QuantEnt / Perspectives

Ideas for accountable enterprise AI.

Exploring business authority, agent governance, and the controls that make enterprise AI useful.

Emerald glass geometric forms and fine orbital trails enclosed by a separate solid metal ring on a charcoal background.

October 4, 2026Trent Walker

The Pain Axis isn’t the right way to think about AI.

A critique of the Pain Axis framing argues for external controls over AI actions, fleet-wide limits, and human escalation when an agent cannot reach its goal within its mandate.

Read article

An agent inside a dashed sandbox sends one action through a gate marked with a check to its target, while a second action is stopped before the gate.

October 3, 2026Trent Walker

Where NVIDIA (and Others) Got It Wrong

NVIDIA's Open Agent Safety Platform, like guardrails, identity governance, vaults and observability, controls what an AI agent can reach. This whitepaper argues the control belongs at the business action instead, sets out six mistakes in the current stack, and credits what NVIDIA and others got right.

Read article