fallibifallibi
An interface that connects opinions to their reasons.
fallibi is a planned platform that organizes a person's opinions from what they have written and said, and explores how those judgments rest on values, norms and axioms. Through cited questions and answers, comparisons between perspectives and conversation across languages, it aims to help people understand each other and to reduce political polarization.
The full service has not launched, and it is not a formal research project. Today fallibi only runs the back office of my personal homepage, doni.page, and an MCP server that lets AI tools like ChatGPT read doni.page's writing. During my PhD I focus on quantifying knowledge measurement, so I am not expanding fallibi into a separate research topic.
The code is planned to be open-sourced. I will do my best to keep the data format backward compatible. If fallibi launches one day, I hope you will visit again.
Use it in ChatGPT
Add one of these MCP addresses to ChatGPT or another AI tool to search the published writing and answers with their sources, and to leave questions.
https://fallibi.com/mcp: All of fallibi. Each question names the person it's for.https://doni.page/mcp: Doni only.
Collaboration
If you are interested in leading the research direction of this project or collaborating on it, please get in touch from the second half of 2027 onward. For now I am concentrating on my research topic, the quantification of knowledge measurement, and I plan to keep this project at the scale of a personal-homepage proof of concept while improving its usability. I am currently on an F-1 visa, and I have decided to engage with political polarization only after finishing my PhD, so I cannot carry this project forward as its primary contributor.