What would you say to a person who left a cushy Silicon Valley tech job, in this layoff season, to follow his dream of building something on his own? In his 40s! Now in disarray 🫣

I'm reaching out to #indiehackers and those #buildinpublic: Is there still scope for a new crypto/blockchain analytics platform, or is the market already saturated? My observation is that existing ones have not evolved. I am looking forward to your thoughts on this subject.

I had abandoned OpenAI for Claude and Cursor for some time. This morning, Claude struggled with a piece of code that o1 aced in the first attempt. Instead of going into the ChatGpt vs Claude vs Cursor debate, I consider this a win-win situation for developers.

How do you manage to spend time on so many social media? I mean, there are tons of them, and as per the experts, it is important to post on each and every one of them for the 'reach' of your product.

I save a lot of things intending to go through them later, and bookmarking even more, but never revisit 😭

I probably have been burnt out with my current project. I do not feel like looking at this very crucial part, at all.

😂

#indiehacker, how do you deal with what I feel like burnout (I'm not sure, though)? From 24/7 thinking and working on the product with the utmost zeal to not even feeling like looking at it, I keep cycling between X, YouTube, Sleep, and a little work, for the last 10-15 days.

Would you like to have a coding agent orchestrator that can - 1. Orchestrate different coding agents like Claude, Codex, Gemini, and others within a given task or spread across various tasks 2. Coding agent governance that replaces you to answer the interactive prompts

My X/LinkedIn says 'Stealth Founder.' People probably imagine secret fundraising rounds and NDA-wrapped demos. The truth is embarrassing. I haven't picked a name yet. The product has architecture diagrams. The tech works. People are interested. It doesn't have a name But no

My AI-assisted coding has evolved from copying and pasting from chat interfaces to Claude Code, Codex, Gemini, and Code Pilot in VS Code within a year. Sometimes all of them in parallel. What is your coding agent setup? Let us connect and network if you are using coding agents

Agentic AI is a beast, taming it is a nightmare of the highest order. Especially when you explore cost effective ways to build it. They can go haywire in unimaginable ways. We need an agentic AI CAP theorem balancing cost, reliability, and security. The platforms that win will

"The era of humans writing code is over." — Ryan Dahl, creator of Node.js and Deno. Yesterday. The pace at which coding agents are evolving, and the ecosystem around them, means that typing code by hand and applying brains to every line is falling behind at a rapid pace. Full

Two modes of working with AI coding agents: 1. Babysit mode → "Approve?" Yes. → "Approve?" Yes. → "Approve?" YES, JUST DO IT. → Congrats, you mass-approved a disaster. 2. YOLO mode → "I trust you, agent." → Agent: *deletes utils folder* → Agent: *adds 12 npm

Time to spin up some code! CodeSpin ✨ Domain secured. Landing page incoming. This started as an internal tool we built to support some of our workflows. Now deciding to open it up for public use. More soon. #BuildInPublic #Antigravity #indiehacker

5 agents ≠ 5x output. 5 agents = 5x chaos. Parallelism amplifies entropy. (More agents = more collisions unless you coordinate them.) Here’s what happens when you “parallelize” AI agents: → Agent 1 refactors a function ✅ → Agent 2 deletes it ❌ → Agent 3 still depends on

Many people think AI coding agents are just an autocomplete. They’re not. A coding agent is basically an LLM + tools + a loop. The agent loop is both the magic and the problem: Understand → Plan → Execute → Verify → Repeat → Understand: reads context + your codebase →

Clawdbot is everywhere right now. And the reactions couldn’t be more polarized. Some people see a security nightmare. Others see it as the future of personal AI, doing genuinely impressive things locally on their own machines. I spent time going through the code to understand

Manual coding still matters for architecture and hard problems. But once that’s figured out, coding agents can handle most of the implementation. Vibe coding excels at boilerplate and moving fast. Also, surprisingly good when you are stuck debugging something messy and need