Black October Proved AI Cyber Threats Are Real | Varin Nair
MTS
November 11, 2026
Factory’s Varin Nair joins Sophia Dew to explain how open-weight models and agent harnesses like Artex and HexStrike have automated cyberattacks at nation-state scale.
Varin Nair
So, I think that MTS, METOS and Databricks really scared a lot of CISOs because they realized that, okay, theoretically, models can find vulnerabilities in your system, but they hadn't really seen a breach from METOS, or after METOS. That changed in October, when you had all of these vulnerabilities show up at the same time on Black October. And after Black October, people suddenly started realizing, okay, this is actually a big deal.
We need to do something with our budgets to actually procure the necessary tools to protect ourselves in this new environment.
Sophia Dew
Hello, everyone, and welcome back to MTS. Today, we are joined by Varin Nair, who leads cyber work at Factory. And last week, Factory launched War Games, offering to simulate a nation state level attack on a company's infrastructure for free. And another cyber product is coming next week. Varin, welcome to MTS.
Varin Nair
Thanks for having me. Great to be here.
Sophia Dew
Excited to have you. I think this entire conversation around cyber security is very prevalent right now. And I want to talk about War Games because this is fascinating what you guys launched. Tell me a little bit about what it is.
Varin Nair
Yeah, so War Games is a program where, with a customer's permission, Factory runs a simulated attack on the customer's environment using the same tools that attackers are using today. And the goal is to point out the gaps and vulnerabilities in the customer's systems and then help them patch those gaps and vulnerabilities.
Sophia Dew
Yeah.
I want to understand what a nation state level attack is, because I think for people who aren't maybe super active on tech Twitter or deep in the weeds, you hear about attacks happening and typically, you'd assume it's a hobbyist or someone who's a hacker in their basement. But that's not what hacks look like now. Hacks are often nation state level led. There's entire countries who are working on hacking into systems. So help me understand what this looks like.
Varin Nair
Yeah. So I think something to point out, which is quite interesting is that in the last few months specifically, I think the tools and the techniques that attackers are using has changed quite meaningfully.
If you notice, attackers have started using agents with custom harnesses to run really large parts of an intrusion all by themselves, from finding a way in to stealing credentials, to moving across systems or the lateral movement pieces, as it's known in the attack life cycle. Now, that matters because the way these agents behave is very different from how human operators or human attackers behave. So just to give you a simple example, agent-driven attacks today launch hundreds if not thousands of operations in a given hour, which is much more than what a human attacker can potentially do. A human attacker at the same time has a lot more gaps in their attacks because they still have to read the response, figure out what to do next, but these agents are acting much more quickly. So I think that's the one big change that has happened. The second big change is these agents are also attacking multiple parts of your infrastructure in parallel. Whereas if you look at how human operators behave, it's usually a lot more sequential where the humans are usually attacking one part of your environment, trying to figure that out, and then moving on to the next. Now, most companies' security practices are really designed around human attackers, right? And so our goal is really to run the most capable version of these new kinds of attacks to show customers where their defenses are failing and then help them pass those fixes.
Sophia Dew
Yeah, you brought up a really important point. Most companies' infrastructure is designed for human attackers. And when you just told me about the parallel attacks that happen from nation states, and all of this is happening in the last three months. This is very recent to just the scale and size for how nation states are able to attack software. So before War Games, how does a company go about trying to understand how to proactively protect themselves from a nation state level attack?
Varin Nair
I think there are, what I would say is I think there are already quite a few tools in the market, which allow you to, that I think I would say are point solutions, whether it be, you know, security operation centers that feed, that take in signals from individual feeds and provide you with a complete picture of what's going on.
But again, these were designed for a very different paradigm where you had human attackers with a very specific behavior in terms of how they operated. And I think over time you have companies like Google and Mandiant that have really pioneered a lot of the playbooks when it comes to defenses. So I think it's quite well understood today how the previous generation of attacks used to occur. But I think it is going to require a complete rethink given the new kinds of tools that are available on the market.
Read the other 12 minutes of this episode here.
$15 gets you this episode and 99 more. Credits never expire. After checkout, one link brings you back here to read it.
Not what you expected? Full refund within 14 days.
Thousands of transcripts fetched by people building searchable podcast archives
Fetch transcripts with code
The same credits work from a script or an AI agent. The demo key returns a sample episode in full, no card needed:
curl -H "x-api-key: pt_demo" \
https://spoken.md/transcripts/1000651996090Markdown with the speakers named, for your notes, your knowledge base, or anything that makes HTTP calls.
From $0.10 per transcript. No subscription. Credits never expire. Prices exclude VAT, added at checkout for EU customers. Not what you expected? Email us within 14 days with 20 or fewer credits used and we refund the pack in full.
Using your own key:
curl -H "x-api-key: YOUR_KEY" \
https://spoken.md/transcripts/1000801153243