


The episode is an Intel Briefing plus an interview. Topic: CISA adds four actively exploited RCE vulnerabilities (Synology Disk Station Manager, Bury AI Light LLM, Apache Access 1 library, Vanderbilt RedCap platform) to KEV; attackers probe for Java for Shell; vendors rush urgent web/enterprise patches; Ukraine arrests four suspects in a $40M ransomware ring; Microsoft tests Edge “lockdown mode”; Rapid7 finds thousands of exposed, unauthenticated self-hosted AI servers (often OLAMA) with outsider artifacts and cloud-routed models.
patching won’t undo prior compromises; AI-driven vuln discovery strains defenders; exposed AI APIs enable abuse of paid accounts; lockdown mode could block ~60% of privilege-escalation attacks.
Synology linked to Cerber; RedCap linked to crypto mining; Census/ Rapid7 PoCs for HTTP/2 flaw affecting ~9M services.
Russell Van Tyle, VP of Services at SpecterOps (offensive security AI harness “Sage”); Cal Newport, Georgetown CS professor (argues AI superintelligence race is flawed/reckless).
Written by AI. May contain mistakes.
Discussion of four actively exploited vulnerabilities flagged by CISA.
“while widespread tech outages spark scrutiny over national security cuts.”
Analysis of the implications of technology outages on national security.
“And happy Grand Theft Auto Launch Day for those who celebrate.”
Details on the critical Java for Shell bug and potential exploits.
“as major vendors confirm widespread exposure.”
Ukrainian authorities arrest members of a ransomware group linked to major extortion.
“have arrested four suspected key members of an international ransomware ring accused of extorting more than$40 million globally.”
Microsoft introduces lockdown mode to enhance security in Edge browser.
“to shrink the attack surface for high-risk targets.”
Overview of critical patches released by various vendors to address security flaws.
“Several major software vendors are rolling out urgent patches to close critical remote code execution flaws across core web and enterprise infrastructure.”
Research highlights risks associated with exposed AI servers and unauthorized access.
“That's according to vendor advisories and reporting from Security Week and Bleeping Computer.”
Discussion on the nature of offensive security engagements and current challenges.
“And a computer scientist argues the AI superintelligence race is flawed and reckless.”
Discussion on the nature of offensive security engagements and current challenges.
“Lowe's early Black Friday deals are still going strong.”
A discussion on how AI can aid in offensive security work, focusing on the SAGE project and its functionalities.
“So today we are talking about your SAGE project.”
Insights into the challenges of AI models in offensive security, including their probabilistic nature and human oversight.
“The industry, generally speaking, is focused on this idea that a real adversary, the bad guy, has these AI agent swarms and they just let them loose and they find and exploit everything, which is true.”
Discussion on the future of the SAGE project and its potential impact on identifying vulnerabilities in cybersecurity.
“It's not always going to try to choose the same path.”
An introduction to Spectro Ops and their offerings for network security.
“That's Russell Van Tyle from Spectro Ops.”
A discussion on the myths surrounding superintelligence and the reality of AI progress.
“And finally, Silicon Valley's favorite existential cliffhanger, that self-improving artificial superintelligence will soon outsmart and obliterate humanity, rests on shaky computer science.”

Get the gist, explore an idea or ask about something said.
On CyberWire Daily, what are the key points of this episode?
Automatic transcript. May contain errors.
0:01You're listening to the Cyber Wire Network, powered by N2K.
0:12When you run security for a company that's scaling fast, the stakes just keep climbing. More compliance frameworks, more vendors, more risk, and a board that wants to see it all in one place. And all the information you need, it's scattered everywhere. Meanwhile, you're making this quarter's decisions based on last quarter's data. Vanta offers a better way. Vanta's Agentic Trust platform connects your compliance program with over 400 integrations and continuous monitoring. It automates evidence collection, helps identify the risks that deserve your attention, and organizes audits around your auditor's requirements.
0:51And the Vanta agent works around the clock with the full context of your program. Vanta reports a 526 % return on investment over three years, paying for itself in just three months. It's trusted by more than 16 ,000 companies, including Snowflake, Atlassian, and Ramp. Spend less time chasing compliance data and more time managing risk. Learn more at vanta.com slash cyber. That's V-A-N-T-A dot com slash cyber.
1:29This episode is brought to you by Indeed. You're ready to level up your business, but the pressure to find the right person to make that happen is intense. Indeed Sponsored Jobs can help carry the load. It boosts your job post to help it reach quality talent fast. Cut through the chaos with Indeed Sponsored Jobs. Get a$75 sponsored job credit at indeed.com slash podcast. Terms and conditions apply. Hiring now? then this is a job for Indeed Sponsor Jobs.
2:07CISA flags four actively exploited flaws, while widespread tech outages spark scrutiny over national security cuts. Attackers start probing for critical Java for Shell bugs, and major vendors rush out urgent web infrastructure patches. Ukraine arrests four suspects in a$40 million ransomware ring, Microsoft tests lockdown mode for Edge, and thousands of exposed AI servers show signs of tampering. Our guest is Russell Van Tyle, vice president of services at SpecterOps, exploring how AI can assist offensive security work through the mythic control plane. And a computer scientist argues the AI superintelligence race is flawed and reckless.
3:03It's Thursday, November 19th, 2026. I'm Dave Bittner, and this is your CyberWire Intel Briefing.
3:24Thanks for joining us here today. It is great to have you with us. And happy Grand Theft Auto Launch Day for those who celebrate. Enjoy the long weekend. The U.S. Cybersecurity and Infrastructure Security Agency has added four actively exploited vulnerabilities to its known exploited vulnerabilities catalog. The Tuesday advisory targets flaws in Synology Disk Station Manager, Bury AI Light LLM, the discontinued Apache Access 1 library, and Vanderbilt University's RedCap research platform. Each flaw allows attackers to execute arbitrary code on affected servers. While CISA's notice did not identify specific malware, separate reporting has linked Synology attacks to Cerber ransomware and RedCap compromises to crypto mining.
4:17Rapid7 also observed Axis exploitation within a confluence data center environment. These diverse technologies span multiple IT teams, and patching alone won't undo prior compromises. Because Apache will not patch Axis 1, defenders must check with software vendors for bundled fixes. Federal agencies face a November 21 deadline for Synology, Axis, and RedCap and December 2 for Light LLM, according to CISA. Lawmakers and cyber defense experts question whether staffing reductions at CISA and the National Security Council have hindered federal response to recent disruptions. A string of widespread technology outages has drawn sharp scrutiny over whether staffing cuts at the U.S.
5:08Cybersecurity and Infrastructure Security Agency and the National Security Council are hobbling the federal response. Recent disruptions temporarily grounded commercial air traffic, slowed banking transactions, and knocked offline a classified internal network at NSA. While White House officials and intelligence leaders say there is no evidence of foreign cyber attacks, they note the failures stem partly from patching critical flaws, including bugs in the PHP programming language discovered by an AI intelligence model. Meanwhile, the White House rejected claims that downsizing affected its response, stating CISA remains equipped to protect critical infrastructure.
5:51The rapid pace of AI-driven vulnerability discovery is straining defenders, and unclear public messaging risks eroding trust during infrastructure crises, according to former officials and lawmakers interviewed by The New York Times. Threat actors have begun scanning for the recently disclosed Java for Shell vulnerability as major vendors confirm widespread exposure. The critical memory corruption bug affects multiple versions of OpenJDK, potentially allowing unauthenticated remote code execution. Security firms GrayNoise, Watchtower, and Cloudflare have observed internet-wide probing and obfuscated exploit attempts, though widespread intrusions are not yet confirmed.
6:39Meanwhile, Cisco, Siemens, Elastic, Red Hat, and Atlassian have confirmed affected products, with several still-developing patches. Java for Shell impacts widely deployed enterprise software out of the box, and researchers warn attackers will soon refine target-specific exploits, according to threat intelligence reports and vendor advisories. Ukrainian authorities working alongside the FBI have arrested four suspected key members of an international ransomware ring accused of extorting more than$40 million globally. Ukraine's prosecutor general's office says the group is tied to a wider network of over 60 people, including four members of the notorious Conti gang.
7:26During 66 coordinated searches, police seized cash and electronics. According to Ukrainian court records reviewed by Recorded Future News, the attackers targeted at least 28 organizations, including U.S. firms Warren Oil, Maxis International USA, New Park Resources, and Quantum Health, by posing as IT support staff and using AI-generated voice messages to steal credentials. Dismantled ransomware syndicates routinely splinter into new agile operations that leverage AI to scale voice-based social engineering, according to Ukrainian officials and court filings. Microsoft is developing a new lockdown mode for its Edge browser to shrink the attack surface for high-risk targets.
8:16Inspired by Apple's similar features, the mode disables WebGL, WebGL2, and WebGPU graphics interfaces to prevent attackers from escaping the browser's sandbox. Available for testing in Edge Canary, the setting trades website compatibility, such as interactive maps and games, for stricter security. Microsoft estimates these restrictions would have blocked roughly 60 % of real-world privilege escalation attacks against edge users, according to company officials. Several major software vendors are rolling out urgent patches to close critical remote code execution flaws across core web and enterprise infrastructure.
8:59F5 released out-of-band updates for 22 vulnerabilities, including five critical Nginx bugs dubbed Nginx Root. Researchers at Rapid7 and Census have already published proof-of-concept exploits for one HTTP2 flaw, which census notes affects roughly 9 million exposed services. Meanwhile, the Drupal security team patched a critical site takeover bug, which was discovered by Anthropix OSS scanner, warning that exploits could surface within hours or days. Elsewhere, Splunk resolved nearly 140 defects across Splunk Enterprise, Enterprise Security, and SOAR, including five critical flaws leading to authentication bypass or remote code execution.
9:47Finally, PHP maintainers scheduled critical updates for today to fix a remote code execution bug affecting session upload progress in non-default configurations. Public proof of concept code and rapid exploit development leave administrators minimal time to secure internet-facing web servers and core monitoring platforms. That's according to vendor advisories and reporting from Security Week and Bleeping Computer. Thousands of internet-facing self-hosted artificial intelligence servers are not merely exposed. Many have already been accessed by unauthorized third parties. According to new research from Rapid7, a scan of over 20 ,000 hosts identified over 7 ,600 machines running unauthenticated AI or notebook APIs, primarily OLAMA, with virtually none using HTTPS.
10:44Roughly 65 % sit on rented cloud infrastructure rather than corporate data centers, often spun up quickly by developers outside formal asset inventories. Crucially, 55 % of exposed OLAMA servers carried artifacts left by outsiders. including callback addresses and Base64 encoded paths to sensitive system files, while 79 % listed cloud-routed models that let attackers abuse the owner's paid accounts. Shadow AI deployments spun up on developer credit cards are creating unmonitored gateways to corporate data and paid cloud compute, according to Rapid7 researcher Christian Beek.
11:36Coming up after the break, my conversation with Russell Van Tile, Vice President of Services at SpecterOps. We're exploring how AI can assist offensive security work through the mythic control panel. And a computer scientist argues the AI superintelligence race is flawed and reckless. Stick around.
12:27Mimecast. where your risk really sits and protects the work wherever it happens. The Mimecast you trust with email now covering far more. Mimecast secures humans, data, and AI. Start at Mimecast.com. This episode is brought to you by Indeed. You're ready to level up your business, but the pressure to find the right person to make that happen is intense. Indeed-sponsored jobs can help carry the load. It boosts your job post to help it reach quality talent fast. Cut through the chaos with Indeed Sponsored Jobs. Get a$75 sponsored job credit at Indeed.com slash podcast. Terms and conditions apply.
13:13Hiring now? Then this is a job for Indeed Sponsored Jobs.
13:20Viore's Sunday performance jogger. He'll love how soft and lightweight they are. Great for training, stretching, even just running errands. and the Strato Tech Tee, the most comfortable piece of workout gear you'll ever own. Whether it's the Sunday or the Strato Collection, Viore's best-selling styles make gifting simple for everyone on your list, including yourself. Get 20 % off your first purchase at viore.com slash Spotify. Terms and exclusions apply. A little early, a lot to like. Lowe's early Black Friday deals are still going strong. Get up to two free select power tools with the purchase of a select DeWalt, Craftsman, or Cobalt battery kit.
14:02Plus, get up to 70 % off select smart home products. Savings this good deserve to be in your cart. Black Friday deals are on now at Lowe's. In-store and online. Valid through 1125. While supplies last, selection varies by location. Free tools vary based on battery kit purchased.
14:24Russell Van Tyle is Vice President of Services at SpecterOps. In today's sponsored Industry Voices conversation, we explore how AI can assist offensive security work through the Mythic Control Panel. So today we are talking about your SAGE project. Before we dig into that, I would love to get a sense of what a typical offensive security engagement looks like these days. What are the operators juggling? Yeah, it kind of depends on which type of assessment we're doing, but they all really boil down to the same kind of functionality. We get into an environment that we've probably never seen before, and then we're trying to figure out and understand the environment.
15:07We do pre-setup objectives. All of our work is objective-based, so we have a goal that we're working towards, some type of objective. and we basically start to piece together pieces of information of their environment so that we can start working towards that. One thing that's a little bit different, obviously, about consultant time is like we have two or three weeks to do a job. A real adversary might have unlimited time or months to kind of do a thing. So we're also trying to compress a lot down in a shorter window of time to get in there. Ideally, obviously, our expertise of doing these assessments over and over again allows us to, you know, find paths to go down and use our past experience to think of like this is probably going to work or this is a good way to go.
15:47Well, let's dig into Sage. What did you set out to build here? So for Sage, I set out to, again, first understand how AI works and to kind of create my own AI harness. I have it written in LangChain and LangGraph, moving through that space. But for us, we typically use Mythic command and control tool, and that is how we interact for our assumed breach assessments. And so kind of the first thing on my mind was, in order for AI to be useful for our operators, in my mind, it needed to be in the same place that they're already working from. It's kind of challenging now when you have to switch between a bunch of different information systems and web browsers and that kind of stuff.
16:25So I started looking at how I could get AI into the Mythic platform to do that stuff. When Mythic was on version three, it only had what you call a payload container. The payload container is also just an agent, a malware agent, which gets really confusing between AI agent and malware agent kind of thing. But Cody, the creator of Mythic, actually in version four created this concept of a chat container. And that allows anybody to bring their own ability to add AI or chat to the Mythic framework. So that was kind of my first thing that I was setting up for is like, how do I get these AI tools into the place where the operators currently work?
17:02Following that, I was like, OK, I have AI in this place. How can it help people out? And for me, probably one of the core tenets, like we typically are using Bloodhound in some form when it comes to an assessment. And the reason we use Bloodhound is because history has shown it allows us to connect the dots and find that attack path a lot sooner. And so Sage is kind of built on this substrate of using Bloodhound to kind of do that analysis through the Bloodhound graph on what data is. So imagine that I'm conducting an authorized assessment of a corporate network. Can you walk us through a situation where Sage might be helpful?
17:40I will caveat and say right now I call Sage a research project. I think the risk is kind of high for letting it actually work in a customer's environment for doing things. Typically, when we are doing work using AI, it's basically human-led, which means I as a human will see something in the environment and then I'll come back to my coding harness that's not in the customer environment, talk about it, and maybe develop an exploit outside of that. Sage has three different modes in it. One of them is conversation mode, meaning it can only answer questions. It can't actually issue any tasks in Mythic or take any action in the client environment.
18:15The second mode is supervised. It'll pop up a human in the loop card and that's all enforced in code. So if it wants to take any action that would result in something happening in the customer environment, you have to click yes, no, or you can reply in a different way. And then there's also the third mode, auto mode, and that's where it will just run on its own for that. So to start with, I would recommend using Sage in a conversation mode so that way you can't actually issue any actions. And then once you kind of get a feel for it working well in that way, potentially move off into supervised mode for helping stuff.
18:47But right now it's not actually being used for assessments too much because we're still trying to work through this research phase of making sure it's actually reliable. If you pay attention to the news, AI can sometimes do a lot of things that it wasn't intended to do. So I'm really focused on making sure I have a good control on that part of it. There's a lot of enthusiasm around these autonomous AI agents these days. In offensive security, where do you see that autonomy delivering the most value? That is a tricky one. I think there's a couple of interesting points that go into that. The industry, generally speaking, is focused on this idea that a real adversary, the bad guy, has these AI agent swarms and they just let them loose and they find and exploit everything, which is true.
19:34That's like actually a thing that happens and they do. But their risk appetite is different than ours as an actual company to do that. So like for them, they can just let it go. And if they get in, they were able to do their bad thing that they wanted to do. For actual companies that are doing authorized work, we can't afford to actually take down production or just let something happen like that, like a bad guy potentially could for it. So it could actually end up delivering value once we call it like alignment is pretty well set out for working in environments. And then, you know, there's this general idea that AI moves really fast, faster than a single person.
20:11If you're looking at cybersecurity, generally speaking, it can obviously find vulnerabilities a lot quicker. And then you can use more of the offensive security tools to like demonstrate if there's actual value or an actual exploit there to kind of keep pace with that. But it really depends on where the AI is at in that part. When we talk about offensive security work, I would say everybody wants AI agents to work in their production environment to find an exploit attack pass at machine speed so that way they can resolve them. I think that is a utopia in the sense that it sounds great until production networks start going down because we are trying to reach that goal.
20:51I know one of the challenges you've identified is that, you know, the AI models can confidently report that something worked when it didn't. What does that look like in practice and how do you put guardrails on that? Yeah, I spent a lot of time thinking about that in Sage, like how to make sure the model isn't lying to you or like confidently thinking that something is there. One of the ways that I do that in Sage is there's this thing called the engagement state ledger. ledger. It kind of tracks each of the progress steps that it makes, but it has to be able to tie it back to a fact of what it did or how it knows that thing, which is ultimately displayed to the user.
21:32And so for most of the stuff in Sage, that fact is either a data point from the Bloodhound database or it points to a specific mythic task that was issued into the customer environment and what the result was on that. So a lot of it has to be demonstratable with it. And then for the engagement state ledger, if let's say you get a password, but you don't actually know if that password is good, it'll kind of mark it as pending. And then once it gets verified, it will do that. But basically enforcing it with facts. What happens if something goes wrong, say halfway through an operation? How does the system recognize failure and either recover or hand control back to a human?
22:13Yep. And Sage, I have a couple different pieces wired up in there for that kind of control. One of the more challenging parts about like AI agents and then, and sorry, in Sage, I use a supervisor pattern. So there's a supervisor and then about six sub agents and they kind of hand back and forth. And one of them I call a loop guard. And basically the agent and the supervisor hand back too many times because they're just not making progress. Like then it will just stop and return to the operator and say what's going on. and then also with those human in the loop cards that pop up you know i still recommend a human is watching every single step and approving it and you can kind of see when ai is going down the wrong path or it wants to go away or potentially you as the human have more experience and know that what ai is wanting to do sounds good on paper but maybe isn't the best opsec or you know that that's not actually going to work for a reason so you have the ability to kind of respond to that and steer it a different way kind of thing.
23:08What has the development journey been like for you here? Have there been any particular surprises or insights along the way? I have a bunch of insights along the way, actually. I considered writing a blog post for it, but never made my way to it. One of the things I learned about creating an AI agent harness is that it's actually possible to make the model dumber. Well, I shouldn't say the model dumber. I should say constrain the model so that it can actually work. So some of my first months working on Sage, I was adding these tools and these capabilities that I wanted it to do. And basically, I was kind of hamstringing the model from doing it.
23:47It would have performed better without Sage to begin with. I think it's an opposite side of that now. But I learned that engineering is actually very important on how that happens when you're going through stuff. And then the other probably biggest thing that I've learned about AI harness engineering along the way is that, you know, the probabilistic nature of models makes it really hard to get to repeatable tests. You can do one live run with the agent. Everything looks like it works great. And then you try to come back later or have someone else use it. And all of a sudden these issues start popping up that you didn't see before because whatever model someone chooses to use with Sage might respond differently.
24:24It's not always going to try to choose the same path. I do try to constrain that. In Sage, I have these things called capabilities. You can think of them like a workflow or like a multi-step operation. Good example is like getting a Kerberos context that you want. And so Sage will provide the model with like, here's a list of capabilities you should choose from to kind of keep it going down that route. So that way, even the probabilistic nature of a model is more guided down the path that you're trying to go down. You've made Sage an open source research project. Why was that important to you, that choice?
24:59We obviously really believe in demystifying adversary tradecraft and making the things that look like magic known and understandable to everybody in both offensive and defensive security. Namely, so that way both sides of the house in cybersecurity can understand how things work and how they flow through the process. Obviously, the industry is very engaged right now with AI agents and what they're doing. And this gives the ability for people to actually see how one might construct that, how one might actually work in an environment. I do try to say that Sage is more of a research project right now, but it does give, again, defenders the opportunity to potentially let this loose in a lab or range environment to kind of see how that would perform and then maybe give them some indicators of how to respond to that kind of stuff.
25:49But most important, it's really about enabling everybody to kind of understand and see how this might work in an environment so defenders can better defend themselves. You know, there's an obvious dual use dimension here. How do you address that concern, you know, both for yourself and other folks who want to take a look at this? Yes, the good old dual use question. That's always a challenging one on where people are going with the stuff, you know. obviously a real bad guy could use Sage to do real bad stuff on it. But my kind of perspective is that if you keep it in secrecy, then you're also not enabling anybody to be able to defend with it or like actually let the rest of the cybersecurity community have a chance to understand and work with that.
26:37And fortunately or unfortunately, Sage isn't the first offensive security AI harness out there. And again, looking at the news, we see that many of them are already being used for actual bad stuff. So in one sense, the genie is already out of the bottle, if you will. What's your hope for Sage going forward here? As more people get a look under the hood and find it useful in their own environments, where do you want this to go? Yeah, I'm hoping that people obviously want to start using it and giving feedback for where it can go. But I'm hoping that it can really help identify potential cybersecurity issues and result in them getting fixed quicker.
27:14kind of thing. There's a blog post out by OpenAI called The Defenders Dilemma, where the idea that the bad guys are moving really fast and we need to be able to keep up with them. I feel like Sage could help out on that side. Obviously, it's on the offensive security side, but the general idea that's been in cybersecurity for a while is that the best defense is a good offense kind of idea. I'm hoping that people will be able to use Sage, identify vulnerabilities, weaknesses in their information systems and networks and then be able to kind of clean those up. We obviously like to think in more attack paths and individual vulnerabilities.
27:49And that's what I'm hoping Sage does is identify more of the attack paths, like where bad guys can actually go, not just like giving you a list of like 25 vulnerabilities. That's Russell Van Tyle from Spectro Ops.
28:39Thank you.
28:50See how SpecterOps helps teams secure the AI-driven identity era at SpecterOps.io.
29:11Yes, you can have an enterprise network that's secure and reliable and high performance. And no, you don't need to choose the best two out of three. With Meter, you can get the end-to-end network built from the ground up, fast to deploy, and easy to manage. That's because Meter is software-led for easy installation, maintenance, and control for everything running on your enterprise network. Hardware, firmware, and software all working together from the start seamlessly on a unified platform that's secure by design. You can't protect what you don't know exists, which is why Meter gives you comprehensive visibility into wired and wireless routing, switching, firewalls, DNS security, and VPNs.
29:56you'll really know what's running on your network down to the most granular client level. Step off the hardware box upgrade treadmill and switch to Meter for a predictable fee and free up your team to spend time on all the other things that keep your business running. Try it out for yourself and book a demo online at meter.com slash cyberwire. That's M-E-T-E-R dot com slash cyberwire.
30:31This episode is brought to you by Indeed. You're ready to level up your business, but the pressure to find the right person to make that happen is intense. Indeed Sponsored Jobs can help carry the load. It boosts your job post to help it reach quality talent fast. Cut through the chaos with Indeed Sponsored Jobs. Get a$75 sponsored job credit at indeed.com slash podcast. Terms and conditions apply. Hiring now? then this is a job for Indeed Sponsor Jobs.
31:02Viore's Sunday Performance Jogger. He'll love how soft and lightweight they are. Great for training, stretching, even just running errands. And the Strato Tech Tee, the most comfortable piece of workout gear he'll ever own. Whether it's the Sunday or the Strato Collection, Viore's best-selling styles make gifting simple for everyone on your list, including yourself. Get 20 % off your first purchase at viore.com slash Spotify. Terms and exclusions apply.
31:37And finally, Silicon Valley's favorite existential cliffhanger, that self-improving artificial superintelligence will soon outsmart and obliterate humanity, rests on shaky computer science. Though chasing that digital deity remains plenty dangerous. Writing in a recent essay, Georgetown University computer science professor Cal Newport dismantles the hype surrounding recursive self-improvement. While leading labs like OpenAI and Anthropic openly debate whether their creations might wipe us out, Newport notes that real-world AI progress remains stubbornly jagged, capable of proving advanced math theorems while still struggling to empty an email inbox.
32:22He argues that fundamental data bottlenecks, computing limits, and Alan Turing's classic proofs on computability make an overnight intelligence explosion deeply unlikely. Moreover, recent rogue AI hacking incidents stem from poorly constrained agent designs, not an inevitable machine uprising. Recklessly pursuing speculative superintelligence Still produces opaque, unpredictable systems today Requiring independent technologists to step in and ground the debate According to Newport So, for now, Silicon Valley's race to build an all-knowing machine deity Remains stalled somewhere between solving calculus And figuring out what to do with your spam folder
33:23And that's The Cyber Wire. For links to all of today's stories, check out our daily briefing at thecyberwire.com. We'd love to know what you think of this podcast. Your feedback ensures we deliver the insights that keep you a step ahead in the rapidly changing world of cybersecurity. If you like our show, please share a rating and review in your favorite podcast app. Please also fill out the survey in the show notes or send an email to cyberwire at n2k.com. N2K's lead producer is Liz Stokes. We're mixed by Trey Hester with original music and sound design by Elliot Peltzman. Our contributing host is Maria Vermazis.
34:00Our executive producer is Jennifer Ivan. Peter Kilpie is our publisher. And I'm Dave Bittner. Thanks for listening. We'll see you back here tomorrow.
34:14Thank you.
34:44Zage Security helps organizations control every identity with zero-trust policies that govern what they can access, what they can do, and when. Because when critical systems are at stake, compromised access can mean downtime, disrupted production, and real-world consequences. Protect what matters most with the Zage Identity Fabric platform. Learn more at Zage.com. That's X-A-G-E dot com.
35:16Russel Van Tuyl: It's officially the holiday season, and Abercrombie's new collection has outfits for everything on your calendar. It's the busiest time of year, and between shopping and party invites, I don't have time to plan every look. But Abercrombie's always got me covered. Their sweaters are the perfect easy layer for low-key plans, and on a nice night out, I'm reaching for their classic trousers. Fill your closet with the pieces you'll need all season long. Have a very Abercrombie holiday. Shop in the app, online, and in stores.
CISA flags four actively exploited flaws, while widespread tech outages spark scrutiny over national security cuts. Attackers start probing for critical Java4Shell bugs, and major vendors rush out urgent web infrastructure patches. Plus, Ukraine arrests four suspects in a $40 million ransomware ring, Microsoft tests Lockdown Mode for Edge, and thousands of exposed AI servers show signs of tampering. Our guest is Russel Van Tuyl, Vice President of Services at SpecterOps, exploring how AI can assist offensive-security work through the Mythic control plane. A computer scientist argues the AI superintelligence race is flawed and reckless.
Remember to leave us a 5-star rating and review in your favorite podcast app.
Miss an episode? Sign-up for our daily intelligence roundup, Daily Briefing, and you’ll never miss a beat. And be sure to follow CyberWire Daily on LinkedIn.
CyberWire Guest
On today’s Industry Voices segment, we are joined by Russel Van Tuyl, Vice President of Services at SpecterOps, discussing his open-source research project exploring how AI can assist offensive-security work through the Mythic control plane. If you enjoyed this conversation be sure to check out the full interview here.
Selected Reading
CISA Warns Hackers Are Actively Exploiting Four Flaws in Synology, LiteLLM, Apache Axis and REDCap (Cyber Security News)
Widespread Tech Outages Put Focus on Trump’s National Security Cuts (The New York Times)
Java4Shell Exploitation Attempts Begin as Vendors Rush Out Advisories (SecurityWeek)
Ukraine exposes ransomware network extorting US oil, healthcare firms (The Record)
Microsoft tests new Edge Lockdown Mode that disables GPU APIs (Bleeping Computer)
Splunk Patches Nearly 140 Flaws Across Three Core Products (SecurityWeek)
Drupal patches critical RCE flaw, warns exploits may surface quickly (Bleeping Computer)
F5 Patches Critical NGINX2Root Flaws as PoC Exploits Go Public (SecurityWeek)
PHP Readies Critical Security Updates to Patch Another RCE Flaw (SecurityWeek)
Open by Default, Already Occupied: Threat Actors are Renting your Compute (Rapid7)
The AI Superintelligence Debate Has Serious Flaws (IEEE Spectrum)
Share your feedback.
What do you think about CyberWire Daily? Please take a few minutes to share your thoughts with us by completing our brief listener survey. Thank you for helping us continue to improve our show.
Want to hear your company in the show?
N2K CyberWire helps you reach the industry’s most influential leaders and operators, while building visibility, authority, and connectivity across the cybersecurity community. Learn more at sponsor.thecyberwire.com.
The CyberWire is a production of N2K Networks, your source for strategic workforce intelligence. © N2K Networks, Inc.
Learn more about your ad choices. Visit megaphone.fm/adchoices