We're submitting a paper about the Community Archive to NeurIPS!
As you know, we’ve been building open tools to map how ideas spread.
One of the most interesting examples has been word trend analysis: both "vibe" and "psyop" were introduced to the community at around the same time, but the graphs show that their introduction & spread are very different.
Vibe was gradual, and psyop had a feverish few days where everyone was writing “X is a psyop” tweets and the word remained prevalent in community’s shared language
A couple of weeks ago, Brandon from Nomic AI was in our Discord and suggested that we could submit a dataset paper to NeurIPS, a machine learning conference. He and others we spoke to explained that this dataset is unique in that it's completely opt-in.
“The Community Archive uses an opt-in model to obtain author consent for every post in the dataset. This enables us to release both the metadata and the content of users’ posts at a previously unprecedented scale.”
Publishing gives us credibility which is great for funding and finding collaborators. We’re also really excited about this because it could be a new template for "self memetic study". We're showing that any community can gather its own data, analyze it, and draw useful insights!
For example, we find that a small core of accounts in the network is effectively exporting culture! This is especially cool if you already buy the fact that the CA contains tweets from a cultural frontier. Below we see the example of the word “solarpunk”, which was adopted 2 years earlier by the core (which includes people like visa, eigenrobot, DRMaciver), before getting adopted by the periphery in a second wave.
The list of new words we studied was adopted from Ivan’s winning submission to our NYC hackathon. The graph that aggregates all of these words looks similar.
We’re publishing a pseudonymized snapshot of the archive along with the paper, if you want to be excluded from the paper, let us know on twitter (@exgenesis, @defenderofbasic) or in our discord.
If you want to be IN the paper, upload to the community archive! Paper deadline is May 15th.
Even more exciting news is coming soon!
Xiq