Risky Business Features

In this podcast episode SOCRadar CISO Ensar Seker and James Wilson chat about the company’s deep dive into the Fortibleed campaign. A small investigation into a curiously open directory on an unknown server expanded into the discovery of an attack that targeted 400,000 Fortinet devices. As Ensar says, each time the SOCRadar team pulled a single thread, it led to a tapestry of AI-enabled cybercrime. They uncovered custom initial access, persistence and packet sniffing tools, as well as direct links to the INC and Lynx ransomware operations. Most interesting though is the use of AI to design, implement and operate all aspects of the campaign across a team of 20 individual actors. Operating more like a modern software company than a traditional cybercrime gang, Fortibleed serves as our first in-depth look at the future of cybercrime. Show notes SOCRadar Links FortiBleed Campaign to INC and Lynx Ransomware Operations

In this podcast episode Brad Arkin joins James Wilson to discuss how defenders can get ahead of the late-running bugpocalypse. While we’re confident the offensive cybersecurity capabilities of frontier and open-weight LLMs are real, attackers don’t yet seem able to fully utilise them. This creates a window of opportunity for defenders to tackle the threat. There are a few well-funded and seemingly overlapping industry efforts under way including Athena, Akrites, and Patch the Planet. But, as Brad says in this interview, there’s too much focus on fixing bugs and traditional vulnerability triage, and not enough on exploring how to make entire classes of vulnerabilities inert. Show notes

In this podcast episode James Wilson chats with Karsten Nohl about his research into using local LLMs to replace cloud AI in security code reviews. In essence, Karsten created a hybrid code reviewing system where both cloud and local models are used to orchestrate, triage outputs, and write reports. In this system, only the local LLMs have source code access, with the cloud models used to manage the local models. In this “source-local” review technique, the source code never leaves the local endpoint, which is a requirement for some reviews. But funnily enough, Karsten was able to use this system to generate findings that were as impressive as when using frontier models directly. In a nutshell, Karsten proved it’s possible to use locally-hosted, open-weight models running on commodity hardware to produce findings comparable to those discovered by frontier cloud models. This episode is also available on YouTube. Show notes Beyond Fable: Can a Local LLM Replace Cloud AI for Security Code Reviews Mythos smythos! How to find 0day with lesser models

In this podcast Patrick Gray and James Wilson chat with Decibel Partners founder and Managing Partner Jon Sakoda to talk about pitching cybersecurity startups to VC firms in the AI age. Coding agents and large language models have made it easier than ever to create software products, but despite this, the bar for what interests an investor is still largely the same. Everyone can run the marathon, but it’s usually the same few folks who finish first. So tune in to hear Jon share with us his wisdom on when to start the conversation with investors, how to leverage the experience of the founder community, and what founders should watch out for. This episode is also available on YouTube Show notes

In this podcast episode James Wilson and Brad Arkin talk about how to safely use open weight large language models in the enterprise. The cost of frontier models was already driving interest in freely available open weight models like DeepSeek, Kimi and Qwen. But now the US government is forcing Anthropic to pull its Fable and Mythors models from the market, the argument for having greater control over your own AI stack is stronger than ever. But as you’ll hear in this episode, the model itself is just one component of the complex tech stack you’ll need to spin up if you want local inference. There’s a lot of moving parts, each of which comes with its own supply chain risks. So whether you’re hosting these models on your own hardware or via a SaaS provider, there’s a lot to ponder! Show notes

In this solo podcast episode, James Wilson breaks down the current state of AI model jailbreaks. If you’ve somehow missed the story, last week Anthropic released its Fable 5 and Mythos 5 models to the public. In the name of safety, both models were guardrailed up the wazoo, but that didn’t stop a bunch of jailbreakers from figuring out how to bypass at least some of their safety restrictions. In response to these guardrail bypasses the White House issued an export control directive on the models, citing national security concerns. But was the Trump administration right to do this? Do these jailbreaks represent a threat to the security of the USA, or was the export restriction overkill? Tune in to find out! Show notes Pliny the Elder on Fable 5 Jailbreak whoJumper's response to Pliny ConfusedPilot: Confused Deputy Risks in RAG-based LLMs

In this podcast episode, James Wilson is joined by Open Source Malware Security co-founder Paul McCarty to talk about the supply chain attack mitigations coming in NPM v12. NPM disabling (by default) auto-run install scripts and dynamic dependencies is a positive step forward… but it’ll take years for this new version to be adopted, and these changes do nothing to prevent malicious packages being imported into projects. Further, Paul thinks disabling these features by default will introduce friction that will cause them to be re-enabled. When the choice is “this builds” and “this is less prone to malware”, the former will always win. Show notes

In this podcast Brad Arkin joins James Wilson to talk about how the fear of being left behind in the AI era means enterprises are taking risks that would have been considered insane just a couple of years ago. Fears around outages or being hacked have been trumped by fears of being labelled an AI laggard. So where are we all going? Say hello to tech debt-riddled, vibe-coded apps, crazy dependencies on AI providers, and an emerging threat landscape that can’t be mitigated by a contemporary SOC. Sounds like fun, eh? Show notes

In this solo episode, James Wilson takes a detailed look at TeamPCP. It started off by launching clumsy attacks against misconfigured Kubernetes clusters in September 2025. But by February this year, TeamPCP had skilled up and was smashing global software supply chains in the highest profile attacks of 2026. TeamPCP upskilled and turned the software development ecosystem into its personal credential harvesting machine. Here’s how TeamPCP did it, and what we can learn from it. Show notes

In this podcast James Wilson chats with Brad Arkin about why software supply chain attacks have gone from rare, once-in-a-while disasters to an operational problem affecting mainstream enterprises almost daily. AI has made attackers faster, and “vibe coding” means the number of environments pulling packages from the internet has gone to the moon. It also means legacy tooling that seeks out the bad packages and cleans them up isn’t enough. Package cooldown windows won’t fix this either. But all hope is not lost! Tune in to this podcast to find out how you can get a grip on the disaster de jour! Show notes

In this episode, Theori’s Brian Pak and Andrew Wesie join James Wilson to discuss why the CopyFail exploit was publicly disclosed before Linux distributions had their patches ready. As you’ll hear in this episode, mistakes were made and lessons learned. It’s worth a podcast, too, because in our opinion this incident foreshadows the inevitable problems that open source software will face in the unfolding vulnpocalypse. Show notes

In this edition of Risky Business Features Ollie Whitehouse, the CTO of the UK’s National Cyber Security Centre, joins Patrick Gray and James Wilson to talk about why “patch faster” will only get organisations so far in the face of the AI “bugpocalypse”. As Ollie explains, organisations will need to reduce internet-facing attack surface and make better architecture decisions as 0day discovery speeds up. This episode is also available on YouTube. Show notes

In this podcast James Wilson and Brad Arkin workshop the advice they think the industry needs to hear when it comes to deploying agentic AI in the enterprise. Relegating agentic AI to non-sensitive and low-risk tasks doesn’t deliver value, and avoiding all risk stalls progress. James and Brad discuss the phases of AI adoption and contrast what a great plan looks like, versus an overly cautious one. Show notes

In this podcast James Wilson chats with Niels Provos about his research into using older AI models to successfully hunt for 0day vulnerabilities. Niels has had a long and prolific career in cybersecurity, having worked as a Distinguished Engineer at Google and then heading up security at Stripe. His interest in AI bug hunting was piqued recently when one of the Mythos 0day vulnerabilities that received lots of attention happened to be in code he wrote for the OpenBSD project 27 years ago. It got him thinking: Are these frontier models really that magical? Or could we replicate their findings with some clever orchestration instead of relying on the model’s smarts to find bugs with a single prompt? As it turns out, this was worth looking into. Niels’ orchestration framework, Iron Curtain, works extremely well. This episode is also available on YouTube Show notes Finding Zero-Days with Any Model Security Blueprints

In this podcast James Wilson and Brad Arkin chat about emerging trends in AI agent identity and credential management. Brad was formerly the CISO of Adobe, Cisco and Salesforce, and is now working with all sorts of companies that are deploying AI. With everyone now in at least a large-scale pilot of agentic AI, the issue of how to manage agent identities and credentials is still an unsolved problem. But, some interesting patterns are emerging. Show notes

In this solo episode of Risky Business Features James Wilson explores how distillation techniques are both a legitimate way to train smaller models, as well as a way to steal model capabilities. It’s not just a problem for frontier labs! Any LLM-based product could have its competitive advantage stolen through these attacks. James covers: High-level concept of distillation Why it matters including close/open-weight/open-source explanation Types of distillation and the prompts used The distillation pipeline end to end Distillation at scale and mitigation techniques Hardware resource constraints for distillation Show notes Self-Instruct: Aligning Language Models with Self-Generated Instructions Alpaca: A Strong, Replicable Instruction-Following Model Vicuna: An Open-Source Chatbot Impressing GPT-4 with 90%* ChatGPT Quality Orca: Progressive Learning from Complex Explanation Traces of GPT-4 Zephyr: Direct Distillation of LM Alignment Stealing Part of a Production Language Model Microsoft probes if DeepSeek-linked group improperly obtained OpenAI data, Bloomberg News reports Detecting and preventing distillation attacks

In this episode, Anthropic’s Nicholas Carlini joins Patrick Gray and James Wilson to talk about advancements in AI-driven vulnerability research and exploit development. Nicholas’ talk at the recent [un]prompted conference demonstrated how Anthropic’s Opus 4.6 could find and exploit vulnerabilities in popular open source projects. In the short few weeks since then, Anthropic announced a new model that’s already identifying hundreds of bug fixes across critical software. Nicholas talks us through the work he does at Anthropic, what’s possible and the limitations with current frontier models, and where this goes from here. This episode is also available on YouTube Show notes

In this episode, James Wilson is joined by entrepreneur and investor Yaniv Bernstein to discuss Anthropic’s Mythos through the lens of startups and growing businesses. Yaniv is Google’s former VP of Engineering, and is former VP Eng and COO at Airtasker. He’s now an investor and advisor to startups and he co-hosts The Startup Podcast. Show notes

In this episode, James Wilson is joined by Brad Arkin who provides a CISO’s perspective on Anthropic’s Mythos. As former CISO at Adobe, Cisco and Salesforce, Brad’s perspective challenges the notion that finding and fixing exploits makes us safer. Show notes

In this episode of Risky Business Features, James Wilson chats to professional hacker Jamieson O’Reilly about Anthropic’s Mythos and the impact it could have on offensive security. Jamieson is CEO of DVULN and co-founder of Aether AI. He’s been hacking into organisations for more than a decade, and knows a thing or two about combining AI and offensive security. Show notes

In this episode, investigative journalist Geoff White joins James Wilson for a look into the complex machine that is North Korea’s IT worker infiltration scheme. They discuss the interview process, what happens once the workers are actually hired, how value is maximised for the regime, and how the money moves around. It’s even more diabolical than the headlines divulge. Show notes Lazarus Heist Podcast

In this episode, James Wilson chats with Brad Arkin (former CISO of Adobe, Cisco and Salesforce) to talk about the mounting pressure that CISOs are under in the AI era. Attackers are operating at unprecedented scale, and internal users are adopting AI faster than security teams can keep up. This requires CISOs to bend on things that would have otherwise been a hard-no in the past, and revisit some fundamental controls that might have seemed too difficult previously. Show notes

In this sort-of-solo episode, James Wilson is “joined” by one of his OpenClaw AI agents for a chat about whether or not an LLM can understand, modify or even create a sophisticated nation-state grade iOS exploit kit. Technically this podcast is James having a conversation with himself, but the exchange is illuminating. It turns out LLMs can really help with finding 0day, even in mature code repos like WebKit. Show notes

In this interview you’ll hear former NSA executive Rob Joyce and former CIA cyber intelligence leader Andy Boyd talk to host Patrick Gray about how AI is changing the state of art in offensive security. Recorded in front of a live audience at the Decibel Oasis side event next door to the RSA Conference in San Francisco, the trio also talk about why a series of iOS exploit chain leaks don’t seem to be stirring up a scandal. Show notes

In this episode of Risky Business Features, James Wilson and Brad Arkin discuss the attack that devastated medtech company Stryker. It turns out the attackers used Microsoft’s inTune to wipe the company’s devices, but what else could they have weaponised? This podcast basically turned into an incident review of the Stryker incident. Enjoy! Show notes

James Wilson delivers his take on the state of the Model Context Protocol (MCP) in this solo episode of Risky Business Features. Despite MCP being the technology that made Large Language Models useful and AI Agents possible, the models have shown us they want to use something else instead. They want to use the shell directly, and that is going to have serious cybersecurity consequences. Show notes

In this podcast James Wilson chats with Brad Arkin about how enterprises can better deal with attackers logging in with valid credentials. Stolen identities, weak special-use credentials, and over-scoped API keys are the new zero-day and they’re abundantly available to attackers. Sadly, the solution here isn’t as simple as deploying phishing resistant MFA. Fixing this takes an enterprise identity strategy. Show notes

Join James Wilson in this solo podcast as he takes a (ridiculously) deep dive into the Coruna exploit kit. James was a software engineer and senior manager at Apple for many years, so he has an intimate knowledge of iOS internals. He even worked alongside the people who wrote the software that the Coruna kit exploits! This long-form solo podcast follows the chain of exploits from watering-hole website right through to full device compromise with many incredible leaps over security boundaries along the way. You’ve heard Coruna described as a sophisticated, nation state-grade exploit kit, and James will explain to you why that description is fitting. Show notes

In this edition of Risky Business Features James Wilson chats with cohost Brad Arkin about what it’s like being a CISO for a global company when a war starts. How do you deal with a branch office full of important key material being abandoned? What about cloud infrastructure that’s in a data centre that falls into enemy hands? And if your staff are okay, are any of your key suppliers going to face problems? As you’ll hear, being a wartime CISO is less about adjusting your SIEM sensitivity because the Iranians are coming to get you, and more about figuring out how to deal with very real threats to life and infrastructure. Show notes

In this podcast James Wilson chats with Brad Arkin about North Korea’s sprawling fake IT worker ecosystem. From fake interviews, to stolen identities, basement laptop farms and IP-KVM tricks, the North Koreans are operating a whole employment fraud industry. Brad and James discuss how the scheme works in practice and the technical detection challenges defenders now face, like dealing with stolen or borrowed identities, bribed verification checks and multi-person operational chains. They also dig into why enterprises are largely on the back foot, and why there’s no single product you can buy to solve this. As the former CISO of Adobe, Cisco and Salesforce, Brad has some firsthand experience dealing with this stuff! Show notes

In this debut feature conversation in the Risky Business Features feed James Wilson sits down with Brad Arkin, the former CSO of Adobe, Cisco, and Salesforce, to talk all about AI pentesting. Finding and fixing bugs is great, but does it materially improve the overall security of a product? What’s the point of a pentest if the tester can’t walk you through their findings when it’s over? Is “bugs per dollar spend” really the measure of value in security testing? We hope you enjoy this podcast! Show notes

AI agents are being deployed with the same trust-by-default architecture the early internet had. Same mistakes, MUCH faster timeline. OpenClaw has hit 180K+ GitHub stars. But in the past week: 341 malicious skills on ClawHub were distributing Atomic Stealer ZeroPath disclosed a Browser Relay vuln enabling cross-tab cookie theft CrowdStrike, Cisco, and Bitdefender all published enterprise advisories VirusTotal partnered with ClawHub to scan uploads Korean tech firms (Kakao, Naver, Karrot) banned it on corporate networks 1,000+ Open PRs, 250+ less than 24 hours old. But how does this thing actually work? Join James Wilson as he explains why banning these types of agents doesn’t work, why browser sessions are now API surfaces, and why your organisation needs to think of these issues early or be condemned to decades of catch-up programs. Show notes