Agents, Version Control and Bunnies - Daniel Siegl & David Payr

AI is writing code and video is moving to the browser. Learn how a classic Git feature and a lightweight library solve these modern development challenges.

#1about 8 minutes

Announcing the AWS Football for Good hackathon

Learn about the multi-day, football-themed AI hackathon at the WeAreDevelopers World Congress, focusing on building agentic solutions.

#2about 7 minutes

Solving browser media challenges with the WebCodecs API

The browser's native video elements lack fine-grained control, but the WebCodecs API provides direct, hardware-accelerated access for advanced media processing.

#3about 11 minutes

Building MediaBunny for efficient in-browser media handling

MediaBunny is a dependency-free TypeScript library that uses WebCodecs to read and write media files directly in the browser with a small bundle size.

#4about 7 minutes

Playing a game of Fake or News

Guests and hosts play a game, guessing whether various tech and AI-related headlines are real news or completely made up.

#5about 10 minutes

Using Git Worktrees to manage AI-generated code

Isolate code generated by AI agents into a separate Git Worktree, allowing for review and cherry-picking commits into the main feature branch.

#6about 17 minutes

Maintaining engineering discipline in the age of AI

AI tools require stronger engineering practices, including maintaining a clean Git history, signing commits for accountability, and writing clear documentation.

Notes and resources

Syntevo’s Daniel Siegl joins us to talk all things version control in the agentic age, and Mediabunny’s David Payr hops on to talk about working with media in TS.