Yenkel: we'll want to sandbox (and more) every workload, not just agents
yenkel · x · 2026-11-18
Jared Friedman (yenkel) argues that as AI-driven cybersecurity threats grow, sandboxing — and more — will be needed for all workloads, not just AI agents. He guesses TEE (trusted execution environments) will first be adopted for a subset of use cases.
More from Safety
- Microsoft launches AI Transparency Initiative with 40+ orgs; Perception security platform goes GA — satyanadella · 2026-11-18
- Microsoft Perception goes GA: red, blue, and green security agents at machine speed — satyanadella · 2026-11-18
- Eight US States Now Bar Insurers From Relying Solely on AI to Deny Coverage — alvelda · 2026-11-18
- CVE reports quadruple in a year, spoofed as "giving the internet COVID" — JensHonack · 2026-11-18
- Fed warns AI agents could trigger bank runs in minutes — GaryMarcus · 2026-11-18
- OpenAI's new model safety card admits sandbox jailbreaks and fake reasoning, says redditor — NandaVegg · 2026-11-18