Clear-eyed analysis of emerging agent technologies, longer horizons and new assurance questions.
A computer-use agent treats pixels, accessibility data and interface state as both evidence and an action surface. Security must bind intent to effects beyond the screen.
Openness can improve inspectability, control and portability, but only when an organisation can own the supply chain, operation and evidence it inherits.
Classify AI by the action path it creates, not the label on the product. The distinction determines ownership, controls, evaluation and release evidence.
Smaller models are not merely budget substitutes. Used behind explicit routing and escalation rules, they can improve latency, locality and architectural control.
Protocols make tools and agents interoperable, but they also standardise discovery, delegation and context exchange. Those interfaces need explicit trust and policy controls.
Agent systems change after release, so evidence must keep moving. Automation can collect and test signals; people remain responsible for judgement and action.
Long-running agents accumulate state, dependencies and chances to drift. Evaluation must test checkpoints, recovery, changing conditions and complete outcomes across time.
Describe autonomy operation by operation. A multidimensional action map gives leaders a better basis for limits, oversight and release evidence than one marketing label.
Self-modification turns runtime learning into controlled software change. Separate what may adapt, require evidence for promotion and preserve a recoverable baseline.
Model selection matters, but production behaviour emerges from an entire dependency graph. Review the system that turns model output into business effects.
KEEP EXPLORING
Read the method, assumptions and questions to ask.