Why can't technology companies avoid central points of failure? A casual chat about the clothes the emperor isn't wearing and an introduction to a real-world battle-tested anti-fragile approach to building and deploying applications: Pear.

Silicon Valley is buzzing again. Dozens of nightly AI meetups, packed co-working spaces, and a lot of hype. But the real action isn’t in a postal code. It’s on GitHub.
This talk breaks down why JavaScript is powering the AI boom globally, and why you don’t need a Valley badge to be part of it. Just open an issue and commit.
AI-assisted development is becoming part of modern workflows. The goal isn’t to redefine open source. It’s to make sure that as tools evolve, accountability, transparency, and maintainership stay intact.

Let's be honest, nowadays most of us see an error, copy it into ChatGPT, and hope for the best. I've done it. You've done it. We've all wasted 30 minutes going back and forth with AI when our browser could've shown us the answer in 2 minutes.

This talk is about remembering that our browsers have incredible debugging tools we've forgotten how to use. Through live demos of real bugs, I'll show you how breakpoints, call stacks, and the network tab can solve problems maybe faster than any AI assistant.

You'll learn:

• How to ditch console.log hell and actually watch your code run
• Breakpoint tricks that pause your app so you can inspect what's really happening
• How to read call stacks without your eyes glazing over
• Network tab techniques that expose exactly why your API is misbehaving
• When debugging tools beat AI (and when AI is actually helpful)
 I'm not saying don't use AI—I use it all the time. But there's something powerful about understanding your own code well enough to fix it yourself. Plus, you'll be way faster, and you'll actually learn something instead of just copy-pasting solutions you don't understand.

Accessibility is often framed around screen readers and contrast ratios. However, in many parts of the world, accessibility starts with unreliable or expensive connectivity.

This talk reframes offline-first architecture as a core accessibility concern. Using React Native examples, we’ll explore how network assumptions silently exclude users in low-connectivity regions and how offline-first patterns dramatically expand who can actually use our apps.

We’ll cover practical strategies for designing user flows that remain usable without a stable network and how these decisions improve resilience for all users, not just those offline.

Most marketing scripts and cookie banners run on the critical path. They block the main thread, delay first paint, and inject third-party code before the page is interactive.
Consent managers often make this worse by synchronously loading vendors, mutating the DOM repeatedly, and re-hydrating UI after page load. The result is slower LCP, higher TBT, and unstable CLS.

The fix is not fewer scripts, it is better orchestration. Treat consent as infrastructure, load vendors after intent, and keep compliance off the critical path.