Build the agent. Don't build its computer
nibzard
Steel argues that teams should build AI agents but choose an existing provider-managed “agent computer” service instead of building their own execution infrastructure. It says in its browser fleet that changing browser VMs from 4 virtual CPUs to 2 improved cold starts under the same host load. As a result, teams can control the environment configuration and agent behavior while shifting capacity, isolation, and recovery operations to a provider to reduce reliability and operations load.
Why it matters
Agent teams should specify their execution environment, test recovery on real workloads, and start with managed infrastructure unless ownership offers a concrete advantage. The evaluation checklist covers persistent files, expired logins, interrupted external actions, access boundaries, and cost per accepted result.