---
date: "2025-12-09T05:36:05Z"
categories:
  - llms
  - linkedin
description: I show how to analyze personal digital exhaust using LLMs to uncover hidden life patterns. By exporting data via Chrome Devtools and using AI code execution, I discovered surprising habits about my own YouTube usage.
keywords: [llms, personal data, digital exhaust, chrome devtools, data analysis, behavioral insights]
---

In my Mining Digital Exhaust workshop on Saturday,

- One discovered that they **cycle when life is unstable**, not for fitness.
- Another found that their **buys are good but sells are bad** trades.
- I learnt that I **watch YouTube most at office** (12-4 **pm**), not at home.

How? A fairly straight-forward process:

1. Export your personal data. (Use Chrome Devtools Protocol to scrape.)
2. Upload to ChatGPT, Gemini, Claude, ... and have them analyze **with code**.
3. Have them narrate in the style of your favorite author.

Models are super smart, but everyone has equal access to them. Your personal data is unique. Combine them to get something **powerful**.

Workshop summary: https://sanand0.github.io/talks/2025-12-06-mining-digital-exhaust/

<div class="video-embed"><iframe src="https://www.youtube.com/embed/rVACVezdDxc" title="YouTube video" loading="lazy" allow="accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope; picture-in-picture" allowfullscreen></iframe></div>

Organized by Hasgeek at the Sahaj Software office in Bangalore.

![](https://sanand0.github.io/talks/2025-12-06-mining-digital-exhaust/sketchnote.webp)

[LinkedIn](https://www.linkedin.com/posts/sanand0_in-my-mining-digital-exhaust-workshop-on-activity-7403991793068892160-ix31)