正在加载视频...
视频加载失败
I tested Gemini Pro 2.5 as my main coding model for 40+ hours. Here're 2 documents that are working brilliantly well with Gemini. "App flow document + App flowchart." This made my Cursor workflow 10x better. Here's why it is working: ↓
11 条评论

Gemini Pro 2.5 has a 1 million token context window. In simple terms it can scan upto 100,000 lines of code instantly to understand full codebase. Claude & GPT models has 200k context window and that's why they struggle with large codebases.

But large context window doesn't mean Gemini doesn't hallucinate or assume things. That issue is still there but I'm able to fix hallucinations by 95% in Gemini with my 'Context Boundary' technique. You need to provide context with coding docs: - PRD - App flow doc - Tech stack doc etc But here's how 2 docs are working really good with Gemini.

I noticed that Gemini Model is understand the App flow document really well. App flow document is like a user journey: "users start its flow from landing page, then they sign in and land on Dashboard. On dashboard my app should have 5 pages.... ) This document creates and end to end blueprint in Gemini's context memory. Tips: Don't use bullet points for this document. AI explains the flow better in paragraphs.

To make it even better I added a MermaidJS diagram (Flow chart) to explain the app flow. This addition made Gemini understand the app from technical and visual aspect. These 2 docs explained AI "what to build."

I generate all my coding documents with @CodeGuidedev which is compatible with all AI coding models. - Gemini Pro 2.5 - Sonnet 3.5/3.7 - GPT o3-mini/o1 Plus it works with all Top AI tools: - Cursor - Windsurf - Replit - Bolt - Lovable Cline and Firebase studio coming tomorrow.

Bottomline: Gemini Pro 2.5 is the best coding & debugging model as of now. But to fix it's hallucinations and context awareness, attach coding docs. Use @CodeGuidedev to generate docs as it also provides a flowchart diagram of the app and 50-step implementation plan.

This is the biggest productivity cheat code right now. Kiss reading documents goodbye. You can get an instant summary of any document with this tool.

How do you update the extensive documentation as the project evolves? So far I've seen if you try to do anything even slightly unusual (non-CRUD) then it cannot be planned out in a waterfall ahead of time. It has to be tested and different approaches tried.

I am testing a custom mode in Cursor that can auto update docs on the go after each successful request. Will ship this soon.

@CodeGuidedev @threadreaderapp unroll

@CodeGuidedev I agree 🤙

