{
  "schema_version": "1.0",
  "id": "s8:https://www.youtube.com/watch?v=2JX6JYyQG4Y",
  "slug": "perception-agents-antje-barth-amazon-agi-lab-1gq6jg4",
  "url": "https://feed7.dev/p/perception-agents-antje-barth-amazon-agi-lab-1gq6jg4",
  "title": "Perception Agents — Antje Barth, Amazon AGI Lab",
  "why_included": "Perception agents share the rendered interface with users, turning visual annotations and design rules into agent input and checks. The open-source harness is an early attempt to close computer-use feedback loops.",
  "summary": "Amazon AGI Lab released **two open-source harness pieces**: annotation captures marked screen regions and style details, while verification checks rendered output against design rules and walks user flows.",
  "practical_implication": "For UI-building agents, give visual intent as shared context instead of translating every change into prose. Add **rendered-screen verification** after implementation so the agent evaluates what users see, not only underlying code or tool responses.",
  "agent_context": "Amazon AGI Lab released **two open-source harness pieces**: annotation captures marked screen regions and style details, while verification checks rendered output against design rules and walks user flows.\n\nFor UI-building agents, give visual intent as shared context instead of translating every change into prose. Add **rendered-screen verification** after implementation so the agent evaluates what users see, not only underlying code or tool responses.\n\nThese components are described as a first step, not a complete perception agent. Visual and flow checks still produce a report for human judgment, and the broader problem of reliable end-to-end work across applications remains open.",
  "source": {
    "name": "AI Engineer",
    "url": "https://www.youtube.com/watch?v=2JX6JYyQG4Y",
    "published_at": "2026-07-23T16:00:06.000Z"
  },
  "source_class": "video",
  "content_type": "Video",
  "layer": "agent",
  "domains": [
    "coding",
    "image"
  ],
  "topics": [
    "computer-use",
    "tool-use",
    "agent-reliability"
  ],
  "verification": {
    "status": "source_linked",
    "label": "Source Linked",
    "method": "source_feed",
    "verified_at": null
  },
  "uncertainty": [
    "These components are described as a first step, not a complete perception agent. Visual and flow checks still produce a report for human judgment, and the broader problem of reliable end-to-end work across applications remains open."
  ],
  "lifecycle": "Current",
  "published_at": "2026-07-23T16:00:06.000Z",
  "modified_at": "2026-07-23T16:00:06.000Z",
  "supersedes": [],
  "expires_at": null,
  "formats": {
    "html": "https://feed7.dev/p/perception-agents-antje-barth-amazon-agi-lab-1gq6jg4",
    "json": "https://feed7.dev/p/perception-agents-antje-barth-amazon-agi-lab-1gq6jg4.json",
    "markdown": "https://feed7.dev/p/perception-agents-antje-barth-amazon-agi-lab-1gq6jg4.md"
  }
}