🔥Telegram İfşa

Day 2: Sorry for going offline. Fell terribly sick. Finally in momentum today. Finalized on 2 ideas for now. Will release MVPs this week :).

Day 3: Work in progress. Building an AI app that takes your paystub (PDF/JPG), extracts the data, and gives you smart AI insights to help you understand it better. If it works, I’ll expand it to balance sheets, income statements & more. MVP mode.

Most bugs tend to appear in the front end, so I think it's smarter to build the basic layout first before adding the app's functionality. Can most of the AI agents always hallucinate when it comes to debugging simple front end bugs. Will build design system from now on.

Day 4: Just launched an MVP! You just saved 10+ hrs of data entry Built an AI that reads any doc and gives clean JSON + insights: → Upload → GPT-4 Vision processes it → Download & done ✅ Free MVP is live → ai-document-parser-gold.vercel.app

I asked a YC video: “How do I get my first users?” youtube.com/watch?v=hyYCn_… The answer? ❌ Not SEO ❌ Not cold email ✅ Talk to people. Manually. 1-on-1. This broke my brain. I’ve been doing it all wrong. Built a tool to ask YouTube videos anything: youtube-transcript-qa.vercel.app

Day 5: Changing my entire distribution strategy. After watching this YC startup school video (youtube.com/watch?v=hyYCn_…) and pasting it in my tool (youtube-transcript-qa.vercel.app) I've realized I have to manually look for customers and do things that don't scale.

Day 6: Using Reddit to test out ideas and make MVPs. I feel using Reddit makes more sense to get initial users. Should be starting meaningful conversations and understanding customer pain points.

Day 7: Feeling super confused and overwhelmed. Don't know where to start, what to build and how to get initial customers.

Day 8: Simple building strategy. Look for AI automations in no code tools like n8n which people raving about. Convert that automation with a killer landing page to a product in Cursor. Niche down to a group and sell the automation product to them. Let's see how it goes.

Cursor is everything to me.

Being in the B2B space is so much better than the B2C space. For B2C build an audience, look for pain points, hope you get viral it's a never ending list. For B2B just cold email, look for business tasks that can be automated and sell an AI automation packaged as a product.

Some people use X, some use Reddit, some use Instagram, some use Linkedin. But everyone uses email. Cold emailing with Instantly.ai seems like a such a easy and fast idea validation and distribution strategy. I wonder why people don't talk about it. It's so obvious.

Deep work session with Lindy. Trying to generate highly targeted lead lists.

Started Mastra 101 Mastra. Love how the course is structured and how AI agents are directing us for this course.

🔥 Everyone uses Lindy for business ops. But what if you could use it to EAT better? Just built a Lindy agent that finds the best restaurants in NYC by cross-checking Google, Yelp, and real review sentiment. This is Lindy for LIVING, not just working. #LindyAgentyAcademy

Holding the future. @mastra_ai

Using Lindy 3.0 version to develop 3 agents that finds leads, researches their pain points, and makes an email copy catered to my services that can help them solve their pain points. With 3 clicks you can get highly highly personalized messages. This is insane.

evals is where its at

Wingstop figured out something most restaurants missed: their customers didn't need to eat inside. They built stores that don't need dining rooms. Some locations have zero seats. ~90% of sales at mature stores are off-premise. Most restaurants pay rent on big dining rooms. 70%