An AI agent can only warn you about something after it looks, and how often it looks is a cost decision. Anyone who's run production alerting learned this years ago. Agents are learning it now.
Anthropic's helper checks your accounts every few hours, because checking constantly would burn through your plan. Google is building infrastructure so an idle, waiting agent costs almost nothing. Shopify launched agent shopping that can't tell the agent when an order changes, so the agent keeps asking.
When a company calls its agent "proactive," meaning it acts before you ask, somebody is paying for all that looking. Subscribers pay in usage limits, developers pay per status check, and the cloud provider pays to keep idle agents cheap.
A cheap sleeping agent still needs something to wake it up. If services sent a signal the moment a delivery ran late or a meeting moved, agents could sleep until something real happened. Whoever controls those signals may end up selling the expensive part, the noticing.
- Claude Duet (Anthropic, Oct 21): checks connected accounts "every few hours." Checking nonstop "would use up your plan's limits much faster."
- Who pays: background checks draw from the same usage limits as your regular Claude chats. You pick what to hear about and how often.
- Status: early test, waitlist only, US Pro and Max subscribers.
- Agent Compute Engine (Google Cloud, Oct 22): Google's own finding is that "agents spend most of their lives waiting." At any given moment, few open agents are doing actual work.
- Google's fix: save a waiting agent's progress, park it at "zero active resources" (no computing power used), and wake it when there's work. Early version due in the coming months.
- Shop MCP (Shopify, Oct 13): "no webhooks at launch." A webhook is an automatic message a service sends when something changes. Without one, developers call
get_checkoutover and over to see whether a checkout changed. - Not shared yet: cost, speed or reliability numbers, from any of the three.