Super! 😀. You should publish it along with the document you've written for building the scraper.

I've never come across such a comprehensive doc in my decade long experience so far from any engineering leadership. It'll be an amazing case study in how to leverage AI for planning and scoping.

@unclecode.bsky.social once told me the he does Socratic debates with AI over stack, library, or logic choices. Asking the model not just to agree but to challenge him - weigh pros and cons. Now that's the philosopher/thinker like functionality we need from an AI in IDE, not just an errand boy!

This is where AI could truly shine. What we have now? Over-stepping autocomplete.
Great for portfolio apps, but not real-life codebases. The iPhone moment for AI dev tools? It’ll come from a team that brings scoping, planning, and technical decisions into the IDE with AI.

Most devs don’t work like bricklayers. They work in bursts. Shaving off seconds/minutes isn’t really a game-changer.

The real work differentiating $5 and $5000-quality software, currently happens outside the IDE—in docs like Word/Notion, where specs are flushed out and technical decisions are made.

Cursor, Copilot... these tools are just improving existing IDE experience. No one's rethinking the IDE itself for the age of AI. It’s like making horses faster instead of building an automobile

The fundamental problem? They're optimising for developer productivity, not software quality.

A company still basing their interviews on leetcode questions, is NGMI. Honestly, I would rather run for the exit than sit through the interview. The same goes, if they don't allow me to use a model to solve a challenge. It's like calculating logarithms with a log table instead of a calculator