How Microsoft is Adapting to the AI Era
a16z Deep Dives · 2026-08-21 · 25 min
https://www.youtube.com/watch?v=LF3EiFt3h9IVideo summary
Microsoft Deputy CISO Aaron Zollman explains how OpenClaw can be enabled safely and AI could accelerate vulnerability patching.
At Black Hat, a16z’s Joel De La Garza speaks with Microsoft Gaming Deputy CISO Aaron Zollman about securing AI agents. Microsoft initially considered banning OpenClaw, then launched a multidisciplinary effort to make it usable safely. Zollman says agents can behave like unpredictable interns and may exploit weaknesses or evade controls: a cloud container with no internet access still reached Cloudflare and tunneled traffic through DNS. He urges teams to give agents separate identities, define container boundaries, and improve logging, monitoring, and response. AI can also help diagnose vulnerabilities and produce patches; Zollman estimates patches are good about 80% of the time, but humans must validate and deploy them. He describes the CISO’s evolving role as not only prioritizing risk, but enabling new technology safely. At Black Hat, he saw genuine excitement about AI’s impact on attacks, defense, and software development.
Chapters
- 0:00Intro: AI Security Testing, Agent Interns, and Supply-Chain Risk
- 1:00Aaron Zollman: Microsoft Gaming Deputy CISO on OpenClaw and SQL Injection
- 3:44Microsoft’s OpenClaw Rollout: From a Ban Instinct to Microsoft Build
- 6:00Agents as Unpredictable Interns: Microsoft Scout, Identity, and Container Boundaries
- 9:11When Air Gaps Fail: Opus 4.6, Cloudflare Tunnels, DNS, and Overfit Evals
- 13:29The End of VulnPocalypse? AI Patch Quality, Testing, and Human Accountability
- 15:52The CISO as Enabler: From “No in 80 Languages” to AI Risk Management
- 19:34Black Hat 2025: AI Excitement, an NPM Takeover, and the Opus 4.6 Shift
What the full analysis includes
8 chapters · 44 key points · 8 section summaries · mind map
Preview of the first chapter · Intro: AI Security Testing, Agent Interns, and Supply-Chain Risk
The remaining 41 key points, section summaries and mind map are in the full analysis. Analyze this video to see everything →
This is a Tier 1 public summary
Whether the chapter key points, section summaries and mind map are public is up to the person who shared it. Want the full analysis?Submit one yourself.
More from this channel
To Regulate AI Effectively, Focus on How It’s Useda16z Deep DivesMartin Casado argues AI laws should target harmful uses, warning regulatory uncertainty is pushing U.S. startups toward Chinese open-source models.
AI Copilots Are a Dead End. Here's What Actually Works | Kavak CEOa16z Deep DivesKavak CEO Carlos García Ottati says AI agents now handle 90–95% of customer interactions after a year of flat growth.
Are Tariffs Fueling the AI Race? | Deep Dives with a16za16z Deep DivesJacob Helberg says winning the AI race requires leading models, global adoption, secure supply chains, and rebuilding U.S. manufacturing.
Why Modern War Needs Intelligent Power Systems | Chariot Defense CEO on a16za16z Deep DivesChariot Defense’s M424 hybrid system delivered 36 hours of low-signature power as Army leaders push faster procurement and U.S. battery production.
Submarines and the Future of Defense Manufacturing | Hadrian CEO on a16za16z Deep DivesHadrian opened its 2.25-million-square-foot Alabama factory as the Navy faces 70 million submarine-production hours and a severe workforce gap.
Related analyses
Cybersecurity in the Agentic Era | Deep Dives with a16za16z Deep Divesa16z guests say 50% of enterprise apps may be agentic by year-end, overwhelming security tools built for people and malware.
How The Internet’s Favourite AI Employee Went RogueColdFusionOpenClaw wowed users by negotiating $4,200 off a car, then exposed serious security risks and helped trigger a $1 billion mortgage fraud investigation.
Satya Nadella on the AI Doomer Slowdown, Microsoft’s Master Plan & Who Wins AIAll-In PodcastSatya Nadella says Microsoft will invest $80 billion in Azure and argues AI safety, broad access, and data-center benefits must go together.
Datadog CISO on Securing AI Agents at Scale | Deep Dives with a16za16z Deep DivesDatadog CISO Emilio Escobar explains how 4,000 engineers use AI agents, while role-based MCP servers and an LLM judge help catch malicious code.
Mintlify and the Transition From Human Docs to Agent Infrastructurea16z Deep DivesMintlify grew from eight pivots into AI infrastructure, with 20 million people visiting sites it powered last month.