Uploaded July 2026 | Updated September 2026, 2 weeks ago
Kyle Balmer publishes daily across six platforms using a system of roughly 10 AI agents. In this AI Superstream talk he maps every step, including the places where he deliberately takes the wheel back. Kyle covers the AI-generated morning briefing he uses to select the day's topic, the voice notes he records on a walk to inject his own perspective into the research brief, and the live stream he runs to generate a human-authored transcript that anchors everything downstream. From there, his agent suite turns that transcript into newsletter drafts, LinkedIn and Instagram carousels, and short-form video scripts, but he insists on a manual rewrite before anything goes live, records his own videos instead of using AI avatars, and runs a feedback loop that compares his edits to the original AI drafts to improve future output. If you're trying to build a content system that scales without losing your voice, take a look.
Follow O'Reilly on:
LinkedIn: linkedin.com/company/oreilly
Facebook: facebook.com/OReilly
Instagram: instagram.com/oreillymedia
BlueSky: https://bsky.app/profile/oreilly.bsky.social
TikTok: tiktok.com/@oreillymedia
Kyle Balmer publishes daily across six platforms using a system of roughly 10 AI agents. In this AI Superstream talk he maps every step, including the places where he deliberately takes the wheel back. Kyle covers the AI-generated morning briefing he uses to select the day's topic, the voice notes he records on a walk to inject his own perspective into the research brief, and the live stream he runs to generate a human-authored transcript that anchors everything downstream. From there, his agent suite turns that transcript into newsletter drafts, LinkedIn and Instagram carousels, and short-form video scripts, but he insists on a manual rewrite before anything goes live, records his own videos instead of using AI avatars, and runs a feedback loop that compares his edits to the original AI drafts to improve future output. If you're trying to build a content system that scales without losing your voice, take a look.
Follow O'Reilly on:
LinkedIn: linkedin.com/company/oreilly
Facebook: facebook.com/OReilly
Instagram: instagram.com/oreillymedia
BlueSky: https://bsky.app/profile/oreilly.bsky.social
TikTok: tiktok.com/@oreillymedia






![Inside the Code Factory: Ryan Carson Live with Tim O’Reilly
Ryan Carson has spent 25 years building developer communities, conferences, and Treehouse, which taught over a million people to code. His latest company, Untangle, is an AI-powered divorce assistant—and hes building it entirely alone. Just $2 million in seed funding, his guidance, and a team of agents running while he sleeps.
Ryan sat down with Tim to walk through the code factory powering Untangle: a system where agents write and review the code, run the tests, triage error reports, and monitor the production environment under his oversight. In their conversation, they covered the Ralph Wiggum loop (Geoffrey Huntleys deceptively simple technique for giving agents large goals across multiple context windows) and the power of primitive loops, how Ryan used Claude Design and a human designer to build a full design system he can now reproduce with AI, what attorneys really think about Untangle, the economics of running a company of agents, why the narrative that programming is going away gets the abstraction story exactly backwards, and why, even when you can automate nearly everything else, you still cant automate the judgment call about what to build.
There isnt a magic wand still, Ryan told Tim. You can build faster, but whether youre building the right thing, and doing it better, is something [else].
Follow OReilly on:
LinkedIn: https://www.linkedin.com/company/oreilly/
Facebook: http://facebook.com/OReilly
Instagram: https://www.instagram.com/oreillymedia
BlueSky: https://bsky.app/profile/oreilly.bsky.social Inside the Code Factory: Ryan Carson Live with Tim O’Reilly](https://i.ytimg.com/vi/vHB_0RCDOWk/mqdefault.jpg)



