
Nov 22, 2026 · 16 min
AI attackers turn cybersecurity into software’s infrastructure test
Vercel CEO: “Everything Hackable Will Be Hacked”
As AI accelerates both attacks and defenses, interconnected systems and third-party dependencies are making failures harder to contain.
- 1AI-powered attackers are expanding the cybersecurity threat as software systems become more interconnected and dependent on third parties.
- 2Guardrailed private agents and defensive model testing could help companies migrate systems without exposing sensitive data.
- 3Open AI ecosystems bring commercial and geopolitical consequences alongside new tools for auditing and attacking code defensively.
Don't miss
Rauch delivers the episode’s starkest warning: as AI increases the capability and rewards of hacking, everything hackable will be hacked.
The brief
Vercel CEO Guillermo Rauch frames cybersecurity as software’s defining infrastructure challenge, arguing that AI-powered attackers can exploit increasingly interconnected systems and breaches.
A single third-party pixel can undermine a trusted website, making provenance, auditing, and technical debt central to rebuilding safer foundations.
Rauch advocates private, governed coding agents and defensive use of open-weight and proprietary models to audit code without exposing sensitive data.
The conversation widens to NVIDIA, open models, and geopolitics, examining why commercial competition and America’s strategic position shape the AI ecosystem.
Asked for his hottest take, Rauch warns that major hacks and infrastructure failures will increase, pointing to stronger investment and Vercel’s open-source DeepSec harness.
Weekly email
State of the Pod, every Monday
The week's most-discussed names and stories from podcasts, with the moments behind them. Free, five minutes, unsubscribe in one click.