Apple’s Foldable Bet, OpenAI’s Enterprise Push and AI Agent Risk

Hey everybody, welcome to the Techstrong gang. And the awesome thing about Tuesday is that there's always lots to talk about, so today will be no exception. Joining us today on the gang, we have Jon Swartz.

John, how you doing, buddy? I'm good. Good to see you, Mike.

Good to see everybody. Good to see you as well. Chris Blask, how you doing up there?

I assume you're still in Canada, are you? As far as we know. I'm loving life.

How are you? All right. North of the border.

All right. Anne Hallward, how are you? Good to see you as always.

Always a pleasure. Thanks for having me back. Always.

Well, you're welcome anytime. Sid Nag, good to see you as well. You're back from your recent travels.

I think you're home. Is that true? Yes, sir.

Back home. All right. A hot seat.

Thank you. And a member of the gang that we haven't seen in a while, Dr. Stacey Thayer is joining us as well.

Stacey, good to have you back. Great to be back. Thanks.

All right. Let's jump in. Apple is in the news again.

There are reports that they are going to build a new kind of, I guess I don't know what to call it. It's a phone that folds. I guess we used to call those flip phones back in the day, but I don't know if they're going to use that terminology or if it's just all passe now.

But John, walk us through what's going on here since you've been tracking Apple all these years and- Oh, yeah. Oh, yes ... correct me if I'm wrong, but I think Samsung has one of these things already, but what- Oh, yeah.

I was going to mention that. I'm glad you said that, Mike, because I remember working for Barron's back in 2018, going to a foldable Samsung event. When was that?

That was in San Francisco, and then they did a whole big show with a number of other vendors at Mobile World Congress in 2019. So back then, there were rumors that Apple was going to work on a foldable phone, like the rumors about the car project. But in any event, this foldable phone, which is evidently called iPhone Fold or iPhone Ultra, is expected to be part of the 2026 iPhone cycle.

So this, I would presume, will be the September announcement that will be the coming out party for John Ternus, who is the new CEO for Apple, who is also a hardware engineer. So that would make sense, right? So this phone evidently, and we had a story in IT by James McGuire, this is going to be unlike the taller, narrower Samsung Galaxy Z Fold, this one's going to be shorter and wider with roughly a four to three display ratio.

And it's going to be closer to an iPad Mini than a stretched smartphone. 8 inches. The starting price, from what I understand, is going to be $2,000.

And this is where I want to ask Anne, who actually is a resident Apple expert. So at $2,000 to $2,500, which is the price range for this, who's the buyer? Is it an executive, second device crowd, creators, road warriors, affluent consumers, power users?

Enterprise. I'm just wondering kind of where this goes. Okay.

I think enterprise. I think this could be an enterprise software story for them, because that's who's going to obviously pony up. But I'm the Apple person because I work for Apple.

It's the only large company I ever worked for. I was a developer/designer in Apple Education Marketing, Enterprise Marketing, and they're just doing what they always do. They're fashionably late, charging a premium, and expecting to wipe the floor with everybody else.

That's the Apple way. And I love also that you call me the Apple person as I carry a Google Pixel phone. I know.

But you know a lot about the comp-- Well, that's the reason I was asking you is, maybe you or Sid or somebody else can answer this, but so would this replace the tablet entirely, or is it a specialized tool? Yeah. I am an Apple person too.

I just won't give them everything. I won't give one company everything. So I think that the play here is enterprise, and I think that this is essentially a $2,500 phone that folds into an iPad Mini.

I think this is maybe going to be propositioned as a productivity device, but it's going to require serious app support. " If the rumors of the new iOS don't match and it's not a reinvention of the form factor, I think they're just selling expensive hinges. Two.

So I think the challenge is there for them in a lot of ways, but I think that they're going to probably capture at least half of the North American foldable sales almost instantly, I think because of who they are and what they're offering. Sid, are you going to sleep out in your car the night before to be first online for this? No way.

I think this is the most un-Apple thing to do in my mind. This is not true innovation. It's simply form factor experimentation.

Now, Apple always led the way in new products. This is like an existing product they're repackaging in many ways. Although, to give them some credit, I think this has some legs because if they're really looking at, as Anne said, the enterprise market, this could open up a whole new world of doing things around spatial computing, AI application, rendering, and a whole host of other high-value capabilities.

So maybe Apple's targeting that market. I don't know, but to me, it's a very- Sort of a me-too thing to do, but- Right. I mean- We'll see where it ends up.

Yeah ... Samsung came out with their foldable phone in 2018, I believe. Exactly.

Yeah. They were my client at that time. I did SEO for developer marketing of Samsung for many years.

So, also look at the Apple Vision Pro, right? You're shipping almost the better part of a decade later than Oculus and all these other headsets that are much more affordable and usable. They're very slow to move, but again, they're expecting to dominate.

They wouldn't be doing this if they didn't expect to dominate. I mean- This actually reminds me of when Apple before the iPod, right? When they were just throwing things against the wall, and sometimes it would work.

Here's the computer in the monitor, that's great for your dorm, right? They're just sort of trying these different form factors and seeing what sticks, and I don't even know that they're considering Samsung. It's like they're, "We're Apple, we're just going to dominate.

We're going to win. It's our flip phone or our fold phone. It doesn't matter what anybody else has done, we're going to do it the Apple way.

" I feel like we're getting that mentality from them. Yeah. Yeah, I don't even think- And they're not wrong, unfortunately.

I mean, who would- I always think innovation is the really- Who would see a market where people stop carrying iPhones and start carrying foldable phones? That may well be. Who knows?

Yeah. I don't even think innovation is that important. I mean, was it a thinner hardware, less visible crease, multitasking on an iPhone, like iPadOS or iOS?

I think with them, and I wonder, and maybe I'm probably wrong, but I wonder if this is an admission, too, that they really need to do something to kind of rejuvenate or refresh the iPhone line, because it's really gotten stale the last few years. I mean, I remember writing the last couple of years on how it's just marginal, incremental developments or improvements in the camera or the memory or something else. And it's always been- It's always glass ...

kind of a mild disappointment with lacking any type of AI or anything even imaginative. So I wonder if this is the way they fill the gap until they finally get their AI stuff together. They can also increase significant revenue as a consequence of new accessories that'll go with this new phone.

Mm-hmm. Mm-hmm. I read something about that as well.

So I'm a little unclear on how this is supposed to work. Am I supposed to be able to surface, I don't know, go to two different websites on two different screens? Or is it one screen that's going to be somehow displayed across a hinge?

And is this because there's some new grand experience, or is it just because we're all getting older and my eyesight is shot and I need a bigger pixel format, or what's driving this thing? What's driving it is our obsession with our phones and how phones are essentially wearables, and this is just something that's going to allow you to watch and consume more of their Apple TV content in every place you go. I mean, I'm not the only person in my friend group that has now a lanyard on their phone because I wear it so I can track my steps.

It's just another way to increase your addiction to their platform. If this does not come with an appreciable update to iOS that enhances the usability of this, this will flop. It will be $2,500 hinges.

I don't know. Chris, I haven't talked to you about this, but I'm dying to know. North of the border in Canada, are you allowed to even buy these, or do you have to go get BlackBerries or something?

How does that work? Oh my God. Chris, I think you're on mute.

He's been rendered speechless- He has muted now ... by your question. I guess he has been rendered speechless.

All right. " And when they see new endpoints and new things like that, how do the security people and the compliance folks just kind of look at all this? So I think I'm going to say with a grain of salt and a lot of cynicism, probably.

No, I think they look at something like this and there's immediately looking at, okay, what is the functionality, right? What is the operating system? How are people using it?

And then what are the security vulnerabilities? What does this mean? And then to play with it, hack it, or jailbreak it, all the things that we get to do.

But I think it is to some point of how much are they really changing? Is this just Apple and now it folds? And then in which case, well, that's not too much fun to play with.

That's too much different. It's just more of the same old in a different form factor. So that's not going to interest anybody in the security field in terms of what it can do, because it's not that new.

I think if all of a sudden there's new capabilities, new operating system, things like that, there's going to be things to play with, and then of course, concerns. If it becomes, as we talked about, being enterprise-focused, what does that mean for enterprise organizations then to be able to secure it? And then how are people using it?

Is this one of those phones where you've got your work fold phone and your regular phone, but then... We feel like we don't know enough about it to know the use cases, but as soon as we get some, I absolutely believe that the security industry will be right in line to figure out how it's being used and what the vulnerabilities are, or what the gaps are. You know, I'm sitting here, I still have a motorcycle helmet with an Apple sticker on it, and my original Apple IIc, right?

But I'm talking to you on an Ubuntu-based machine. I have an Android phone right here that's Motorola right now, but I don't care. Right?

And so I get it, the usability of the Apple ecosystem, but Anne, I basically have to agree that this is aIf this is your big foray, then maybe if you're inside the Apple ecosystem and can't go anywhere else, maybe this is important. I don't know. It doesn't really matter anymore.

It's like hardware is real, and as we talk about on this show every week, right? Well, a lot of this is, like John, this stuff is moving into agentic AI on small devices, and, and, and. Apple has a big ecosystem.

Does this device offering change usability and so forth? That's important, but I'll just repeat my normal rant, that without the transparency into how these devices are being used and the agentic capabilities, it's just more vendor lock and churn. So we'll see.

I will say this. I have lost any sense of appreciation for having the latest and greatest phone. I seem to remember there was a time that I cared, but now I'm not even sure what I have.

I have- Yeah ... a Samsung phone, but I couldn't tell you what version it was off the top of my head. And frankly, I'm much more careless with these devices because they are getting a little bit older, and I leave it out on tables all the time, and I walk away from it.

And I'm just generally probably, it's a bad habit and something eventually bad will happen, but I notice other people too are just kind of leaving their phones around now as almost like they're disposable and it just doesn't have the cachet it used to have. I don't know, Anne, is that just me, or is that the general population now is just less enthralled? There's always this sort of status symbol bits of having an iPhone, and digital tribalism is a very real thing and I have been given grief by a number of people for being an Android primary user.

However, I do carry an iPhone too. I just have an older one, because I like to see how they throttle it, as I don't update it every year. I like to let it go till it's on its very last legs because I want to see how Apple treats its older users, its poorer users, and it's not very well.

The last version of iOS throttled the phone, and it's basically unusable, but I'm not buying a new one until I absolutely have to because it's a lesson for me, and I think in empathy and accessibility because in the US we are iPhone-centric. The rest of the globe is not that way. Statistically- I'll take- Yeah.

Oh, sorry. Go ahead. That was wonderfully said.

I got to stop there, but devices, the handheld devices to me are disposable. I have an N number of tablets and phones and things, and I don't care if any one of them particularly dies. And yeah, the logo on them doesn't make any difference at all.

So, I don't know about you, Mike, and your kids, but I have young adult children, and it doesn't really matter to them as much about the iPhone. The thrill is gone. It's just like it's a functional device.

They don't get excited by the new models. They haven't been excited by it for a while. It's just something they use.

It's not a status symbol for them in the very least. Now it's like cigarette lighters, right? You just need enough of them around.

It doesn't matter what they are. Yeah. They're just a commodity thing.

Yeah. Sorry. But there are people that love to get that new iPhone.

New iPhone came out, I just got it, and I remember, God, I don't even know what version I have right now, but it almost doesn't matter. It's like I got all excited, got the new phone, and then was like, "Okay, this feels exactly like my old phone. " That's been the experience for the last few years and actually- That's where fanboy experience is diminishing.

Like you really don't want- Yes ... the latest iPhones anymore. So we'll see.

I just want to know who has cigarette lighters lying around their house? I don't get that, but you know. There's some of us still out here.

And I know who a lot of the rest of you sneaky smokers are too, so. All right. I will tell you who gets excited about this, though.

My kids get excited for this for all the wrong reasons, because then they're like, "Oh, so that means that the previous generation of gear is going to come down in price, and maybe I can afford that," and they'll go upgrade to that because they'll be like, "Oh, there's a new high-end," which means the middle tier is going to get less expensive. And they're like this is their time to go upgrade to the thing that's two or three years old rather than the thing that's in their pocket that's probably four or five years old at this point. But I don't know.

Is that a thing, Anne? Absolutely. Absolutely, especially to those younger users who maybe don't have a job yet, and they're dependent on the ecosystem of mom and dad upgrading.

I've seen it with my brother and his kids where, oh, dad's upgrading, everyone else below gets the new... And that's the secondary market that they've created. And the thing is unless there's an appreciable camera update or something, I'm not going to rush to get it.

There's not been a killer feature. I keep saying it. Liquid Glass was, to me, the biggest joke ever.

Like, ooh. It was harder to read. I don't think that they appreciate the people that can afford their phones are 35, 40, and up.

I don't know. It seems to me that I like your idea. There's going to be a lot of hand-me-downs for phones coming up in the end of the year and probably around the holidays, and people who are still fanboys will buy some more, and then they'll just take whatever they have and hand it down to whoever else is in the family.

And maybe that's just the new order of things, but we'll see. But folks, if you are an Apple fanboy, I'm sure you'll be first in line and you're still out there in the hardcore, but the rest of us, yeah, we're kind of starting to shrug a little bit. All right.

Let's jump gears here and shift into, well, the enterprise. And it turns out that OpenAI has launched a new subsidiary company that basically is dedicated now to... putting together the consulting expertise needed to build out AI applications and agents and all the things that go with that, and they've bought a company to help them go do that.

Sid, I'd love your opinion about this because I could swear when this whole AI thing started out, it was one of these things where everybody said, "Well, it was so easy as the commercial says, a caveman can do it," but here we are. Yeah. Another myth-busting exercise for AI industry, right?

It's not about buying the latest models. It doesn't equal transformation, right? What I think what OpenAI is signaling here is folks are looking for folks being the real buyers who have money in their pockets, are really looking for deployment models, the operational excellence.

It's not about the AI stack that all the vendors keep talking about, but how does the AI stack actually get implemented to yield a real business outcome or scalable intelligence that I keep talking about. So it's more about operational execution, it's more about access to intelligence, overcoming integration complexity, and all of that. Now, having said that, historically, providers that build the technology have not dabbled into the services arena.

" If you look at the history of cloud computing, you've got Amazons, Googles, Microsofts, they are all playing in the AI space. So it's kind of getting very interesting there. But historically, these folks have all had very robust partner programs, right?

Relying on the, from the McKinseys of the world, to the Deloittes, to the Accentures, and even at the bottom layer, the Infosyses of the world, et cetera, right now. That is shifting with this acquisition of Tomorrow. I think that's the name of the company that- Yeah ...

OpenAI is buying. And I see this sort of OpenAI recognizing the fact that they've got to own the whole entire experience of making their technology operational, because honestly, they're struggling. They're investing 500, 100, 200, $300 billion in data centers and GPU technologies, but they're not seeing the revenue come through on the buyer side.

And I think this is really a move towards accomplishing that. So we shall see how this plays out. I think the big four should be reading this and being very concerned, as you mentioned, Accenture, Deloitte, I think they should be quaking in their boots because- Mm-hmm ...

OpenAI just vertically integrated into their lane. They are formally now admitting that they're selling API access, and that's a commodity play, and that the real money's in the services. So, it's a conflict of interest as well, right?

Your vendor Salesforce is also the system integrator. Yeah, there's another angle to this. I think OpenAI has been sitting by the sidelines and watching these guys.

When will they step up to do what they need to do? And I think these guys have been struggling to change their business model because it's so labor intensive and labor arbitraged. So, they're struggling with deploying AI to deploy AI, right?

It's kind of interesting. So OpenAI said, "Enough is enough. We're going to take charge of this whole situation because we can't wait for these Accentures of the world to go do it because they see AI as, they may not admit it, but they see AI as some sort of a threat to their core business," right?

So I think there's a lot of that going on as well. I think, there's this whole channel conflict angle here, and I'm wondering what Anthropic is going to do. " So maybe they'll recommend something else more aggressively, whether it comes from Microsoft or Anthropic or somebody else.

And they'll be counting on the fact that the global SIs especially, they have their deep relationships with a lot of these organizations, and they have the relationships. And just because OpenAI has a subsidiary now that talks about system integration doesn't mean that those customers are going to walk away from their long-held relationships with those SIs. So, I think it's a fairly brazen move by OpenAI to kind of, shall we say, disintermediate a lot of partners out there.

And historically, that's not played out very well. So I don't know. I don't think this goes in very far, but I could be wrong.

I don't know if you guys read this, but Anthropic actually invested $100 million into the Claude partner network. This is an announcement they made on March 12th. Mm-hmm.

So they're moving forward in this direction as well, but they're not acquiring a company to go do this, but they are building a significant partner network. And that is interesting because you think that the ones that would go build these partner networks would be the guys that own the hardware, like the Googles and the Microsofts and the Amazons of the world, while they use Claude as a frontier model or whatever model that is being supplied by Anthropic. But Anthropic is doing something very similar.

They're building their own partner network, and they invested $100 million, and they made this announcement on March 12th. So, we're seeing some interesting shifts in the industry by the model owners, like OpenAI and Anthropic. So that's interesting.

Oh. Chris, what's your take on all this? Because it seems to me that it's also signaling that the building out of these AI agentic workflows is a lot harder than people have been letting on.

And I think you need some expertise, and some folks are saying maybe you shouldn't be doing this at home by yourself, or is this just the big boys are playing games, but at the end of the day, organizations are going to have that expertise themselves and not need these consulting partners? Well, so the big players, as a rule, you should pick on them. That's what you get for owning that much of the landscape.

Yeah, been there, seen it, done it. So, at the same time, you've got to give them credit where due. And Sid, I agree with you, right?

This is them recognizing, "Oh, yeah, chatbot is not an employee. " Pardon me. AndHowever, right?

Anne, Mike, yes. So let's be clear about this. We were just talking about Apple.

This is unlike, Sid, what you say, building channel partners and so forth, what we've always done traditionally and, getting the values spread out. This is another example of no, we're one company with one board of directors and one CEO and one pipeline, and you have to put not just all of your data, but all of your expertise and workflow and everything else inside our corporate boundary. Maybe.

Apple, like we were just talking about in the last segment, that could work for some folks. I'm not saying they're going to go out of business on this, they're bright folks and they can work it out. But I think there's a huge chunk of the world that doesn't want any of that, and plus every channel partner that would like to have a business and any sort of resilience in systems.

So good luck. See where you're going. It matches your corporate philosophy.

I publicly disagree with that approach, but we'll see how it works. Anne, we had a show where you were talking about this with NVIDIA as well, which also seems to be wanting to be very vertically integrated and own everything as well. Is that the in vogue thing for all these AI companies?

They want to own the entire vertical stack, or were they absent that day in kindergarten when we all learned how to share, or what's going on? I think that this is a deliberate choice to own the ecosystem. They want to control every aspect of it.

There's too much money at stake to play fair. I think there's a procurement government conversation that should probably be happening that isn't. But none of this is a surprise to me, and it's actually one of the major reasons, if a company comes to me for help at my agency and they're a ChatGPT wrapper, I'll say no.

You're playing in their sandbox. You're going to have to expect that they're going to want to control all the sand. That's just how this is.

That's why there are great battles being fought over centralization versus decentralization. That's why I'm a ride or die crypto person, because anytime there's centralized power, just like there is with OpenAI and Anthropic, there are going to be those who they're going to seek to control the entire ecosystem because they can. And I agree with you, but again, I think that where we're going in this is that's just a brittle system approach, right?

You can get a lot of power and a lot of control and everything else, but historically, there's not any examples of that lasting long term. And the current technical environment is crashing all the time because we have those unilinear systems. So again, good luck, but I don't think it's the right way.

Anne mentioned a very interesting point around using the word government. And I think some of this is being driven by OpenAI's desire to get government contracts. They've already replaced Anthropic in the government contract space because of all the politics and other reasons we don't want to get into.

Maybe that's a separate show altogether. But if you Google Palantir and its partner program, you find that Palantir has built a massive partner program. And some of the names in the partner list are not the Accentures and the Deloittes of the world.

They're like Foresight and Foxtrot and all these companies that are very specialized in doing AI implementation in the world of government federal business, right? So I think OpenAI is probably realizing that if they don't keep up with that same momentum, they're going to fall back, and that's probably another reason why they're doing this. So OpenAI becomes, in a sense, kind of a service-based consulting firm, and Tomorrow is actually an impressive company.

Some of their clients include some pretty big companies like Virgin Atlantic, Mattel, Tesco, Red Bull. But what OpenAI is trying to do, it kind of underscores this idea that they are desperately trying to figure out, how do we amp up our revenue? And I think Sid referenced the $500 billion, or however much money they've committed to building out the infrastructure.

But I wonder, if you're a CTO and you're looking at OpenAI, would you be concerned about a vendor lock-in where your core infrastructure becomes so dependent on OpenAI and their specialists and you're locked out from switching to a rival model? I just wonder if that blows up. That's a great point because a lot of these enterprises that OpenAI is targeting probably already have long-term contracts with these SIs, and they have preferred SIs.

So I don't know how that's going to play out, because if OpenAI is partnering with Tomorrow, what's the name of the company? Tomo, right? Tomorrow, yeah.

Tomorrow, yeah. Tomorrow. So then they're going to bring in Tomorrow under their hood as their own sort of blessed SI.

And if I'm an enterprise that has already chosen and have a long-term contract with a Deloitte, then how does that sort of conflict play out, right? So if I play out this scenario, John, to the most extreme paranoid degree that I can, which I will. You're appealing to me.

Yeah. I know. Get on your tinfoil hats, guys.

Clearview AI. Clearview AI is, to me, the worst case scenario of implementation of AI for facial recognition. They went around and sold to law enforcement all around the country.

Look at what that company's doing. Look at their stories. Look at how they own the market on facial recognition and law enforcement.

There's a story that just broke last month about someone's grandma was jailed because of facial recognition. This is where this technology comes to pass in the worst possible way, and all of these law enforcement small agencies are using it. All of them So I would agree that about it.

I will go back to the math on this, though, right? So historically, for every $1 somebody spends on a platform, they typically spend $7 to $10 on services to make that platform work. So I think that OpenAI is, with the tacit admission here to your earlier point, Anne, that the back-end models are rapidly becoming commodities.

Here's the greatest advance in computer science in recent history, and it's already starting to feel like it can be swapped out in an instant for anything else, and it's just another thing that sits on the back end of an API somewhere, and they're disposable. If that's the world we're going to live in, then OpenAI needs a new model for business because that's not going to be sustainable to the level of investment that they have now gathered. So now they've got to be in the services space, and that is going to be a hard space to fight in because it's not like you can sit there and say, "I'm the only option in town," right?

Yeah. They've admitted as much that they face pressure and that they're very concerned about Anthropic in the enterprise market. So, they had to do something.

But there's also another key word that's missing in this whole discussion, the A word. There's no mention of agentic anywhere. Isn't it all supposed to be done by agentic?

Why do you need an SI to do these kinds of things, right? So what is OpenAI's strategy? They're buying this company, but they're not promoting agentic as a mechanism to orchestrate, run, manage, operate, all of that in an enterprise environment.

So ideally, that's the end state you want to be at, right? Where you don't even need humans to implement or orchestrate these AI technologies and models and training, inferencing, whatever. So, hello, where is the word agentic in this announcement?

None. Nothing. Zero.

I think- I have... They went ahead, what I've seen when they go ahead and they built the technology, and then they somewhere along the line also forgot, though, that at the other end of the technology, for now, are still humans that need to understand it, that need to know how to use it, that need to know how to make it make sense. And so many people that I know are still trying to make sense of AI.

At the university, that's like the number one thing that we see is students asking professors, "Well, what does this mean? How do we use it? " And they haven't really answered that, and I don't think it's really being adopted at a human level yet.

We're still in that technology level, and there's a breakdown. And I almost wonder if they're- Yeah ... essentially doing this so that they can connect that because there's still a human on the other side that they need to connect to, to make it have it make sense to them.

Well, if I can throw directly to the next block, Mike, because what you both just said last, right, this is the point. But I'll start by challenging the end goal said, right? Now that it's not to get humans out of the loop, because again, we're humans, right, Stacy, and we're actually here at the end of the day.

And I will argue that stewardship, this word that again, on this show over the last, I think the last two, three months has been the last of the primal functions that we look at, that I look at. But it doesn't go away. And I don't think it's possible.

I'm a big agentic, we'll take this block and I can get really enthusiastic about specifically where we are right now. This week, last week, last couple of weeks, really fascinating things going on. But I'm more and more convinced that the humans, we never get pulled out of the loop, right?

And stewardship, for all sorts of things in agentic-driven systems, I think is persistent forever. So whether OpenAI is making the right choice by trying to be a monopoly or not is a different question, but I think the human stays in the loop. I think we've seen this with every emerging technology, and we use it to do whatever we have been doing the same way a little bit faster, and then it takes us years to figure out how to actually redo the entire thing using that technology, and I think that's where we are with agents and the whole nine yards.

Because we're basically inserting a bunch of agents alongside people to replace processes we created for people rather than AI agents, and we got some years to figure out how to actually reinvent the actual process, but that's the next thing, I think. Yeah. But let's jump to our next topic here.

There's a report out from folks who did an analysis of AI agents, and they took a good look at thousands of these things, and what they determined was that the security protocols for this stuff, even the basic authentication stuff isn't really being followed, and hardly almost zero of any of these AI agents have any kind of guardrails in place. And, so the folks who did this report were looking at all this stuff and concluded that, at least the researchers at Capsule Security said that there are no protocols in place for securing AI agents, and they're basically all naked out there. I don't know, Stacy, is this a surprise to you?

Oh, right along up there with water is wet. I'm going to take a chance because any time I can quote Jeff Goldblum is a good day. " Right?

And I feel like that, in so many ways, is what I'm seeing with AI, where it's like, we can do all this. We can have these amazing technological advances, and here's what we're trying to do. But no, it's not secure.

They don't necessarily have clarity on how to use it. So is it going to be a surprise when there's major security gaps and things start to fall apart? And hopefully, the AI doesn't start to eat people, but I'm not going to put anything past it at this point, right?

And I feel like that's just where we are. It's all about what can we do, not what should we do. What can the technology do?

" Well, okay, let's take a step back from that first. But people don't know what to make of it. And I even say, okay, well, when we're looking at development, when we're looking at code, are we able to keep up with the security gaps?

Security in itself is already a challenge. Now we're adding this entirely new component to it, and there are security gaps, and companies are so pressured to be AI. You walk around the show floor at RSA or Black Hat, AI, AI, AI.

And again, as you said before, it's like we don't even necessarily know what that means or how we're really using it, which is maybe why they're hiring these companies to show, again, how to use it. So how do we know how to secure it if we don't really know the use cases even yet? Chris, you've been on this show a lot, and you've always advocated for innovation, responsibly moving forward, and we can do some awesome things.

And I look at this report, and I happen to feel like maybe, I don't know, you're a minority of one out there, and everybody else is saying, "The hell with it. " I think I might be with the majority on this one if we're talking about leaning into it. I was just thinking, we should have, and maybe Taylor in the background, you can pull this up, like a one-year moving window, right?

Reflecting what was our conversations a year ago. And a year ago, Assault Typhoon was a big deal, right? And my explanation at the time was that's just the gap between what's deployed and what you should have in infrastructure.

And right now, a year ago, with available AI tools, anybody with intent can just exploit all that just fine. We're talking about now in this context, and again, as I mentioned a minute ago, us as an example. Me.

Our systems, we do agentic stuff right now that systems of AI systems, it's fascinating. Code development acceleration, and just hold my coffee. But in OpenAI, again, to give them some credit, their new agent structure thing in ChatGPT is really good, and we've been using that in ways as well.

But basing it on the kind of ways we do things, which is keep track of what the hell you're doing. But oh yeah, I could sit down at OpenAI right now and use an agent system to do all the bad things, and it would work, and there's no guardrails inside that structure that would stop it from doing anything. So we are on that path.

The results when you use them properly are amazing, and the results if you don't is whatever you can imagine to achieve that you don't want. I feel like the security people are kind of absent from this conversation, and I can't help but wonder if they've just been told to get out of the way. Or is it that they're just kind of waiting for the next disaster or two so they can say, "I told you so," and then they're going to step up and be more appreciated and maybe get a raise?

But I don't know, Anne, do you think that's feasible, or where are these folks in there? I honestly don't know a lot of security people, but I would say they're lurking in the shadows as usual. But I would say this is a structural crisis hiding behind the excitement of everybody playing with a new toy.

I think that the Microsoft attack really is an eye-opener, at least for me, that there was no special access needed, no exploitation of any vulnerabilities. They didn't have to have any advanced technical skills. They just had to understand how LLMs process instructions.

And so I think it's pretty common of the security people I do know that they're the Cassandras of the world. They tell us what's going to happen, and then we don't listen, and no one really cares about security until they need it. I think that's the tale as old as time.

But I think that this is an eye-opener, and I hope that people will pay attention. I have clients in this space, and I really have to wonder what their response will be to this. Stacy, you know what I can't figure out is, is there going to be some sort of event that is so catastrophic that everybody wakes up and pays attention, or are we just going to have a death of a thousand cuts, and then sometime in 2027, somebody's going to wake up and say, "We should do something about this"?

My hope is the latter, because I'm not a big fan of the big catastrophic events, though. " I fear it's going to be death by a thousand paper cuts, kind of bit by bit by bit, with a little bit here and a little bit there. AI just has so much excitement and momentum and the potential of everything.

It's almost like people slowing down on the side of a road when there's an accident where, these are good things. I don't mean to sound like I don't appreciate them, but we kind of look through like, "Oh, look, there's the cute dog dancing. Yay, AI," right?

Meanwhile, we don't really necessarily think, well, what is it really being used for? Mm-hmm. What do we know?

And then there's always what's behind the curtain of what we don't know. And hopefully, there won't be a big catastrophic end. " But I think it's going to have to really be a pushback because corporations, and we're pushing this as money makers and advancements in technology and the United States being the leader in it.

Whatever you want to say, whatever agenda that people have with AI, that's what I think is really going to be the determiner because there's the technology itself and then there's the people who either have the stewardship of it or have an agenda with it, and I think that's going to impact what happens in the future. Chris, I have a developing theory, and it kind of may be the opposite of what this report is pointing at, but I wonder-Are we trying to secure the AI agents to... And there's no point and avail to doing that, and maybe we should just secure the infrastructure and the data that the AI agents are trying to access because, well, if we get that part right, it may be simpler to do that than it is to kind of try to put guardrails into tens and thousands of AI agents.

I'm going to have somebody, wherever you are, check what's in your coffee, because you sound like me, and that's not right at all. But yes. Again, look at any resilient infrastructure.

bin or something, and I'm going to make it secure. It's like, at a certain point, it's systemic. The infrastructure has to have the basics.

And I think at this point, we can go back to old school cybersecurity, and there's a lot of stuff, provenance and custody and evidence and attestation and structures to adjudicate and failure states and recovery states and return states. There's nothing really new under the sun, it's just that we accelerated 50 years into the future and built an infrastructure that's been lovely, but it will not function a lot longer when we try to make individual corporate entities good. Right?

Is OpenAI and Anthropic, are these good companies, are they good people? That's not a plan, right? The plan is to have an infrastructure that survives when mistakes or maliciousness happens.

Mm-hmm. And LLMs aren't the point. Yeah, I think it's also the agents, the power of the agents today, right?

It's like the agents are really more than humans. They have credentials, they've got access to databases, they have access to APIs, they can trigger workflows. So it's not just the infrastructure, there's a lot of the application layer, at the middleware layer, that's causing a lot of headaches when it comes to security.

So I think that's kind of interesting. Why no? Let's just deploy them without any guardrails against prompt injection attacks.

Let's not have any actual security. Well, you're starting to see the cracks. Microsoft, today, they came out with a research report about the fallibility of the AI agents, and they threw the words liberally around catastrophic, unreliable.

So you're starting to see cracks, but I think, Mike, I think all of us who go to these trade shows, the narrative is just full steam ahead. We'll worry about things when they're broken, and hopefully there are not going to be major breaks. They'll just be little fissures.

They won't be catastrophic events. And I think it's just kicking the can down the road until it happens. Maybe we're all just optimists.

Maybe we all just want to believe the best, and we don't think they're bad actors. This is fine. Everything is fine.

Yeah. Everything is fine. Isn't that a game that kids play about the Oregon Trail, and everybody dies?

Is that... Jesus. My kids used to play in grades school.

But, all right. I'm going to leave it here because I think it's fair to say that the sins of the past are coming home to roost, but the sins may not necessarily be in the AI, as much as it is all the stuff that we did previously is now going to be exposed for what it is. Hey, I want to thank everybody for sharing their knowledge and insights.

As always, guys, thanks for being on the show. And I want to thank everybody else for watching and paying attention to what we're saying. And hopefully, you learned something because, well, I learn something almost every show.

TV lineup, which is coming up right behind us, and that too is a place where you learn all kinds of fascinating and interesting things. The question is what's the latest and greatest, and how to apply it. That's always been the trick in my mind.

Hey, everybody, thanks for being on the show, and we'll see you all tomorrow.