Prediction Markets, AI Drones and the Future of Cloud Computing

Hey, everybody. Welcome to the Techstrong gang. We're gonna have a little chat today about all kinds of fascinating things, starting with these predictions markets, but let me introduce our guest today.

Steven Foskett is here as usual. Steven, how you doing? It is great to be here, and it's, warming up here in Ohio, so I appreciate that.

Exactly. There you go. Also joining us is Garima Bajpai, who's also become a regular these days.

Garima, how you doing? Good. I'm in the middle of spring and sum- summer and winter, as I'm in Ottawa, right?

Yes. That's how the weather plays around here. We're all caught in the middle of everything here, as they say.

And then finally, we have a new member of the gang today, Blake Hendricks. Blake is joining us with, with some, knowledge of what goes on in the land of the cloud and AI and all kinds of fun stuff. Blake, welcome to the show.

Yeah, thank you for the warm welcome. Super blessed to be here. All right.

Cool. Well, let's just jump in. It's Monday, and there's all kinds of fascinating things that are gonna happen this week, but one of the weirder things that has started to occur with IT is the emergence of these things called predictions markets, and people are...

Well, you know, people have been betting on weird stuff since time began, right? I mean, how many times have folks been sitting in a bar somewhere betting somebody that something would happen and putting money down? But now we've kind of taken it to another level, and there's platforms, and there's whole back-end IT systems that people are using to bet on these things, and it's...

Basically, it's FanDuel has come to everyday life. And I guess there should be some rules here, and there's a couple. Apparently, you're not allowed to actually bet on somebody dying because, well, then somebody might just go out and kill somebody, and apparently there were some bets made about that as it relates to the, former head of, the theocracy in Iran, but nobody got paid on those bets.

But I guess that's relevant and also good. But Blake, what's your take on what's going on here? Is everything about our lives gonna be up for a bet?

Um, can you elaborate on that a little bit? And- Well, I'm just wondering, is this the way society is, and we should just get used to it? Or is there something fundamentally changing as we start to think about the way IT is evolved?

I, I, I mean, so I think there's always been predictions, in, in the market, right? Um, I think it's kind of came from a pinnacle of, you know, when it was used to be just ones and zeros, and people were placing things and, and kind of, contextually drawing information from what was around them without technology and, and still doing predictions. But I think it's kind of evolved now into, you know, these, these subsets of AI, right?

You started with, like, machine learning, and you had... You can predict trend analysis. You could do machine learning al- algorithms like gradient analysis, all these different algorithms that came out to help people, you know, sell businesses, help enterprise and IT infrastructures become more resilient and also pr- provide enterprise-grade, performance.

And, you know, it kind of took a turn lately, you know, when ChatGPT and OpenAI came out, when, you know, the COVID virus happened. Um, and it's so interesting, the timing, right? Like, everything was silent, and then, you know, these, these, you know, these mad, interrupted technologies emerged, and, no one kn- knew what to do with them.

Um, you know, it was almost like Bitcoin to me at first. Um, and, you know, I got my hands on it, and they were really hallucinative at first. This, this AI technology that people harnessed and, and, and, and basically advertised as something that was better than what it used to be, it is really great now, wasn't what it used to be.

So, you know, predictive analysis wasn't as sharp as it used to be. Um, and with the, the help of, you know, these, these large language models and, you know, robo- robust, you know, research capabilities that they have to, to build predictive analysis markets has, you know, created a large generation and market of, of, of, of, of adoption. You know, like you said, it's in, it's in certain departments like, you know, war.

Um, you know, it's in enterprises. It's in people's personal lives. Um, people are using it to predict, you know, based off the weather if, you know, what should I wear?

Um, you know, it's being used in a vast amount of ways. But at the end of the day, I think that our lives aren't really predictive on a number, right? Because now that everyone knows that things are so predictive, it's, it's almost like people will influence against it, right?

'Cause people know nowadays what the outcome is, you know, whether that's you're, you're betting or in the stock market, and gambling and, like you said, you know, you can, you can predict the weather or certain outcomes of events. Um, you know, so with that knowledge in hand, it- it's kind of turned the scales of justice, right? Like, in terms of weighing what can actually happen, now there's things that have predictive outcomes, but also people will try to engage against it to, to kind of control what's happening.

So- Yeah. I mean- I mean, even the article, the article opens up with a physics concept Yeah. Let me, let me, let me bring Alan in here for a minute, who authored that article.

Um, Alan, good to see you, but I guess the, the premise of- Nice to... Sorry I'm late, guys. Yeah.

The premise of the conversation- Yeah. Hi, guys ... though is that with predictions markets, I mean, we, we've been able to bet on everything and anything since time began, but the tools we have at our disposal for analyzing the likelihood of something, whether it's using predictive analytics tools or machine learning algorithms or whatever it is, is much better than ever.

And so the question is, is can we... You know, is there such a thing as gambling anymore if we can predict all these things, and are we betting on things that maybe we shouldn't be? Well, what, what you're really saying is we can predict better than we could before.

But- Mm ... you know, but as the saying goes, they still let the horses run, or that's why they let the horses run, right? No matter how good your predictions are.

But the point of my article was almost like a, a, almost a quantum kind of argument, where the fact that you're even making these predictions are somehow influencing the outcomes, right? Uh, you know, it- it's, it's, it's metaphysical and Eastern and whatever you wanna call it, but, you know, i- it's one thing if you're playing numbers and you just pick a number, right? It's another th- what we're doing now in the...

And, and we're, and we're not talking about who wins a soccer game or a baseball game or a football game. We're talking about what are the likelihood of, of something getting destroyed in a war, someone getting killed, or, you know, like, with all due respect, a whole different level of things we're betting on and predicting on. And, and in doing so, are we in essence kind of gaming the system or influencing this, you know, influencing it and, and in essence steering history, if you will?

Mm-hmm. You know, steering, steering. Stephen, what's your, what's your take here?

Because to Alan's point, the more people bet on something, the more the odds change, right? Or at least they should theoretically in a predictions market. And so does that influence the ultimate outcome, do you think, as more people start to lean one way or the other or, or have some butterfly effect that we can't measure?

What do you think? Yeah, I, I would say so. I mean, I think it's pretty obvious that, that the predictions markets are, being used as a source of data for, many, you know...

I, I mean, certainly in, in sports, you know, sports booking is frickin' everywhere and it's really annoying, to me, to see, odds, on everything all the time everywhere. But, you know, now that it's become part of the discourse on politics, for example, and on technology, on war, it is, it absolutely does risk, changing people's perspectives. Um, you know, if, if you, somebody said the other day that, you know, the predictions markets said that the, you know, the Democrats had a good chance of taking, you know, the House and the Senate.

Well, the, the predictions markets, you know, that's gonna change, that's gonna change people's perception. It's gonna change whether people show up. It's gonna change whether people push harder or softer.

But I think there's another insidious factor here too in, in these predictions markets in that it gamifies everything. And frankly, whether there's going to be a nuclear war in Iran is not something that I wanna see gamified and that I wanna see people gambling on. Like you mentioned as well, you know, y- you know, I don't want there to be a gamification of people, predicting whether somebody's going to be murdered.

I don't wanna see a gamification of a lot of things that are happening in this world, because I feel like it is, it really runs counter to the way that these discussions should happen. We should, we should be talking about things on the merits, not just looking at them in terms of financial gain and loss. And similarly, you know, some things are just universally negative outcomes, and we shouldn't be cheering for a negative outcome because we're gonna make some money off of it.

Yeah. Yeah, that's- And this is again to, to your point, Alan, I, I think the whole issue here is also about passive predictions and how passive predictors become active stakeholders, and the blurring line between how forecasting becomes influencing outcome, right? And that also gives rise to these fakes, you know, superficial news.

These are other negative outcomes of the whole kind of problem statement. It's not only politics, sports, it can be anything. Like build and buy decisions in technology, it is also influenced by a predictive market, right?

Mm-hmm. Yeah. A- absolutely.

So you- Yeah, I mean, that's exactly where it's getting at. But is this something we can regulate or is it just gonna be a new fact of life? Is there something that we're supposed to do about all this, or is it we just need to be aware that this is going on, but it's perfectly legal?

I think it's just something that it's best to just be aware, right? Like, be prepared and, and not kind of be ignorant to what's going on and not ignore it. Um, I think awareness is a, is a lot of the problem.

Uh, people don't really know what's going on. I don't think anything here that we've been talking about, like, is in our control unfortunately . So you kind of have to either adapt or figure out ways to use it and guide it that's in, you know, in line with, you know, a, a good morality compass to, to, to put it short.

Um, you know, there's a lot of ethics that go into AI. There's been a lot of people that have, have quit certain companies because the ethics have gotten into certain hands of groups of people that have larger influence over certain groups of people. Um, and it, you know, that also causes other problems.

And then, you know, you have these groups of people that now, criticize AI because of who's in control of it. And, you know, so, it's, it's very sk- it's a skeptical conversation, and AI, it took a while for it to get adopted. Like, I, I remember, like, I started off with the API developing, like, as soon as it came out.

Like, I remember how, like, I was almost like, it was like something I kept in the closet. Like, I wouldn't even mention it to my company. I wouldn't even mention how I was using it.

I still don't even talk to them about how I use it personally, 'cause it's, it's still so, such a sensitive thing. Like, so like I said, you know, it's in the hands of people that are way greater than us-Uh, Sam Altman, Anthropic, I mean, you name it, and it's kind of up to them, right? I mean, and they also-- It's so interesting 'cause they garner data on how people use it.

So I've seen lately, like, certain companies, OpenAI, obviously they sell- sold out to, to the Department of War recently, but they've been selling out lately, like throwing ads out on their, on their platform and stuff. So now it's like you feel like your data is being sold out. So now they're probably making predictions, like you said, is our lives in the hands of AI?

Well, it depends on who is in control of it, right? Like, our data shouldn't be made or used to, to predict certain markets. Like, and I feel like that's kind of where I draw the line in, in terms of that ethic right there.

Because, I mean, I mean, you've seen how it's been used in certain countries. I mean, you've seen how people get pillaged and by these drones and certain, and factions just lose so much in, in a, in a little amount of time. You know, it's, it's, it's, it's ridiculous, honestly, if you ask me personally.

But, the, the way I use it is healthy, right? Like, I mean, predicting markets is a, is a healthy way to do it, but I don't think those markets will ever open up to, a commercial point like you mentioned, where, oh, are people gonna be betting on who's, you know, winning or losing X, Y, and Z wars, or are our lives and longevity at stake? And, you know, I don't think AI should be making those decisions.

And if it is, it's, it's, you know- Well, well, well, so Alan, Blake mentioned moral compass. We seem to be short on that little particular skill set these days. So, what would prevent the following scenario, right?

And Steven kind of alluded to it as well. Let's say that I am, part of one party versus another, and I want to motivate my people to go out and vote. Might I not start betting the other way in the predictions market so that more people get outraged and show up at the polls?

And can I manipulate this stuff where I bet one way to drive and affect the other way? Agreed. Well, so there's morality and there's legality.

Don't confuse them, right? One, one is not necessarily the same as the other. I, I think from a mor- a morality point of view, we can all agree it's probably there's gonna be a lot of, you know, immoral kind of acts that are done for people's own enrichment.

From a legality point of view, though, I just don't know what we could do to, in essence, legislate or adjudicate morality. And this has been, this has been a problem long before we had AI, right? This has always been a problem in human society.

" But th- that was a simpler time. Life's more complicated now. So, you know, I, I don't know if it's a moot, a moot point, Mike.

I don't know. Steven, you wanna jump in here? Yeah.

I'll, I'll just point out that, we, we can do things about stuff. I mean, you know, this, this is a, a- the- that's how, you know, governments work and stuff. And, you know, Polymarket absolutely was, outlawed.

Um, it was fined, one point four million dollars by the CFTC in twenty twenty-two. It received a cease and desist because of their, their practices of trying to use prediction markets, il- illegally in the, the finance industry and stock markets. And, and it was blocked and outlawed for, two years until it was, reopened to the US market, by the current administration.

Um, I, I don't think it's fair to say we can't do anything about this, and this is just how the world is. We absolutely can do stuff about stuff, and it's called the law. We are trying to do something about the law and the lawmakers, but apparently the lawmakers can engage in insider trading- That's, yeah ...

and so why can't they just engage in prediction markets, right? Yeah, just that's a, that's a good point. So, like, even like like the, the, the market, the predictive market has become so...

It's always been there, but it's, like, all, all this digi- digi- di- digifying everything. Everything became online. Like, everything's so accessible, so you can, you can literally technically make a market for anything if you wanted to.

But, to your point, like, there's people so high up now, I mean, that no one's, no one's not able to influence a market anymore. Um, and I mean, even you said it in the, in the article. C- CFTC was already evaluating how these, these prediction market platforms, like they're saying they should be classified and some lawmakers, should start imposing, you know, government practices on, on, on how they should be, used because certain policymakers can bet on these outcomes and directly influence them and ultimately corrupt them to their favor.

Um, it's early. I know it's, like, early, but it's, like the regulatory conversation should be probably addressed, and, and that could probably solve a lot of the problems because, you know, it's also whose, whose hands is it in at the end of the day? Like, it's not AI that we're doomed by, right?

Like, people are like, "AI, AI, AI can't do anything unless it's maliciously told to, unless it's maliciously trained to," and that's, that's where people need to understand it. Like, it's, it's clear as day. Like, it's like, it's like a baby, right?

You-- A baby's not born evil. You have to train- You have to, you have to, you have to train this, this, this, this, this, this life to do certain things and act a certain way, right? Like, yeah, it's, it's the same thing, you know?

Um, so that, that's how I look at it, black and white there. Yeah. All right.

Well, we're gonna have to shift the gear here because we are running out of time, and we have two other topics to get to. And well, we might be talking about some usage of AI for good, but let's see where we go with this thing. Apparently, the Ukraine is now saying that it will make available its knowledge about combating drones to AI folks who wanna build algorithms, and I guess the Chinese government also said, at least signaled, that it is researching similar capabilities.

So Garima, can we, like, maybe use AI for good here to kind of combat all these swarms of drones that are now the new state-of-the-art for warfare? So I'll actually, di- ... parts.

One is what Ukraine is, trying to kind of do, which is, like, the quick win now and here kind of an opportunity. And then when you refer to the, Chinese, you know, strategy on how, they are thinking about it, it's more advanced, right? So think about this.

Ukraine would be able to provide you, like, re- real-time data, and, you know, they are betting on edge autonomous drones. So are orchestration and layered autonomy with, if, they can extend this with field robots, right? So they have the data, they have the technology at hand, and they can provide that technology.

So that's the here and now situation. I think more or less, more advanced countries would look at this opportunity and try to kind of foster some kind of a collaboration where, you know, they can in- inculcate some kind of de- defense strategy. But more, like, worrisome is how, you know, this technology's developing in terms of advancement.

So if you see what, China is doing is they are working on sensors, and they are working on cognitive radar and RF sensing. So this is very interesting because if you combine radar and radar s- sensing technology with RF and AI and cyber telemetry, it becomes kind of an opportunity, you know, space. And there's a lot of kind of, opportunity in terms of how defense w- technology would develop from here.

I would see that a lot of, investment would go into live labs, for example, cyber labs, for example, and, there is some kind of innovation coming out of these, you know, four technologies which we are discussing, RF sensing, cyber telemetry, as well as AI on the battlefield, right? So it's more than what you see right now. And I feel that, you know, there will be some kind of a privacy preserving, techniques also in the loop.

So the AI stack for defense is getting more and more interesting, and I think there will be a lot of investments in that space. Mm. Steven, what's your take on this?

I mean, I think I could just hook up a bunch of AI models to some lasers and just knock out all those drones as they're coming by, and, you know, we'll all just sit there and it'll be like a giant 4th of July party. What do you say? Yeah, that sounds like a lot of fun.

I'm sure that'll definitely work. Um, yeah, no, I, I, honestly, that sounds like some of the suggestions that we've heard from, coming out of Washington for dealing with this war. Um, I, I think that it is going to be a lot more nuanced than that.

Um, apparently the Department of Defense/war is already, using AI to evaluate, bombing targets inside Iran, and that has apparently already resulted in some pretty unusual, and pretty bad choices. Um, I think that, it's inevitable that when new technologies come out, they are accelerated, especially in times of war. Uh, we've seen that for the, all, all, all, all of human history, right?

You know, we invent something that we think gives us a, a leg up, and we deploy it to kill people, and then we later figure out how we're supposed to use it or, what, what its limitations are, and I think that we're seeing that as well here. Uh, when it comes specifically to drones, I think that one of the interesting aspects that we're seeing is using AI to build autonomously piloted drones rather than, remotely piloted drones. I think that's a, a, a key, technology that has shown itself to be needed in this world of, you know, drone defense.

Um, you know, RF interference, all these other ways that they're, that, people are trying to block drones has, has caused people to want drones to be autonomous and autonomously select their targets, which is, pretty fraught, because if they're selecting their targets autonomously and they're piloting themselves autonomously, then maybe they'll decide that a, a school or a, a hospital is a legitimate target. Uh, and that's really not, I think, where we wanna go as a, as a society. So yeah, it, it, there, there's a lot to unpack here, between drones and AI.

Um, and I'm not sure yet if any of us know the answer. All right. No.

Alan, is this pro... Is this prolonging wars? And I ask the question because, it's just simpler to launch drones, and if we could combat the drones, then we'd have to resort back to traditional, you know, boots on the ground, and nobody likes that, so the war would be over sooner.

Perhaps. Perhaps. I mean, but you know what?

When, when are wars sensible? So is it, is it prolonging? Only, only if you're on the losing side, it's not, right?

And, and... But let- let's, you know, be honest about this. I think the biggest thing that comes out of the Ukrainian-Russian war is the fact that- Piloted billion dollar manned aircraft are done.

There's not much you're gonna get with that, that you don't get with some drone the size of a Volkswagen Beetle that costs 10 grand or something like that, right? 10 grand versus a billion dollars, it's a pretty easy thing to, to decide, and it's an equalizer, right? Go, go to the Iranian situation.

What, it's not a situation, it's a war. Go to the Iranian war, right? You've got the US and Israel using the greatest technology, you know, and the most expensive weapons, bombing these people into the Stone Age, but they're still able to get these $10,000 drones up and attach oil facilities and, and banks and, and other strategic targets at a fraction of the cost, and we can't stop 'em.

So this, this is a game changer. Now, we start adding AI to the mix to make them more autonomous, right? And that's part of what Ukraine is adding to the mix here, is by, by donating the, the real time data of how these d- drones, what they're seeing, what they're doing, how they're, how they're behaving, is really...

Think of it as training data for the next generation of drones. And, and so they're gonna get better, and as they get better, the idea of buying that sixth generation fighter jet becomes less and less appealing. So the, the, you know, AI's gonna make the drones better, better drones are gonna change the, the equation, and when you're the country spending 10X more than everyone else does on, on defense and they're getting better bang for their buck, well, that's, that's important.

That's game changing. I don't know. Garima, what's your take on this?

Is this gonna be... You know, you live in a country where the spending on defense isn't quite as high as it is here in the United States, and would that kinda make sense to you guys or, you know, would you think about it differently? I also think about, like, m- you know, from a Canadian economic perspective, we rely on NATO, right?

And, we see a policy shift in NATO as well, like how the data sharing would happen, what kind of, national versus, you know, alliance, policies would come into life with these kind of, you know, war strategies. This is macro versus micro war strategy, right? So if you think about more cost-effective ways of defending, this is definitely, on the radar for economies like Canada, and, we are pretty kind of advanced in AI and ML kind of technologies.

So we probably will be, having some kind of, you know, edge on top of, these, technologies, how we can foster that collaboration with organizations like NATO, right? So I, I see that a multi-fold opportunity, to kind of ensure that we have, the right kind of defense strategy moving forward with technology. I don't know.

Steven, are we shorting the stocks of the providers of these very expensive aircraft out there? I mean, 'cause, you know, make a small fortune building those things for various governments around the world. " But at the same time, I think that if we look at, at the war in Iran objectively, what we see is that w- probably the most significant factor in the current situation with that war is the ability of Israel and the United States to domin- dominate the, air, over Iran.

And if it wasn't for that, and that's actually one of the significant factors that's held Russia out of Ukraine, and that in many f- you know, in many ways, that's, that's directly dependent on those manned expensive fighters and bombers. And so I would say maybe... You know, I'm not a military analyst, but, maybe these expensive, manned aircraft still do play a, a role considering that they are the deciding factor so far in this war, not drones.

All right. So here comes the fun part. So let's go back to the A block and the predictions market starting with Alan.

Are they gonna make a remake of Top Gun one more time? What do you think? Um.

Top drone. You know, with a drone, though, and, and... Well, that'll be good, 'cause Tom Cruise is getting a little old to be up in a fighter jet anyway.

So maybe he sits back somewhere at the base and, and he's the human in the loop for an autonomous drone or something. I, you know- All right. 5.

Garima, what do you think? I think, one important point which we have missed out is how much upskilling is required in defense. This is true.

So, but, but, you know, it doesn't make for a good movie. That's the problem. Steven, what do you think?

Top drone. Uh, AI Tom Cruise. Uh, he already looks like AI.

You might as well just, go full in and make it AI. There you go. So we got an AI version of Tom Cruise managing an AI fleet of anti-drone weaponry.

Blake, what's your bet here? Uh, no comment. No comment?

All right. Yeah. No, I'm good with that.

I'm, I'm saying that Hollywood's gonna take a pass on this one. It's just too expensive, and they're gonna move on to the next storyline, and nobody wants to pay the salaries for those real life actors anyway. Third block, we are coming over to Stephen here, who had a tech-filled day around the cloud last week.

Had a bunch of folks come together and talk about the latest trends. Stephen, you wanna give us kind of a summary of what you guys were talking about at this event? 'Cause I swear, you know, just when I thought the cloud was getting boring, all these things start changing.

Well, I think that's the most interesting thing is, you know, coming out of this event, was the more things, change, the more they stay the same. Um, one in... one of the aspects of this event that I, really enjoyed was having a good friend of mine, Tom Lyon, present, basically how did we get here, and he went all the way back to punched cards.

Uh, he, he went before mainframe. And, you know, he talked about, you know, Ada Lovelace and all that and, you know, Turing, a- and all the way up till, you know, Sun Microsoft, he... and Microsystems, he was one of the founders, and, and today.

And, I think that one of the, one of the things one of the delegates said to me was that it was really interesting to think about these things in that kind of context, that kind of really big historical context, because it really isn't all that different today from back then. And, you know, one of the sort of truisms that mainframe folks taught me early in my career was that all of IT is just about moving bottlenecks around. It's, you know, you, you, you can never make a system that is, really just ultimately high-performant.

It's about figuring out where's the bottleneck and how can we overcome this? And whether that's the history of compute, you know, CPU, storage, GPU, memory, system design, scaling, you know, software, everything is about moving around these bottlenecks. And so we saw some companies specifically focused on addressing some of the key bottlenecks in the current stack.

Uh, you know, Hammer Space, VMware, you know, both of them talked about, essentially about this topic, about how can they, how can they get the bottlenecks out of the way? How can they make more efficient use of resources? That was another, a big trend and something that people are certainly talking about.

We've talked here on The Gang about the current, crunch in terms of memory and solid state, you know, flash memory storage, production and pricing. Well, one of the challenges there is that every company in the memory, industry knows that it's a boom and a bust cycle. And, and, and it really is a bust.

Because what happens every few years is there's a memory crunch, not enough, memory to power, you know, whatever the applications are of today. Uh, a lot of investment happens, new factory's built, new production, new machinery purchased, and then suddenly there's a glut of memory. The price goes through the floor, and everybody's trying to figure out, you know, how do we stay, in business through this glut?

And typically there's, you know, some companies go out of business, some companies get purchased, more consolidation, and then suddenly, guess what happens next? Exactly the same thing happens again. We're in that...

currently in that bust cycle in term... or the b- boom cycle, I guess, depending on your perspective, in terms of flash and RAM. And so everybody is trying to figure out how do we make the most of the flash and RAM that we have in order to, to keep up with the fact that suddenly everything's much more expensive.

Well, it's gonna turn, unless, you know, people figure out, some way to get off this train, but I don't think we will. And then suddenly it's, you know, we're gonna have a glut of these things again. Um, in, in many ways, for example, the glut of GPU capacity after the, the crash of cryptocurrency was what led us to the current AI world.

Um, and so when you look at cloud infrastructure and what this means for enterprise IT, a- and frankly all of us, what it means is that these boom and bust cycles and this challenge of trying to address bottlenecks in the stack, it, it, it affects the choices we're able to make. And so if you wanted to start a company today, you would find a, a huge challenge if you needed a lot of computer memory and storage to make that thing happen, whereas, you know, maybe you try to find a different way to use other people's resources or to optimize resources or something. Maybe that's a better business approach.

So i- it's, it's interesting to see how the more things change, the more they stay the same. So think... I'm gonna read that a little bit differently.

I'm gonna say that rational thought is finally prevailing here, and that IT folks are looking at the cloud and they're looking at on-premise and wherever else they might have at hand and saying, "What is the nature of the workload? " And then deciding where they're gonna run it accordingly, which is kinda how they used to do things, versus I think there was a l- you know, almost a full decade there where it was like the answer is the cloud, what was the question? But Garima, am I crazy or what?

I think you're absolutely right because, we also have to look at, the economy, right? I mean, if you think about critical infrastructure, I talk about tele- telecom domain, for example. So who's paying for all these GPUs?

So we have to make a case where, you know, you can commercialize ki- some kind of AI centricity into radio access network, for example. And this would only happen at the edge, right? And when it happens on the edge, it is basically a question of how efficiently we can deploy these, advanced or emerging technologies, and where it should run, and how much we can invest on this.

Where is the return of investment coming from? So it's not only like, the, the, the data centers, it's the entire AI stack. How do we build that stack with front ends, back ends?

You know, what kind of, you know, technology, orchestration technology we have? What, goes in at the edge? What is on the cloud?

How many, how GPU-centric that technology can become, and where is the return of investment? All right. Blake, you wanna jump in here?

What's your take on what's going on? Um, yeah, I think it's kind of like a, it's a shift from on-prem to on cloud right now. I think we're over that lift and shift, era where everyone was kind of getting onto the cloud.

I feel like either people have, like disparate systems where they have like a little bit of both. Um, and I think that's kind of wh- where everything's going right now. Um, I think the hybrid cloud is getting more sophisticated.

I think, I think, I think, the, the tech behemoths are starting to work together now because like you said, stuff is getting expensive. Like, why reinvent the wheel when certain companies have what we need? Let's just expose APIs and/or create dashboards together so that customers, whoever, whomever for, for whatever use case can interoperate efficiently and save time and money.

So there's ways that are people-- you know, these bottlenecks are obviously inevitable. Technology is changing by the minute. Um, I mean there's, there's so many things to learn.

Um, I can't even keep up with all of it. I just stick to like my favorite things. Um, but, yeah, I think that's where it's kind of going right now.

I don't-- And I think what's also an issue right now is data transfer and data mobility because like you said, RAM and memory and speeds are important in which things need to operate nowadays, which is the infrastructure that runs off of AI for the most part. Um, and if not, most companies are falling behind, because you, they're static to put it short. They're, it's, it's you can't be as elastic if you're not adapting to the whatever's going on in front of you.

Um, so I, I think the cloud's great. Um, I work for a cloud company, so I, I, I, I'm picking up what you're put-- you guys are putting down in that terms. Um, but yeah, I think, I think as technology and the days go by, I think these companies are gonna get more and more complex and they're gonna have to figure out ways to cut costs, save costs, and, and, and also meet the demand and needs of, of, of enterprise, customers and, and small, medium, large businesses.

Mm-hmm. Alan, you're on mute. Here's where I am.

Been in technology thirty plus years, thirty-five years. Been on the media side of things, I don't know, fourteen, fifteen years. Um, so often the hype runs ahead of the reality.

The media message runs ahead of what's going down, boots on the ground, so to speak. We've been hearing about first it was hybrid multi-cloud, then we started hearing about the edge, right? When, when did you hear about the edge?

Four years ago, maybe five years ago. The edge, we're gonna put containers near your cell phone towers. What exactly is the edge?

All of these things. What I think we're seeing now, and it's being hastened by AI, is the real birth of the edge, and the edge is a Mac mini computer today, right? That's the edge.

People, people are recognizing, "Hey, I don't want to pay the token usage necessarily to the big LLM that's up in the hyperscaler land. And I don't want to pay for that heavy... I don't wanna have all my stuff in that hyperscaler who charges me 'cause it's expensive.

And though I'm doing FinOps and I'm doing this, and I'm trying to turn it down, right? I work at home mostly, I work from anywhere. " The edge is getting really, really real now, right?

And so what I think you're gonna see is reality runs face first into hype and, and the cloud providers and, are gonna learn to, to adapt to that and live with it. I think we're seeing, if I could call it, the next version of the cloud hyperscalers. They're called Cloudflare and Akamai and so forth.

But the, the big three are gonna adapt too where they're gonna give you a new hybrid cloud, which isn't just public-private cloud, it's core data center and edge data center and, and, or edge, you know, data or edge compute. We've spoken about it. Stephen, I'm sure your past Tech Field Days have, have hit on this model, but I think AI is bringing that model home.

Y- yeah, absolutely. And, and, and that's the other thing that, I guess goes unsaid, is that A- AI is everywhere and it has, you know, it's not just the elephant in the room, it is, you know, the entire room. And so you go to Cloud Field Day, you talk to these companies, they're AI companies now.

Everything is being built to support AI applications. And, you know, our, our joke at, at the Tech Field Day is that we could basically rename every Field Day event AI Field Day, and it would be pretty accurate because that's what it is, you know? So we've got, for example, you know, we're gonna-- we were gonna be at RSAC.

Um, I bet that the most of the conversation there is gonna be around, around AI. Uh, you know, Karim, I'm sure that you have, have seen that too. Uh, you know, we're doing networking Field Day, most of it is AI networking.

We're going to Qlik Connect. Um, I just recorded a podcast, Tech Field Day podcast with Qlik Connect, and the entire conversation was on AI. Um, because of course it is, because that's where we're, that's where we're at.

You know what I mean? Blake has been talking about AI this whole episode because of course it is, because of course that's what, what, what everything is, is i- in the industry. Um, but you know, that being said, I think the fundamentals still matter.

We still have to think about compute. We still have to think about storage. We still, still have to think about software stacks, and we still have to make this stuff work.

And, one final thing I wanna say, by the way, before we end, we also, we had a special guest presentation. Again, this was a, a really special thing for me, and I think that it'll appeal to the audience here. We had the Internet Archive come in and talk about, present to the delegates a- about what they're doing to preserve the history of the internet.

I know many of us use the Internet Archive Wayback Machine, but they pointed out that, that it's, it's much bigger than that. So, just a little shout-out for a great, nonprofit that's really helping so many of us, preserve everything, movies, books, software, they're really all over the place. So, we're gonna be posting that, video probably later today, so check that one out.

Hey, I think Blake touched on it exactly, and I think part of the issue here is regardless of what makes sense, people just go with what they know first. That's their first preference, and then if that doesn't work, they might figure out how to learn something else later, but there's always this human bias that goes into all these decisions. Hey, I wanna thank everybody for being on the gang today, sharing their knowledge and their insights.

As always, stay tuned for the rest of the replay for the Techstrong TV lineup. It too is awesome once again, and we'll see you all tomorrow.