Today, we're proud to finally bring @RemoteThreat out of stealth with the first integrated, end-to-end offensive cyber operations (OCO) platform. Much of the focus today is on patching as fast as possible against noisy AI-driven attacks. We expect the next phase to arrive soon: models optimized for stealth, connected to sophisticated payloads and nation-state-tier offensive tools. Those attacks will be designed to evade detection, maintain access, and adapt as defenses respond. Enterprise security teams need to start testing against that level of sophistication now. Evolving U.S. policy is also opening new opportunities for commercial offensive cyber. We see an emerging market for dual-use technology that strengthens enterprise security and supports authorized operational missions. That convergence of capability, policy, and operational demand makes the timing consequential. RemoteThreat is built for internal enterprise red teams and warfighters. Its composable capabilities let operators select, combine, and customize tooling around their environment and objective. Internal red teams can vary sophistication to match the adversary they’re emulating. Warfighters can configure capabilities for the mission and adapt as conditions change. These capabilities are available to human operators, hybrid human–AI teams, and AI operators. The platform connects planning, capability development, command and control, execution, and analysis- with existing tools, human oversight, and approvals integrated throughout. It’s been a significant journey building this alongside my co-founder, Shawn Jones @anthemtotheego . We’re now 15 strong and growing, with some of the most talented security researchers I’ve ever worked with. Given the sensitive nature of our work, some of our people will remain out of the spotlight. I couldn’t be prouder of this team and what we’ve built together. If you lead an internal enterprise red team or build capabilities for the warfighter, I’d welcome a conversation. businesswire.com/news/home/2…

Big personal news: my labradoodle officially is cancer free as of yesterday! Naturally we had to celebrate with bubbles, rawhide and a new toy lobster over a sunset last night. Here’s to another 13 years!

I spoke with Lorenzo Franceschi-Bicchierai at TechCrunch about frontier-model verification programs and cyber guardrails. My view is that the current approach is still creating too much friction for legitimate offensive-security researchers. The same community that has spent decades finding, validating, and helping remediate critical vulnerabilities is eager and ready to apply frontier models to that work at far greater speed and scale, yet they're currently being ignored, under utilized, and hindered with the current guardrails, even when they are "verified" and given access to slightly less restrictive, yet inconsistent, guardrails. What frontier labs fail to recognize is that their models were built on decades of community research, late nights finding bugs, and exceptional open-source work, yet they have not meaningfully brought that same community into the fight. The answer lies in broader, verified access with materially reduced restrictions for legitimate researchers, clear rules of engagement, appropriate monitoring, and real accountability for abuse (aligned to existing law). By keeping capable researchers constrained or pushing them toward less accountable foreign and open-weight systems, we risk sidelining some of the people best positioned to strengthen critical infrastructure before AI-enabled attacks accelerate.

The sun is setting on submissions for OAIC 2026. The Call for Papers closes July 31. Selected speakers receive: • A complimentary three-day OAIC conference ticket • Three hotel nights at a 4 star beachside resort • An invitation to our Michelin-star VIP speaker dinner • Meals, drinks, and official parties throughout the event • Travel stipends for qualifying academics and seed-stage startups • The opportunity to present alongside leading offensive AI researchers in an intimate, vendor-free environment Join us at a luxury seaside resort in San Diego and help define the next generation of offensive AI. Submit before the sun sets on July 31.

OAIC's CFP is now open! The first conference dedicated to the cutting edge of the offensive use of AI is returning for its second year. Speakers will enjoy three nights at a four-star beachfront resort, which includes all meals and drinks, three exclusive parties, and a Michelin-star welcome dinner. Please see sessionize.com/offensive-ai-… for accepted topics.

The frontier labs don’t want you to know this, but stickers add +10% inference speed to local AI racks 🤫

After 13 incredible years at IBM, I’m stepping away to start building something new. IBM has been a defining part of my life. I’m especially proud to have founded and grown the Adversary Simulation team and to have served as a Distinguished Engineer focused on AI and offensive security. What began as a small idea became one of the greatest red teams, pushing the boundaries of what real-world adversary simulation can look like - made possible by an exceptional group of people I learned from every day. I’m deeply grateful to the X-Force leadership and to the teammates, mentors, researchers, and operators who made the journey what it was. The team is in great hands and will continue to do important, impactful work. Looking ahead, I’m excited to focus on building what’s next - continuing to raise the bar for offensive cyber and exploring how AI can enable high sophistication offensive cyber operations and capabilities at a speed, scale, and intensity never before seen. It’s still early, but I’m looking forward to sharing more soon. I truly appreciate everyone who’s been part of the journey.

We're look to expand some local meetups in DC and UK in the future, hit me up if you want to help out!