AI Didn’t Kill the Web, It Moved in! — Olivier Leplus (AWS) & Yohan Lasorsa (Microsoft)
TL;DR
AI has permeated every stage of the web development lifecycle, from coding agents that use customizable "skills" to browser-native AI debugging tools, fundamentally shifting how developers build, test, and optimize applications.
🤖 Agentic Coding & Custom Skills 3 insights
Skills extend agent capabilities
Lightweight text-based plugins following open specifications allow coding agents to access domain-specific expertise and external tools like GitHub CLI or Playwright.
Workflow automation via agents.md
Developers can define repeatable workflows in a standard agents.md file, enabling autonomous implementation of features, video recording, tunnel creation, and mobile notifications.
Custom skill creation
Developers can build custom skills (e.g., sending Telegram notifications or creating public tunnels) to integrate personal workflows and external services into the agentic pipeline.
🔍 Browser Control via MCP 3 insights
Chrome DevTools MCP server
The Model Context Protocol (MCP) enables agents to programmatically control Chrome DevTools, allowing automated clicking, form filling, screenshots, and performance tracing.
Automated performance testing
Agents can simulate various network conditions (2G, 3G, fast internet) and generate detailed reports with Core Web Vitals metrics (LCP, CLS) and optimization recommendations.
Seamless IDE integration
MCP servers integrate directly into development environments, giving agents access to browser capabilities without manual intervention.
💻 Native AI in Browser DevTools 3 insights
Contextual error debugging
Chrome DevTools now features built-in AI assistance (when enabled) to explain console errors like CORS issues and suggest fixes directly within the browser.
Performance optimization insights
Developers can analyze network requests and performance traces using AI chat interfaces to identify render-blocking resources and optimize LCP scores.
Integrated debugging workflow
AI assistance eliminates context switching by allowing developers to debug failing requests and analyze CSS directly within familiar DevTools panels.
Bottom Line
Developers should adopt "skills" and MCP servers to automate their specific workflows while enabling AI features in browser DevTools to streamline debugging, as these integrations represent the new standard for efficient web development.
More from AI Engineer
View all
How Google DeepMind Runs Agents at Scale — KP Sawhney & Ian Ballantyne, Google DeepMind
Google DeepMind engineers Ian Ballantyne and KP Sawhney demonstrate their internal "Antigravity" agent platform, revealing how the organization manages massive-scale deployment through strict quota controls, hybrid model architectures, and collaborative multi-agent workflows while grappling with token consumption costs and evaluation complexity.
Your Agent Is an Infinite Canvas — RL Nabors, Dressed for Space
Rachel Lee Neighbors argues that chat interfaces are merely a transitional phase like the CLI was to GUI, demonstrating how HTTP-based MCP servers and interactive MCP apps can turn agents into an 'infinite canvas' for rich web experiences while eliminating inefficient DOM scraping through emerging Web MCP standards.
Prompt to Pipeline: Building with Google's Gen Media Stack — Paige & Guillaume, Google DeepMind
Paige from Google DeepMind demonstrates how Gemini 3.1's native multimodal capabilities and AI Studio enable developers to prototype complex media pipelines—from video analysis to code execution—that can be deployed to production with a single click, while advising against building infrastructure that frontier models will soon absorb.
Fast Models Need Slow Developers — Sarah Chieng, Cerebras
As AI coding models like Codex Spark reach 1,200 tokens per second—20x faster than current standards—developers must abandon bad habits formed during the era of slow inference. This talk outlines a practical playbook for "slow development": orchestrating fast models for execution while using slower, smarter models for planning, and treating AI as a real-time pair programmer requiring constant verification and strict context management.