AI Attackers Are Here: Threat Research, R&D Labs, and A16Z on What’s Actually Changing
The threat actor your security program was built to stop no longer exists.
Over the last two years, something fundamental changed in offensive capability.
AI didn’t just give attackers better tools. It removed the constraints that shaped how attacks used to work — skill, time, and human effort. Campaigns that once required sophisticated operators can now be executed by a single person running AI-assisted tooling. Targeting that took weeks now takes minutes. Entire phases of the attack lifecycle can be automated.
The adversary has structurally evolved. Most defense programs have not.
This session brings together three perspectives that rarely share the same room: real attack telemetry, AI security research, and the venture ecosystem funding the next generation of security companies.
Not predictions.
Not marketing narratives.
What they’re actually seeing happen.
Michael Clark — Head of Threat Research, Sysdig
Lives in the telemetry every day, watching attacks against cloud-native infrastructure evolve in real time.
Iman Ghanizada — Founder, Eqlipse (Former Global Head of Autonomic Security at Google)
Runs an AI R&D lab. His team recently used AI-driven offensive tooling to red-team organizations — and discovered just how trivial many attacks become when human expertise is removed from the equation.
Patti Degnan — Partner, Andreessen Horowitz
Former CISO of Notion, where she built the company’s security organization during its growth to over 100 million users. Now at Andreessen Horowitz, she works closely with security founders and CISOs across the ecosystem, giving her a unique vantage point on how both attackers and defenders — and the companies building security tools — are rapidly evolving.
The conversation will explore what many security briefings avoid:
• How AI is collapsing the skill gap between elite attackers and low-skill operators
• What happens when powerful offensive tooling inevitably spreads beyond research labs
• Why most detection models are calibrated for attacker behavior that no longer exists
• What the startup ecosystem is building in response — and where the industry still has no clear answers
Who this is for
Security leaders who want an unfiltered look at how AI is changing offensive capability — and why the next few years will force a rethink of how modern security programs operate.
Attendance is limited.