Show the limits of a feed-based story shortlist

Provide synthetic post summaries, sampling dates and methods, and the intended audience. List story candidates with duplicates, missing perspectives and claims needing original sources.

Execution environment and usage

AI / software

Text chat (ChatGPT, Claude or Gemini)

Where to operate

Optional text entry: start a new conversation at chatgpt.com; the supplied-text exercise can also be adapted to Claude or Gemini.

Getting started

  1. Start a new conversation at chatgpt.com and paste the original exercise prompt; do not use retired agent-mode onboarding.
  2. Provide synthetic post summaries, sampling dates and methods, and the intended audience.
  3. Review the requested result against its original inputs: Separate popularity, factual evidence and editorial judgment. Do not perform actions excluded by the prompt.

Setup and plan requirements

Complete, authorized text/sample inputs specified in the prompt; no external account access required.

Prompt portability limits

Portable as a planning/drafting exercise; the historical product is provenance, not a necessary execution environment. This is an adapted exercise, not equivalent execution of the historical agent workflow.

Usage limitations

API client toolset exists; distinguish from separately configured Cowork computer use. Historical tool/source attribution is retained. Not hands-on tested; verify plan, region, workspace and site access separately.

Environment sources

適用情境 / When to use

Provide synthetic post summaries, sampling dates and methods, and the intended audience. List story candidates with duplicates, missing perspectives and claims needing original sources. Analyze the sample without accessing a private feed or publishing. Source record: Only the readable author account is used; neither his account nor original social posts were accessed.

步驟 / Steps

  1. Prepare synthetic post summaries, sampling times and editorial goals.
  2. Separate visible samples, engagement, unchecked claims and editorial judgments.
  3. Check reposts, missing perspectives and uncovered sources.

提示詞 / Prompt

Using [synthetic post summaries/sampling times and method/audience/editorial goal], prepare topic candidates. Label the sample limits and distinguish engagement from factual evidence. Flag duplicates, missing perspectives and claims needing original sources. Support editorial selection without requiring a fixed count or publishing.

留意事項 / Cautions

  • Popularity is not fact verification.
  • Do not access private feeds or publish articles.

來源限制 / Evidence limits

Personalized feeds and engagement do not establish overall importance or truth.

流程說明 / Workflow note

These steps and prompts are original, untested exercises by this site, not the author’s verbatim workflow. Reported results do not guarantee exercise outcomes.

原始工具組合 / Original tool stack

Claude computer use; Dispatch; browser feed

Source publication date: 2026-03-26

Source checked: 2026-10-20

Claude Computer Use: I Used It to Research This Post