Emm Letchindjio

It's actually insane how open source voice models have reached s decent level for a while. VoxCPM and Vibevoice are impressive. Playing around with them is surprisingly good. Although for conversation, Vibevoice is way better than VoxCPM

3 revenue streams. 3 ways to disappoint a customer at the same time. Dine-in. Takeaway. Delivery. All pulling the kitchen in different directions. WeOut Order Management puts every order in one place — placed, routed, and tracked in real time. 📲 Send a DM now to get started

Nobody seems to talk about how time consuming it is to set up your home AI lap. Choosing the right model for your use cases, setting up your workflows (I don't use claws) etc... I got my home lab for product building and coding. It's taking some time, but I like the direction

Successful robust softwares have never been the result of one-shots. Even MVP were scoped. SCOPE is the keyword. Sure, you can build a community around a buggy demo and improve as you go... IMPROVE

@rdbuilds7 @GoalFrameApp smart journaling, project/goal tracker with offline AI coach and journaling assistant. Work without internet and free

Startup founders, this one's for you Share your project link Let’s help each other grow

Android, and it just doesn't work. All the flutter plugins pretending to work for MLKit are just loading forever though. I still have gemma as my backup, but I wanted to spare users from downloading something they have built-in their device. Hopefully I find a solution soon

That moment when Claude code tells you "Bro, you're on your own this time" 😔. As I'm implementing native AI model support for @GoalFrameApp , after a successful and smooth implementation for iOS(I'm actually impressed with the performance of apple native model), I moved to

This approach is easier for people with some tech background, but the gains I had proceeding that way make sense. Rushing to generate an app full of bugs is no good

For a while I was bombarded with ads and videos about online platforms for vibe-coding. How you could one-shot your way to a fully functional app. The narrative is changing a bit towards one-shot demo, which is more realistic.I decided to give it a try. And like almost everyone,

by doing this on a new project I'm releasing soon. I spent two days just to make sure I rounded the idea very well by text. On the third day, I generated the specs documents and features list. Then I generated the specs for each feature.

I decided I will build an interface, that will push me to make sure I have all the specs artifacts before I open my ide. The requirements and acceptance criterions should be defined at feature level, each feature is tested. Tokens will be saved massively that way. I saw gains