🔍 Read the full analysis: Want Better AI Agent Decisions? Try These Four Rules on ThorstenMeyerAI.com
Get hardware and tech essentials delivered free with Prime
- Fast, free delivery on millions of items
- Prime Video, Amazon Music and more included
- Member-only deals all year
TL;DR
An October 27 article on ThorstenMeyerAI.com recommends four ways to manage AI agents: count the human work their outputs create, test ideas without committing to all of them, set stop rules for recurring tasks, and limit what coordinating agents bring to a person. It cites Anthropic research reporting that employee-directed agents spent about 80% of session time waiting for human input, while warning that the study reflects one company.
Thorsten Meyer published four rules for managing AI agents on October 27, arguing that their output can create more review and coordination work than users expect. The article draws on Anthropic research about its own employees, which reported that agents spent about 80% of employee-directed session time waiting for human input.
Meyer’s recommendations are to count the human work that follows an agent’s output, run small experiments without turning every promising result into an ongoing project, set an end point before making a task recurring, and give coordinating agents a limited attention budget. He suggests that a coordinating agent should default to silence and bring forward no more than three decisions a day, each with a recommendation, supporting evidence and the consequence of taking no action.
The article says Anthropic’s study found that by August 2026, employee-directed agents logged six times as many hours as the employees directing them. It also reports that reading all agents’ output would take non-technical employees an estimated half of their working time. These are findings and estimates from the company’s study, not evidence that every workplace will see the same pattern.
Meyer’s first rule is to ask who will read an agent’s work, who will act on it and whether the resulting work is worth doing. The article uses examples from Anthropic customer stories, including a contractor who checks blueprint measurements and asks the agent to show its work. The examples illustrate possible uses; they do not independently establish how reliably agents perform across different businesses.
The new bottleneck
Agents don't get tired. Your attention does.
Anthropic measured how its own staff work with AI agents. The agents now do far more hours. The limit is the person who has to read, check and decide.
An agent drafts the report, the code or the summary in minutes.
- Reading
- Checking
- Deciding
- Fixing
- Explaining it to others
Illustration of the idea, not a measurement.
“An army of Claudes, with every one asking for my attention.”An Anthropic engineer, quoted in the study
Source: Anthropic Economics team, How employment is changing at Anthropic, 16 October 2026. Employee-directed interactive sessions; human time counted as 40 hours a week per head. One company studying its own staff, who chose to adopt the tools.
Powered by Thorsten Meyer AI
The Human Cost of Agent Output
The article addresses a practical problem for people adopting AI agents: generating output may be cheap, but reviewing and acting on it still takes human time. If employees receive more reports, proposals or alerts than they can check, automation can shift work rather than remove it. That matters for managers deciding whether to deploy agents and for workers expected to supervise them.
The four rules focus on controlling this downstream workload. Counting reviewers and follow-up tasks can expose costs that are hidden when a tool is judged only by how quickly it produces a result. Limits on recurring tasks and notifications may also help prevent an accumulation of low-value work. These are the article’s recommendations, rather than tested findings that the rules will produce a particular productivity gain.
Products Worth Considering
LG 24G414B-B 24-Inch UltraGear Full HD IPS Gaming Monitor
LG 34GX90SA-W 34-inch Ultragear WQHD OLED Curved Gaming Monitor
LG UltraGear 45GX900A-B 45-Inch Curved OLED Gaming Monitor
As an affiliate, we earn on qualifying purchases.
What Anthropic’s Study Reports
The Anthropic study, published on October 16, 2026, examined how the company’s own employees used AI agents. According to the source article, the study reported that the share of sessions devoted to meta-work, largely managing agents, rose from 10% in fall 2025 to 22% in spring 2026. It also said written standing instructions for agents grew from a few hundred to 8,500 over a year.
Those figures have limits. Meyer notes that Anthropic employees chose to adopt the tools and that aspects of the company’s work may not apply elsewhere. The results therefore offer a case study, not a universal measure of agent use. The article’s four rules are Meyer’s response to the attention and management burdens he says he encounters in his publishing and software work.
“an army of Claudes, with every one asking for my attention.”
— An engineer quoted in Meyer’s article
How Far the Findings Apply
It is not clear how representative Anthropic’s findings are of other companies, occupations or agent systems. The source describes a study of Anthropic’s workforce, whose employees adopted the tools by choice; it does not establish that the same time use or management burden occurs across the wider economy.
The article also presents four operating rules, not results from a controlled test of those rules. It does not provide measured evidence that limiting notifications or adding end dates will reduce workload by a specific amount. The customer examples are attributed accounts, and the reported business earnings have not been independently verified in the material provided.
Testing the Rules in Daily Work
Meyer presents the rules as practices readers can apply to agents they already use, rather than as a product launch or formal workplace standard. The next step for readers is to track who reviews an agent’s output, what actions it generates and whether recurring updates are actually read. The article does not announce a follow-up study or a formal evaluation of the rules.
Further evidence would be needed to determine whether these practices reduce supervision time across different teams. Until then, Anthropic’s figures remain a company-specific account, while the article’s recommendations are practical proposals whose results will depend on the work and people involved.
Key Questions
What are the four rules for managing AI agents?
Meyer recommends counting the human work after an agent responds, testing several ideas without committing to all of them, setting stop conditions for recurring tasks, and limiting how many decisions a coordinating agent sends to a person.
What did Anthropic report about agent use?
As described in Meyer’s article, Anthropic reported that by August 2026, employee-directed agents logged six times as many hours as their users. The study also found that agents spent about 80% of session time waiting for human input. These figures relate to Anthropic’s employees, not all workplaces.
Does the study prove AI agents create more work?
No. The reported findings describe agent use and management work at one company. They do not prove that agents increase work in every setting or that the same results apply to other organizations.
What attention limit does Meyer suggest?
He recommends that a coordinating agent default to silence and raise no more than three decisions a day. Each should include a recommendation, evidence and what may happen if the user does nothing. This is Meyer’s proposed rule, not a research-backed universal limit.
Source: ThorstenMeyerAI.com
