Claude is for vibe coding. Codex is for shipping. Made the switch this week: - Generated working auth in one prompt (Claude wanted to discuss the philosophy of auth first) - 3x rate limits = 3x faster launches - Interface that looks like 2026 not 1996

Imagine your Net Worth increases 14% Congrats! maybe you got a promotion, or made a lucky investment For Elon Musk, that 14% percent is $107,000,000,000 9 ZEROS Jesus Christ.

I started posting/replying on X less than a month ago Yesterday, I got almost 300,000 impressions on one reply. ~7000 likes Fun fact: it converted into -> O new followers Why? Probably because it was completely out of niche You speak your thoughts, it resonates, and it bounces around the timeline Lesson: It doesn't always translate to direct growth, and that's okay!! Some are just for the love of the game😎

my first month posting on X just crossed 200k impressions on the day😅 it really just sky rockets when you least expect it. quality > quantity

Crossed 100k impressions in one day! No viral threads. No engagement pods. Just consistent, quality replies. The secret wasn't volume, I averaged 30 replies since I started. The strategy is boring: - Find good conversations. - Add actual value. -Saying quick 'Hello' once in a while doesn't hurt:) - Repeat until the algorithm notices. And it will. It's a slow grind to build momentum, but once the algorithm tags you as 'high signal,' you will be rewarded!

reached 200 followers! 0 to 100 was a grind 100 to 200 felt like it happened overnight. huge thanks to everyone who hit follow back to fixing (breaking) things

Regardless of whether “AI WILL REPLACE US” happens soon, the real take is that your work doesn’t become pointless because AI is better at it. The scary part is the economy, not meaning.

People think they have time. But the window to build leverage is shrinking fast. If DeepMind’s CEO is even close to right, the next 5 years decide everything: careers, companies, power. It’s up to you to decide whether you own assets, or rent your life.

‘I'll start posting when I have an audience.’ No. You post to get the audience. 190 followers 76.5k impressions in ~2 weeks You don't need to be an influencer. You just need to show your work. The algorithm doesn't care who you are. It cares about what you build!

There are only two things guaranteed in life: 1. Taxes 2. Technology breaking when you need it most. I focus on 2. I don't need to predict the future. I bet on things breaking, fees getting hidden, and subscriptions getting messy. As long as AI/Computers make mistakes, I have a job. And they make a lot of mistakes. Always will. While everyone seeks the next big thing, I'm betting on what never leaves.

time for a long walk to clear the head and back to work!😄

This is what the '48-Hour Audit' looks like when delivered. No fluff. No complex graphs. Just a ledger of exactly where your billing logic is broken, linked to the specific Stripe customer ID. • The Issue • The Revenue Impact • The Fix Stop guessing where the money is going.

This is what the "48-Hour Audit" looks like when delivered. No fluff. No complex graphs. Just a ledger of exactly where your billing logic is broken, linked to the specific Stripe customer ID. • The Issue • The Revenue Impact • The Fix Stop guessing where the money is going.

Silent revenue leaks in Stripe are surprisingly common. Often, you don’t catch them until the end of the quarter. I am building a 60 second diagnostic to surface the details: Overdue invoices still open, "Active but unpaid" accounts, Old discounts that never expired This isn't a recovery bot. It’s a clean audit report for your finance team. Delivered in 48 hours.

Biggest mistake I see: People ask AI for permission to build. You’re seeking validation from a probability machine that has never paid for anything or felt the problem. AI predicts the average. Startups are outliers. What did AI tell you ‘won’t work’ that you built anyway?

100 verified followers! It's not a viral number, but it matters to me because: - No bots - Real builders - Genuine feedback loops I'd rather have 100 people who read every word than 10k who scroll past. Thanks to the early crew. Back to building :)

Most people spend 6 months building a failed startup. Thanks to Perplexity research, I can now fail in 30 seconds flat Me: I just thought of a million dollar idea. Perplexity: here are 3 competitors and a lawsuit waiting to happen

Day 9 Building LeakFix in Public ✅ Spent the last 6 hours refining the dashboard. I wanted the feedback to feel like a senior engineer looking over your shoulder, not a robotic error log. • The Issue: "Plan names detected" • The Why: "Clear labels help users self-select." • The Fix: "Keep plan names distinct." It’s not just about finding errors; it’s about explaining why they matter. Appreciate any thoughts on the dashboard layout!

Having GPT-5.2 and Claude Opus 4.5 in the same dropdown seems like a cheat code. But does anyone think that the models get slightly nerfed inside a wrapper vs using them directly on their own sites? Or it’s just a placebo

Day 10 Building in Public analytics🚀 The algorithm doesn't care who you are. It only cares if you have something to say. Stop waiting for permission to build in public. The leverage is open to everyone!

Day 7 of Building LeakFix. The engine is officially running! ✅ 1. URL Input 2. Playwright captures full-page screenshot + HTML 3. Node.js analyzes distinct conversion rules 4. Generates a prioritized "Fix Report" Who wants a scan?🔍

The Indie Hacker "Source Code" is actually stupidly simple: 1. Attention 📢 2. Trust 🤝 3. Cash 💰 4. Time ⏳ (The Asset) 5. Scale 📈 We love to overcomplicate this with tech stacks and growth hacks. But usually, if you're fail, it's because you skipped Step 2. You can't withdraw cash from an account with zero trust deposits.

Saw this horror story on Reddit today. Company spends 6 months hunting a "unicorn" candidate. • Drags out interviews • Ghosts her for weeks • Finally offers... 10% BELOW asking. She declined immediately. I’m glad. The worst part? The recruiter's boss said, "We don't get involved in salary." If you aren't protecting the talent from the client's ego, you aren't recruiting. You are just forwarding emails. Question for the founders/devs: What is the fastest you have ever turned down a job offer?

It's weekend, but we keep pushing, Swapped the “demo form” vibe for a proper scanner layout: Still rules-first. No AI magic. Just “pricing URL in → screenshot out” done properly. Next up: leak rules that actually move revenue (CTA clarity, plan framing, trust gaps), ranked by impact.

Day 2 shipped. LeakFix is live ✅ Pricing URL in, full-page screenshot out, scan record linked. No steps skipped. No AI, just the plumbing. It has to work before anything else matters. Capture, upload, link, done. No magic, just steps. Next: rules that flag the leaks founders actually care about. CTA clarity, plan framing, trust gaps. Then a ranked fix list. No noise.

Day 1 of LeakFix shipped. ✅ Paste a pricing URL ✅ Click Scan ✅ It creates a scan record in Supabase (id, url, timestamp) No AI, no audits, no pretending. Just proving the pipe works before I build the leak checks. Reply with your pricing page URL if you want me to use it to build the first 10 “conversion leak” rules. #SaaS