Subarna Basnet

Fully busy building Operator. I have 2 Claude Code accounts. Both hit the usage limit. So I came to X just to post this. 😂 A month ago I wasn't even into AI. Now it's 10 AM → 3 AM, every single day (17 hours/day I work everyday). I'm not learning how to write code anymore. I'm learning how to build systems. The coding? Claude can have that job. The marketing? Soon, Operator will have that job too. Launch is close. Just give me a few more days...

@RamJonchhen thank you, ram.

@subarna253 Insightful

Don't hire anyone if you're building the first version of your product. No one will understand your product the way you do. The problem isn't that they can't understand your vision. The problem is... it's not their dream. They have their own goals. Many developers dream of working at Google, Microsoft, or other big companies... not spending months building an idea-stage startup with uncertain outcomes (u hired them, they watch movie. because its remove job right :( ) And that's okay. But as a founder, you can't expect someone to care as much as you do. If you don't know how to code, use AI. Build the first version yourself. Launch it. Get real users. Get real feedback. Then hire. Your first version doesn't need a big team. It needs obsession. This isn't advice for everyone. It's the reality I faced while building Operator (@try_operator)

Training a small language model (SLM) for Operator on my MacBook My MacBook: "babe... I'm a hottie now" 🥵 Me: "Don't worry, I got you" puts ice on the MacBook 🧊😂 Pretty sure I'm the first AI founder running a liquid-cooled MacBook

For the past few months, I've been building something I truly believed in. I gave this project everything I had. -> 17+ hours every day. -> Invested my own savings. -> Rebuilt the product more times than I can count. -> Learned entirely new skills just to keep moving forward. -> Spoke with developers from different companies to understand how they solved difficult problems. -> Started talking to potential customers before the product was even ready. -> And sacrificed many things in my personal life to make this work. I wasn't building it alone. I had people who believed in the vision and were helping me bring it to life. But two days ago, everything changed. They decided to step away from the project. There will be no further support, and the product won't launch the way we originally planned. I'm not writing this because I'm angry. I'm writing this because I don't want months of work, learning, mistakes, and late nights to end up sitting on a private GitHub repository that no one will ever see. So I've made a decision. On July 20, 2026, I'll release the entire project as open source. If I can't continue building it the way I imagined, then I'd rather let other builders learn from it, improve it, or take it further than I ever could. Not every project becomes a successful company. But every project can become someone else's starting point. I hope Operator becomes that for someone.