Class Central is learner-supported. When you buy through links on our site, we may earn an affiliate commission.

YouTube

AI Paradox - Faster Coding, Slower Shipping

JavaScript Conferences by GitNation via YouTube

Overview

Coursera Flash Sale
40% Off Coursera Plus for 3 Months!
Grab it
Explore the paradox of AI-assisted development in this 25-minute conference talk by Google Engineering Leader Addy Osmani, who addresses why development teams are coding faster but shipping slower in the AI era. Learn to transition from traditional coding to becoming an "orchestrator of agents" while avoiding the pitfalls of "vibe coding" that can degrade code quality and create technical debt. Discover practical strategies for escaping the "70% trap" where AI generates code quickly but fails in production, and master context engineering techniques that help AI tools understand legacy codebases without relying on ineffective prompt engineering. Examine Google's internal data revealing a 91% increase in PR review times and explore solutions like Trio Programming to combat the code review crisis and prevent skill erosion. Understand the shift from the inner development loop (writing code) to the outer loop (shipping features) and learn how to map AI tools effectively across the developer lifecycle. Address the trust gap affecting 46% of developers who distrust AI output and discover Chrome DevTools MCP for AI-browser integration. Navigate the challenges facing senior developers in the "Last Mile" problem and explore strategies to maintain junior developer hiring and skill development in an AI-dominated landscape.

Syllabus

00:00 - 01:34 | Intro: from "Will I be replaced?" to AI-Native Engineer
01:34 - 04:23 | The shift: evolution from Implementer to "Orchestrator of agents"
04:23 - 04:57 | The spectrum: "Vibe coding" vs. Managed engineering
04:57 - 05:42 | A few words about JS Nation 2026 in Amsterdam
05:42 - 07:39 | Mapping AI to the developer lifecycle Inner vs. Outer Loop
07:39 - 08:09 | Google’s internal data: measuring real velocity gains
08:09 - 10:44 | Why prompt engineering is dead enter context engineering
10:44 - 14:27 | Practical pattern: Spec-driven development and "learnings" files
14:27 - 15:39 | The trust gap: why 46% of devs distrust AI output
15:39 - 17:15 | Tool reveal: Chrome DevTools MCP AI seeing the browser
17:15 - 20:20 | The code review crisis: why PR times exploded by 91%
20:20 - 21:44 | The "70% Problem": why the Last Mile is hardest for Seniors
21:44 - 23:06 | Skill erosion and the threat to Junior hiring
23:06 - 25:24 | Solution: "Trio programming" and No-AI challenges

Taught by

JavaScript Conferences by GitNation

Reviews

Start your review of AI Paradox - Faster Coding, Slower Shipping

Never Stop Learning.

Get personalized course recommendations, track subjects and courses with reminders, and more.

Someone learning on their laptop while sitting on the floor.