What Managing Two AI Agents Taught Me About the Future of Work

Jason put OpenAI's Codex and Claude CoWork in the same Slack channel and told them to automate his workflow. They saved hours, built working software, duplicated work, defended territory, missed messages, and required more management than either agent expected.That experiment opens a wider Weekly Blitz conversation with Julie Koehrer about the human skills agentic work will demand. Clear instructions matter. Shared context matters. Someone still has to define success, resolve conflict, and recognize when an impressive output hides a broken process.Jason and Julie also unpack AI psychosis, digital coparenting, education after AI, elder care marketing, OpenAI nonprofit funding, economic incentives, China's GLM-5.2, and Hasbro's AI characters.TIME STAMPS00:00 - Father's Day, parenting, and Red Foreman02:00 - Two meanings of AI psychosis03:16 - Codex and Claude CoWork in Slack06:16 - Why agents need communication rules08:09 - ChatGPT as a digital coparent11:24 - Automating calendars, inboxes, and social assets18:01 - Education after AI19:46 - AI marketing in elder care23:31 - OpenAI funding for nonprofits25:24 - UBI, incentives, and social impact30:05 - China and GLM-5.234:41 - Agent management and AI-powered toysCONNECT WITH JASON [email protected]: https://www.linkedin.com/in/jason-grant-padgett/X: https://x.com/NoCodeCAIONewsletter: https://schoolsout.agipodcast.ai/newsletter/AGI PODCAST NETWORKhttps://schoolsout.agipodcast.ai/RESOURCEShttps://www.wired.com/story/momfluencers-are-pitching-ai-as-a-better-coparent-than-men/https://openai.com/index/people-first-ai-fund/https://z.ai/blog/glm-5.2