AI, AppSec and the Meaning of Life: The Answer is 42
With Chris Romeo, Izar Tarandach, and Matt Coles
What are the core competencies that matter most for modern application security teams? Today we discuss understanding code and systems thinking and the crucial ability to assess risk in context - plus why your AppSec team might eventually get absorbed into engineering (and why it could be a good thing).
About this episode
What are the core competencies that matter most for modern application security teams? Today we discuss understanding code and systems thinking and the crucial ability to assess risk in context - plus why your AppSec team might eventually get absorbed into engineering (and why it could be a good thing). We debate the role of developer mindset in security, the importance of technical depth over tool knowledge, and how to build teams that truly enable rather than gate development.
Follow AI Security Table:
➜ Home: https://securitytable.ai/
➜ X: https://x.com/AISecurityTable
➜ LinkedIn: https://www.linkedin.com/company/ai-security-table/
➜ YouTube: https://www.youtube.com/@AISecurityTable
Transcript · 45 min conversation
0:09Chris Romeo Hey folks, welcome to another episode of the Security Table. My name is Chris Romeo, joined as always by Matt Coles and Izar Tarandash. If you're watching on video, Matt went for a haircut, something I don't recommend in life at this stage. So don't do it.
0:29Matt Coles They will grow back, right? Knock on wood.
0:32Chris Romeo You know, over enough time. If you have a, also if you have enough, if you have an infinite room full of monkeys typing on typewriters, they will create the great works of Shakespeare after some infinite amount of time. So.
0:44Izar Tarandach Now it's AIs, not monkeys.
0:47Chris Romeo Then I create an AI. Yeah. Could you generate and to tell the LLM you are a monkey? How long would it take you to create the works of Shakespeare? No, I'm curious. I really want to go figure that out. I'm just going to push pause on the record. We'll be back shortly, folks.
1:02Matt Coles I mean, you know, you could also have the AI, you could have an AI jump, uh, you know, a model to a robot that can physically type on a keyboard.
1:12Chris Romeo I like where this is going.
1:14Matt Coles Yeah.
1:14Izar Tarandach Just think, just think about the prompt. Just think about the prompt. You are now an army of infinite monkeys.
1:25Chris Romeo Wait, if you, no, hold on, you just changed the problem. If you had an army of infinite monkeys, it would be instantaneous, right? They would create the works of Shakespeare instantly because there's an infinite number of them.
1:36Izar Tarandach Only if you have a quantum computer.
1:39Chris Romeo Oh, so we finally found it.
1:41Matt Coles Alternatively, you would, you would crush the processing capability of, of whatever LLM you're using because infinite monkeys would consume infinite resources. And it would just stop.
1:52Chris Romeo Oh, so we need—
1:53Matt Coles It would be a giant denial of service. It would be a giant denial of service.
1:56Chris Romeo So there's more conditions. You need infinite— Matt always threat modeling.
1:59Matt Coles Come on.
1:59Chris Romeo Where's the software? Just try to have fun with an example. And you're like, no, that's not going to work. Typical security person.
2:05Izar Tarandach That is a security person answer right there. I can already hear the marketing people from JCP and Azure and Amazon on our door saying, did we hear you need infinite resources? We have them. We have them.
2:20Matt Coles Well, now that we've solved that problem, do you, do you, I mean, so obviously the answer is 42.
2:26Chris Romeo Of course. It's the, uh, the meaning of life, the universe, and everything, right? If you don't understand the reference, put it into an AI.
2:35Izar Tarandach If you don't understand the reference, just turn off the podcast because you're not going to get the reference.
2:40Matt Coles That's true.
2:40Chris Romeo You're just not going to, you're not going to get us. You're not going to get us if you don't understand that reference.
2:45Matt Coles All right.
2:45Chris Romeo Don't forget your Don't forget your towel.
2:47Matt Coles I'm going to go, I'm going to go ask, I'm going to go ask. What is the meaning of 42?
2:54Chris Romeo Are you asking what is the meaning of life?
2:56Matt Coles No, no. What is the meaning of the quote 42?
2:59Chris Romeo Okay. There'd be a lot of different things. Now I'm curious as to whether it gives you a different, there's a couple of different answers I can think of. Because I think 42 is Jackie Robinson's baseball number.
3:12Izar Tarandach So many questions.
3:14Matt Coles The code for the Q is famously from Doug Adams' comic sci-fi series, The Hitchhiker's Guide to the Galaxy.
3:21Izar Tarandach Nice.
3:21Chris Romeo Win for the AI. I tell you what, we can just stop talking now and have AI generate the rest of this podcast.
3:29Matt Coles It's also, it's also the ASCII code for, for the asterisk.
3:33Izar Tarandach Asterisk? Oh, you just reminded me that I left something cooking on NotebookLM and I have to see what came up.
3:41Matt Coles While fans have concocted numerous elaborate theories about why Adams chose 42, it happens to be ASCII code for asterisk.
3:50Chris Romeo I think that's what—
3:51Matt Coles he just chose it randomly, he says.
3:53Chris Romeo I think that's what Izar was using on his notebook LLM thing. It was like, what's the meaning of life is what he had asked it.
4:01Izar Tarandach No, actually, this time I'm looking for something serious. I want to better understand The connection between LLMs and knowledge graphs.
4:12Matt Coles I had the same question.
4:14Izar Tarandach Oh, cool. I wonder if they cheated.
4:16Matt Coles I wonder if they talk to each other.
4:17Izar Tarandach I'll tell you offline why I'm asking that. Shoot.
4:24Matt Coles Wow.
4:24Chris Romeo So this is the podcast.
4:27Izar Tarandach Yeah. Yeah.
4:28Chris Romeo This is FYI. It's, it's not just the beginning of a conference call before a work meeting starts. We get bored talking about the weather. Hey, what's the weather where you live?
4:40Izar Tarandach You call this life?
4:43Matt Coles Yes.
4:44Chris Romeo That's a whole other podcast episode right there.
4:46Izar Tarandach Whole other.
4:48Chris Romeo Okay.
4:48Izar Tarandach All right.
4:48Chris Romeo Well, the question of the day is, if I were building an AppSec team today, I'd focus— what skills, what technical skills would you focus on? Matt, are you putting this question into an LLM?
5:00Matt Coles I am.
5:01Izar Tarandach Just kidding.
5:03Chris Romeo Every time you hear typing in the background, someone is talking to an AI to try to sound more informed about the topic.
5:11Izar Tarandach Hands up.
5:12Chris Romeo Hands up. Let me see your hands.
5:15Matt Coles Same for you, Chris.
5:17Chris Romeo Listen, I don't believe in AI. Actually, I used AI to fix a construction engineering project. Of all things. So we built a, uh, like a platform deck and we were having some sway problems where it was moving too much at the top. And I literally went to ChatGPT and I gave it the parameters and it gave me a solution, which, uh, which we've installed and it worked.
5:43Izar Tarandach It was like, it works.
5:46Matt Coles If you ask that same question 10 times, you may not get a working functional answer.
5:50Chris Romeo They'll, you know, one of them was just knock it over. That'd be one of the answers, but. Okay. So, on a serious note, trying to find a way to pivot this conversation. You're building an AppSec team today. What are the technical skills that you would focus on? So, this was a post I saw come by on LinkedIn. There's a few different answers that the author posted, but I'm not going to share those yet. We can talk about those after, but I'm curious how the two of you would answer this question cold.
6:20Izar Tarandach It needs to be technical.
6:22Matt Coles Yeah, I was gonna say, that's the, that's an interesting parameter change. Not a parameter change, I see it here from the original discussion, but, uh, yeah.
6:32Chris Romeo 3 core technical skills.
6:35Matt Coles Right. So not the soft skills. We'll ignore that for the moment.
6:38Chris Romeo Okay.
6:39Izar Tarandach So here's an interesting thing. Technical. You hear technical in our, in our area, at least you immediately think code, lines of code, and being able to connect two things together and make them talk and stuff like that. But I'm going to offer that in security, something like being able to properly assess risk is as technical as anything else.
7:08Matt Coles Yeah. Okay. What does that mean?
7:09Izar Tarandach Hold on.
7:10Chris Romeo Hold on. What does that mean? I want to know what that means. —is a loaded term.
7:15Izar Tarandach Taking the stuff that comes out from the 3D scanner and actually understanding how it applies to your environment and interpret that in light of a number of other factors that's not necessarily a part of the calculations of the scanner, that that is as much a technical skill as writing lines of code or reading lines of code.
7:39Chris Romeo I almost agree with you, Azhar. Yeah, I almost agree too.
7:42Matt Coles Sure. Well, we're done. So almost is good enough, right? Horseshoes, hand grenades, and what else? Anyway, so the only, I think the only thing I'm sort of disagreeing with you on, I don't think, I think you implied it. I don't think you've made it explicit. That's a very reactive thing, what you described. So you're talking about scanning, you know, taking the output of scanners and identifying risk. By the time you have the ability to run scanners, you've already designed and implemented a system.
8:21Izar Tarandach No, no, no, no, no, no, no. Sorry. You went the other way. I'm talking about, I'm talking about interpreting some kind of risk in light of what you have in where you have. Doesn't, doesn't depend on the scanner or whatever.
8:33Chris Romeo Yeah, this scanner part threw me too because I thought like all of a sudden you went right to tools. No, no, no, no. Sorry.
8:39Izar Tarandach Sorry. As a vehicle. Let, let me try to recover that. Yes. Recover. Recover is good.
8:47Matt Coles Um, what I hear you saying is, uh, I, I think that, yeah, people, you need context.
8:55Izar Tarandach You were, you weren't there. You had to be there. Anyways. Um, what, what, what I mean is that, uh, the ability of looking at something that's given to you as an absolute, as this is the risk because somebody said so, and being able to understand that risk in the context of the place where you are, the place where you live, the place where you play, that's a technical skill. That does not translate to reading or writing lines of code. Sure.
9:29Chris Romeo Why do you think that's a technical— like, what do you— I agree with you, but I want to explore deeper into why. What makes that a technical skill? What do you have to understand that makes you categorize that as a technical skill?
9:44Izar Tarandach I think that you will have to have a basic understanding of systems to know where the input is coming, where the output is going. And what's transforming there in the middle. You're going to have to have a bit of understanding of how bad things can get. So if somebody actually pulls that cord, what's the worst thing that could happen and how that's going to impact the rest of the system. So it's a bit like, I'm thinking in terms of something that I heard years ago, that there was a movement to, in medicine schools, to create technicians that could interpret results and actually understand the systemic part of just that bit of the body. You know what you're saying, right?
10:39Matt Coles The technical skill, I think you're, well, a technical skill, one of the technical skills I think you're looking at is Data science.
10:46Izar Tarandach No, no, no, no, wait, wait, wait, wait, wait. Let, let me keep that one open for the future. But this one specifically, this one specifically, I think that I'm, I'm looking for the technical ability to understand the system that you are responsible for and interpret risk in the context of that system. Does that make sense?
11:08Matt Coles It does. And it requires some analytics it requires technical understanding of a system, as you said. Well, what is that role? That sounds like a security engineer, but it's not, right?
11:25Izar Tarandach But that's the thing. I think that it's a technical ability that lives outside role. I think that every role in security should have some of this to some extent.
11:40Chris Romeo Simply because you have to understand how things work. If you don't understand how things work, you can't secure. I've always, I've said that for 25 years at this point. If you don't know how it works, you can't secure it. It's just not possible.
11:52Matt Coles Which means, right? So GRC teams look at compliance to controls, but do they have that understanding of how the system works?
11:59Chris Romeo Matt, stop, stop cussing on this program, please. You said GRC.
12:03Matt Coles Where's the, where's the jar? That's not in the jar.
12:08Izar Tarandach I have to bank some money.
12:09Chris Romeo I'm about to draw some banned hate mail here, but—
12:11Matt Coles I have to bank some money in that jar for this episode.
12:15Chris Romeo I mean, you mentioned compliance, which is just— I mean, compliance is— I don't even know if I want to go there because we'll lose like 3 listeners and we only have 5.
12:28Izar Tarandach Look, compliance is a funny thing, right? I keep thinking myself. And don't misunderstand me, I know some great compliance people who really, really know their security. I keep asking myself if compliance lives in the security team or not. And I keep going back to yes, they do, because they need the same visibility that the rest of the team has. But would the compliance person need this specific technical skill? I think that it would probably make their life, uh, no, I won't say easier, but it would probably give them a bigger picture than they have otherwise, but it's probably not a must.
13:19Matt Coles Can we take a step back actually? I'm wondering, you're a trainer out there. What's the role of an AppSec program? If today, like, is it changed, has it changed that will inform what the technical skills are?
13:40Chris Romeo At the end of the day, the role is to build secure stuff, ensure that the team's building secure stuff and private stuff. Yeah, but how, how much, right?
13:49Matt Coles So there's been this shift, I think, in AppSec to, from being the security team, you know, you're outside the development team and you're, You know, basically you own certain things and you take control of certain things and this is wall. And then over time, that's become more of, well, we're here to enable the business and therefore more, we get more integrated. At one time, it very integrated. Now it's sort of, we are here if you need us, but, you know, and we're gonna give you stuff to enable you, but it's really up to you, right? That, and so the, the role of AppSec has, has evolved over time. So if we look at it, what does it today, or more importantly, what should it be today if we were building it today?
14:31Izar Tarandach So, separating what it is from what it should be. What it is is a machine to transform budget into dashboard.
14:43Chris Romeo Oh, that's a quote right there. That should be a t-shirt.
14:46Matt Coles So, great. So, what you just said was replace the AppSec team with AI and be done with it.
14:53Izar Tarandach No, no, no, no, no, no, no.
14:55Chris Romeo Have you seen this? You can vibe code some dashboards, no problem.
14:59Izar Tarandach No, no, no, no. Have you seen the AI, the dashboards that come from AI? They are ugly.
15:04Chris Romeo But, uh, I mean, have you seen the dashboards that come from regular people?
15:08Izar Tarandach Then the people are ugly. But, uh, okay. Let me put it like this. So jokes aside, uh, almost.
15:18Chris Romeo Why? Why? What jokes?
15:20Izar Tarandach Oh, we're being serious. So the, the, the, the role today, the role today is definitely one of support, right? I think that most AppSec teams are going to come out and say, first and foremost, we are here to serve the business and help everybody do their business better. Yep. So I definitely think that that's an evolution. From when we used to say that it was a gate and whatever we say doesn't go. So I would say that today the first role of an AppSec team is to use marketing speak to enable the business to do what it does in a more secure way.
16:08Matt Coles I'm gonna object to it just a moment there. I think there is an evolution back to The security team needs to provide a little bit of weight. They don't, they should not just roll over on that enablement.
16:21Izar Tarandach But wait, wait, but there, there's a difference on what the function is and how it's achieved. A lot of teams, and I've seen this written in a couple of places, went all the way from the department of no to the department of yes, dear, why don't you do it just the way you described?
16:43Chris Romeo That's useless. That's useless to have that.
16:47Izar Tarandach That's useless. That's useless.
16:48Chris Romeo I mean, you might as well just, you might as well just, you know, reallocate those resources to something else if you're just going to let people do whatever they want.
16:55Izar Tarandach Which is basically what's happening with all kinds of artificial things. But I still think that the golden place for the AppSec team is yes, but. Yes, but you're going to have to do a couple of things to the right way so that you can do your thing the right way.
17:13Chris Romeo Yeah, that's always been the right direction. And I've heard people say that for maybe the last 10 years. You know, when we came out of the department of no, it was, you're not the department of no, you're the department of yes, comma, but here's the things you need to do.
17:30Matt Coles We've had this debate, I think, on this podcast about is it yes, but, or If yes, but we can, we don't have to re-debate that.
17:41Chris Romeo That's just a whole different algorithm. You just created an algorithm for this.
17:45Matt Coles So, but, but, you know, so this is the, this is the making security, the AppSec team basically the function of guardrails and paved roads, right? We said.
17:57Izar Tarandach Yes. Which brings us to the second.
18:01Chris Romeo Let me, let me, let me finish one more thought just to, to. I want to address something you talked about, what it is, what it should be. I'm going to now mention what it's, what I think it's going to be in the future. And that is, I think the AppSec team is going to get eaten by the engineering team. I think it should get, other than some enforcement condition, the AppSec team is going to morph into become part of the engineering team. Because think about where, think about what the ultimate goal, like you're the, the AppSec team is trying to influence the engineering team ultimately. And so if it, I think, I think it's a logical progression that's gonna happen over the next 10 years. I actually— With AI, AI mixing together. You don't even, you don't even need AI for this.
18:53Matt Coles If you treat it, if you had original, if, if we collectively in the industry had treated security the way we treat quality engineering, we would probably be in a better shape. Quality engineering has a seat at the table in every engineering discussion. They have weight to pull or stop releases, and they usually have a separate but connected function.
19:18Chris Romeo And we do— But under the same leadership, right? They typically sit under the same leadership. At the very top.
19:22Matt Coles At the highest, at the VP level. Right, the highest level. But, but, but at the, You know, the engineering lead responsible for making a delivery and the quality lead responsible for making sure it's not crap going out the door sit as peers, right? And so, you have a check and balance, right? And with the exception of the notion of shadow IT, the quality team manages a tool suite responsible for checking the work that comes out of development team, which has their own tool suite. And if you put security in that same bucket, then you have natural built-in check and balance with the same report structure. And it's, with the exception of those exceptional cases where you need some of that governance, that works for the most part.
20:10Izar Tarandach But Chris, when you speak about that, I hear in your voice the educator coming out, the trainer and the teacher. And you know very well that I am a huge proponent of developers doing their stuff the right way. But I think that you are being a bit too optimistic because this assumes that from a certain point in time, developers will know and will have all the tools to do security the right way. Going forward. And so that's—
20:47Chris Romeo I think that's going a step further than where I went. That, that would be a very optimistic future. What I'm describing is not that every developer is an AppSec person, but the function of AppSec is not going to exist as a separate team. I don't think so. I don't think it will in the future. I think eventually those resources will just be pooled, but I'm not saying that the AppSec team goes away. I'm just saying that they're not a separate entity that sits over under the CISO and it has the opportunity for a more combative relationship with the engineering team.
21:24Izar Tarandach So the, the world belongs to security champions?
21:29Matt Coles Uh, a little bit more than that. They need to be more than that because security champions are developers with some security work. As opposed to, I think, Chris, what you're saying. And again, similar to my analogy with the quality team, they need to be given the flexibility to be experts in their, in their areas because they bring that function to the business within the business. So security champions are not, I mean, unless you, unless you would have your security champions doing security research and, and, you know, being presenting at DEF CON and things like that, they're not That's not the— not necessarily, not yet at that skill level, necessarily, right? They may be really good.
22:10Chris Romeo The ultimate goal, I mean, like, to, you know, kind of where you went with your conclusion there, Yezar, has been my ultimate goal for the last 15-plus years of, I'd love to get to the point where developers have a foundational understanding of security. Now, have we made a lot of progress in the last 15 years? Yes, I think we have. It's better than it used to be. It's better than it used to be from a, from, from people that approach it and educate and invest in educating their developers is there's, you know, there's, I've got case studies, but I can't, I can't share them because those people haven't given me permission to talk in public about what I've seen in, in some different organizations. Industry-wide, yes. Okay. Maybe we haven't moved the needle, but for those people that have dedicated time and resources, you can move that needle.
23:02Matt Coles You can. Yeah, I agree, but it's, it is, uh, you're, it's supplemental learning, right? The, the, again, this, this analogy of, of quality, developers get taught quality in school as a base, you know, if they're going comp sci or they're doing many programming classes, they generally get taught. I don't know about that.
23:20Chris Romeo I heard some horror stories about comp sci graduates can't get jobs. I don't know if you guys saw this.
23:25Matt Coles Yeah. But that's a different, that's a different bubble.
23:29Chris Romeo I saw some people's comments that were talking about the difference between, now I'm going on a whole different tangent, but the, the difference between what's taught in comp sci now versus 20 years ago. Like 20 years ago you were writing device drivers and, and working with moving things around in memory, and now you're building web apps and cloud, you're building, you're deploying cloud services on AWS.
23:52Izar Tarandach But again, that's the whole divide that has always been there between computer science and software engineering, right? When I went into college, there was no software engineering, only computer science. And a couple of years, well, not a couple, but years later, the place where I went to added software engineering. Looking back, if I had to go today and if I was more cognizant of my lack of abilities, I would definitely have gone into software engineering and not computer science. But for a computer science graduate today to come and say, I can't find a job, my first reaction, right or wrong, possibly wrong, would be you're not looking at the right place.
24:46Chris Romeo And if you're looking— There's a lot of them that are— There's a lot of them that are stuck, but I want to draw this back all the way around to the first thing you said. Your computer science education is what gives you the ability to measure, to understand and measure risk because you have the foundational pieces for how things work. No, no, no.
25:03Matt Coles Oh, really? What kind of programs did you go through?
25:06Izar Tarandach My computer science education is what gives me the arrogance to suggest that just by looking at a system, I completely understand it.
25:16Matt Coles You have credence, dude.
25:17Izar Tarandach You have credence. Okay. Because when I, when I took my computer science education and Andrew Hart, it was one of the best the world could offer back in the time. It was imprinted into us that, and actually I saw some people say that later on in life in, in, in job interviews. Uh, I'm a comp sci graduate. I don't write code. I create algorithms.
25:49Chris Romeo Oh, that is, that is arrogant to say that.
25:52Izar Tarandach Yeah. Yeah. There was a hubris there in the, imbued in the, in the function.
25:57Chris Romeo Didn't they teach you how operating systems work?
25:59Matt Coles Of course they did. Yeah. And, and how programs are built.
26:04Chris Romeo When someone's trying to secure an operating system, if you don't understand Memory management, file management, uh, you know, various types of access control and the security functions that exist in those subsystems. That's what I'm saying is what CompSci gives you as a foundational thing to be able to understand risk. Yeah, but the— you understand how the pieces work under the hood.
26:24Izar Tarandach Yeah, but real, real, uh, real story. The moment you have somebody being explained how a RAM on disk works, And their first response is, well, great, now I have much more memory. Then you understand that, yeah, you can teach operating systems to a lot of people, but that doesn't mean that they learn, which goes back to what you were saying about developers. Yeah, we've been teaching security for a lot of time to developers. It does not actually mean that they learn.
27:01Matt Coles That's a pretty, uh, now you're talking about education systems and measurement of, yeah, which, which we have, which we have talked about a lot.
27:12Chris Romeo Look, it's, it's not pessimism.
27:14Izar Tarandach This is not pessimism. This, this is actually, I'm gonna throw it there. This is actually optimism because recognizing the problem.
27:23Chris Romeo Now you're in denial now.
27:25Izar Tarandach I can't see canoes. I can't see, I can't see crocodiles. I'm probably not in the Nile. But, but, uh, my point is that if we recognize that, uh, yeah, we're teaching and not everybody's learning, then, but somebody is, some people are, then just to connect to what you said before about the the team disappearing, but the function remaining, if I understood you right. The security champion does get a boost. And the developer teams that I grew up in, you always had the figure of the guru, the developer that knew the whole thing inside out, the guy that you asked before you started anything and read all the manuals. That was the guy that you went to and sort of got a blessing to start doing the thing that you wanted to do. I don't think that we're going to get there with security champions, especially because I don't think that security champion needs to be a developer. I think that there are other functions in the team who can make absolutely great security champions. But that's the limit of the involvement that I— that's the limit of the disappearance of the AppSec team that I see happening. In teams going forward. But you will still need somebody somewhere in the organization who will have those skills to look over the horizon and say, that's the next thing coming. That's the next thing that my— That's the office of the CTO.
29:07Chris Romeo I see. That was for fun.
29:11Izar Tarandach Look at Matt. Look at Matt. Look at Matt. It's like they got some ideas in there, they're just not coming out because the black hole doesn't spit light back to you.
29:51Matt Coles So the big thing about having an AppSec team part of an engineering organization is having a seat at the table and visibility on the views. So they either need to have, and I'm sorry, I'm stuck on this, they're part of quality engineering or a parallel organization at all. 'Cause quality and engineering have two different line items in a program plan, right? How does engineering feel about this release? How does quality feel about this release? And then leadership can make a decision. If security was one of those functions within the engineering team, but not part of the engineering line item, what does security feel about this release? And now you have a proper informed risk decision.
30:32Chris Romeo Now I'm hearing something in the way you're describing that, that's, I don't believe is universal. Sure. So what you just, the way you just described that requires an organization of a certain size and maturity to say they have a quality team. So take a 100-person startup right now. Do they have a quality team? Hopefully they have a qual— hopefully they know about quality, but they may not at a 100-person startup, right?
30:58Matt Coles Wait, but they don't think there's, I mean, although if you're, although if you're building an AppSec team, you're probably not that 100-person startup, right?
31:07Chris Romeo I don't know about that. I've seen teams building AppSec functions before. They would say they're building a quality function, a separate function. I mean, because, I mean, think about it, like, I'm somebody who's gone through the procurement hurdle a number of times on the vendor side, right? Like, you have to demonstrate various AppSec, let's call them minimum bars of entry, to even go through the procurement and sell your thing to somebody. So AppSec is something that's being forced into, you at least have to, have to have a story for how you're doing it to sell to companies of any size.
31:46Matt Coles But you can have a, you can have a GRC function or a good legal team and a pen testing company on retainer. You don't need an AppSec program at that level.
31:57Chris Romeo The things that I've seen people asking about, I would have to lie if I didn't have an app, if I hadn't at least considered them, which I didn't because the advantage of being an AppSec person and starting a company is that you can apply. You can be like, we got this. I know all about this. I read a post on LinkedIn. Let's deal with the other two things. So these are, you were trying to go to the next one and I took us back on a —10-minute rabbit hole. I think you were going to the developer mindset. Were you going to AI or the developer mindset portion?
32:32Matt Coles Wait a minute. So we weren't actually quoting that original thing, right?
32:36Chris Romeo I wasn't referring to that. I'm sorry, I thought that's where Yezar was going when we—
32:39Izar Tarandach No, we didn't quote it, but just to get a little buzz on that one because we're running out of time, but well, not quite. Yeah, sort of. But, uh, AI and developer mindset. I think that that post, to my point, the mistake that they made was that they were depicting an application security team that it's definitely turned all the way to automation and to serving the business with automation. And the tools that they found was that everybody in the AppSec team has to be a developer. With the developer mindset, which is a very different way that I interpreted where they were going with developer mindset, and throwing everything to AI because that's what people do today. So the developer mindset, I would agree that having a developer mindset is an important technical skill for an AppSec person, but not because of the reasons that the original post Mentioned. They wanted people to have a developer mindset so that they act as developers. I want people to have a developer mindset so that they can understand the developers that they work with.
33:54Matt Coles Well, also, I think you mentioned earlier, you said there's some—
33:58Chris Romeo They work together, the two things you just said, though.
34:00Matt Coles But I think you also said that from the system designs, you need to know how developers make the things that they do. That they make because all those, all those interactions, all the stages that development goes through. And by the way, uh, we probably should mention, right, that there's more to building and releasing something to a market than just the developers, right? There's the supply chain. There's integration, there's tooling, there's a release management or packaging or testing or manufacturing if you're doing physical products, right? So there's more than just developers in there that needs to be involved in AppSec.
34:36Izar Tarandach But I think that today things got to a point where most good developers, they are basically doing their stuff in a very Zen way.
34:53Matt Coles Like vibe coding way?
34:55Izar Tarandach No, no, no, no, no, no, no, no, but supply chain. Happens. Well, of course, deployment happens. Happens.
35:02Matt Coles Because there's other processes and other people who do all of that. Right. Right. And developers get input or they get the output of the supply chain into their thing and then they produce output for the next phase.
35:13Izar Tarandach So what I'm saying is that while you, you write that these things are part of the developer mindset, what I'm positing is that they don't take part of the developer mindset today. The developer mindset today is more of a, uh, uh, what's the size of the box of Legos that I can use to build this thing that I'm building?
35:33Matt Coles I wasn't suggesting it was just the developer mindset. What I'm suggesting is it's the system lifecycle mindset. Oh, okay. Because you need more than just the development view. Yeah. Yeah. Yeah.
35:44Chris Romeo And I like how they, I like how they went the other direction though. Normally when someone says developer mindset, it's about. How do we get the developers to think about security more? But that's not where they went. And so where this individual went with the definition, it was more about how can we prepare the secure AppSec team to understand and have empathy for the developers, but also understanding for how they're building things. Wait, wait, wait, wait, wait.
36:09Izar Tarandach They didn't say that I saw anything about empathy in there.
36:14Chris Romeo They weren't talking about empathy. I just added that in because I'm me.
36:17Izar Tarandach But that's the whole difference. That's the whole difference because What they suggested was that the AppSec team had to have a developer mindset to be developers themselves of that automation that's going to save the day at the end of the day.
36:28Matt Coles That's not the way I read that. No? I think I agree with Chris that— Wait. So the very first statement is, so I'll just quote it, right? Someone who can build, not just break applications. Yeah. And in order to do— so if you— so knowing how to build an application means you can, you can, you know where the touchpoints are. To be able to secure it.
36:48Izar Tarandach Agree? Yeah, but then, okay, so I'm cheating a bit and looking at the third technical skill that they support, and that's DevSecOps, the ability to actually build security automation pipelines, not just talk about it. So they built a progression in here. AI, developer mindset, understanding how things get built, And then DevSecOps. Why? Because they want the AppSec team to be an automation factory of security things.
37:21Chris Romeo I mean, when I see DevSecOps on this list, like, did you guys, did you guys used to read Wired magazine when it was in its heyday? Oh yeah. You used to? Okay. I stopped reading it now because I don't know. I just don't read magazines anymore. Oh, look at this. Do you remember? Do they still do the Tired and Wired?
37:37Izar Tarandach Oh yeah, they do, they do.
37:39Chris Romeo Okay, so I see DevSecOps these days, I think of that as the tired. It's like DevSecOps, tired. AI automation to build software, wired.
37:48Izar Tarandach Like, DevSecOps at the end of the day, I still remember when the, oh my God, I'm reminiscing a lot today. But, uh, the discussion was DevSecOps or SecDevOps.
38:04Chris Romeo Or ops, dev, sec, or there was a bunch of different permutations.
38:10Matt Coles Remember when it used to be dev and sec and ops separately? Yeah.
38:16Chris Romeo The good old days. The good old days.
38:17Izar Tarandach But at the end of the day, what did we get? We got the DevSecOps, right? Because if it happens, it happens. If it doesn't happen, then we throw more automation at it. And we hope that good things happen. And when they don't happen, we just create another dashboard. So it's, to go back to what I was saying, sure, both aspects of the developer mindset are technical skills. Both of them are necessary for the creation of a good AppSec team. But I, I still say that the meat is in how you're going to use that AppSec team. If it's going to be a security engineering team, if it's going to be a true, and sorry when I say true AppSec team, somebody who goes out and helps security exist in the environment, as opposed to somebody who goes out and creates tools and creates scaffolds and stuff for security to emerge in the environment.
39:27Matt Coles Yeah, actually, I'm gonna, I think I'm agreeing with you maybe a little bit here that they, because they went with the AI first and then the developer mindset, probably should reverse those. The last one is, as a security team, we're gonna inject tools into the process, as opposed to we're going to have the developers when they do their DevSecOps, They're going to know what security tools to include. The function's different. So here they're, I think what they're, what they're suggesting is the security team is still distinct. The AppSec program is still an outsider from the development process. Yes, they can talk the language. Oh, I know what, you know, all the development, fancy development tools are, or I know how, you know, what, what AI coding agents you guys are using. but I'm still on the outside. I'm just still gonna drop stuff into your environment for you to manage. And that's, so that's still removing the AppSec team from the general organization.
40:29Izar Tarandach The, the other day I had, uh, the other day, a couple of years ago, I had, uh, uh, a team come to me and say that, uh, they would like to onboard one of those nifty, uh, pipeline tools, you know, the thing that manages the whole build pipeline for you. Yeah. CI/CD. That I would be very, very happy because this one had security built in. And looking at it, yeah, really cool product. But then I just reminded them that 2 months before we had released an internal pipeline that had exactly the same tools, had exactly the same functionality, and that they could use. Without any cost. And no, the other one was better. Why? Because they found it. That's it. So it went from not written here to not found here. And that to me illustrated exactly how much what you're saying is right, that the AppSec team is so external to the thing that you can bring exactly the same functionality blessed by security, saying, this one is good, guys, go for it.
41:36Chris Romeo That's just bad budget management, because like, they should have used that money for a Hawaiian vacation for the team.
41:43Matt Coles It's bad, it's bad a lot of things. This is a repeated problem though, right? It happened in the early days with static code analysis tools. I think it's a very consulting mindset. Oh, we could have the AppSec team do it, or we'll hire consultants because they're better. They're outside. Right? There's a, there's a not invented here problem.
42:04Izar Tarandach Yeah, let's talk about seagulls.
42:08Chris Romeo We're coming towards the end of our time here. We gotta, we gotta wrap, we gotta put a bow on this thing, as they say in the business.
42:14Matt Coles All right, so rapid, rapid, uh, what's the, uh, lightning round? Uh, 3 core skills, 3 core technical skills. Chris, go. How come I gotta go first?
42:24Chris Romeo Because usually you're the one asking the questions.
42:27Matt Coles It's 3 core.
42:29Chris Romeo I think understanding code, I'm going to go differently. I'm not going to say developer mindset. I'm going to say understanding code. Being able is one that I just think is crucial to AppSec success. If you don't understand, if you can't write code and listen, I'm not a very good coder, I'll admit it, but I understand how code works and I can look at your code and say, hmm, this seems strange. This is coming in here and something's missing or Whatever. So that's, so, so, uh, everybody can give, everybody give one, that'll give us three. I'll say understand code. Matt, calling on you next since you called on me.
43:04Matt Coles Well, I'm going to go with, I'm, of course I'm going to go with, with the systems thinking and understanding, being able to integrate into the lifecycle, understanding the aspects of what happens from the time I have requirements to the time I ship something and that beyond.
43:18Chris Romeo Okay.
43:20Izar Tarandach Is our final thought understanding things enough to be able to know when to ask for more details without the fear of people thinking that they don't know what they're talking.
43:39Chris Romeo So you crossed into the soft skill.
43:41Matt Coles I was going to say, that's not a technical skill.
43:44Chris Romeo Uh, I would say that it is.
43:46Izar Tarandach I would say that it is because understanding what you don't understand is very much a technical assessment of your knowledge. It's true.
43:53Chris Romeo It's a humility as well. Like one of the, you know, it just, one of the things I think all of us have experienced the longer we've been in this industry is the more we realize what we don't know versus what we think we know. And when you're early in your career, you're like, I know everything. I'm on the security team. And then you get around for a while, you're like, I don't know anything.
44:14Matt Coles That's it. That's not even—
44:15Izar Tarandach that way I didn't get paid! Yeah, but th-the real skill is in understanding that things that you thought that you knew may change the way that they're used over time? And that's the time to ask questions. Absolutely. And those questions have to have the technical depth to actually bring good answers.
44:34Chris Romeo You just described wisdom being able to sss— yeah. Well let's— you're getting older so... You're becoming wise. You got gray hair. So welcome to the club that we're all in. All right. That's all the time we have for today. Hopefully this hodgepodge of words provides some value to you as you're running an AppSec team somewhere and stay tuned for our next rant on AppSec. I'm sure coming to a podcast feed near you.
Transcript supplied by AssemblyAI. It may contain transcription errors.
Originally published as The Security Table. Part of the AI Security Table archive.
