My friends are keeping me VERY busy these days. They send me websites, screenshots, names, organizations, and random links with some version of, “Can you look into this?” I always think I’m going to spend twenty minutes checking something and send back a sentence or two. That almost never happens.
This time a friend sent me Doikayt Mobilization Labs.
The name comes from doikayt, Yiddish for “hereness.” It’s associated with the Jewish Labor Bund and its belief that Jews should build their political and cultural future wherever they lived rather than look toward a national homeland. Doikayt pulls directly from that history and describes the Bundists as non- or anti-Zionist.
I started clicking around the website and eventually reached its projects page. That led me to GitHub, which presented a slight problem. You see…I have no clue how to read GitHub. I can click on things and probably break them. But repositories, commits, schemas, deployment files, branches, and code…none of that is in my wheelhouse. Way above my paygrade. So I pulled in my AI partner (I call it Atlas. Long story, maybe another time) and asked it to translate. What am I looking at? What does this do? What changed? Who changed it? And, most important, where else should I look?
A public GitHub repository can contain far more than the finished product people see on a website…things like source code, working documents, planning notes, old versions, funding research, and a recorded history of changes called commits. Doikayt’s repositories contain hundreds of them.
That’s where this became much bigger than the website my friend sent me. It may also be way, WAY bigger than the Wikipedia problem.
I’ve written before about what happened inside Wikipedia, where a relatively small number of highly active editors were able to reshape articles about Israel, Zionism, Jewish history, and the Israeli-Palestinian conflict. The problem doesn’t end when someone closes the Wikipedia tab. Wikipedia has become part of the information supply feeding artificial intelligence. Its content has been widely used in language model research and training, while Wikimedia now provides structured commercial access to its data to companies including Google, Microsoft, Meta, Amazon, Perplexity, and Mistral AI.
Doikayt points toward what comes next.
Instead of editing an existing encyclopedia whose contents may eventually make their way into AI, Doikayt is building a curated knowledge system with AI already helping create the material. It has published plans to make that information directly available to AI systems via the Model Context Protocol. It is also building and cataloging technology for political movements, has an AI tool for Palestine advocacy in its Field Guide, is recruiting AI-focused developers who share its politics, is researching funding networks, and has developed a governance model for technical infrastructure designed to remain functional amid deplatforming and government action.
The Wikipedia investigation was about people changing information at one of the sources. This is about building the infrastructure that can create, organize, distribute, and eventually feed curated information directly to machines.
And? It’s being built in public.
