project screenshot 1
project screenshot 2
project screenshot 3
project screenshot 4
project screenshot 5

Twitter Aukaat

A peer to peer negative reputation protocol

Twitter Aukaat

Created At

ETHIndia 2023

Project Description

Each of us has a unique life experience which shapes who we pay attention to in real life. But on social networks we pay same attention to everyone. Our attempt is to capture this gamut of complexity with a simple user action: reward/punish others by changing font size. Thereby controlling how much attention we pay to each other. The higher the font size the more likely you value their ideas/opinions. Additionally we also let users publish these aukaat graphs i.e font-size graphs in a peer to peer network powered by GunDB.

Our choice of GUNDB ensures we won't fall into the trap of centralisation and censorship on this network whilst keeping it almost gasless. We've implemented this as a browser plugin on Twitter instead of looking to start a new social network. Twitter being a cesspool of trolls, it's a great place to put a negative reputation protocol in place. That's also the reason we've a key feature in the plugin called "LoserBoard" to highlight users who are given most distribution by Twitter algorithms but got least attention from users.

Any negative reputation exchange in general is done discretely. A broadcast even if it's against trolls must be done fully privately. Otherwise users risk retribution. We also have to avoid the spam and ensure safety from revenge or targetted attacks someone may be subject to on such a decentralised network. We plan to implement ZK credentials to avoid spam and revenge attacks in future versions of this plugin.

How it's Made

Challenges:

  • Extraction of handles from Twitter feed in real-time
  • Updating DOM in near real-time for font size mutations
  • Making Gun DB JS work inside browser plugin

I had initially considered Waku for reputation broadcast. But I ran into a thorny issue to bundle the Waku SDK to use inside the plugin. It opened up a whole of bunch of dependency issues. I abandoned that exploration after spending a few hours. And started going deep into GUN DB.

To be solved:

  • Identifying spammy negative broadcasts
background image mobile

Join the mailing list

Get the latest news and updates