Mozilla CTO: Open Source AI Agents and the Fight for Control

Thank you to Cornerstone onDemand for underwriting CXOTALK.

Cornerstone Workforce AI helps organizations connect workforce data, workforce signals, skills intelligence, labor market insights, and agentic AI into a unified decision-making layer for workforce readiness.

Explore how organizations are moving from static workforce planning toward continuous workforce intelligence with Cornerstone Workforce AI.

AI agents now operate inside enterprise systems, accessing data, executing workflows, and making decisions on behalf of the business. But the organizations deploying them often have limited visibility into what those agents are doing or whom they serve. It’s a security and privacy nightmare waiting to happen.

Raffi Krikorian, Chief Technology Officer at Mozilla, argues that the platform choices executives make today will determine whether their organizations own their AI infrastructure or become permanently dependent on a small number of technology companies that do.

Krikorian brings an unusual mix of credentials to this conversation. He rebuilt Twitter's core infrastructure at scale, led Uber's first commercial deployment of autonomous vehicles, and served as CTO of the Democratic National Committee before joining Mozilla, the organization that built Firefox and challenged Microsoft's control of the early web. Mozilla is now applying the same open-source strategy to AI agents that it used to break open the browser market two decades ago.

What we will discuss:

Why AI agents create a different category of enterprise risk: unlike every prior software tool, agents initiate action rather than wait for instructions
The strategic cost of building agent infrastructure on closed, proprietary platforms, where the vendor controls behavior, memory, guardrails, and roadmap
What Claude Mythos revealed by finding vulnerabilities in critical open-source software, and why the most exposed organizations are not always the best equipped to respond
The governance and permissions gap: no mature standard yet exists for defining what an agent can access, what actions it can take, and how those boundaries get enforced and audited
Why the frontier model race may be the wrong question: how small, locally deployable models running on enterprise hardware change the build-versus-rent calculus for CXOs
What Mozilla is building and what a credible open-source alternative to Big Tech's agent platforms looks like today

JOIN LIVE AND ASK QUESTIONS

Friday, May 29
10am PT / 1pm ET

ADD TO CALENDAR

Watch LIVE on CXOTalk → Watch LIVE on LinkedIn →

EPISODE PARTICIPANTS

Raffi Krikorian is CTO at Mozilla. Previously, he was VP of Platform Engineering at Twitter, Director of Uber's Advanced Technologies Center, and first-ever CTO of the Democratic National Committee. He also writes Owners Not Renters, a Substack on open source AI, and contributes regularly to The Atlantic.

Michael Krigsman is an industry analyst and the host of CXOTalk. He has interviewed hundreds of senior executives on enterprise AI, IT strategy, and the organizational conditions that determine whether technology investments succeed or fail.