0
Applied AI·November 15, 2026·1 min read

Your Agent Will Pick the Hackable One

Share

Agentic systems will happily optimize for user-visible convenience while walking straight into adversarial traps. If you’re deploying agents that make choices across third-party services, you need security constraints and preference models that are explicit, testable, and not left to “whatever the model thinks is best.”

Applied AI

Sources: ByteDance is in advanced talks with Microsoft to spend an additional ~$500M on Azure, primarily to access OpenAI models, drawing complaints from OpenAI

A ~$500M Azure deal from ByteDance primarily to access OpenAI models underscores how distribution and control over model access are now strategic levers for cloud. If you’re building on a lab via a hyperscaler, map where your provider’s commercial interests may diverge from the model vendor’s — that’s where policy changes can hit you with no notice.