A machine learning librarian at Hugging Face just released a dataset composed of one million Bluesky posts, complete with when they were posted and who posted them, intended for machine learning research.
Daniel van Strien posted about the dataset on Bluesky on Tuesday:
“This dataset contains 1 million public posts collected from Bluesky Social’s firehose API, intended for machine learning research and experimentation with social media data,” the dataset description says. “Each post contains text content, metadata, and information about media attachments and reply relationships.”
The data isn’t anonymous. In the dataset, each post is listed alongside the users’ decentralized identifier, or DID; van Strien also made a search tool for finding users based on their DID and published it on Hugging Face. A quick skim through the first few hundred of the million posts shows people doing normal types of Bluesky posting—arguing about politics, talking about concerts, saying stuff like “The cat is gay” and “When’s the last time yall had Boston baked beans?”—but the dataset has also swept up a lot of adult content, too.
I mean isn’t this what we want? The data is Public. It’s already being done behind closed doors. I’d rather this transparency. Especially because there’s such a large % of the population tuned out to how large companies with 8-10 figure r&d departments focused on marketing psychology manage to control them.
Even as aware as I think I am I’m certain there are 10s of thousands of strategies being employed that take advantage of my “immunity”. At least with FOSS and public records steming from that, the average Joe gets a peak behind the curtain and sees what is possible.
I had a conversation recently about mcdonalds app surge pricing and they never heard about surge pricing which is totally fine but they fought me on the premise “there’s no way they can do that” on technical feasibility to just “no one would do that”. I’m not sure what they were defending but I digress. My biphenton has kicked in and I should stop typing right n
It depends on who you mean by “we”. I don’t think the “we”, as in the users of Bluesky would want their posts to be used this way. Yes it’s transparent and accessible, but only to the technically inclined who can make use of it.
I wonder how much more The Matrix references become relevant. We become food for the machines whether they tell us so or not.
The real question here is why the researcher “librarian” didn’t even attempt to anonymize the dataset before making it available. Full anonymization isn’t a trivial task, but at least removing unique identifiers or replacing them with randomly generated ones would be good practice.
That dataset is going to contain a lot of furry content
Terminator was almost correct. The real terminator will be Arnie walking around in a fursuit.
I’ll be back. UwU
If you post something publicly, that thing will be used to train AI. Nevertheless the privacy speaks of the company.
Yes, it absolutely will. That’s why I fragrance the pandas. Just a little here and there so that some Howard will need to sort through it. The lime really comes through clearly.
Be super fucking foul and un advertiser friendly to make it less useful, OUTLAW COUNTRY
I don’t know why social media are used for training. It’s like the worst quality of data ever and it results to answers like « go kill youself » when prompted about something sad…
They are used because they are “real life” (not really but you know) conversation example
So? What’s with the hoopla?
Its a social maedia platform anybody can see and access.This type of news makes me want to step away from the interner entirely.
I keep noticing situations on social media and wikis were the only way I can frame it now is that it’s just data entry for AI models.
Wish I could. My job has me knee deep in it everyday trying to keep up with all this.
Modern day life over here requires the internet, our government has taken away options that use paper and replaced it with websites. Same goes for banks.
