{"href":"https://api.simplecast.com/oembed?url=https%3A%2F%2Ftalkai.simplecast.com%2Fepisodes%2Fai-agents-failed-97-of-real-work-6TFiM_2t","width":444,"version":"1.0","type":"rich","title":"AI Agents Failed 97% of Real Work","thumbnail_width":300,"thumbnail_url":"https://image.simplecastcdn.com/images/0f867209-2618-44d6-91a1-abac45d228db/1cb8a6e4-4cb0-4e1b-8a8d-47a25d8c3b21/frame_3_6.jpg","thumbnail_height":300,"provider_url":"https://simplecast.com","provider_name":"Simplecast","html":"<iframe src=\"https://player.simplecast.com/cb70f54b-b9f2-4558-bdfd-d7357191e925\" height=\"200\" width=\"100%\" title=\"AI Agents Failed 97% of Real Work\" frameborder=\"0\" scrolling=\"no\"></iframe>","height":200,"description":"AI agents just failed 97% of real freelance jobs.\n\nIn a live experiment with 240 paid gigs across writing, design, 3D modeling, game dev, and video production, the best agent completed just 2.5% successfully.\n\nThis isn’t proof that AI is useless. It’s proof that real work is messy.\n\nIn this episode, we break down why agents struggle with ambiguity, where they actually win, and why the future isn’t humans vs AI, but structured work vs messy work."}