How AI Ate My Website
summary
A system ingests 28 years of content — 2,000 articles, 100 videos, 370 PDFs — transcribes and diarizes media, breaks everything into cleaned chunks, embeds them into a vector index, and retrieves and ranks pieces for GPT‑4–style summarization plus a fine‑tuned keyframe selector. The UI uses a collapsed FAQ‑like Q&A with inline suggested questions and in‑line document/video viewers with citations. PDFs cause noisy chunks so cleanup and a diversity score are required. The deployed prototype at ask.lw.com generally surfaces useful, citable answers.
AI-written summary · about
chapters (10)
- 0:00Welcome and intro
- 1:54AI and large language models
- 6:50Rethinking software with AI
- 11:07AI transforming websites
- 13:58Designing better interactions
- 22:55Processing content with AI models
- 30:26Impact of AI on responses
- 36:49Live demo of ask Luke
- 40:31Adding AI to products
- 44:06Q&A session
related talks
24mDocumentation and AI: How to write right now
01h 00mThe Future of Content: Context-Aware, Centralized, and Written by Robots
45mSEO for AI: Be found by AI search apps like chatGPT
27mBuilding AI applications: Successes and Lessons
39mDocs AI Tooling is Better (and Better for Us) than You Think
37mBest AI Tools for Web Developers / How to Use AI in Web Development
35mThe Podcast Pipeline Nobody Sees: Behind the Scenes with AI Agents and AI gen
39mWhat I Learned Building an AI Documentation Auditor (And What It Found in 500+ Pages of Docs)