---
title: "Slides: Building Agent Interfaces: Lessons from Chrome DevTools (MCP) for Agents — Michael Hablich, Google"
category: "slides"
video_id: "_B4Pv9ttFgY"
sourceLabels: ["Public YouTube video frames", "Public YouTube metadata"]
---

# Slides: Building Agent Interfaces: Lessons from Chrome DevTools (MCP) for Agents — Michael Hablich, Google

## Source Video
[Building Agent Interfaces: Lessons from Chrome DevTools (MCP) for Agents — Michael Hablich, Google](https://www.youtube.com/watch?v=_B4Pv9ttFgY)

## Relationship To World's Fair 2026
These slides are extracted from a public AI Engineer YouTube video connected to World's Fair 2026. Speaker-matched clips are supporting context unless later confirmed as exact session recordings; official livestream recordings are day-level/event-level source material.

## Related Scheduled Sessions
- No individual scheduled session mapping has been assigned yet; treat this as an event livestream deck.

## Extracted Slides
![[assets/slides/_B4Pv9ttFgY/slide-002.jpg]]

- Recreated text/layout view: [open HTML recreation](/assets/slide-recreations/slides/_B4Pv9ttFgY/slide-002.html)
- AI slide classifier: `content_slide` confidence `0.94`
- Text source: agent_vision.

Slide text:

> Engineering the future of AI

![[assets/slides/_B4Pv9ttFgY/slide-003.jpg]]

- Recreated text/layout view: [open HTML recreation](/assets/slide-recreations/slides/_B4Pv9ttFgY/slide-003.html)
- AI slide classifier: `content_slide` confidence `0.96`
- Text source: advanced OCR `rapidocr-live/bright-screen/contrast`.
- OCR decision: ready — dense code/browser screenshot with small labels

Slide text:

> hrntad ids at: Iia i e ：
> 3+t3+tar: ve+AV+tala teuteretin PR LRS9 34 CINEMAI
> AIE eT PITSTO
> wdetheLiryeut
> M (rlesl vs
> Flnelea. and get performance insights!
> Welcometo thc track!
> 2O28 AlEngineer AI Engineer EUROPE

![[assets/slides/_B4Pv9ttFgY/slide-004.jpg]]

- Recreated text/layout view: [open HTML recreation](/assets/slide-recreations/slides/_B4Pv9ttFgY/slide-004.html)
- AI slide classifier: `content_slide` confidence `0.98`
- Text source: advanced OCR `rapidocr-live/bright-screen/opencv-adaptive`.
- OCR decision: ready — small bio lines and QR code are better OCR targets

Slide text:

> 腺胶
> MichaelHablich
> AIE
> Tester, developer. project manager. Product Manager @ Chrome Developer Tooling 20+ years of experlence in tech linkedin.com/in/mkhaelhabich Guest lecturer @ Unlversity of Applied Sclence St. P&lten
> B028 AEnginace AlEngineer EUROPE

![[assets/slides/_B4Pv9ttFgY/slide-006.jpg]]

- Recreated text/layout view: [open HTML recreation](/assets/slide-recreations/slides/_B4Pv9ttFgY/slide-006.html)
- AI slide classifier: `title_card` confidence `0.99`
- Text source: agent_vision.

Slide text:

> We built it wrong

![[assets/slides/_B4Pv9ttFgY/slide-007.jpg]]

- Recreated text/layout view: [open HTML recreation](/assets/slide-recreations/slides/_B4Pv9ttFgY/slide-007.html)
- AI slide classifier: `content_slide` confidence `0.98`
- Text source: advanced OCR `rapidocr-live/right-72/contrast`.
- OCR decision: ready — Code-heavy and multi-panel screenshot with small terminal/text content.

Slide text:

> We built it wrong
> All the data × Metrics(lab/observcd): Metrics (fie Available ir URL:https://hablich.dev/ 8ounds:(min:878787192835.max:878792338739} -LCP:135 ms,event:（eventKey: r-3682,ts: -insight name:Cache -LCPbreakdown: -CLS:e. examplequesrion:wnat can I aoto reduce relevant example descript example description:A long cache lifetine can spee insight relevant trace bounds:(min:878787261166, -Rende -TTF8:2ms,bounds:(min:878787192835, summary Semantic ase was nost s:(min:8787 specific impr 78787192835， nize my LCP nodatafor
> example question:what caching strategies
> Engineering the future of Al

![[assets/slides/_B4Pv9ttFgY/slide-008.jpg]]

- Recreated text/layout view: [open HTML recreation](/assets/slide-recreations/slides/_B4Pv9ttFgY/slide-008.html)
- AI slide classifier: `content_slide` confidence `0.98`
- Text source: agent_vision.

Slide text:

> Agents are a different user class
> Intent
> Identify errors on the page
> GUI for humans
> Schema for agents

![[assets/slides/_B4Pv9ttFgY/slide-009.jpg]]

- Recreated text/layout view: [open HTML recreation](/assets/slide-recreations/slides/_B4Pv9ttFgY/slide-009.html)
- AI slide classifier: `content_slide` confidence `0.97`
- Text source: advanced OCR `rapidocr-live/bright-screen/opencv-adaptive`.
- OCR decision: ready — Product/dashboard screenshot with small metrics and labels that OCR can extract more reliably than manual transcription.

Slide text:

> Practical example
> ★ + AIE Test O Averaoe upLift full-2026-03-30T08-23-38 Mr 30, 2:00 AM LCTON OATE geminl_ch TOrh? 37 WODEL gemird-pro-latost SEAYINO HODE SkMt (CLI)
> 4794 84%
> ASSIRTNONS SASSEO ACTIVADON 100%
> 71 101+ 283 (26/26) uk)
> GUIOANCE CONSUUED 96%
> (25/26]Lb
> AEngner. Engineering thefuture ofAl

![[assets/slides/_B4Pv9ttFgY/slide-010.jpg]]

- Recreated text/layout view: [open HTML recreation](/assets/slide-recreations/slides/_B4Pv9ttFgY/slide-010.html)
- AI slide classifier: `content_slide` confidence `0.97`
- Text source: agent_vision.

Slide text:

> Concern 4: Trust boundaries


### Hidden Non-Slide Evidence
- [`slide-001.jpg`](/assets/slides/_B4Pv9ttFgY/slide-001.jpg) — `speaker_stage` confidence `0.99`; speaker on stage, no readable slide
- [`slide-005.jpg`](/assets/slides/_B4Pv9ttFgY/slide-005.jpg) — `speaker_stage` confidence `0.99`; speaker on stage with projected slide in background

Classification audit: `raw/sources/slide-ai-classification/slides/_B4Pv9ttFgY/audit.json`

## Slide-Derived Subjects To Review
Subject extraction uses video title, related session titles/descriptions, transcript context, and OCR text when available. OCR is best-effort and should be reviewed against the embedded slide images.