I talked with people about these thoughts and concerns, and added my personal observation. I built this instead:
- Chronological feed only. No ranking, no algorithmic shuffling. You follow people, you see all posts from all users, or choose to see posts from friends, pages you follow, and groups you are part of.
- Built-in fact-checking and bias detection across text, images, and video. With Human-in-the-loop & AI combined in the process, AI detects, analyzes, and moderates posts, and every contested moderation decision is reviewed by a human against a published set of principles (not just personal judgment only).
- Per chat AES-256 encryption. We designed a way that currently, if a chat is attacked and compromised, it doesn't have any effect on other chats or the platform in any way - this is constantly being improved, and security is my top priority.
- No ads currently. Your data is never sold or shared (minor data is stored, all in encrypted strings)
- Child safety filtering is built into the content layer - Pseudonymous mode for speaking about sensitive topics - A built-in news feed
The moderation system is designed to evolve; it is more like a constitution than a rulebook. The principles are fixed but can be amended as the platform grows and severe cases emerge. The goal is to detect as accurately as possible, systematically, over claims of neutrality and move on.
I'm still working on figuring out monetization. I'm thinking of contextual ads that work based on the viewing page's contents, and not on your behavioral profile. Additionally, other features like the job board and commercial tools are also an option. None of this is live at the moment, and tbh, I'm still figuring it out.
The platform is currently LIVE with a small early community of active users. I'm building this solo.
I'm happy to answer your questions/feedback/suggestions...
www.manob.bd