DEV Community: Xor

My Current Vibe Stack: The Step-by-Step Guide

Xor — Sun, 26 Oct 2025 16:14:16 +0000

My new vibe coding workflow: Use a powerful web UI with my full repo context to create a plan.Feed that plan to a fast Cline CLI agent for surgical execution.This "dual-brain" approach is 6x faster than using Gemini CLI alone.

Read full article

Blind Conductor and Amnesiac Agents-problems no one talks about

Xor — Thu, 16 Oct 2025 15:04:06 +0000

After months testing every major AI coding assistant, I discovered a fatal architectural flaw burning tokens and time. My workflow split between JetBrains for real work and VS Code for AI experiments.
https://xor01.substack.com/p/blind-conductor-and-amnesiac-agent?triedRedirect=true

My War with AI Agents: When 1000-Line Prompts Met Reality

Xor — Thu, 09 Oct 2025 11:27:03 +0000

Imagine hiring 10 specialists, giving them 1000-line instructions, and getting chaos instead of coordinated work. Welcome to my month of building an AI agent framework.
My goal was ambitious: a fully autonomous system where an army of AI agents a Researcher, an Architect, a TDD-tester, and more would take a task and handle everything from planning to deployment. I designed a complex, multi-phase workflow with protocols like escalation and detailed "Mission Briefs". On paper, it was a perfect, self-managing machine.
In reality, it was an expensive nightmare. The system was plagued by constant file editing errors, infinite loops that burned through tens of thousands of tokens, and "phantom executions" where the orchestrator would mark a task as complete without writing a single line of code. My job turned from developer to full-time prompt debugger.
In desperation, I posted on Reddit, and the solution wasn't a better prompt. It was a single comment that led me to disable two "experimental" checkboxes in the tool's settings. Miraculously, 90% of the file editing problems vanished.
This led to a painful but crucial experiment: what if I removed all my carefully crafted, super-detailed prompts and went back to the default settings? The result was disheartening: the system performed almost exactly the same.
Read the full story with detailed architecture diagrams and my final, simplified workflow: https://xor01.substack.com/p/my-war-with-ai-agents

Hello Devs! My journey from AI-powered coding to a live Chrome extension.

Xor — Sat, 13 Sep 2025 15:12:18 +0000

I've been following the stories and insights here for a while, and I'm finally ready to start sharing my own journey. I'm currently working on a couple of AI-powered tools, and I thought this would be the perfect place to document the process and learn from you all.

My Path So Far: Using AI to Build AI

My development process has been a bit meta. Instead of just building AI products, I've been building my own team of AI sub-agents to help me code. I've created an "orchestrator" that manages a sequence of agents for tasks like research, architecture, coding, and testing.

For example, my refactoring workflow involves setting up pre-commit hooks, letting an LLM agent fix all the warnings, and then doing a final manual pass. It's been a fascinating experiment that has improved my code quality, even if it sometimes breaks things in unexpected ways.

The First Milestone: The Prompt Improver

After a lot of this AI-assisted coding, I'm excited to say my first product, the Prompt Improver, is now live in the Chrome Web Store. It's a simple extension designed to help users write better, more effective prompts for any LLM. Getting that "approved" email was a huge moment for me.

Of course, the journey wouldn't be complete without a hurdle. We're currently navigating a rejection from PayPal for our business account, which is the next problem to solve.

Why I'm Here

I'm really excited to be a part of this community. My goal is to share my path transparently, including the technical details of building with AI agents, the business challenges like payment gateways, and everything in between. I believe the best way to grow is to learn in public.

I'd love to connect with others who are also building in the AI space. Has anyone else had interesting experiences using AI in their own development workflow?

Thanks for reading, and I'm looking forward to the conversation.