0
Applied AI·August 10, 2026·1 min read

Brex assumes its AI agents could do anything — so it watches the network, not the code

Share

Brex’s stance that AI agents are unbounded and must be governed at the network and permissions layer — not by trusting code paths — is a blueprint for enterprise-grade agent deployment. Security and platform teams should be instrumenting traffic, identity, and data access first, and treating agent logic as inherently untrusted.

Applied AI

Mark Zuckerberg’s Answer to Growing AI Safety Concerns Is to Just Trust People to Do the Right Thing

Zuckerberg explicitly framing AI safety as a question of trusting end users — alongside releasing Muse Glimmer as open weight — draws a clear line between centralized control and distributed responsibility. If you deploy open-weight models, assume you’ll need your own governance layer and documentation; relying on upstream policy won’t be enough for regulators or enterprise buyers.

Applied AI

Sources: Apollo, Blackstone, BlackRock, KKR, and other financial groups are partnering with Nvidia on a $500B funding package for AI infrastructure development

A proposed $500B AI infra funding package with Apollo, Blackstone, BlackRock, KKR and others around Nvidia turns data centers into a structured-asset class, not just cloud capex. For operators, this means more creative financing options for large-scale compute — and more pressure to lock in long-term capacity commitments earlier.