Descriptions:
Bloomberg Technology speaks with Erica Brescia, partner at Redpoint Ventures and Linux Foundation board member since 2016, about the governance challenges posed by increasingly powerful AI models in the cybersecurity domain. Brescia argues that the so-called AI cyber race is fundamentally a governance problem — one requiring structured collaboration between private AI labs, enterprises, and public institutions rather than purely competitive dynamics.
Brescia names Project Glass Wing as a specific initiative she supports, in which AI model vendors and the Linux Foundation are working together to give vetted organizations controlled access to frontier models — allowing them to understand capabilities and build defensive tooling before broader release. She draws a direct analogy to responsible disclosure norms in traditional cybersecurity, where zero-day vulnerabilities are reported privately before public disclosure. She also highlights Jetstream, a Redpoint portfolio company building AI governance and visibility tooling that maps AI activity across enterprise systems and enforces controls.
The interview also covers enterprise AI adoption more broadly, with Brescia pointing to portfolio companies like RAM and Railway as examples of organizations deploying AI across engineering, finance, HR, and go-to-market functions simultaneously. The open-versus-closed model debate — whether smaller open-source models or restricted frontier models better serve the security research community — frames much of the conversation.
📺 Source: Bloomberg Technology · Published April 16, 2026
🏷️ Format: Interview
