Inside the Code Factory: Ryan Carson Live with Tim O’Reilly
Why It Matters
Carson’s AI‑only startup proves that sophisticated SaaS products can be built and scaled without a human engineering team, reshaping cost structures and accelerating innovation across legal and tech sectors.
Key Takeaways
- •Ryan built an AI divorce assistant using only AI agents
- •He raised $2 million seed while remaining sole full‑time employee
- •Developed a “code factory” where AI writes, reviews, tests code
- •Switched from desktop Cursor to cloud‑based Devon for automation
- •Uses design.md specs and AI designer to generate UI from Figma
Summary
In a Code Factory episode, Tim O’Reilly talks with veteran entrepreneur Ryan Carson about his newest venture, Untangle, an AI‑powered divorce assistant built and operated entirely by artificial‑intelligence agents.
Carson recounts how he re‑learned full‑stack development with ChatGPT, launched a minimal viable product on the Cursor IDE, and then raised a $2 million seed round while remaining the company’s only human employee. He describes the “code factory” model where AI writes, reviews, tests, and deploys code, and how he migrated from desktop‑based Cursor to the cloud‑native Devon platform to run nightly cron‑jobs and automated smoke‑tests.
A vivid anecdote illustrates his motivation: his parents’ messy 30‑year marriage and his sisters’ recent divorces inspired him to create a cheaper, less painful process. He also details using a Designjoy contractor to produce Figma mockups, which are converted by Claude‑Design into a design.md specification that AI agents then implement, even generating video recordings of end‑to‑end tests.
The interview signals a broader shift: AI can now replace many traditional engineering roles, allowing solo founders to launch capital‑intensive services with minimal payroll. For investors and incumbents in legal tech, the model suggests faster product cycles, lower burn rates, and a new competitive frontier driven by prompt engineering rather than headcount.
Comments
Want to join the conversation?
Loading comments...