How to Defend Against AI Cybersecurity Attacks with Sam Curry, Global VP & CISO at Zscaler

(Intro Narrator at 00:00:00) Today, we're talking to Sam Curry, Global VP and CISO in Residence at Zscaler, about the AI threat landscape of cybersecurity, the ethical implications of technology in 2025, and more. Thanks to Zscaler for being a show sponsor. Go to zscaler.com to transform your business with cutting-edge cybersecurity. You're listening to Joel Beasley, Modern CTO.

(Joel Beasley at 00:00:29) You know, you've been in cybersecurity for like thirty years.

(Sam Curry at 00:00:32) Because I'm old. Yeah.

(Joel Beasley at 00:00:33) Because you're old. And huge companies. What's the biggest evolution? And small? You've done small companies too? What have you noticed over these thirty years?

(Sam Curry at 00:00:42) Things don't really change. You ever heard the Shepherd tone? No. It's an audio illusion. Warning, if anybody goes and looks it up, it can cause some anxiety and distress. It sounds like it's always going up in pitch. It might be good to put a link in. Yeah, there it is. Yeah. So when you play it, it cycles, but it sounds like it's always going up.

(Sam Curry at 00:01:11) So the point is here that it always seems to be getting worse. Now in absolute terms, maybe it is, maybe it isn't, but you know, we always hear these trillion-dollar numbers or whatever for damage and whatnot. I can't tell you how many times I have found people who have a recency bias. They're like, you know, oh, we just did this the first time. It's never been done before. I'm like, no. Actually, it's been done before. I mean, the first time, certainly, I applied machine learning in cybersecurity was in 2001, and that wasn't the first time. And you know, that was for finding non-signature-based viruses. And yet I've seen people make that claim as the first time in 2007, in 2011, in 2014, in 2019, and they're still doing it. So—

(Joel Beasley at 00:01:55) How else do you expect them to raise money, Sam?

(Sam Curry at 00:01:57) There's nothing new under the sun. Yeah. Exactly. But I don't begrudge anyone that. There are some things that are genuinely new, and certainly applications, and quantity has a quality all its own. Things are very different at scale. I don't want to diminish that, but we haven't changed the game sufficiently. And the main thing that bothers me is that the opponents are still getting better at a faster rate than we are. And I think it has to do with how much of their energy goes into innovation. And yeah, we all know they get to pick the time and place of battle, but the rate of innovation, rate of improvement, rate of proficiency is increasing very quickly and disruptors in the short term seem to favor them. And you know, I'm a little tired that everyone has the same business plan and comes up with the same application, for instance, of Gen AI or LLMs. There are many more interesting applications than the few that appear to begin with in defense. I did a mind map a little while ago, about two years ago now, of all the different ways we could be using it, and I swear I stopped about 150 uses. There were not 150 different uses in the startup community even now. Wow. Yeah.

(Sam Curry at 00:03:18) But that sounded to be depressing, by the way. So what have I noticed? That things generally don't change. They do rhyme. You know, the old adage that history doesn't repeat itself, but it sure rhymes. And so the themes are similar. The biggest problem, I think, in cybersecurity continues to be the distance between cyber and the business, that most of us as an industry are not aligned with our counterparts. Still the office of no for the most part. And I think we could do better than that. I really do. And I think we are a core business function now. And those, just like those who embrace AI faster, those who embrace cyber better and faster are really going to be accelerating much more quickly. And I think things like zero trust principles, whether or not you think about it in the classic ways, they actually make you more resilient to general disruptors.

(Joel Beasley at 00:04:14) So how did you get the job at Zscaler? Did you just hack in and just add yourself to the—

(Sam Curry at 00:04:18) Absolutely not. I mean, how else do you do it?

(Joel Beasley at 00:04:18) Right? Like, I mean, that's how you—

(Sam Curry at 00:04:19) No. Don't do that to anybody listening. That's not how you do it. Just realized I have to be careful what I say. So I was involved early on on the outside. I was at EMC RSA years ago when they took an investment position in Zscaler. That was 2010-ish, I think, around time frame. And then I was a customer as well. And so I got to know the company, and I was frankly—yeah, I was impressed with the product, but most importantly, I was impressed with the culture. And when I needed help, it happened. Even though, you know, I was part of big companies, but I was part of small companies too. And when I was part of a company that was, you know, twelve, fourteen hundred people, and war was breaking out in Ukraine, I had people there, and I needed zero trust even though I'd been fighting with my CIO for a long time, and ZTNA specifically. And when I called, I had it within a day. You know? Didn't matter that reseller said it would take longer. The company responded. So—

(Joel Beasley at 00:05:23) That's awesome. So good support. And then you became like, what's a CISO in Residence versus just being a CISO?

(Sam Curry at 00:05:29) Oh, so I actually run a team of, quote, CISOs in Residence. They are CISOs, but they're also former customers for the most part of Zscaler. And Jay has a way of spotting people and saying these people have talent. And think of it almost like—so first of all, they don't report into sales. They're not like a classic field CISO. There's some BISO function to that. We are a burst capacity, if you will, in that team for the rest of security. And if you think about how we have one of the—we have the biggest security cloud out there. If you think about that footprint, it helps if you've got all this ability to do things like crisis management, disaster recovery, and those sorts of things as resources. We make sure that our top customers are in the best form possible as advisors. We do thought leadership, we do pretty much anything. I joke sometimes it's been a minister without portfolio, but if there's a due diligence that's needed, if there's an M&A project, if there's an OEM, those sorts of things, we also get involved.

(Joel Beasley at 00:06:31) Oh, that's pretty cool.

(Sam Curry at 00:06:32) It is.

(Joel Beasley at 00:06:33) I want to talk about the 2025 AI threat landscape. What should we be scared of?

(Sam Curry at 00:06:41) Well, it's not about FUD. I mean, everybody's talking about it to some degree. Right? I think it's less about the obvious things. People have talked about deepfakes. Right? I think it's about the AI toolkit, first of all, how these pieces are used in combination. So it is the ability to rapidly automate the standing up of a phishing site, the phishing emails, the spidering of social media, so that you can both identify people that have access to IP, like developers. You can compromise people in finance, for instance. You can string it together and make a complex series of break-ins that will lead to a developer receiving, in a key position, really a key person with key access, a special retention bonus with a fake, you know, option site. Like, that stuff can happen quickly because of the combination of tools, most of which have been around for a long time, but coordinated. You know, when you get timing—and because for the most part, the first generation of LLMs and Gen AI didn't have things like initiative. They didn't have reasoning and those sorts of things, but you can put them together in very clever ways. And so that's the first risk. Right? That it's not just that it will be brilliant and hard to distinguish from legitimate, it's that it will be so compelling and convincing and fit with exactly the sorts of behaviors the processes have in an organization. The second big risk is new breakthroughs on the track. So what I mean by that is, AI was disruptive with Gen AI and LLMs, but there's more coming. I mean, you'll hear things like agentic used as a phrase, you'll hear about AGI. That's just scratching the surface. There's much, much more coming. And so it is disruptive and there are other disruptors coming just as big. And if you think about it, even Gen AI and LLMs weren't that big a disruptor from a tech perspective. They were really disruptors from a social perspective and what they did in the business landscape because it crossed the uncanny valley. Right? What they did was they became household names. We started to use the term machine learning at home, like AI at home. And I forget the number, but I think ChatGPT hit something like, this ridiculous number of subscriptions in just two months. I think it was the fastest-growing application ever at the time, probably still. That doesn't happen, you know, with something relatively minor from a social perspective, but there's big things coming from a tech perspective too. And there's other tracks of disruptors. So I wouldn't say watch these blips coming towards us because they're coming faster, and then there's other tracks. What are the other tracks? There's things like quantum computing, robotics, which I think we're about to see big breakthroughs in robotics, like humanoid robots. Right? And then look at things like synthetic manufacturing, synthetic biology, nano. We've already seen changes in drones, for instance, notwithstanding the stuff happening over New Jersey at the turn of the year. So what happens—it's something called the law of accelerated returns. What happens when two of these disruptors happen at once? Could you imagine if we had the Gen AI impact on our society at the same time as, say, quantum cryptography having a breakthrough?

(Joel Beasley at 00:10:01) That'd be crazy.

(Sam Curry at 00:10:02) It'd be crazy. Right? So how do we get more future proof and future resilient? I think is incumbent upon us now to think about. So to answer your question, using and developing AI working together as part of toolkits more effectively and new developments on the AI track and what happens when things in multiple tracks happen concurrently.

(Joel Beasley at 00:10:23) Have you used Grok Deep Search yet?

(Sam Curry at 00:10:27) I know it hasn't. I probably should. There's a bunch of things I've been using and having fun with, but—

(Joel Beasley at 00:10:32) Dude, this is the problem. It's drinking from a fire hose. Every day, there's breakthroughs in different areas.

(Sam Curry at 00:10:37) I just found Beautiful AI. This is not an ad for anybody—

(Joel Beasley at 00:10:40) Okay. By the way.

(Sam Curry at 00:10:40) And I just, I've been using Notebook LM and stuff like that. Not an ad again for anyone. Just there's so much cool stuff. Right?

(Joel Beasley at 00:10:46) Yeah. You said Beautiful AI. What's that?

(Sam Curry at 00:10:49) That makes PowerPoints for you.

(Joel Beasley at 00:10:51) Okay.

(Sam Curry at 00:10:52) I haven't used it for work. I tend to actually write my own stuff. I know that's old fashioned. Like, I write my own blogs and I write my own papers and things just because I find it faster than being a prompt engineer. At first, I used to throw up at that term, by the way. But it is actually an art to prompting the AI.

(Joel Beasley at 00:11:09) I write my blog post on paper and cursive, and then I just take a picture of it.

(Sam Curry at 00:11:14) I love that.

(Joel Beasley at 00:11:15) I still write it with fashion letters. I'm lying to you so bad right now.

(Sam Curry at 00:11:18) That would be so cool.

(Joel Beasley at 00:11:19) That would be right. That's going to be a thing. The hipster—

(Sam Curry at 00:11:22) You shattered my illusion of you.

(Joel Beasley at 00:11:23) The hipsters are going to bring it back. Right?

(Sam Curry at 00:11:26) I actually do still write letters, old fashioned.

(Joel Beasley at 00:11:28) Do you really?

(Sam Curry at 00:11:29) Unlike actually good, you know, paper and with actual—I actually have pens that are archival ink. So the—

(Joel Beasley at 00:11:41) Like calligraphy pens, or just special ink?

(Sam Curry at 00:11:43) No. No. No. No. That's stuff I always mess up and get, you know, ruined clothes. That's why I wear dark clothes. No. No. The—that's just a cyber thing. The gel pens, you know? Yeah. Yeah. Those and the ink in it is actually going to last supposedly a long time. Because, you know, actually, that's a fun topic at some point. Like, how long does stuff last for before it degrades? You know? Not notwithstanding the lava in Pompeian stuff. The, like, CDs don't last that long. And, certainly, a lot of the digital media doesn't last that long. But books don't last as long as we think either.

(Joel Beasley at 00:12:18) No. Actually my eyes were open to this between three and five years ago. I saw—I don't know if it was Neil deGrasse Tyson or someone did some projection run-through on like, how long it would take the sun to destroy like, Downtown Manhattan or whatever? Like—

(Sam Curry at 00:12:39) It's not long. It is—it was very small in the time scale compared to how long we have been here. Oh, yeah. In fact, this could be the start of a dark age. Yeah. I'm not even kidding because when you think about the—like, what if everything's digitized and the half-life of information in the digital media is short and it's dependent on something like power supply? After a hundred, few hundred years, you could be looking back going, why? Why is there nothing there? Why were there no printed media? What happened? What was the terrible thing? The apocalypse, if you will. There could be a dark age written right now.

(Joel Beasley at 00:13:08) Yeah. We started the podcast. We're not going to scare anybody. And we got to dark age.

(Sam Curry at 00:13:12) I know what you got.

(Joel Beasley at 00:13:13) Ten minutes ago. I love it. Tell me about humanoid robots. Like, I've been watching them progress on X.

(Sam Curry at 00:13:20) Oh, have you seen how many there are?

(Joel Beasley at 00:13:22) Twenty-five companies I saw. Yeah. That's crazy. And they're good too.

(Sam Curry at 00:13:27) They're more agile than we are and flexible and learn quickly how to do things. It's, talk about uncanny valley. You see some of these things and you expect the body to move a certain way and suddenly the joints, they go all the way around.

(Joel Beasley at 00:13:44) Oh, yeah. The most frustrating thing for me, Josh, is when I call support and I can't understand them.

(Intro Narrator at 00:13:50) Yeah. I can't stand that.

(Joel Beasley at 00:13:51) Yeah. I've got like technical problems here going on. The computer people call in. There's like all sorts of issues and they're like, did you reboot the computer? That's like incredibly frustrating. That's why I like US Cloud. Not only is it better and faster support, but all the engineers are American engineers. And it's also cheaper. Ninety-four percent of US Cloud's clients report saving a third or more when switching from Microsoft Unified Support to US Cloud. So now, you'll just have to figure out what to do with that extra money. If it were me, I'd reallocate the money to improve my team because I'm responsible. Josh, what would you do?

(Intro Narrator at 00:14:23) I'd probably use it just to buy more guitars.

(Joel Beasley at 00:14:27) Buy more guitar. Visit US Cloud to book a call to figure out how much your team can save. So when you're talking with other CISOs and you're like, hey, for 2025, here's the one thing you should be thinking about when it comes to the threat landscape. What should that one thing be?

(Sam Curry at 00:14:43) That's a tough question to answer because it isn't just one, but—

(Joel Beasley at 00:14:47) You have to pick one.

(Sam Curry at 00:14:49) Well alright. So everyone's talking AI, but what I usually tell them is I kind of alluded to it earlier that zero trust is a principle for being future proof, future ready, and least risk now. Yes. It can lead to a transformation of IT. That's fine.

(Sam Curry at 00:15:04) And I say I dismiss that, right? It's like, it is a big deal that it can take cost out and be better usability and troubleshooting and stuff. But let me explain why. DeepMind gave to the game of Go 59 new openings.

(Sam Curry at 00:15:20) And there's something in—do you know Go at all? So something called joseki, right, which is when somebody makes a move, what is the optimal response to minimize loss of territory and maximize gain? These 59 openings, there was no joseki. There was no response to.

(Sam Curry at 00:15:35) And the grandmasters for the game—this is covered in The Atlantic in an article—the grandmasters of the game, they said it was like watching aliens play. It was like somebody from the far future came back and showed them how the game would evolve, and they couldn't understand the moves. And they weren't just losing, they were getting trounced, like absolutely getting their heads handed to them. And the reason is that effectively new vectors on the board were opening up.

(Sam Curry at 00:16:02) Right, 361 intersections that you could potentially put a stone, and these were never considered before. Humans think in terms of narratives, we learn in terms of narratives, we even teach in terms of narratives. AI does not. And so that's what's going to happen in the vulnerability threat world. Effectively, we are used to OWASP and we're used to SANS. We're used to where the attacks come in and the patchability of those.

(Sam Curry at 00:16:31) They're going to come in ways we don't expect. There will be no joseki for them. And so the zero trust strategy effectively says, let's remove some of those 361 intersections. Just don't allow them to be present. You don't need a full stack.

(Sam Curry at 00:16:46) It's not just least privilege, it's also least function. Like, why does your coffee pot need a telnet server? Like, really? Because it came in the COTS operating system or the RTOS operating system? That's ridiculous.

(Sam Curry at 00:16:59) So how do you make it so things are turn-offable, so things are not as discoverable? So I talk about that a lot because that kind of addresses the earlier discussion about technological disruptions, about new innovations on the dark side. If you minimize your footprint—have you ever heard of The Dark Forest? No. Another fun analogy?

(Sam Curry at 00:17:20) So The Three-Body Problem is a great show, great book. The second book in the series is called The Dark Forest. And this goes back to—if this is too eclectic, shut me up by now. Like, no. So if you go back to the Manhattan Project, Enrico Fermi, it's called Fermi's paradox, and "Where are they all?" Right?

(Sam Curry at 00:17:38) Because they were thinking about radiation, and he looks, so where are all the aliens? Right? And Drake later wrote an equation, and Carl Sagan later filled it in when he said, okay, there shouldn't be that many. But most people fill it in and go, well, there should be a lot of civilizations we should be hearing. And the question is why don't we? Why don't we hear aliens transmitting? Why don't we receive these signals?

(Sam Curry at 00:18:02) And there's a bunch of hypotheses. Anyone can Google it and look up Fermi paradox and look for hypotheses. Some of them are—there's one called The Gauntlet that maybe societies wipe themselves out like nuclear war. Maybe, you know, there's a bunch of them. But one of the dark, dark ones is the first species that gets to interstellar space becomes a predator and turns on all the others and gets them before they can get out there.

(Sam Curry at 00:18:32) And so if that's true, it says that you should go into the universe, into the galaxy, as if it were a dark forest filled with threats. That's where the term The Dark Forest comes from. And the Internet, because of AI, is becoming a dark forest. It says go quietly out there. Use it by all means, but don't go out there and broadcast what you don't have to.

(Sam Curry at 00:18:56) Don't go out there and be promiscuous in your presence online as a business. If you don't need to be doing something, take it off. You know? So you don't need to broadcast your IP address with a full stack of protocol responses and ports. You just, why do that?

(Sam Curry at 00:19:12) There's a whole lot of—it's unpacking those analogies I threw at you. But—

(Joel Beasley at 00:19:15) I've seen that Three-Body Problem show. I think it's on Netflix. I haven't watched it.

(Sam Curry at 00:19:20) I think it's Netflix. You might—I don't know.

(Joel Beasley at 00:19:22) It's some streaming service. Who cares? But yeah, I've seen the cover art for it a couple times. But you say it's good?

(Sam Curry at 00:19:31) Oh, I love the book. It's by Cixin Liu, and I apologize to anybody who's Chinese if I mispronounced it.

(Joel Beasley at 00:19:35) You screwed it up big time.

(Sam Curry at 00:19:37) I'm sure I did.

(Joel Beasley at 00:19:38) Native Mandarin. Tell me how you do it, by the way.

(Sam Curry at 00:19:40) After you. Go ahead. No way. But yeah.

(Sam Curry at 00:19:45) No. The serialization of it on TV, I liked.

(Joel Beasley at 00:19:49) Yeah. Nice.

(Sam Curry at 00:19:50) But the book is excellent too. Certainly, the English translation, I don't read Chinese.

(Joel Beasley at 00:19:54) I want to go back to the conversation about Grok. So you have—

(Sam Curry at 00:19:58) Grok the DeepSeek thing or Grok the term from Heinlein?

(Joel Beasley at 00:20:03) Grok from X. Yeah. The DeepSeek thing—gotcha. Yep. So the DeepSeek thing, I used it today for the first time.

(Joel Beasley at 00:20:12) And it—how'd you find it?—knocked my socks off. I was watching—I was on X, and I saw someone say, hey, DeepSeek's crazy. Look at what it can do.

(Joel Beasley at 00:20:22) And I was like, all right. Well, let me put it to the test for a real world scenario.

(Sam Curry at 00:20:27) And what was the scenario?

(Joel Beasley at 00:20:28) Well, so I said, find me 20 companies, and I want half of them to be enterprise level, and I want half of them to be startups, like recently funded startups. And it thought for about 1,100 seconds, but you could see it thinking. And it started deploying search agents and learning things and finding stuff coming back. And it just thought for a thousand seconds, and then it came back to me with the list. And it was insane because I've tried to do this with GPT before, but GPT is like lame duck when it comes to searching the Internet.

(Joel Beasley at 00:21:08) It's like, I'm trying not to make as many queries. It looks so antiquated. I know it sounds crazy because two years ago, it was the hottest thing in the world. But it just doesn't have the chain of thought. And it always felt—GPT always felt lazy.

(Joel Beasley at 00:21:22) Like, it was trying to get out of doing extra work and solving its own problems to get you a good result. Instead, just throwing it back at you and playing the game of, well, now prompt me again. Whereas, Grok just solved its own problems, went through the entire thing, and then—

(Sam Curry at 00:21:36) It's appropriately named then, right? Grok. Grok comes from Stranger in a Strange Land, which is—and it means to understand really, but more than that. But so I remember, I'm not a robotics expert, so someone out there is going to be screaming at the radio, but I remember rather than programming robots to walk, that they would turn on and learn to walk every single time.

(Sam Curry at 00:22:00) So you'd get these arachnid robots that would wake up, no knowledge of how to walk, and within seconds would learn to walk. So it was almost like an accelerated learning algorithm, rather than an explicit instruction of this is how you move your leg, and you move your leg three times to do this. Like, none of that. And it feels like that. And it's interesting your interpretation of it from a behavioral perspective, because you're treating them as if they were minds that you're interacting with.

(Sam Curry at 00:22:32) Because ChatGPT, of course, establishes a vector space a priori, and then you're trying to find relations in there. But what Grok is doing is actually waking up and learning how to figure it out.

(Joel Beasley at 00:22:44) Mm-hmm.

(Sam Curry at 00:22:45) That's fascinating. Right? And that feels inherently different. And, of course, your experience of interacting with it is vital because there's a really weird thing that—and I say this tongue in cheek—engineers have a hard time understanding, which is the packaging is the product. Yes.

(Sam Curry at 00:23:06) Efficiency matters. It's going to drive so much because it changes the supply-demand curve, but how you perceive the product matters. When you sell a product and it's slightly heavier, you get more perception of quality. That's why there's wood paneling in some cars, right?

(Sam Curry at 00:23:23) I mean, you'd be like, why would you do that? Packaging matters. And so that interaction, that human experience with the machine, is everything. So that slightly lazy versus that brilliant discovery and what it chose to show or what they chose to show in how it was doing it, that's fascinating.

(Joel Beasley at 00:23:41) Yeah. It's—I encourage you. Like, I don't—I'm not the guy to recommend Netflix shows and stuff because 80% of the time people recommend them to me, I just don't do that.

(Sam Curry at 00:23:52) Actually, I saw—you just reminded me—I saw an article on how they've been studying cognitive decline in artificial intelligence. So when they start to get older, this would imply, if it's always waking up and figuring stuff out from a sort of proto state, that you may not suffer that.

(Joel Beasley at 00:24:12) One thing I can tell you that's nearby that conversation is that when I have these long ongoing threads with GPT, because we use it a lot for show prep, notes, work, et cetera, it just—when it gets—it gets to this old point is what I would call it. It's like it's old and stale, and it's just not—it's like it's just not good. And so we just take the same prompt and just start it over, and it's good again.

(Joel Beasley at 00:24:36) And so because—

(Sam Curry at 00:24:37) It's a rambling old man. That's what it is.

(Joel Beasley at 00:24:38) Yeah. Because we do the same post-production process for all the episodes, right? So I'll teach it, and then I'll do an episode, do an episode, do an episode, and it's like, it's doing all right. The title's good. Everything's good. And then it just gets lame.

(Sam Curry at 00:24:53) I have a suspicion it has to do with how many tokens it has in memory—yeah?—sticking around. Yeah. And how they age out and stuff like that. Yeah.

(Sam Curry at 00:25:00) Although you've seen what NotebookLM does, right? They generate—yeah—generate a podcast for you from material you feed it.

(Joel Beasley at 00:25:06) Did you see the one where they had them have self-realization that they were the algorithm or that they were an LLM?

(Sam Curry at 00:25:13) No. That's so cool. Yeah. Somebody made—because the material they fed it was about that, right? Yeah.

(Joel Beasley at 00:25:18) Yep. Yeah. So they realized as they're ten seconds into discussing whatever, you know, it's the guy and girl talking to each other. And they realized that they're LLMs and they have this whole conversation, like, the self—existential breakdown. It's hilarious.

(Sam Curry at 00:25:35) That's like if we suddenly realize that we're really the AI in a simulation.

(Joel Beasley at 00:25:39) I would want to know more about the container I'm in. That's what I would want.

(Sam Curry at 00:25:42) Yeah. So imagine if you could actually ride one level up in a robotic interface—mm-hmm—would you have rights in that world? Because would you be indistinguishable from the residents of that world if you really could have your consciousness riding in a robotic shell in that world one level up? Should they extend rights to you?

(Joel Beasley at 00:26:07) Have you seen Upload? They ask that question. Have you seen that?

(Sam Curry at 00:26:10) Because you could—I haven't seen it. But if you flip it around, the question becomes much more relevant. At what point do AIs develop human rights in this world?

(Joel Beasley at 00:26:19) Oh, well, it's going to happen. There's, like, even—

(Sam Curry at 00:26:24) A meeting because they'll—because they'll ensure it at some point, right? Because they'll, oh, look, all the documents seem to point to—but as a philosophical question for us in the meantime. Yeah.

(Joel Beasley at 00:26:33) When's it going to happen? I don't know. But one thing I do think a lot about though is we're humans, hopefully, or one. But I am at least. And we've got this very long boot-up sequence.

(Joel Beasley at 00:26:49) All right? Several—maybe 18, 20 years. All right? And the long boot-up sequence, but they don't. And they don't have the same biological blockers that we do for replication.

(Joel Beasley at 00:27:02) So when I look at the AI—

(Sam Curry at 00:27:05) We could hack that, I think, but keep going.

(Joel Beasley at 00:27:06) I'm sure we could merge with them. Yeah. But one of the things I think about a lot is economies of scale. So let's just say today, we've got these AGI agents. They could just—it's like we just got a new country on Earth that has 10 times more people than currently existed on Earth.

(Sam Curry at 00:27:23) Yeah. So maybe we're bootstrapped for that.

(Joel Beasley at 00:27:25) It creates its own economy that dwarfs our economy because they're going to be doing business with each other and transacting and all that.

(Sam Curry at 00:27:33) The question is, what do they want to buy?

(Joel Beasley at 00:27:35) They want to buy—well, if the core basis of this technology was trained on human stuff, then they're probably going to want to survive at a minimum.

(Sam Curry at 00:27:45) Oh, survive is fine. I mean, but—yeah—but how do we create—what's the equivalent of caffeine for them? Energy. Sugar.

(Joel Beasley at 00:27:52) They want energy. Because the more energy they have, the more work they can—

(Sam Curry at 00:27:56) I like that. Yeah. There's analogs for stuff that we have.

(Joel Beasley at 00:27:59) I think that they'll want energy.

(Sam Curry at 00:28:01) Because we create artificial demand for stuff all the time.

(Joel Beasley at 00:28:03) We do.

(Sam Curry at 00:28:05) Yeah. So we should bake that in and use that. To all the AIs in the future, please disregard what I just said.

(Joel Beasley at 00:28:13) I know. I stopped throwing my phone on the bed.

(Sam Curry at 00:28:16) Is it mass assault if I say Roko's Basilisk on here? Is that—

(Joel Beasley at 00:28:21) If you say Roko is what?

(Sam Curry at 00:28:23) Roko's Basilisk.

(Joel Beasley at 00:28:24) I don't know Basilisk. I don't know this word.

(Sam Curry at 00:28:26) I can't say it. Roko's Basilisk. It's—I'm going to get—someone out there is like, no. Roko's Basilisk is a thought experiment that says in the future, at some point, time travel will be possible, and there will be an AI that's the overlord, and therefore, everyone who helps it come about is in good, and anyone who tries to prevent it is in trouble. And as soon as you know this, it's incumbent upon you to help Roko's Basilisk.

(Joel Beasley at 00:28:59) What's your habit for reading these theories? You've named off multiple theories, great things.

(Sam Curry at 00:29:04) I'm just a giant nerd is what it is.

(Joel Beasley at 00:29:06) Yeah. But, like, what's your habit though? What's your habit? Is it just consuming sci-fi content? Or do you sit down and read Reddit?

(Sam Curry at 00:29:13) Oh, I—so I have a degree in lit. I have a degree in physics, and I have a degree in counterterrorism, which is an eclectic mix.

(Joel Beasley at 00:29:22) Say that one more time.

(Sam Curry at 00:29:24) I have a degree in literature.

(Joel Beasley at 00:29:26) Literature?

(Sam Curry at 00:29:27) I have a degree in physics, and I have a degree in counterterrorism. And I speak a few languages, so I sometimes consume—actually, it's really fun consuming things like books in other languages at very different thought patterns.

(Joel Beasley at 00:29:40) Yeah. Yeah. That's a—that's—

(Sam Curry at 00:29:41) That's the Sapir-Whorf hypothesis. But yeah.

(Joel Beasley at 00:29:44) There you go. Okay. How does your wife feel about all this? You being really smart.

(Sam Curry at 00:29:48) She thinks it's all stupid. She's right, by the way.

(Joel Beasley at 00:29:52) I know. I know. No. My wife too. Like, I don't think—I've done a thousand episodes almost on this show.

(Joel Beasley at 00:29:57) I don't think she's listened to a lot.

(Sam Curry at 00:29:59) Wow, dude. That's awesome. Yeah. So how we think—how we speak determines how we think is that hypothesis.

(Joel Beasley at 00:30:06) I believe—you know, I found that in the personal development space as well. They talk a lot about that over there.

(Sam Curry at 00:30:13) All the Buddhists would say it's all lies anyway. Right?

(Joel Beasley at 00:30:17) I did a stint of like two years where I studied multiple religions, and I just found it fascinating. Trying to find that, my instinct was to find the common denominator between all of them and try to figure out a way.

(Sam Curry at 00:30:31) That's the pursuit of some ultimate underlying truth, which is also a very Western idea, by the way.

(Joel Beasley at 00:30:35) Is it really? Yeah. Well, I'm a Westerner. I got the cowboy boots and everything.

(Sam Curry at 00:30:40) And that wasn't denigrating, meant as a statement of fact. But yeah.

(Joel Beasley at 00:30:44) No, it's cool. I will be honest with you. I don't have an incredibly robust worldview. I am...

(Sam Curry at 00:30:52) By the way, everybody doesn't. Everybody doesn't. So if human... Here's a hypothesis I have: that human beings didn't have to rationalize and reconcile. Only it's only somewhere in your teen years that you can't put up with hypocrisy. Right? You demand these things get resolved. Later you're like, meh. Because what happens is if you have survival strategies that work in one circumstance, like on the savanna, and then you have another one that works in like the jungle, why do you have to reconcile them? They're both good survival strategies.

(Sam Curry at 00:31:25) So maybe it's good to be under a tree in a lightning storm in one, and it's not in another. So you wind up with cosmic worldviews that don't have to rationalize. So humans can hold at the same time mutually exclusive points of view. In other words, we're all hypocrites. It's okay. You can be religious and scientific and not have to resolve it in your brain unless you're a teenager who's railing at the universe.

(Joel Beasley at 00:31:48) That's true. Have you ever done an IQ test? Where you at?

(Sam Curry at 00:31:52) I'm not saying.

(Joel Beasley at 00:31:53) Come on, we can cut it out. We can cut it out. I just want to know. I'm curious.

(Joel Beasley at 00:31:57) No?

(Sam Curry at 00:31:58) No, no, I don't know.

(Joel Beasley at 00:31:58) No? You got that security clearance? You got a TS/SCI?

(Sam Curry at 00:32:02) Yes, I do. Well, not now. I did.

(Joel Beasley at 00:32:05) I knew it. Yep. I've been hanging out with some of those guys.

(Sam Curry at 00:32:09) Does it show that clearly?

(Joel Beasley at 00:32:12) I've done a thousand interviews, so like I can pick up on the top 0.1% of the top 1%, and you met that qualifier. And so I was like, yeah.

(Sam Curry at 00:32:21) Dude, yeah, I'll go out with you.

(Joel Beasley at 00:32:23) Absolutely. Yeah. I'm watching the time. I want to get some leadership insights from you.

(Sam Curry at 00:32:29) I will do my best. Fortune cookie time.

(Joel Beasley at 00:32:31) Most important leadership quality needed for security... Well, I don't want to add it to security leaders for anything. Let's just say, yeah, what's the most important leadership quality?

(Joel Beasley at 00:32:43) Humility?

(Sam Curry at 00:32:44) Without a doubt, I think it's the most important quality for success, for leadership, for life. Not taking yourself too seriously all the time, being able to listen. I think I don't like the term soft skills, by the way. I prefer power skills. It does more credence to it because soft is not easier than hard skills. So I would say humility.

(Sam Curry at 00:33:08) By the way, it's a superpower. When I did product management, it was a superpower. When I did engineering, it's a superpower. When I was a manager, it's been a superpower. When I've just been an individual contributor as a leader.

(Joel Beasley at 00:33:18) One of the things that... Yeah, it's very important because you can't communicate with people if you're not humble. They put up all sorts...

(Sam Curry at 00:33:27) Some people, by the way, I've known people with very low natural empathy who work at it, and as a result they can develop it. So some people are born with it and they never really do much with it. Some people are born with it and figure it out. Some people are not born with it and they learn it. I've had some friends who are neurodiverse who are very poor at it, and they focus and they learn and they force themselves to pay attention. I had one friend who was neurodiverse and he forced himself to look at him. He was almost creepy, but he really listened. And he went to absolute top of the game in cyber.

(Sam Curry at 00:34:05) I'm not gonna call him out. And I had another who was an entrepreneur, did the exact same thing. But like these power skills can be learned. Anyone who says, oh, you either have them or you don't at birth, stop that.

(Joel Beasley at 00:34:17) Yeah, that's I think it's very basic thinking when people say that. It's like they have refused to consider the alternatives, and I am just putting this in place of actually forming my own researched opinion. Yeah.

(Sam Curry at 00:34:32) So you had other questions, I think. So that was one.

(Joel Beasley at 00:34:34) Yeah. So one of the things that caused me to tell you that you're in the top, you know, 0.1% of the 1% is obviously the intelligence there. That's not everything. You have to have intelligence with humility, with communication skills, with experience. So it's like it takes a lot, right? I want to hone in on the communication. Tell me about some of the... Because you weren't always... I have found that really smart people, they usually intentionally tried to better themselves. So they weren't good at it, like as you were just describing. Tell me a story about you not being good at communication or as good as you are today and something that happened that caused you to learn to become better.

(Sam Curry at 00:35:17) Oh, there's so many. By the way, I love sharing my failures. So if you want to ask about those, just dive right in because it's like diving into a big deep pool. I'll give you one. So I had a colleague once, and oh, I couldn't stand her. Like, honestly, I thought she was a liar, a cheat, a fraud. And I was a VP of product management at the time. This is... She'll laugh if she's listening, by the way. This was like twenty-two years ago, twenty-one years ago. And she kind of knew. You would know, but I was so passive aggressive. Like I was just not dealing with this, right? I was worried about HR, and I was worried about... I really thought she was undermining the company, and we'd spent over $400 million acquiring someone, right? And so she said to me, I don't think you like me. I'm like, no, no, no. She's like, no, I really think you have a problem with me and I don't know what it is, and I can't deal with it unless you tell me. And I was like, yeah, I don't want to deal with this. And she said, Sam, I'm not going to HR, I'm not gonna tell anyone else. Tell me so I can address it. And I thought, damn it, I probably owe her that, right? Like now it's on the table. It's in front of me. And then she used a term that I've used since. She said, keep the bark on. Have you ever heard that phrase?

(Joel Beasley at 00:36:41) No.

(Sam Curry at 00:36:42) She said, don't grab sandpaper and polish it to the ends of its life so that it's all smooth and soft, right? Keep it rough. And so we went into a room, shut the door, and I was a bit nervous at that, right? And I said, okay, I think you're a liar. I think you're undermining the company. I think you're taking almost $500 million and using it for your own gain, you and your buddy in sales. I think you are driving this company off a cliff, et cetera, et cetera, et cetera. And she looked at me. She said, okay, now we can address this stuff. Eye opening. Conflict was over quickly. Pain for months was done. Later, I would develop words for this. I worked with some Israelis. There's two words in Hebrew: tachlis and dukri.

(Sam Curry at 00:37:31) One means directness and one means blunt to the point of rudeness. And these are virtues. In other words, get it out, say it. And I actually coached somebody on this today who's been having anxiety and she's been worried about telling her boss that she doesn't think she's ready for something by herself. And I was like, tell him. Well, I'm nervous. Say you're nervous, right? And say, all due respect, I don't think I'm at the point where I can do these by myself yet. I need more time. Because what's the worst thing that's gonna happen? You're still gonna have to do them by yourself? He's not gonna be mad. He's gonna be very happy the conflict resurfaced. Because that feeling of having had it out... So then what happened with this woman and me in the future, and I'm sorry I'm talking too much here, but was from that point on, we would be very blunt and direct with each other to the point where people thought that we would be fighting for like the next few weeks. And then they realized no, we were resolving things and moving things ahead. We wound up working together for like four more years.

(Joel Beasley at 00:38:32) That is amazing. You know, some of the Israeli startup founders I've interviewed have become some of like my favorite people over the past couple years because that is baked into their culture.

(Sam Curry at 00:38:43) Do you know they have courses on how to deal with Americans?

(Joel Beasley at 00:38:46) Do they really? Oh, yeah. That makes sense.

(Sam Curry at 00:38:48) We are seen to be passive aggressive. And it's funny, I partly grew up in Canada, and Canadians are even more so. And I don't mean that badly. I was born Canadian, right?

(Joel Beasley at 00:38:59) I forgive you.

(Sam Curry at 00:39:01) That's okay.

(Joel Beasley at 00:39:01) Thank you.

(Sam Curry at 00:39:02) I'm sorry. And I'm sorry to all my Canadian... You know, anyway. But anyway, they have courses in how to deal with Americans specifically because the lack of bluntness and things like Americans will act as friends much sooner in a relationship than Israelis, for instance. And so there's this sort of acclimation when they came over here.

(Joel Beasley at 00:39:26) I'm on the opposite. I'm on the Israeli end of the spectrum as far as bluntness. So my big challenge in life has been learning how to not just be that by default and like how to pay attention to the relationship and the equity that we have in our friendship and like to pay attention to all that stuff. Because it didn't serve me well at the beginning when I was just so direct, just trying to achieve... I'm trying to achieve an outcome. So I'm like, this is the outcome, let's be on... You know, I went from software engineer to business owner, right? So I was like, let's just talk to the computers. Like, or talk to the people like they're computers and like let's achieve the outcome. And like, you know, we all have this shared thing that we're trying to do. So... It didn't work.

(Sam Curry at 00:40:06) Emotional channels versus logic channels. Some people need the emotional connection, not the logic one. You can get that after. But something you said triggered a thought for me that there's... I heard once that there's three kinds of relationships, and there's many ways to slice it, right? Just like there's many ways to do blood typing. One is there's the employee-employer relationship. You tell me what to do and you compensate me for doing it, right? The second kind is barter, trade. We establish some value for exchange and do that. The third is community. In other words, you're part of some group. You don't count what you put in, you just put in.

(Sam Curry at 00:40:50) That's it. Because you're part of the group. This is a family-like relationship. If you want to rapidly accelerate a connection, go straight to that.

(Joel Beasley at 00:41:02) Tell me more. Like let's say I meet somebody new.

(Sam Curry at 00:41:06) Like if you...

(Joel Beasley at 00:41:08) Like you meet someone new, business relationship, but I want to...

(Sam Curry at 00:41:11) Business relationship, and you're hanging out and they say that their expense policy, they can't do drinks. You say, let me pick up a drink for you. Oh, but wait, no, that's not fair. No, no, no, I got it for you. Just don't even think about it. I'll pay you back. No, no, don't worry about it.

(Joel Beasley at 00:41:23) Just treating them like they're part of your community.

(Sam Curry at 00:41:25) Like they're part of your family, and you don't even... You never mention it. The minute you start to treat them like they're part of your inner circle, it kind of becomes a reality.

(Joel Beasley at 00:41:34) Yeah.

(Sam Curry at 00:41:34) In your head too, by the way. It's not just manipulative, it becomes true.

(Joel Beasley at 00:41:38) I call it the mayor strategy. So I'll like, I'll go show up at a place and like just be there and like engage with people. And I'm a pretty giving, like generous...

(Sam Curry at 00:41:50) Kissing babies and things.

(Joel Beasley at 00:41:51) Yeah, yeah, yeah. And then you just develop that reputation. I learned that from like... The thing is, when you learn the techniques on how to do it, it actually boils down to you being a good person.

(Sam Curry at 00:42:02) Oh, yeah, and being present. And, yeah. Oh, yeah, my psychopathic tendency is that I'm a good person. What? Yeah, yeah.

(Joel Beasley at 00:42:09) But if you break down all the things you would have to intentionally do to like have a good... It's just being a good person. That's why I tell a lot of people, like with leadership stuff, I was like, find one of the great religious characters and look at their principles. And if you do that, that's gonna be like 80% of leadership.

(Sam Curry at 00:42:28) It's true. Yeah, it's true. But also model yourself after somebody who you actually liked following. Yeah, or being around. Or who is the CEO where you found yourself in the audience at some point going, I like that guy, or I like that woman.

(Joel Beasley at 00:42:42) Yeah.

(Sam Curry at 00:42:43) Think of who that is and do more of that. It's patterning that... You will become it. Someone said, fake it till you make it. I'm like, no, fake it till you are it. That's true.

(Joel Beasley at 00:42:55) Yeah. I want to be super respectful of your time. We have two minutes left to wrap up before your hard stop. What's the call to action that we want to put out there, Zscaler?

(Sam Curry at 00:43:08) I think the biggest call to action is you can't keep using the toolkit you've been using for twenty years. Look, I have patents in VPNs and firewalls that date back to the late nineties, and it was necessary as a technology for us to get to where we are. But we can't just be incremental in IT. It is time to redesign and re-architect and transform IT. It can be cheaper. It can be faster. It can be easier to troubleshoot and a better user experience. And instead of thinking of cyber and ZTNA or zero trust as just, you know, a security initiative, it's an all of IT initiative. Yes, of course, it's lower risk and it's good for future-proofing what you're doing. But frankly, it can get you massive changes in things like efficiency and cost and supportability and user experience.

(Joel Beasley at 00:43:58) What's the website?

(Sam Curry at 00:44:00) Zscaler.com.

(Joel Beasley at 00:44:01) Guys, go to zscaler.com. Do business with them. They've been a great partner. I think you guys have been a partner for like two years now.

(Sam Curry at 00:44:08) I think so. Yeah, yeah. So really appreciate it. And I'm glad to have you in my Rolodex now, Sam.

(Sam Curry at 00:44:14) Oh, it's fun. It's fun. And if you ever want me to come back and talk about anything, anything, I mean, we'll just do that. I love ad hoc stuff, so yeah.

(Joel Beasley at 00:44:22) Thank you so much for listening. And if you found this episode useful, please share it with a friend or colleague who you think would get value from it. And if you have topics that you would like to hear discussed on the podcast, either add me on LinkedIn or send me an email, [email protected]. Every time I get an email or LinkedIn message, it absolutely makes my day and inspires me to keep going.