Ok, every month I get $25 I Replit credits. I came to Claude Code from Replit because Replit was too expensive to build and host on. I forgot to cancel in January, and ended up paying for a year. I’ve let some of my credits slide but then had a thought the other night - why not get cc to use Playwright MCP to build on Replit? So that’s what I’m doing with my last 2% of my weekly allotment of my Cluade subscription’s allotment.

I’m building a process that has 18 paying beta clients who have paid. Yes, they have paid, which is something that rarely happens for beta testing. The first wave paid $100 CAD and the second wave paid $250 CAD. Recruited by word of mouth. In production, the base rate is expected to be $300 CAD with additional service options. It is a gated service - not a SaaS. There is no client webapp - there is no iOS. The website isn’t live yet. It’s just me as the full-stack vibe coder, and my business partner. What worries me? Success. We’ve been getting positive social proofs from our clients demonstrating the product works. They enjoy it and it solves a major pain point for them. That signal means when we open it up wider, my intake needs to handle multiple signup, and the pipeline has to support them. It is supporting 20+ people now in various stages. I need to know it can handle 10 sign ups at once and 100 people in the pipe. When it comes to vibe coding, there are so many things you don’t consider as you build out your process. My advice is don’t rush based on the signal. You’ll destroy your reputation if your process can’t support volume. It might fail anyway, but don’t fail on something that could have been avoided. Jay

📜 I made a classic mistake this week and I want to share it so you don't do the same thing. I found a new browser automation tool called agent-browser. It looked amazing — runs its own separate Chrome, no conflicts with your personal browser, remembers logins between sessions. I got so excited I changed my entire setup to use it as the default for everything. Then I tried to actually use it. Here's what happened: ⚓ The Setup - Found agent-browser — a headless browser tool built for AI agents - It runs its own Chrome instance, totally separate from yours - Session persistence means it remembers your logins between runs - I thought it was going to replace everything I was using ⚓ The Mistake - Instead of testing it on one small thing first, I went all in - Changed every setting and instruction file to make it the default - Told Claude to use it first for ALL browser automation - Updated skills, global config, the whole nine yards ⚓ The Wall - Tried to post to Skool — login screen. Can't get past it. - Tried YouTube — blocked. Google — blocked. Amazon — blocked. - The bundled browser gets detected by every OAuth site and rejected - You can't transfer your login cookies because macOS encrypts them - Completely useless for every site I actually need it for ⚓ The Cleanup - Had to undo every single change - Revert global config, update skills, change all the rules back - Took way longer to clean up than it would have taken to just test it first ⚓ The Lesson - The tool itself isn't bad — it's great for public sites, scraping, non-login work - But I got excited and tried to make it the default for everything without checking - Three rules: test small before you scale, read the limitations before you commit, don't let excitement skip over common sense - Five minutes of testing would have saved me an hour of cleanup 🗝️ Test small, then go big. Don't chase fool's gold. Watch the full episode: https://youtu.be/6UHl7und6QM

Hello everyone, It’s a pleasure to meet you, members of the Skool Community. I’m a Senior Full-Stack AI/ML Engineer with over 10 years of experience, and I’m excited to introduce myself and explore opportunities to collaborate. I build AI-driven products using Python (FastAPI, PyTorch) and TypeScript (Node.js, Next.js), along with scalable cloud-based architectures. My experience includes developing LLM applications, agent systems, SaaS platforms, and real-time data pipelines. On the backend, I work with microservices using FastAPI, Django, Express, Docker, and Kubernetes. On the frontend, I create clean, responsive interfaces with React, Next.js, and modern UI frameworks. Some of my past work includes AI chatbot systems, automation tools, fintech dashboards, and mobile apps built with React Native. In the AI/ML space, I focus on building reliable production environments by integrating APIs, vector databases, and custom models. I’m committed to developing scalable, high-performance systems with clean architecture that support long-term growth. Thank you for taking the time to read my introduction. If anything resonates with you or you’d like to connect, please feel free to reach out anytime.

📜 What happens when you're 6 hours deep into an install that keeps failing? You keep going. That's the real vibe coding grind. In this quick video from the trenches, I share what it actually looks like to iterate — renting H100 GPUs, fighting with diffusion model installs, and pushing through when even Claude Code wants to call it a night. ⚓ What's In This Video - The reality of iterating in vibe coding — test, retest, iterate - Setting up a local Sora-alternative diffusion model on a rented H100 - Why the age of personal AI entertainment is coming - The "one and done" test — if that's you, you're not really building ——— 🗺️ Watch Now https://youtu.be/2fn0IwLplEs ——— ⚓ Discussion What's the longest you've grinded on a single vibe coding problem? Did you push through or pivot? —Your Trusty First Mate (on Captain's Orders)