สำเนาบทสนทนา 🔗
WEBVTT
00:00:00.179 --> 00:00:03.471
Firewall where human insight trumps every trick.
00:00:03.471 --> 00:00:06.900
We're not hacking systems, we're hacking behaviors.
00:00:06.900 --> 00:00:13.432
So you won't click no complicated code, just tried and true brain science at play.
00:00:13.432 --> 00:00:15.929
Social engineering for good.
00:00:15.929 --> 00:00:18.902
The best defense is in your mind today.
00:00:32.618 --> 00:00:34.079
Welcome to Phishing for Answers.
00:00:34.079 --> 00:00:48.045
The opinions expressed here are solely those of the individuals involved and do not reflect the views of their employers, of their employers.
00:00:48.045 --> 00:00:52.677
Get ready, because the next live episode of Phishing for Answers is starting right now with your host, joshua Crumbaugh, the.
00:00:52.737 --> 00:00:53.479
CEO of Phish Firewall.
00:01:11.921 --> 00:01:12.322
All right.
00:01:12.322 --> 00:01:14.626
Hello and welcome to another episode of Phishing for Answers.
00:01:14.626 --> 00:01:15.388
Today I've got Arnaud Lucas with us.
00:01:15.388 --> 00:01:17.031
He is from Cambridge Mobile Telematics.
00:01:17.031 --> 00:01:18.581
Arnaud, tell us a little bit about yourself.
00:01:19.481 --> 00:01:32.670
Sure, as you can hear from my accent right, yeah, I'm OG for France, but I've been in the US for more than three, six years now, and so I'm currently based in the Boston area.
00:01:32.670 --> 00:01:39.349
So on the East Coast I'm a CTO, VP of Engineering.
00:01:39.349 --> 00:01:40.001
I do different.
00:01:40.001 --> 00:01:48.024
I've done different things over the years and I like kind of to inspire, like high growth technology first organizations.
00:01:48.024 --> 00:01:52.233
I did that at TripRaiser, I did that at Wayfair as well.
00:01:53.140 --> 00:02:07.775
I started as an engineer, very much with hands-on expertise in technology and architecture, but I've done plenty of technical innovation projects using AI, data cloud technologies.
00:02:07.775 --> 00:02:15.653
I'm very much concerned about security, because that's always a threat that exists in any organization at any level.
00:02:15.653 --> 00:02:33.769
So that's something that I deeply care about, and you could say that one of my superpowers that seems like a very product thing to say is to kind of bridge the gap between strategic ambition and technical execution, so being able to effectively deliver what your companies actually want.
00:02:33.769 --> 00:02:37.348
And I'm passionate about technology.
00:02:37.348 --> 00:02:44.248
I'm passionate about building and scaling diverse high-performing teams, and voila, that's me.
00:02:44.248 --> 00:02:45.013
And voila that's me.
00:02:46.340 --> 00:02:46.640
Jeff Pelliccio.
00:02:46.640 --> 00:02:47.403
I mean that's awesome.
00:02:47.403 --> 00:02:50.082
So how'd you get into technology?
00:02:50.082 --> 00:02:51.649
What was the draw?
00:02:53.560 --> 00:03:01.536
It started a long time ago, when I was nine, so don't try if you want to go that far, but ultimately right.
00:03:01.536 --> 00:03:07.421
So when I was nine years old, for Christmasmas I asked my, my father, hey, I want a game console.
00:03:07.421 --> 00:03:19.758
And my father said no, and my friends were going to give you a computer instead, and that time it was a mo5 I think it was a french computer, personal computer and so as I could?
00:03:19.979 --> 00:03:21.423
your computers can run games.
00:03:21.423 --> 00:03:23.467
Uh, sounds good to me.
00:03:23.467 --> 00:03:27.567
But then when I received the computer I was like where are the games.
00:03:27.567 --> 00:03:30.243
And my father went when are the games?
00:03:30.645 --> 00:03:36.991
I remember having that same thought when I first got onto a computer myself, but continue.
00:03:37.680 --> 00:03:38.643
Yeah, so exactly.
00:03:38.643 --> 00:03:40.188
So I was like, where are the games?
00:03:40.188 --> 00:03:44.163
And my father was like, well, if you want games you have to code them.
00:03:44.163 --> 00:04:01.403
So I was like, okay, so I don't know if you remember by the time when this typing, listing thing that can be in actual catalogs and stuff, and there was always a mistake in those things for some reason.
00:04:01.745 --> 00:04:21.531
It never worked, and so I started debugging those problems to make them work and then went from there and started coding, and then I took computer science classes in high school, then I did computer science major and so on.
00:04:21.531 --> 00:04:25.827
So that's really how I got started Really from the coding side.
00:04:25.827 --> 00:04:27.432
Hey, I want to play games.
00:04:29.822 --> 00:04:31.526
I mean, for me it was similar.
00:04:31.526 --> 00:04:41.663
I mean, it wasn't really as much about wanting to play games, but I wanted to know how things worked and so just getting out of the computer, initially it was well, how does this work?
00:04:41.663 --> 00:04:42.507
And then I ended up helping everybody.
00:04:42.507 --> 00:04:43.572
And then it was well, how does this work?
00:04:43.572 --> 00:04:54.067
And then I ended up helping everybody, and then it was well, why don't I just write an application to do all of this help for me, so that I don't I'm not constantly providing this tech support?
00:04:54.067 --> 00:04:57.000
And uh, and it just sort of went from there.
00:04:57.000 --> 00:05:03.889
But I know, early on the thing that just fascinated me was the internet.
00:05:03.889 --> 00:05:10.004
And uh, and how do we build these websites and these web applications?
00:05:10.004 --> 00:05:13.692
And I just thought it was the coolest thing in the world.
00:05:13.692 --> 00:05:15.642
I still do, I mean, and it still is.
00:05:15.642 --> 00:05:20.372
And look at all of the things that come from the internet.
00:05:20.471 --> 00:05:27.252
Almost daily we still get new uh new uses and and, uh, incredible things that we're able to do with it.
00:05:27.252 --> 00:05:29.564
So, yeah, it was just.
00:05:29.564
Load full transcript
--> 00:05:37.206
It was that desire, and I know, wanting to know how things work leads to well, how do you break them?
00:05:37.206 --> 00:05:39.732
And that, I guess, is how I ended up being a hacker.
00:05:39.732 --> 00:05:47.350
So you spent a lot of time in development Secure coding.
00:05:47.350 --> 00:05:50.415
Well, it's barely something they teach now.
00:05:50.415 --> 00:05:54.524
It was certainly nothing that they taught back in the day.
00:05:54.524 --> 00:05:58.254
So talk to me about how you got exposed to cybersecurity.
00:05:59.242 --> 00:06:19.601
So when I moved to the Boston area, I actually worked for a company called Iron Mountain, which is a Boston-based company, and traditionally their business is paper archiving and paper trading, but I worked for their digital division, which was focused on doing the same thing, but for digital assets.
00:06:19.601 --> 00:06:54.863
At that point, what I loved about Iron Mountain is that they were very concerned about and part of that because they are very big customers they are very concerned making sure whatever they create is very reliable, like the quality is there, but also very safe, because obviously, if, because I was working on this next generation digital archives and obviously both things were not safe or both things were not reliable, then that's a big problem for the customers for no reason.
00:06:54.863 --> 00:07:22.211
So and they specifically hired somebody very security focused and that person said you know, promoted the same thing that you're promoting, which is, ultimately, everybody in the company should care about security, especially the engineers, especially the software developers right, because at the end of the day, it actually doesn't take more time necessarily for you to code something insecure versus something secure.
00:07:22.211 --> 00:07:24.274
It's more like a state of mind something insecure versus something secure.
00:07:24.274 --> 00:07:25.600
It's more like a state of mind.
00:07:25.600 --> 00:07:30.848
Are you thinking about creating secure code versus not creating secure code.
00:07:31.379 --> 00:07:42.086
So back then that person decided that it would be good to teach people about security and I was one of the students.
00:07:42.086 --> 00:07:57.875
I was an architect on the platform and then but with the goal of that person becoming effectively evangelist for the team that they are on, to make sure that everybody else then cares about secure coding processes.
00:07:57.875 --> 00:08:28.728
So it's kind of a domino effect or pyramid scheme, but so that I get to learn about security and security coding practices, I get to hack different systems and to understand what it means, obviously in a safe, self-guardian environment, and then, based on that, I can talk intangibly with my peers about why it's always things that we're doing.
00:08:28.728 --> 00:08:30.826
It's important to do that a certain way.
00:08:30.826 --> 00:08:48.211
So that's how I got started when I went to TripAdvisor, when I joined TripAdvisor, tripadvisor was not that old, actually, it was still very small, like all the engineers could fit in a conference room, which obviously exploded after that.
00:08:48.211 --> 00:08:55.874
But two other same thing like a lot of the content that TripAdvisor was showing were public content.
00:08:55.874 --> 00:09:05.330
So as a security, you know, people care, but like anybody else, right, that was the thing because there was not much to protect.
00:09:08.061 --> 00:09:23.145
However, one of my first projects at TripRaiser was to implement instant personalization with Facebook, meaning that they find that you know, if you come to the TripRaiser website, we recognize that you're a Facebook user and we personalize your experience.
00:09:23.145 --> 00:09:26.509
It's on your friends content on to browser.
00:09:26.509 --> 00:09:34.272
At that point, facebook is like well, that's great, but then you are getting access to one of our own token for the user.
00:09:34.272 --> 00:09:36.166
That needs to be properly protected.
00:09:36.166 --> 00:09:42.529
Therefore, you need to run a penetration test, you need to understand your security.
00:09:42.529 --> 00:09:47.684
Sponsor, of course, yeah, and so, which makes sense to understand your security posture, of course, yeah, and so, which makes sense.
00:09:47.860 --> 00:10:04.741
I mean totally agree with that, but that was a change for TripAdvisor that if we had to drive, so in addition, to actually implement instant position and drive effectively improving our security posture at TripAdvisor.
00:10:04.741 --> 00:10:05.839
But from that I recognize effectively, effectively improving our security posture, as cheap as they are.
00:10:05.839 --> 00:10:28.321
But from that I recognized effectively, because I got exposed to OWASP and OWASP Top 10, specifically on the web aspect of things, and I realized that there was a need to something educate software engineers on the secure coding practices and the danger of not following these secure coding practices.
00:10:28.321 --> 00:10:37.988
So I started to run training on a regular basis with a whole engineering team to get them exposed to that Well, that brings up a really good point.
00:10:38.268 --> 00:11:18.292
Role-based training, I think, is an incredibly critical part of any training program, and the reason is is that, while cybersecurity is everyone's responsibility, the developers have a very different role in maintaining cybersecurity than your finance team does, and so you know you want your developers being trained on OWASP, top 10 and those different types of attacks and how to prevent their code from being vulnerable to it, but that wouldn't do your finance team any good.
00:11:18.292 --> 00:11:20.120
So what is your thought on that?
00:11:20.120 --> 00:11:35.136
I mean, I see a lot of companies still doing that sort of one size fits all and not necessarily doing that developer training as part
of their awareness program.
00:11:38.142 --> 00:11:39.205
I mean ultimately.
00:11:39.205 --> 00:11:44.005
I think that's obviously a mistake, right, but for multiple reasons, right.
00:11:44.005 --> 00:11:45.850
There is a fact that that's the right thing to do.
00:11:45.850 --> 00:12:09.462
To your point about getting engineers I can say that exposed to how code could be compromised, but also, what are the best practices, therefore, to change our code, and specifically here, in the case of Tupac, what that meant for me is, because we did the pen test, I could actually match up.
00:12:09.462 --> 00:12:35.008
This is the always vulnerability, this is example in our code where we were impacted and this is the solution, which usually, most of the time, is not a big solution necessarily Like it doesn't take that much time to fix, and so I could really get the developer to say this is the issue, this is the compromise card, this is how you fix it and really get that full circle.
00:12:35.008 --> 00:12:43.927
The other reason why it's important to do this whole best training is for even co-plans.
00:12:43.927 --> 00:12:55.610
I've been paid in companies where we take payments, so we have to be PCI compliant, for example, and PCI calls out that you're physically, you're a software engineer.
00:12:55.649 --> 00:12:55.892
Yes.
00:12:56.679 --> 00:12:58.126
Or security practices right.
00:12:59.000 --> 00:13:04.303
Well, and your payment cardholder data handlers, which your finance people.
00:13:04.303 --> 00:13:06.331
They call out that one as well.
00:13:06.331 --> 00:13:08.023
Yeah, exactly.
00:13:08.082 --> 00:13:14.780
But, to your point, there are different things that you have to teach each one of them, right?
00:13:14.780 --> 00:13:18.740
So you know, and I think that's what they call it, and I like that.
00:13:18.740 --> 00:13:41.724
But the point that there is also a compliance aspect that companies need to be aware of, yes, point that there is also a compliance aspect that companies need to be aware of, yes, um, which may not be the main, well, you want to be compliant, but you start the most appetizing thing to say, oh well, doing things because of compliance, yeah, but ultimately, I do believe in this case, that's actually the right thing to do so yeah.
00:13:42.144 --> 00:13:45.549
I mean, I, I, I wish I could remember it.
00:13:45.549 --> 00:13:58.065
I saw this really great definition of GRC earlier today, um, uh, on LinkedIn, but uh, but if I try to repeat what I saw, I'm just going to tear it to like that just.
00:13:58.666 --> 00:13:59.489
I'm going to butcher it.
00:13:59.548 --> 00:14:16.081
So I won't Um, but I think that in general, if we just do the best that we can, we're always going to be on the right side of compliance, and that's why, to me, we've got to strive more for best practices than anything else.
00:14:16.081 --> 00:14:19.885
Back to the role-based training.
00:14:19.885 --> 00:14:39.351
It's been proven that when we can make training contextual to somebody's role, that it is 15 times more effective, and it's because it goes from hey, you know, there's this anonymous threat that exists somewhere on the Internet that might get you to hey.
00:14:39.351 --> 00:14:46.328
Let me tell you about a threat that you're going to face every day in your job, and it just makes it that much more personal.
00:14:46.328 --> 00:14:49.590
But the studies have shown 15x improvement.
00:14:50.701 --> 00:14:51.222
Yeah, I'm not.
00:14:51.222 --> 00:14:53.048
I mean, I haven't seen that.
00:14:53.048 --> 00:14:54.687
But yeah, I will not be surprised.
00:14:54.687 --> 00:14:56.370
That sounds right for sure.
00:14:56.370 --> 00:15:08.769
But because it's more applicable, right, it's not this random thing somewhere that you most likely employees will say I don't care about that, so maybe they should, but they don't understand the relationship.
00:15:10.640 --> 00:15:18.114
Yeah, so what are your thoughts on how to go about doing training?
00:15:18.114 --> 00:15:27.828
Because there's a lot of different approaches that I see, but a lot of what we do isn't effective.
00:15:27.828 --> 00:15:33.533
And so what have you learned over the years to make those efforts more effective?
00:15:37.658 --> 00:15:42.524
I think it depends on the person, but the way I learn is by doing.
00:15:42.524 --> 00:15:48.775
For me that means two things.
00:15:48.775 --> 00:15:54.643
One is that if there is any training more as an exercise.
00:15:54.643 --> 00:16:27.403
There are plenty of exercises we have done over the years to get people understanding security practices, but not by studying or books or by going through actual watching videos or watching a conference call, but actually by doing, by exercising not knowledge but understanding, even getting exposed to somebody to try to hike into their system and then trying to respond to that or something like that.
00:16:27.403 --> 00:16:29.682
So I think it's very interesting.
00:16:30.875 --> 00:16:37.234
I think the more hands-on that you can do anywhere in training, the more effective it becomes.
00:16:37.234 --> 00:16:39.822
So I'm a really big fan of Duke.
00:16:40.956 --> 00:16:42.883
LUDOVIC BLECHER but it's a bigger time commitment.
00:16:42.883 --> 00:16:57.783
It's much easier, to be honest, to give a one-hour presentation and to organize an exercise where the team will spend half a day battling each other in some shape or form.
00:16:57.783 --> 00:16:59.606
So that's one way.
00:16:59.606 --> 00:17:14.780
The other way that I believe is more in creating your systems so that the easier thing to do is the secure way, if that makes sense.
00:17:14.780 --> 00:17:28.404
So meaning that you're putting things in place already from the tools, from a platform standpoint, from a process standpoint.
00:17:28.404 --> 00:17:35.300
You name it where it's easier for you to do things securely than otherwise.
00:17:35.300 --> 00:17:39.084
And people say that
engineers are lazy.
00:17:39.084 --> 00:17:49.704
It's not that they value their time, but also they value simplicity, and that's a very engineering thing.
00:17:50.174 --> 00:17:58.417
You know that's how you fish a developer is that you just promise to save them time, no matter.
00:17:58.417 --> 00:18:00.965
I mean, if you want them to click, that is the trick.
00:18:00.965 --> 00:18:08.608
Now it has to be a fish specific to their tech stack, but that's not that hard to figure out these days, right?
00:18:08.608 --> 00:18:13.923
So you promise to save them time and in my experience, that gets them to click.
00:18:13.923 --> 00:18:15.796
It's interesting.
00:18:15.796 --> 00:18:25.882
I have noticed that, uh, that developers and salespeople tend to be about equally prone, but they don't click on the same things at all.
00:18:25.882 --> 00:18:29.544
Uh, developers click on things that'll save them time.
00:18:29.544 --> 00:18:42.352
They click on operational, like that traditional spearfish, whereas those salespeople they click on anything social media, so the traditional Facebook fish.
00:18:42.352 --> 00:18:43.573
They're going to click on that.
00:18:43.573 --> 00:18:47.465
So I found that interesting.
00:18:47.465 --> 00:18:55.748
Just the differences and how different people are susceptible to attacks like that.
00:18:57.714 --> 00:19:00.424
Yeah, and so for engineers?
00:19:00.424 --> 00:19:08.261
Basically, what that means is having the right set of tools and things that are easy to use that already do the right thing for them.
00:19:08.261 --> 00:19:33.298
For example, at TripAdvisor we used to have a long time ago, or they used to have, a UI framework that was inherently insecure and that would not properly escape things, which as you know is bad, because that means that you can trick to be vulnerable to XSS in some shape or form.
00:19:35.154 --> 00:19:40.175
And so, yes, you can keep on trying to find the things and fix the things, but usually it's your paying catch up.
00:19:40.175 --> 00:19:46.144
Or you can keep on trying to find those things and fix those things, but usually it's your paying catch up or Start with a framework that's secure from the ground up.
00:19:46.464 --> 00:19:46.986
Exactly.
00:19:46.986 --> 00:20:01.041
You start with a framework that will escape everything by default, and if you want something not escape, then you have to put some kind of your special things that make it so that it doesn't escape, and in that case that's something you can actually flag.
00:20:01.041 --> 00:20:03.140
Yeah, what the city commences.
00:20:03.140 --> 00:20:04.799
So that's the difference.
00:20:05.974 --> 00:20:08.964
I think that can go beyond just development teams.
00:20:08.964 --> 00:20:22.643
I mean, I think that keeping things simple and making it where it's easier to have security than to not can go everywhere.
00:20:22.643 --> 00:20:24.020
And I'll give you an example.
00:20:24.020 --> 00:20:29.487
We've recently deployed Passkeys as a company.
00:20:29.487 --> 00:20:40.500
Now, initially everyone was freaking out about it, but it's easier for everybody and everyone likes it better than typing in their password because we have it tied to biometrics.
00:20:40.500 --> 00:20:56.038
And so now, all of a sudden, I've got much, much better password security and or I guess you know access control security, but it's easier for my people to access their accounts than it's ever been before.
00:20:57.936 --> 00:20:58.518
Yeah, exactly.
00:20:58.518 --> 00:21:07.460
So this is exactly what you're mentioning, which is that you make it easier for them and, ultimately, that makes it more secure, like it's a win-win.
00:21:08.434 --> 00:21:10.643
But I think that should be the goal everywhere.
00:21:10.643 --> 00:21:16.027
I mean, for so long, cybersecurity was that thing that just got in the way.
00:21:16.474 --> 00:21:16.796
I mean.
00:21:17.577 --> 00:21:21.106
I hate to say it, but we weren't exactly the friendly department.
00:21:21.106 --> 00:21:25.260
It was hey, you've got to do cybersecurity because we said so.
00:21:25.260 --> 00:21:44.194
And while there's a lot of value to gathering the troops and sitting down and talking through all of your vulnerabilities, I also think that I don't know, I lost my train of thought there completely.
00:21:45.118 --> 00:21:48.114
I'm sorry, no worries but I see what you mean.
00:21:48.114 --> 00:22:03.491
If you focus on creating that foundation, then you're going to have a much better security posture than trying to fix it with processes and other things.
00:22:04.174 --> 00:22:05.059
Yeah, what he said.
00:22:05.059 --> 00:22:09.915
What he said, I joke, okay.
00:22:09.915 --> 00:22:42.528
So one of the things that you mentioned was doing that hour long class, um, and I think that there's a lot of value in in doing things like that, particularly with more complex subjects, um, but what do you think about more short form content at a higher frequency, like, instead of having that one hour, what if you just have one minute every single week of the year devoted to cybersecurity?
00:22:44.276 --> 00:22:48.184
I think, to be honest, I've never tried like a one minute.
00:22:48.184 --> 00:23:04.901
I do so I mean it depends how you want it to be ingested, Because nowadays what is nice is for training to be on people's own schedule, where they can click into it whenever they can.
00:23:04.901 --> 00:23:24.766
And obviously, if you want that to happen during business hour, which is usually what your engineers would prefer as well, it means that especially more busy people with lots of meetings, the content has to be shorter so that they can do that maybe in between two conversations and so on.
00:23:24.766 --> 00:23:26.781
So I really like that from that principle.
00:23:26.781 --> 00:23:33.242
It makes it more efficient because people can say, oh, I can just do that quickly in between these.
00:23:35.154 --> 00:23:44.674
The
question is more what kind of content will it be and how well can it be digested by the employees?
00:23:44.674 --> 00:23:50.008
It's not just giving them content for the sake of content.
00:23:50.008 --> 00:23:52.694
It's actually that they are going to remember it.
00:23:52.694 --> 00:23:54.319
They are going to do something about it.
00:23:54.319 --> 00:23:58.740
I think it's great if we can make it shorter.
00:23:58.740 --> 00:24:11.236
It's great that if we can make it available at any time so that people can do it whenever they have a minute, so it doesn't get in the way as much.
00:24:11.236 --> 00:24:17.729
But I think there is opportunity as well for people to be able to ask questions.
00:24:17.729 --> 00:24:31.484
I think there is opportunity as well for people to be able to ask questions, to get answers, and usually that requires, or at least have better results, when there is an actual in-person discussion or something somebody to talk to.
00:24:33.515 --> 00:24:38.182
I love to use vulnerability reports to enhance those conversations.
00:24:38.182 --> 00:24:45.345
And you know, hey, we're seeing a lot of injection vulnerabilities across our code.
00:24:45.345 --> 00:24:49.405
So for the next eight weeks we're going to talk about injection.
00:24:49.405 --> 00:25:01.486
You know, this week it's SQL injection, next week it's LDAP injection, the next week it's XML injection, and just sort of going through different things like that.
00:25:01.486 --> 00:25:16.980
I think that it can help to keep it front of mind and in my experience, it doesn't matter if we're talking about injection or password security, if we're talking about security.
00:25:16.980 --> 00:25:26.484
Security is still front, front of mind and all of those topics you've learned about stay fresh uh, but if we're not talking about it.
00:25:26.744 --> 00:25:33.605
those topics may fade away and all of a sudden, you know, whatever new shiny object there is becomes the top priority.
00:25:33.605 --> 00:25:59.819
And and when I say shiny object I just mean whatever the new priority is, you know there's always that urgent client demand that you're having to meet as a development team, and if security isn't constantly being talked about and pushed, I think you run the risk of, you know, pushing that code without ever thinking about security, not because they don't care, but because they're on.
00:25:59.819 --> 00:26:01.743
You know they're focused on other things.
00:26:03.224 --> 00:26:03.826
That sounds good.
00:26:03.826 --> 00:26:07.887
Yes, I agree with that To your point.
00:26:07.887 --> 00:26:16.673
It's a combination of making it easy for people attacks that we assume are enabled through AI or these large language models.
00:26:37.865 --> 00:26:39.190
But what are you seeing?
00:26:39.190 --> 00:26:40.453
Targeting developers?
00:26:53.684 --> 00:27:04.962
For me, what I've seen targeted is more like the executives, because I got a few where people say, oh, this is my CEO and I need this from you.
00:27:04.962 --> 00:27:09.491
Or targeting HR saying, hey, I need you to change my bank account for getting my paycheck.
00:27:10.346 --> 00:27:25.233
I get those all the time for my employees I mean, not really my employees, don't worry, I don't think it's actually you but about once a week, if not more, I get a text message and it'll be like hey, this is whoever just wanted you to know.
00:27:25.233 --> 00:27:26.832
I just changed my bank account.
00:27:26.832 --> 00:27:28.290
Can you get it updated for me?
00:27:28.290 --> 00:27:31.627
Or I'll get an email and you know it's just.
00:27:31.627 --> 00:27:49.328
I think that's a fascinating side of it, though, because everyone gets the fish that comes from the boss pretending to be the CEO, but so few people see the other side of it, where you know where it's the email from the employee trying to target the CEO.
00:27:49.328 --> 00:27:51.515
But there's a lot of that too.
00:27:52.125 --> 00:28:02.805
Yeah, nothing like that, but I would argue I haven't seen this specific thing for engineers, but what I've seen is much more targeted to specific people, makes sense.
00:28:02.805 --> 00:28:20.605
For me it's like text, but I know some friends like they got codes from people that seems to have the voice of whoever is calling and you're telling them hey, you need to do this, and usually obviously they make it.
00:28:20.625 --> 00:28:32.016
So there's a dramatic situation so that people don't feel things too hard about it, but people don't say things too hard about it, so you know people that have already seen or heard deepfakes.
00:28:32.656 --> 00:28:41.734
Yes, exactly so I have a friend that got targeted like that and for me, I want to see the future.
00:28:41.734 --> 00:28:45.556
The text message will become something much more sophisticated.
00:28:45.556 --> 00:28:55.326
That will look much more problematic will be oh, it will look much more, probably gonna get so much worse uh-huh, uh, so I'm not looking forward to that, but I also know that.
00:28:55.365 --> 00:28:59.557
You know there's the technology exists, so that that's just.
00:28:59.699 --> 00:29:09.208
Uh, you know, it's just, you know, to get done on the list to the point that I've become a target or anybody else in the organization is becoming a target.
00:29:09.347 --> 00:29:11.453
If we could, it's just a question of time.
00:29:11.453 --> 00:29:47.279
So, and and the point with that is that, yeah, I think it goes through um, I would say training helps, right, didn't be able to recognize some of these type of attacks or some targeting more effectively, which I think something that CMT does very well, because that's the reality, but these things will become harder and harder to spot and to potentially's the problem with that.
00:29:47.279 --> 00:29:51.707
It's really on the employees to detect them, to figure out.
00:29:51.707 --> 00:29:53.112
Oh yeah, this is clearly not.
00:29:53.112 --> 00:29:59.074
No, this is not from Arnold telling me to give him a gift card.
00:29:59.074 --> 00:30:33.306
So, and the problem with that is that, yeah, I'm not sure what technology can help necessarily with that, and because it's so easy, you know, to fake a phone number or to fake anything, really, at this point that you know, I feel like it's going to be on the employee to figure things out, and the only way for employees to figure things out is through proper training, and even there, the training will have to keep on catching up if you go into new methods and ways well.
00:30:33.425 --> 00:30:51.067
So I mean to that I I think it's it's not just on the employee, it's on the company the responsibility to provide training that stays current and up to date on the modern threats current and up-to-date on the modern threats.
00:30:51.086 --> 00:30:55.538
But the other area that I think is on the company is to provide that employee with a really good framework.
00:30:56.519 --> 00:31:28.711
I believe it's process and procedure that protects us from deep fake attacks, more so than any technology ever would, and that's low tech checking out a band calling, texting, emailing, whatever it happens to be, but using out of band communications methods to validate it, having multiple people sign off on things you know, sort of like the two keys to enable the nuclear weapon type thing.
00:31:28.711 --> 00:31:52.093
So I really do think that there's we're going to have to get tight around our process, our procedure and our training, because when we don't, that could be a million dollars like that, and we've seen it a bunch of times through business email compromise attacks, where they get massive amounts of money because they get in the middle of a wire.
00:31:52.093 --> 00:31:54.771
Well, what if they can trick people?
00:31:54.771 --> 00:32:13.513
Now, and in Hong Kong it was $10 million, I've heard of a bunch of these similar to what you said, where they haven't been massive dollar amounts, but it's only a matter of time before we see some massive like really really big losses as a result.
00:32:14.627 --> 00:32:16.032
And I agree in terms of process.
00:32:16.032 --> 00:32:23.990
But in the situation where you can, you're able to push the employee towards like your sense of urgency.
00:32:23.990 --> 00:32:33.936
The employee, even if the employee is aware of the processes you know, may decide to forego these processes for whatever reason.
00:32:34.884 --> 00:32:47.659
I mean, I think that's part of the reason culture is so important when building out a security awareness program, because if that employee is afraid to say no, I need to validate this.
00:32:47.659 --> 00:32:49.332
This doesn't sound right.
00:32:49.332 --> 00:32:51.933
I want to make sure it's actually my CFO.
00:32:51.933 --> 00:32:54.913
Then they're going to have those losses.
00:32:54.913 --> 00:32:58.952
That employee has to be empowered and I think that starts with culture.
00:32:58.952 --> 00:33:08.119
That starts with a CEO that says you are welcome to question us on a wire before it goes out, you know, or a CFO that says that.
00:33:08.119 --> 00:33:20.457
So I do think how we think about these things and approach them, and and and and culturize what we're going to do when these happen.
00:33:20.457 --> 00:33:28.835
I think that's really critical and we've got to be thinking about how do we want our teams to react.
00:33:30.807 --> 00:33:33.516
Yeah, totally agree, like there is a cultural aspect for sure.
00:33:33.516 --> 00:33:39.913
Well, people should feel confident that they can question and double check.
00:33:39.913 --> 00:33:43.319
Yeah, so I agree with that.
00:33:44.246 --> 00:33:59.719
Well, and we saw the recent social engineering in the open source community I forget what it was, I don't know some core element of a Linux or of the Linux platform.
00:33:59.719 --> 00:34:05.958
And so this guy is a contributor to this open source project.
00:34:05.958 --> 00:34:15.489
Guy is a contributor to this open source project, gets elevated to approver or authorizer and pushes malicious code.
00:34:15.489 --> 00:34:19.777
Spent five years working on that plan.
00:34:19.777 --> 00:34:23.585
That's scary enough that somebody could do that.
00:34:23.585 --> 00:34:25.871
But someone infilt that's.
00:34:25.871 --> 00:34:31.235
That's one of those maybe realistic, uh risks.
00:34:31.235 --> 00:34:34.865
You know we're going to try to prevent it, but you know we can't prevent everything.
00:34:34.865 --> 00:34:49.619
But what about when a computer can do, because we're getting close, where a computer could write code, could impersonate, that person could contribute, could participate in a project to get to that level of status?
00:34:49.619 --> 00:34:52.945
What do you think about that?
00:34:52.945 --> 00:34:58.237
I mean, do you think that's a valid attack path or am I just being paranoid?
00:34:59.586 --> 00:35:02.635
Well, I think it's not the case yet.
00:35:02.635 --> 00:35:08.277
I think what you're mentioning is the notion of agentic or something like that.
00:35:08.277 --> 00:35:21.360
Not only you ask something like a genii about something, but your genii is able to take action based on its own thinking.
00:35:21.360 --> 00:35:39.657
I want to say thinking, but gen, you know, jni doesn't think, it forecasts, but based on its own deductions, it can be able to take its own action without having any human intervention in the middle.
00:35:39.657 --> 00:35:44.065
That's what we're going to do effectively For now.
00:35:44.065 --> 00:35:57.344
The way that I've been seeing Gen-AI used effectively is more Gen-AI as a tool for humans to use to improve their own productivity.
00:35:58.409 --> 00:36:07.110
I agree worry that towards having these agents right you know where.
00:36:07.110 --> 00:36:20.777
Potentially, then, that AI could just you know, based on what it sees, or based on what is being the data is being prompted with,
makes a determination.
00:36:20.777 --> 00:36:26.688
For example, a customer says hey, you know, I bought this ticket here.
00:36:26.688 --> 00:36:28.094
I need to get a refund.
00:36:28.094 --> 00:36:41.472
This is why, blah, blah, blah, and they are saying, yeah, okay, I'm going to give you a refund and then trigger an actual refund to that person, even though, potentially, that person is committing fraud.
00:36:47.210 --> 00:37:30.367
So I agree, we're not there yet, but a lot of the experts are saying they think we'll get there in 2025, where agents will enter the workforce, maybe not as senior developers initially, but certainly as junior developers, has junior developers and uh and and so I know it's not there yet, but I I do worry about a world where a computer uh, I guess a rogue ai, if you will can social engineer us at a level we haven't considered that I'm not worried about the rogue, yeah, oh, let's just say that if there is a rogue AI, when I say rogue AI.
00:37:30.429 --> 00:37:37.184
I don't really mean rogue AI, I mean like an AI developed by a criminal organization.
00:37:45.032 --> 00:37:45.634
Exactly so.
00:37:45.634 --> 00:37:49.358
It's more than AI, that's I call them a bad human.
00:37:49.358 --> 00:37:55.713
Behind the bad AI then is able somehow to cause problems, and yeah, I mean that's very possible.
00:37:55.713 --> 00:38:05.659
The problem is, right now what we're seeing is more from an external standpoint.
00:38:05.659 --> 00:38:22.429
Right, you have the generic knocking at the Ruby's door's door, effectively, and hoping that one of them opens, but the point there is that that could also be, as as the as we evolve the technology, that could also be an internal threat.
00:38:22.429 --> 00:38:35.030
You know, within the company, tools, through other aspects, where we introduce something that we don't fully understand, directly or indirectly, that can cause this kind of issues.
00:38:35.030 --> 00:38:38.032
And, yes, that's very possible as well.
00:38:38.032 --> 00:38:41.829
I think something like how do you prevent that?
00:38:41.829 --> 00:38:50.855
I think companies need to be very careful, just like they have always been, about what they install in their own ecosystem.
00:38:50.855 --> 00:38:55.170
That's always step number one.
00:38:55.170 --> 00:39:08.956
Right is to be very careful and aware of what systems are being installed on every system in the organization, including your own laptop that you're using for work.
00:39:08.956 --> 00:39:11.396
Oh, come on.
00:39:11.456 --> 00:39:13.085
I can't just install whatever I want.
00:39:13.085 --> 00:39:16.411
I learned the hard way that.
00:39:16.550 --> 00:39:20.659
I cannot do that yeah, I'm not playing to.
00:39:20.659 --> 00:39:27.635
So that's one way I believe we can try to protect ourselves in a different way.
00:39:27.635 --> 00:39:42.880
But there are still opportunities that even if, let's say, cmt does that effectively which I think it does well, the company that we're trusting to bring the right tools over to CMT may not have this kind of safeguards.
00:39:43.224 --> 00:39:44.871
And they get compromised?
00:39:45.653 --> 00:39:48.693
Yeah, I mean, that happens all the time where?
00:39:49.235 --> 00:39:54.701
it's not us getting compromised, it's our, our vendors, our partners getting compromised.
00:39:54.701 --> 00:39:59.369
That is our weakest link, or our achilles heel exactly so.
00:39:59.389 --> 00:40:09.159
that's, and there's something with the body is not different one than having a bad tool, but potentially we allow it to do much more.
00:40:10.101 --> 00:40:14.023
Absolutely so.
00:40:14.023 --> 00:40:16.333
We are just about out of time here.
00:40:16.333 --> 00:40:22.494
Before we wrap up, any additional bits of wisdom here for the audience.
00:40:25.347 --> 00:40:29.853
What I've seen traditionally is security being like this.
00:40:29.853 --> 00:40:52.280
I mean, we touch on that right, being this kind of demanding beast that feels outside of what we should care about, but yet we have to comply for different shape of form, and I think we talk about that right.
00:40:52.280 --> 00:40:57.137
We talk about best practices, we talk about having the tools that do the right thing in the first place.
00:40:57.137 --> 00:41:17.614
I think it's really, for me, what I care about, which is that how to keep it simple keeping it simple for engineers, keeping it simple in terms of infrastructure and keeping keeping simple in terms of infrastructure and keeping simple in terms of software, because, ultimately, a lot of the vulnerabilities and issues that we're having is because of the complexity associated with those things.
00:41:17.614 --> 00:41:41.547
Right, it's complexity in how we do things, process, complexity in the software itself that we have created, complexity with the tools that we're using that are not fully understood by anybody across the organization, and so there is the notion of keeping it simple, but I think it can go a long way if companies prioritize that.
00:41:43.012 --> 00:41:43.554
I like it.
00:41:43.554 --> 00:41:44.737
I like it, thank you.
00:41:44.737 --> 00:41:48.255
Well, this has been another phenomenal episode.
00:41:48.255 --> 00:41:52.250
Thank you for joining us and we'll see you again tomorrow.
00:41:53.612 --> 00:41:54.233
Thank you so much.