Get live statistics and analysis of Alex ⚡ Fixing AI Context Loss's profile on X / Twitter

Does affiliate marketing work? Like % fee on MRR for all customers affilate bring style? In theory sounds great, but how to find affiliates?

MRR for my little SaaS finally is growing! - I don't have a big account - I don't don't paid ads - I don't have a lot of reach But I do have a great product so I kept trying and trying different strategies until something finally seems to be working! I am so happy!!!

Docs & Rules We can overcome the context window limitations by getting these 2 things right. I built a vibecoding protocol that allows me to reduce hallucinations by 99% regardless of the size of the codebase. I am currently using this protocol to vibecode a product to make it available to everyone. The first piece is almost ready. Can't wait to push it out

If you want to use Webhooks and SSE events in your express.js app, don't do it. I spent 2 days trying to figure out why I was not able to update my onboarding flow in realtime using SSE events. It was working in DEV, but not in PROD. Turns out that SSE and HTTP/3 are not good friends. This was a sneaky bug, but now it's fixed, so hopefully I can save some other people from the pain of debugging this

First day since launch with 0 new customers 🥲 But at least I am done with the freemium model today, so I'll take this as the win of the day! LFG 🚀

Just tested Claude Code's planning ability WITH vs WITHOUT the Context Engineer MCP The ask: "Plan an auth system for my app" Then I Asked Claude Code: "Which plan is better?" Claude's response: "Context Engineer MCP Plan is Significantly Better... My plan was more of a high-level sketch, while the MCP plan is a production-ready implementation blueprint."

Do you use any planning tools before jumping into code or do you just go prompt by prompt?

I am about to launch the Context Engineer MCP. It's packed with a lot of cool features. - it can detect when the user asks for a complex feature and triggers structured planning. - it gathers requirements by asking questions and analyzing your code base - it generates PRD, technical blueprint and implementation plan. All these docs are super tailored and full with both business and technical details and charts that allow the coding agent to build with a very high degree of precision. Oh, and it doesn't have any dependencies. All works in the ide like cursor etc. all the AI tasks leverage the LLM used in cursor. So no extra bills for AI tasks!

The Context Engineer MCP is ready - fully tested and working 🔥 Now I just need to make a nice landing page for it! This is what it will be able to do: ✅ Zero External AI Costs - Runs directly in Cursor ✅ Universal Codebase Analysis ✅ Intelligent Feature Categorization - Auto-detects single vs multi-category features ✅ Complex Feature Intent Detection - Recognizes complex asks & triggers structured planning instead of simple Q&A ✅ Dynamic Question Generation - Creates context-aware questions based on your actual code ✅ Comprehensive Document Creation - Auto-generates PRD, Technical Blueprint & Task Lists ✅ Enhanced Context Gathering - Deep LLM analysis of your specific architecture & patterns ✅ Autonomous Workflow - Complete planning pipeline with minimal human intervention ✅ File-Specific Integration Guidance - References your actual files, functions & patterns ✅ Multi-Category Feature Support - Handles complex features spanning multiple domains ✅ Fallback Systems - Baseline questions ensure workflow never fails ✅ Session Management - Resume workflows across multiple sessions ✅ MCP Protocol Compliant - Native Cursor integration, no external tools needed

What is Context Engineering and Why You Should Care Prompt engineering is dead. Well, almost. We’ve moved past the era of just “talking nicely” to LLMs. Now it’s about designing the right context. Here’s the hard truth: Most people throw random prompts at AI and hope for the best. And it kind of works, until it doesn’t. Hallucinations, buggy code, vague PRs: these aren’t just model issues. They’re context issues. Enter context engineering. It’s not about asking better questions. It’s about giving the AI the right lens to look through before it answers. Want an agent to write a PR? It better know: - The repo structure - The affected files - The related feature spec - Established code patterns - What not to touch Without this, you’re not getting a smart AI, You’re getting a confused intern with no onboarding. Why it matters now We’re moving from single-shot prompts to multi-agent workflows. From toy demos to real autonomous dev systems. That shift demands rigor. Context engineering helps you: - Reduce hallucinations - Increase trust in output - Scale autonomous workflows - Stop wasting tokens and retries The best outputs don’t come from smarter prompts, They come from better inputs. Think of it like this Prompt engineering = telling the AI what to do Context engineering = showing the AI where it is and what matters If you're building with LLMs, context engineering isn’t optional anymore. It’s the difference between agents that ship and agents that stall.

Look mom! I am making money on the internet!!! 🚀🚀🥳🥳

Do you use any planning tools before jumping into code or do you just go prompt by prompt?

Look mom! I am making money on the internet!!! 🚀🚀🥳🥳

Imagine this All your code on GitHub automatically documented every time you push something new in prod. Very detailed docs that would allow: - better use of AI agents - better understanding of the codebase without the hassle of documenting yourself - faster onboarding of new devs The only requirement You need to give access to your private repo on GitHub for processing only. Your code is never stored outside your GitHub. Would you do it?

I tracked every hour spent building Context Engineer... The results shocked me 🤯 Here's the real data on AI-assisted development vs manual coding: 🧵 Thread with actual numbers ↓

If you want to use Webhooks and SSE events in your express.js app, don't do it. I spent 2 days trying to figure out why I was not able to update my onboarding flow in realtime using SSE events. It was working in DEV, but not in PROD. Turns out that SSE and HTTP/3 are not good friends. This was a sneaky bug, but now it's fixed, so hopefully I can save some other people from the pain of debugging this

Does affiliate marketing work? Like % fee on MRR for all customers affilate bring style? In theory sounds great, but how to find affiliates?

Here's the stats for my product since launch 2 months ago. Total visitors: 2380 Total signups: 240 Total customers: 20 Total subscribers: 12 MRR: $100 I have not done much for distribution besides posting here and reddit. Are these numbers good and worth putting more effort?

I am about to launch the Context Engineer MCP. It's packed with a lot of cool features. - it can detect when the user asks for a complex feature and triggers structured planning. - it gathers requirements by asking questions and analyzing your code base - it generates PRD, technical blueprint and implementation plan. All these docs are super tailored and full with both business and technical details and charts that allow the coding agent to build with a very high degree of precision. Oh, and it doesn't have any dependencies. All works in the ide like cursor etc. all the AI tasks leverage the LLM used in cursor. So no extra bills for AI tasks!

if (yourExperience === 🟥 && whatYouWant === 🟩) { console.log("Check the solution in comments!") } else { console.log("Tell me why you hate AI coding !") }

Seems like the waterfall approach to product development might have a case to come back in the AI era. the main reason it got replaced by agile is that the risk of building for too long without quick feedback loop was too high especially for startup. but now with AI agents I think it would be actually better to dump your full vision, make a big fat roadmap with all features planned in detail and then have the agents smash it ASAP

I am proud to announce that I was recently hired by the top tech companies thanks to my vibe coding skills to harden their security systems.

99 people on the waitlist. Feels like bad luck to launch without hitting 💯. Who’s gonna be the chosen one? Tomorrow we open the gates to the Context Engineer MCP.