Conference affiliation: Founder/YouTuber · T3 Tools & YouTuber · 2026
On this page
Theo Browne is the founder and chief executive of Ping Labs, founder of T3 Chat, and creator of the T3 Stack. Better known online as Theo or t3dotgg, the former Twitch engineer builds developer tools and argues that AI is fundamentally changing what small software teams can accomplish.
From streaming infrastructure to developer tools
After working at Twitch and another startup, Browne founded Ping Labs, which joined Y Combinator’s Winter 2022 batch. Its initial product, Ping.gg, helped livestreamers collaborate through broadcasting software such as OBS. He also started producing YouTube videos to better understand creators and continue the technical conversations he had valued at Twitch.
Browne created the T3 Stack and create-t3-app, an opinionated, modular toolkit organized around TypeScript, Next.js, tRPC, and full-stack type safety. In 2023, he released UploadThing, addressing the disproportionate infrastructure required for application file uploads. His other projects include QuickPic, which converts SVGs into high-resolution PNGs, and Stripe implementation guidance.
His more recent products extend that practical orientation into AI: T3 Chat provides a chat interface, while T3 Code supports AI coding workflows. His joke about T3 Code’s public repository reflects the irreverent style that also characterizes his developer commentary.
What AI changes about building software
Browne argues that advances from reliable tool use to longer-running tasks and model orchestration require developers to rethink both their tools and ambitions. His AI Engineer conference talk crystallizes several positions:
Project ambition should expand with model capability. Work that once required a startup can become a side project, allowing small teams to attempt products previously considered impractical.
Inherited developer interfaces deserve scrutiny. Browne compares familiar AI coding interfaces with early smartphone applications that imitated physical objects: attachment to terminals, programming-language identities, and established source-control conventions can obscure better workflows.
Markdown-driven AI automation can replace bespoke software. He replaced a more elaborate pull-request review and prioritization service with scheduled instructions that inspect repositories, assess outstanding work, publish a report, and generate his daily priorities.
Product breadth becomes more achievable. His proposed Lakebed cloud platform would combine capabilities such as hosting, authentication, and databases. He emphasizes extensible architecture while acknowledging that broad initial functionality does not equal the reliability of established infrastructure providers.
Judgment matters more when code becomes disposable. Browne challenges the practice of merging unsuitable pull requests simply because substantial effort went into them; cheaper code generation makes rejecting weak solutions and starting over more viable.
