DtSR Episode 729 - Xavier Ashe 1 A.M. at Sprint: The Night Yahoo Went Down
31 minAudio
TL;DR: Xavier Ashe was there when that first DDoS attack by MafiaBoy went down. He's got stories, and we're going to tell them. Recorded live at BSides Atlanta in 2026. Guest Xavier Ashe Description: A DDoS attack used to be something you could “just block” once you saw the source. Then defenders learned the hard way that thousands of small senders can knock over a giant, and the playbook changed for good. From the stage at B-Sides Atlanta at Georgia Tech, we sit down with security veteran Xavier Asht for story time from the defender’s seat during one of the earliest recorded distributed denial-of-service incidents and what that night still teaches security operations teams today. From there, we follow the thread into the modern SOC: the messy gap between “something is broken” and “we are under attack,” the endless tooling arms race, and why AI makes some problems easier while creating brand new trust issues. We talk candidly about non-deterministic systems, why “AI will triage everything” keeps falling apart in practice, and how AI may be most useful when it helps glue together fragmented enterprise data the pain point SIEM tools have tried to solve for decades. We also shift to the human side of cybersecurity: neurodiversity in the workplace, why so many of us feel more at home in hacker spaces than in “normal” environments, and how strengths like hyperfocus can be a superpower only in the right context. We close with a grounded five-year outlook on AI economics, regulation, and what happens when today’s cheap access meets real-world business pressure. Subscribe for more conversations from the Security Rabbit Hole, share this episode with a friend, and leave a review so more defenders can find it. Have something to say? Let's hear it. Support the show >>> Please consider clicking the link above to support the show! -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= YouTube home: https://youtube.com/playlist?list=PLyo0dkKRvfVtWXjRxNISrhme1MgBj3C2U&si=scHDiTuLXSEQ9qHq LinkedIn Page: https://www.linkedin.com/company/down-the-security-rabbithole-podcast/ X/Twitter: https://twitter.com/dtsr_podcast
Audio and video play straight from the show's own hosting. PodSubs stores and serves no media.
Transcript
From the show. Tap a time to play from there.
SPEAKER_03They say they say we should have known better than two also deep down, deep down into this rabbit hole we found. And uh this is Nikolai and you're listening to my rabbit revol, co-host James Jardy, and featuring Mr. Jim Tillers on the one and only Down the Security Rabbit Hole podcast. Plug in, listen up, and enjoy.
SPEAKER_00Welcome down the security rabbit hole to the fourth episode of the Down the Security Rabbit Hole podcast, recorded live and in person on October 3rd at the B Sides Atlanta here at beautiful Georgia Tech. And this is episode four. Mehu Ravancar got him down first. Phil Brass talked about pen testing and LLMs. Wes Weiner talked about unlocking open away models. And man, that one got philosophical. Don't get too mad at me, but we we went we went sideways on that one and enjoyed every moment of it.
SPEAKER_00And then now I got Xavier Ash, who's gonna do some story time. And uh you know what? Uh I had I had my AI assistant do research on on people for this, so I could you know get an idea of like I said, find me anything interesting. And two pages later, I was like, no, seriously, that's too much. So we cut it down, but I'm gonna let you introduce yourself because I could never do your career justice, boss.
SPEAKER_01Thanks a lot. I you know, to summarize it, it's it I always say I've been hacking since the late 80s and I've been in security since the 90s. Uh so you know, you know, put that together. Yeah, I I happened to just grow up when all this stuff was was being built. And so I I I I've been doing a little bit of everything. I've been in large companies, I've been in small companies, I've done startups. You know, worked for IBM. I was a worldwide subject matter expert for a couple of tools for a while, and then you know did some startups and and doing volunteer work. Uh now I'm I'm doing a good good corporate job.
SPEAKER_01I got back, I was selling security for a while, but now I'm actually doing it again. Every time I would go and try to sell security and say, hey, here's how you should do it, and I'm gonna tell you why. And then I realized it's been like 10 years since I've actually had hands on. So so I I went back in and decided to be a practitioner again. And so now I'm working in the financial space, financial services, learning how banking works and and all of that. And it's it's it's great.
SPEAKER_00But I I I I did uh I did a bunch of financial services stuff, uh late 90s, early 2000s. And it is it is the space that I can only describe as the fastest moving and slowest moving at the same time. That's interesting. All the stuff up front is moving light speed. Yep. And there's still a lot of stuff that you're like, I'm sorry, what year was this built?
SPEAKER_01It nothing really goes away. We just keep adding more complexity. Uh yeah. And so, like when we talk about you know this new AI stuff and all the infrastructure around it, like we just we're adding that to the fact that we still have mainframes and you know, tape libraries and all sorts of things.
SPEAKER_00Yes, yes, mainframes, which we and COBOL, which we you know nixed decades ago. And yeah. So listen, uh, one of the things that AI clearly found was interesting about your background is that you were defending one of the first, or if not the first ever, that recorded DDoS attacked.
SPEAKER_01Yeah, I mean how much fun that was. So yeah, this was uh early 2000s, and I was working for Sprint eSolutions. It was a pre-cloud type offering, basically web uh somebody else's data center, an early version. Exactly. And and so uh we were running a lot of really big companies eBay, Yahoo, you know, some a lot of government sites. And Mafia Boy was the threat actor here. Uh he has a Wikipedia page. I'm still not cool enough to have a Wikipedia page, but Mafia Boy does.
SPEAKER_01So you can look him up. Mafia Boy had this new tool. It was the middle of the night, so he's gonna give it a try. And he tried it on Yahoo to see what would happen. And so uh, and and yes, this was the first attempt to do a distributed denial service attack. So basically, you know, looking at instead of just sending tons of data from you know a bigger person on the internet, getting lots of little guys to kind of uh to come in. You know, this was well before we had huge botnets and all of that. So this was it was pretty novel.
SPEAKER_01And yeah, it took us a while to figure out what we were uh defending of.
SPEAKER_00He's one of those like I hate giving stupid people, and I don't mean like you know, not intelligent. I mean just like because you can be intelligent and still be stupid, but I hate giving stupid people credit. But like this one was interesting because you know the the the issue with the dollar service attacks was you're like once you discovered the pattern of where it's coming from, you just simply blocked off block it. Right. And that was what would happen.
SPEAKER_01We would block it and it would come right back. You're like, why is it going right back? This is you know, this is this is interesting. And so Yeah, we ended up having to do some interesting things with both uh TTLs and and so the the defense was in the in our routers. We were able to to give our routers the right enough code to be able to say, like, nope, this can be dis discarded.
SPEAKER_00And yeah, we were able to just hoping at that point to be far enough upstream with big enough hardware.
SPEAKER_01Yeah, I mean you would think being why I mean Sprint was this was one of the biggest data center you know things that you know, so we were the big guys. And so for us to get attacked like this was uh significant. And so it proved that a whole bunch of little guys can take out a big guy.
SPEAKER_00And yeah, and the game changed forever, as they say. Still dealing with it. And we're yeah, we're still doing dealing with it. What what was the most I don't know, that that that having that as part of like your early, I say early, but like early on, early enough in your career, because now it's been 24 years since then? Yeah, good lord, really? That's terrible. Okay, so now we're very old. Yep. And in that moment of introspection, moving on, what was the like what was the big oh crap moment for you there?
SPEAKER_00Like, you know, like because I would think you're standing in front of a screen that says, hey, everything's running at 100%. Well, that's not good. Right. Right? What was the what was the thing for you that really got you thinking? It really took us a while to figure out that we were being attacked, right?
SPEAKER_01I mean, you know, it's it's because this is not something we would know I I was a third shift manager, right? So this literally was the middle of the 1 a.m. and the alarms start going off and websites going they start going offline, and we're like, what the heck's going on? So I mean we're doing our normal stuff, we're rebooting things, we're doing so yeah, it took us hours. And I, you know, I think that I was one of the first to kind of raise my hand and be like, I think this is an attack, and uh, you know, we go back and rebooting routers and everything like that. And so it it took us a while to even you know realize that there's that we were under attacks.
SPEAKER_00How things how things have changed, although not not really.
SPEAKER_01No, I would say we still have that same similar problem. I, you know, I'm I'm I'm still in operations, so when the crap hits the fan, you know, I'm I'm on the calls, yeah, and we sit there and spend their time. It's like, all right, is a server down because somebody screwed up the server or is it an attacker? I mean, that's still a a tough thing to get to quickly. Is is this an attack or is this just an IT problem?
SPEAKER_00Has the has the uh evolution in security tools, I'm not even we're not gonna talk AI yet, but like has it seven evolution of security tools over the last 30 years done more harm or good? And I've seen that from like every perspective.
SPEAKER_01Yeah, I mean it it it I think it's done well. It's it's interesting because the same tactics and techniques that were discovered at some point in the last 30 to 40 years are still in use today, right? The things I used to do when hacking on Prodigy and CompuServe, right? We still, you know, we talk about you know the the phishing now.
SPEAKER_00CompuServe.
SPEAKER_01Yeah, CompuServe. I mean it's it was just phishing back then. We didn't necessarily do it via email, but it's the same basic concepts. It's is that you know it's it's tricking either computers or people into doing what you want. And and so we get better tools to to manage that, to even be aware of that. So, like for example, the distributed denial service stack. There is there's no way that we would have been able to defend past that first initial attack. Yeah, you know, if he had changed just a little bit of things, he could have continued to attack, but because it was rudimentary, we were able to put that together.
SPEAKER_01And so to be able to defend some of these attacks, you have to have those tools to defend against.
SPEAKER_00Yeah. Uh I I ask that because as a defender, and I and I did defense for a bit and then just joined a dark side and never looked back because it it was a thankless job that on a really good day nobody noticed that you were there. On a on a reasonably bad day, nobody everybody noticed that you were there. And like there was never the notice of good, like, wow, you guys are doing such a great job. It was like, oh, who are you? Yeah, right.
SPEAKER_01Why not? And why are we paying you? Yeah, why are we paying you? Nothing's wrong here. If things are good, why are we, you know, you're in a you're a cost center, you you, you know, this is something that we can minimize and and uh automate or right up until right up until you know you do need them, and then when it's not working, yeah, so why isn't you know why isn't this magically fixed and five?
SPEAKER_00Yeah, but security tools, I think, specifically evolved quickly to adapt. So we we've been chasing, we've been like we talked with uh uh Wes about Spy vs Py. Like we've been we've been playing Spy vs. Spy for for a long time. I go back to those old cartoons every once in a while, and it's and it's more relevant now than I think it's ever been. But it's super interesting to me because the attacker, we find the attackers build something, and it's generally like, oh, it looks like something that we you know that that we weren't prepared for.
SPEAKER_00We'll build a defense for it. They go build a better offense, we build a better defense, and we're always trying to get just get that little look around the corner. You're like, oh, but and and now we're standing, uh, I would argue we're pat we're off the precipice of of AI. We we've jumped in, we're head first cannon, or maybe cannonballing into what I hope is not the abyss, but it's certainly not anything that we're really well prepared for. And the notion seems to be oh, AI is giving the attackers a leg up.
SPEAKER_00Great, we need to give defenders AI in the same unfettered, you know, just give them AI, it'll fight AI, go AI. There was this hilarious thing on LinkedIn yesterday, or like a video and that somebody's like, okay, well, we're being attacked by AI, but we built this AI that's gonna detect that AI, but that AI could go rogue, so we build a second AI that's gonna detect when the first AI goes rogue, and then the third AI is gonna detect when they all shut all of them down. And I'm like, that's hilarious, except that's literally how we built our tool chains.
SPEAKER_01And and to measure those AIs, you need another AI to make sure that that AI is doing it right. And it is, but you know, I think that the concept of of having a basically, you know, an engine that doesn't give us the same answer every time is is the new concept. That unsettles me. It is the new concept that it makes it so hard for us that has been, hey, when you run this script, it does the same thing every time. Input A equals output B. And now it's like input A equals We're not sure. Right.
SPEAKER_01But if you've been like I've my career has jumped between practitioner and and leadership. So when I'm you know running teams, like I've run offshore teams, I've had crap spaghetti code thrown at me for decades, right? I've been dealing with bad code and vulnerabilities and all this. It's like we've been dealing with humans not producing good technology for so long that it's just conceptually now now. We've got computers that also write bad code.
SPEAKER_01And you know, we we would like it to be you know all magical and better, but the but the bottom line is that this is kind of a a new tool in our tool belt that we're not not sure how to use just yet.
SPEAKER_00So that this you hit on something that I said it, you know, the whole non-deterministic thing unsettles me a bit. But I get the argument from people that, well, humans are like that too. I'm like, I know that's why I was we we were like that's why machines for me for a long time were the bastion of like you know, at least if it was wrong, it was predictably wrong. Yeah, now it's unpredictably wrong, potentially, but we're not sure because you have to double check it. I get it, the precision part of that is getting better.
SPEAKER_00We're getting closer to repeatability and recalls getting better, precision's getting better, all those things. But I'm not willing to accept the argument that, well, humans are flawed, they sometimes give you a wrong answer, so you should give AI that same leeway.
SPEAKER_01Like, well, no, it's it's I mean, that's taking an extra leap. It's more I say it to more as a comparison to put it into the right frame of thinking. It's like if I've if I'm running a team of uh, you know, I I I've got a grant and they've given me 10 students, and I'm I've got to come up with a project and we've got to deliver this project. I do this all the time, right? You know, it's just like you you get you get folks that have a lot of skill, a lot of knowledge, very little context, and not necessarily the wisdom to apply all of that. How do you guide that into producing something?
SPEAKER_01I'm using those same skills in AI than I am with people. Now, the inverse isn't true. I can't necessarily replace all of those students with AI. That's not the argument. The argument is to say, hey, this is the situation we have comparisons to. We can leverage our experiences here and say, yeah, you know, if if I'm gonna deal with bad code, here's the controls I put in place, right? If I hire an expert coder, I know exactly what I'm gonna expect. And I could probably take that code, compile it, and throw it up on you know on the web.
SPEAKER_01But if it's you know, if it's not necessarily trustworthy, I I I do stuff to make sure that I can trust it. So it's it's not necessarily brand new waters, right? We we dealt with untrusted code, untrusted whatever, uh possibly bad. You know, this is this is not brand new. It's just new in the application.
SPEAKER_00So you're a big thinker, right? You you you're you're one of those, you've been around long enough that you see kind of the bigger picture than someone, somebody that would be relatively new in the industry or somebody that's hyper-focused. Where is AI for us right now as defenders giving us the greatest benefit versus the most heartache?
SPEAKER_01Most defenders still have to deal with partition systems and and and and networks that don't talk with one another, right? Yeah. You know, we've been, you know, we've been trying to build data lakes, we've got sims, we've got all these other tools, and we've been trying all this the same goal is give me all the data that I need to make a decision on whether or not this thing is a bad thing or good thing, right? It's like I've got an alert, I need to know if the alert is good or bad, I need data to do that. And so we've been trying to solution problems, solution that problem for decades now.
SPEAKER_01And I I SIEM has been one of those that's been really fun. I that I I got into that probably, we I got into that, you know, when when IPSs and firewalls started becoming boring, right? It's like everybody has one, they're just a box, right? Yeah, and if I'm gonna be a value-added reseller or consultant, right, I need something that's hard, complicated, and takes a lot of time to set up, right? And that's why I got into SIM. It's like this is perfect, right? And and we're still haven't solved that problem.
SPEAKER_01Chaosity. You know, it's not it's not like, hey, uh Azure that's fixed it, right? It's like, no, you know, we still have in most enterprises, the most biggest challenge is gonna be getting all of those systems glued together. Now, if that's with a sim or it's with AI, if you can solve that underlying problem, you're better. Now, does AI change that game a whole lot? Probably. I think AI is gonna take us a lot farther than sims did because sim is I know everything about everything and I can organize this data and produce it perfectly, right?
SPEAKER_01Uh-huh. And and we started off with relational databases. I remember building one sim for a customer. The database alone was a million dollars of Oracle licensing, nonetheless. I mean, it was before we even got to the hardware to run their SIM. And this is a security budget. Security guys don't have a million dollars to run a database. So we then we got to you know second gen uh SIMs that that use you know file-based technologies, and and you know, we can scale Mongo database, things like that. And so now we're able to scale, but we still don't have we're still trying to figure out that information.
SPEAKER_01AI is now step in. It's like now we've got the ability to better contextualize that data, better understand, and better uh parse all of that data. And so I think that's just this is my like you know, version three or version four sim is is where I think of a lot of this, at least for the defender tech.
SPEAKER_00I think we're on the next, next, next gen sim. So I think we're on like version nine at the moment, somewhere Anton's just perked up because he knows we're talking about uh no talk about sim. Um where is the uh where's the where's the space to go? Like where's where are we where do we got a lot of work to do in AI right now?
SPEAKER_01You know, you know, I I think that AI generally is making the the making the the things that already exist easier and more accessible to more people, right? Is is and we so so just as we were fighting script kiddies, and you know, when when the there was different technologies that got released at different parts of you know in the last couple of decades where you know, hey, our our automated bone scanning, right? I mean that changed pen testing you know greatly, right?
SPEAKER_01We used to have pen testers that would everybody has their own way of doing it. And here comes I can run a Nessa scan and pretty much give you the same thing. And we had this wave of of AI slop, but it was just you know people running Nessa scans. And and and so we again we've been dealing with this. And so I I think that we're still just in a situation that we've occurred been in before, and we can establish that you know the found the fundamentals still are important, yeah. That's great, right? It's is that we still have to secure things in the same way.
SPEAKER_01AI is making it harder to get there because they were getting attacked more, but it doesn't really change the equation.
SPEAKER_00Yeah, that that's I guess that's fair. We've we've seen a lot of really interesting advancement in AI being able to uh fair disclosure, so I I work in you know, day job is in kind of the general AI sock space. Hate that, but it is what it is. You know, people like their uh acronyms. The general thought that AI is gonna simply disposition and be able to tell us disposition triage, analyze action on all of our security alerts.
SPEAKER_00I I've seen that pitched, I've watched that implode on itself. And if you ask me, I think it's it's a pipe dream, at least for the foreseeable future. Give me your take on that.
SPEAKER_01I don't know. It it could be. I mean, we we've been trying to do this in Sims for, like I said, for decades. Yeah. I mean, it's the same goal, and we're just using it, we're applying a newer technology to get it done. I mean, we didn't try, but there's also you know, blockchain was there, right? And that's one of those technologies that had some you know possible benefits, but ended up really flopping in the for the most part, right? There's still niche cases out there of cryptocurrency, but it didn't change the world like AI is actually doing, right? And so so I think that in you know what we're gonna see is you know continual refinement on what the best use case is for AI.
SPEAKER_01Instead of just, hey, let's throw all the data together and just ask it general questions, right? We're gonna learn how to use this technology better. So I I don't know. I think that it's it it's it's gonna have a lot more promise than some of our other stuff we've tried. Is it gonna fix security? No. I mean, this is we keep being sold that this is the fix. Yeah. I I don't know. I I I think that I am so confident in that the fundamentals is still where it's at, that anything new I I generally dismiss.
SPEAKER_01Like if you if you have just built a new security product in the last you know year, right? I I'm I I seriously doubt it's gonna move the needle, right? Because the thing that the the pressures that we have from attackers using AI, the way to fix that is the fundamentals, right? It's it's IAM, it's segmentation, it's it's you know it's it's the you know zero trust model, it's it's knowing that you've built a solution that has the ability to, you know, even when these controls fail, you still are able to you know protect the environment.
SPEAKER_00Yeah.
SPEAKER_01And you know, we've been building layered controls for decades.
SPEAKER_00Yeah.
SPEAKER_01We need to continue doing that. And like I said, AI just is is not in and these new security products. I I haven't seen anything new that come out built by AI or for AI that's really changed anything. So nothing has got nothing has gotten here like, oh, this is all fun, this changes the game. Just AI in general, like using it and figuring out. We still, I mean, these are still Legos that we Have like figured out three ways to put them together, right? I'm I'm here besides and I'm hearing the way that people are solving it, and we're still all doing it kind of in a vacuum, right?
SPEAKER_01Yeah. Because there's not this is the way to do it, right? It's it's hey, we had this new technology, and you know, go figure it out. I'm thinking about the days when you know TCPIP was new, right? And this new thing called the internet, right? There was all this excitement. People were figuring out how to how to put websites up the same way. We're still figuring out how to use AI, and that's just gonna take us some time.
SPEAKER_00All right. Well, that's interesting.
SPEAKER_01What are you getting into these days? You know, I've one of the things you know I've been focused on is lately is is really looking at neurodiversity in the workplace and specifically in security. I I noticed that when I go to security events, I'm comfortable. I'm I I find I find like this scene is my, these are my people. And I wanted to put that into word of words. Why do I feel more comfortable in the hacking scene than I do at sporting events or at you know you know school things?
SPEAKER_01Nerd. No. That's that was that it's it was there was the freaks and the geeks, and then there were the the the normal people. I I remember that in high school. Exactly. I now am of the firm opinion that this is really just the bell curve of all the the way our human brains work. We're just down here on this end. Yeah. You know, and and that the more I discover about how uh what research we're doing on neurodiversity and in talking about that with people in this industry, it's nothing is stuck more when I give talks.
SPEAKER_01I give a talk about you know how to hack this or how to do this, you know, this vulnerability. It's exciting, you know, but the passion that I get when I give neurodiversity, neurodiversity talks, it's it's amazing because this is life-changing. These are people, you know, some people that you know have these aha moments and go from masking full-time, completely burned out because you know they're trying to live a life that they you know can't really maintain. And you know, to say, hey, if you just do this, you're able to not die at the end of every day, right?
SPEAKER_01You're not, you don't, you don't have to do these other things that are self-coping mechanisms. You can actually, you know, and you can get into industries like security that have high stress, high turnover, you know, a maybe a good place for neurodivergent folks, but only if we can, you know, you know, understand how our superpower before you can. And yeah, and and and what you know is the talk that I gave today was specific around not only knowing what you know that you have a superpower, but that the superpower is only a superpower in the in a certain situation.
SPEAKER_01Hyper focus is a great example, right? If I am hyperfocused and doing something I'm supposed to be doing, that's great. But if my boss says, hey, I need you over on this other project, but no, no, I can't. No, well, I gotta solve this problem before I can move on. Yeah. And so knowing where you're at in that, and so I I give other people a lot of skills, and and it's it's it's just great to kind of be able to expand my impact on the security, you know, field and give back and not have to be a super uber elite hacker to do so and and to bring forth, you know, it's not that you know I'm old, I can talk about all the fun things we did back in the uh 80s and 90s.
SPEAKER_01No, we can't. We can't talk about any of that, Doug. Well, I mean, that's most of it. Well, so we can talk about it. Some of it the the the the uh the it's been too long for them to come to the statue of limitations. Statue of limitations is a run out. That's why I can talk about that stuff now. But no, it but it's it's you know having a place to give back and say, I've you know, not only am I neurodivergent, I've raised you know eight wonderful neurodivergent kids, and then I work in an industry that's mostly autistic ADHD folks, right? It's like we like we have extreme personalities that in if if you were to go and take you know the hacking group out and go put them in you know in a normal mall somewhere.
SPEAKER_01You say normal, yeah. We are the freaks and the geeks, right? Yeah, 100%. And and and and and there's and that's good, right? It's good to say that and be able to say there's a reason for that. And it's not because we're broken, it's because we're different. Yeah, and that we're just a another part of the bell curve. We don't have a disability, you know, we're just different.
SPEAKER_00Yeah. All right, so I don't I don't ask many people this because this is a tough question, but I'm gonna ask you. Look or look down the road for me. Give me give me five years from now, now that AI is most certainly here, it's it's never leaving, that racehorse is out of the barn and down the field. And in some aspects, we're standing there going, what just happened? And and you know, I think the best we can do is chase it down and see if we can make it heal.
SPEAKER_00Give me your five-year outlook. What's out there?
SPEAKER_01So one of the things that I think will be the more of a big change from the technology itself is the companies running this technology and the the economics behind you know this big build-out and whether or not that's sustainable, right? We're currently, you know, got you know fairly cheap tokens. You know, my$100,$200 a month claw subscription gets me a whole lot of clod. That's not gonna be forever, right? They're gonna go public. It's the realities of this uh this financial situation is gonna either come crashing down or just get readjusted.
SPEAKER_00I, you know, we were just talking with Wes about this. Like there, there's uh there's two options we can go here. Like either either we start, we put somebody somebody wins, and we know who that somebody is, and regulation happens, and the open way, open source community catches up and like the entire thing comes down. Right. In a way that like makes all of us way less wealthy.
SPEAKER_01Exactly. And then that's that's the thing, is like, you know, my company has you know investments in things that have investments in AI. Oh, yeah, all of us do, all of us do, right? And so that's what I'm saying. It's like I think that that is gonna be more impactful than necessarily the evolution of the technology, yeah. Because it is, it is, and the other thing that I'm I'm hoping will happen is that biological research will continue, and we've either made something that approximates how the brain works and we're completely wrong, or maybe we're right.
SPEAKER_01And this is how our brains actually work. And again, going back to neurodiversity, I have a lot more people that can relate more on how Claude does its processing than I can, you know, than they can we can in being able to say, you know, you know, look at how the brain works in general. And so again, I I think that we're going to continue to see a lot of of investment in the ability to do AI and not be tied to these large companies, right?
SPEAKER_01And it's like that is, you know, there is so much control, so much money, so much going on there. And then there's the social aspect of it is that AI is not very popular right now, at least the AI companies, right?
SPEAKER_00Well, can't blame people for I don't blame it at all.
SPEAKER_01In fact, some of my kids are on that that same situation and they refuse to use AI, and I respect that. It's like it's like but at the same time, you say that you're not supposed to use AI at school, yeah, and then you go to get a job and you have to use AI at the school.
SPEAKER_00Like that's a requirement, right?
SPEAKER_01It is a requirement in a lot of places, and so you know we do have to fix that. It's either cheating or it's cheating. Yeah, I know. It's like, can can I just ask AI to do this at work? Oh, yeah, that's right. They want us to do this at work, it's like it does feel that way.
SPEAKER_00Well, I think that I think that speaks more to our broken education system in this country than anything else. But that's a episode for a different time. All right, I've run out of time. It's been awesome having you on. I I need to get you back for more story time at some point. And the fact that it's kind of embarrassing that we're both locals and we don't get together at like at all, and I gotta start making it to some of the Atlanta functions because now I have no excuse. But it's been great having you on the show. Thank you so much. Good luck and uh day job, and enjoy the herding of the of the sheep or squirrels in our case.
SPEAKER_00Yep, that are sometimes in your head. All right, folks. This has been another of the episodes from the B-Sides Atlanta 2026 conference here downtown Atlanta. October 3rd, that was Xavier Asht. We had a wonderful conversation, and I think this is one we'll we'll pick up again. Thanks for listening, you guys. Look for one more episode coming, and that's gonna be another good one. We'll talk to you soon. Bye-bye.
SPEAKER_04This is Bella. Thanks for listening. Don't forget to leave my dad a review and share this with your friends. Bye.
TL;DR: AI asymmetry is real. While AI is clearly helping attackers find weaknesses across the enterprise attack surface, is it helping defenders just as much? Where is the benefit of AI felt the most in cybersecurity?...
TL;DR: AI doesn't care about your firewall policy. We dig into the uncomfortable question many security teams are quietly asking: Are firewalls still relevant in the age of AI, agent swarms, and automation? Guest: Anton...
TL;DR: Stop vibe-coding your detections! This week's pod features Matt Coons at GitLab as we talk about detection engineering, but with AI agents. Guest Matt Coons Description Your detection backlog is endless, your...
TL;DR: This week's pod is a conversation about AI and "Open-Weight Models", and if you aren't intimately familiar with what these are this is an episode for you. Wes has a lot of insight, and this is a discussion rather...
TL;DR: Phile has been penetration testing for a long time. The difference now is that he's having a much easier time with it thanks to AI. It's an interesting world, and he's here to explain the offensive side of it....