomware or extortion or phishing or whatever it is, but, um, it's hard to be one of them people when you're getting this kind of attack.

Mm-hmm. That means if you don't know who it is, don't answer the message. Right.

Wait. But you may. But to Steven's point, it's his voice.

How do you, you know, how do you not know who it, Yeah, I mean, I almost bought myself some Best Buy gift cards, to be honest. And I, and I feel compelled to make one other point. It's not just phishing.

Uh, it's how this ransomware is getting into the systems. We are seeing an uptick of, let's call it direct malware injection, whether it AI poisoning, you know, garden variety o or top 10 type type of stuff. It, it's not, phishing is still the predominant ingress point, but there, there are others.

And it, you know, just goes to the point where security, it's real. We, we, we gotta be vigilant. You know, Alan, I, you know, I'm glad you brought that up because it's not just phishing.

Um, there are new AI and others, there's more scale now, but I am, whether it's proud or not to say that I was involved in what I think was the first public ransomware. Um, it was in 2003. I worked at a hosting company.

I had just started, and on day two I noticed what was going on and we worked for five months to extricate ourselves from it, which we did successfully. But it boiled down to we were hosting a million websites and over half of them were taken over. And it was very simple.

We got a call from an Eastern European entity saying, Hey, we're a security consulting company. We see you have a lot of problems. If you pay us a Lengthly fee, I'm sure we could take care of them.

Oh, really? Um, so, and that was in 2003. So when you get down to it, in my opinion, it doesn't feel very different.

It's just scale and spew. Now Abso but, you know, no, there is a difference. 'cause back then they wouldn't stand up and say, we've Ransome Jew and we're gonna send you the, you know, you need ear of the victim.

No, we really got 'em or not. Right? Back then they posed as a security company that was gonna help cure you.

Now they don't have any, you know, gumption about, they're saying, Hey, we're bad guys. Pay us the money. Or, or you're done.

Right? It, it's, you know, it's, it's, How is, how is that any different than it would be a shame if anything bad happened to your nice restaurant, Right? That that's right.

You know, if you elect this guy, mayor, we may not, uh, be able to send you any federal money. Yeah, there's something about that too. Hey, green Mind, don't we ask one question about this?

You know, I gotta say about this whole as a service thing, these people seem to really have adopted some really awesome best DevOps practices for building out this software. So, you know, you gotta hand it to 'em. They read the manual.

Exactly. So, and I, uh, I was also reading about this a couple of days back, like warm GPD as a service. And you know, when these kind of softwares, uh, hit the market, you, you have to be very careful.

They're not targeting enterprise customers. They're targeting, you know, people, you know, and this is like a small ransomware or small, you know, things which you do in the ecosystem, and it is like, uh, spreading the ransomware at scale, right? So this is like a democratization of ransomware through a software as a service model.

And this is a shift which is happening to the point what Alan mentioned, that, you know, enterprise are more cautious. They have these legal, uh, advocates and they have insurers, but like the common people, you know, these 8 billion people, we don't have insurers for these kind of, uh, you know, malicious activities. So I think we need to be careful here and we need to educate as we go along the people in the ecosystem not to fall track of this.

I, I, I couldn't, I couldn't agree with that more. And I have to say one thing on this. Cyber crime as a service, if you take a look, and there is some research on this from the dark web, not only do they get DevOps, not only do they get presentation, they get customer service.

The these guys get rated and they want their ratings to be high, so they get more customers. So if you have any issues, they'll support you or give you a refund, no questions asked. They're kind of like the Costco of, of the cyber world.

Um, and they get it. And, and, you know, on, on the, on the normal side of the web, that really doesn't happen very much. But Na, dark Web Boy, they own it.

They're, they're gonna make you happy and want you to come back. Uh, yesterday, actually, to close this point off, I was discussing this with John Willis, the same point. Like, you know, these, uh, malicious software as a service, think about your teens, you know, teenagers in your home.

You know, they can easily fall trap off these kind of things because, you know, if you don't educate your kids and you know the ecosystem around you, it's that kind of audience they have or they, that kind of customers they have. Absolutely. Hey guys, we're over time though.

I, I'm, I'm afraid I gotta end this one. Um, Reem or Jeff? Steven.

Mike, thanks for joining us. Thank you for joining us here on Dextro Gang. We hope you've enjoyed it.

I believe we, Steve, don't we have Steven, Don, we have a, a field day on today, don't we? We do actually, uh, you know, we, if you, uh, watch, uh, basically right after this, uh, we're gonna be going live with networking Field day, uh, day two, uh, we're gonna be hearing from, uh, a big company you may have heard of called Cisco today. Excellent.

So stay tuned for that on your favorite tech strong TV network here. Whether you're watching it on LinkedIn or Facebook or X or YouTube or, or if you're not watching the Stream, you can still check out Tech Field Day on their Tech Field Day YouTube channel, or on the, uh, tech Strong OTT app, which is on Android and iOS and Amazon and Apple TV and Roku and all that good stuff. But in any event, we will be back tomorrow for our Friday show, DGIF and, uh, we'll have more good stuff then.

But until then, this Alan Ware app up.