What if your livestream layout was just a web app? Sergey will show how HTML, CSS, and JavaScript can power interactive live-streaming experiences inside OBS’s embedded Chromium browser and how WebSockets and Cloudflare Workers can turn viewers into active participants.

Coding agents are only as effective as the context they can see—and for long-lived software, the repository is their real prompt. This practical talk explores how clear specifications, enforceable constraints, observable acceptance criteria and fast feedback loops can turn repeated AI mistakes into improvements that compound over time.

What if an AI agent could customize its own UI outside the chat and update it on the fly? As humans, some tasks that we work on are better done visually than in text format. That's why I started this experiment, where I give an agent a skill and teach it how to work in a highly visual workspace it can extend and update with custom UIs. I'll walk you through how we designed a skill, a CLI, and a front-end app that work in tandem to make it possible. How we spent weeks figuring out the best and most minimal design system an agent can use to generate "slop-free" designs. The project is completely open-sourced and available on https://moi.computer/

Most teams hear "AI design system" and reach for the thing they already know: a component library, now with a chat panel bolted onto it. I think that's the wrong shape. A design system in an AI-assisted codebase isn't a catalogue of components; it's a set of architectural contracts that a coding agent can actually be held to. Tokens, composition rules, and constraints that are legible to a model, not just to a designer in Figma.

The talk walks through what that looks like in practice: what an agent gets wrong when your system is only documented in prose, how to encode intent so the generated code is boringly consistent instead of plausibly wrong, and where the boundary sits between what you let a model generate and what you pin down by hand.

What attendees can expect to take away:

It's aimed at frontend and product engineers, and at design-system maintainers wondering what to do about all this. No prior AI-tooling experience needed.

This talk explores the architecture and implementation of OAuth-based authentication for remote-hosted MCP servers. I'll walk through the specific challenges I encountered while building an MCP server at Tigris — including managing dynamic client registrations, token lifecycle management and scope handling across distributed environments.

I'll also cover UX choices like reducing friction in the auth flow, handling token refresh gracefully, and managing session state.

Our coding agents are getting smarter, but the workspace we drop them into is probably not enabling them to use their full potential. The decisions, failed attempts, conventions, and product knowledge are still crucial context scattered across people’s heads, Confluence, old conversations, and forgotten pull‑request comments.

During this talk, we'll walk through the shift from the “good old” prompt engineering to agent experience, then I'll show a practical AX experiment at my current company that led to a better workspace for both agents and the people using them.