It's pretty sad to see the negative sentiment towards Hugging Face on this platform due to a dataset put by one of the employees. I want to write a small piece. 🧵
Hugging Face empowers everyone to use AI to create value and is against monopolization of AI it's a hosting platform above all.
Hugging Face empowers everyone to use AI to create value and is against monopolization of AI it's a hosting platform above all.
Comments
as for people's projects, if you don't like them, feel free to put a takedown request on why it should be taken down with the legal basis. for the link you sent you can file for NFAA masking for instance.
People are afraid. From the outside, it is one big machinery where profits land in the hands of a few researchers & corps while all they get is "don't be upset, we're giving you a chatbot".
Then someone tells them "we're really ethical, trust me bro" after the same person's boss just stole their data.
And all of that is not even completely untrue.
I'm actually using HF, running/training models, I'm a dev -
I'm way closer to you than the people writing those hateful replies, and I don't agree with those.
I just wanted to paint a picture because I see this so often and I tried to understand.
Call it that, but as I've said, it makes it sound like it's unreasonable -- for the 'bandwagon', it is reasonable, and I'm not sure these messages will come across as you had hoped.
But I don't know how to fix that, is what I meant.
I get the most "I'm
in a bubble, *completely* different view" is when I talk with people outside of AI or tech twitter about AI - Bluesky is majorly that.
If you want to understand better, I'd say talk & listen to them.
i also believe it's especially hard to reprogram the "ai = bad" messaging that's floating around for a while now so stay safe out there
It's nearly all AI hate posts & hot takes on there.
For some people, largely on the far left it seems, AI is a hot button issue where there is *no* ethical or reasonable use for it.
You can't engage meaningfully with people who disengage from reason & logic.
For me, Bluesky has zero privacy, no clear monetisation strategy, and "Federation" is dubious.
Mastodon has a far left/alt/counter culture, longer posts & editing, but you have to get past the terminolgy & concepts.
However, unless you're on one of the fringe instances that tends to get blocked, you still have access to the core fedi-space, and your follows determine your personal experience.
It attracts some, repels others.
I've found *very* few pro-AI voices for example, and most have left due to a constant background anti-AI sentiment to anything posted.
(This is a genuine question, not intended as a gotcha - it's something I have trouble answering myself)
Gotta have thick skin out here.
but people are targeting my colleagues personally, who are amazing people who had nothing to do with the incident, I care about them
I dig what they are doing (Hugging Face).
you need to address that you were the *perpetrators* here.
I’m sure something was learned from this.
That doesn't feel intuitively right to me, because B has always been about decentralization and open APIs - look how easy it is to tap into a firehouse for example https://bsky.app/profile/simonwillison.net/post/3lbe3bmumjs25
https://bsky.app/profile/bsky.app/post/3lbvgvbviak2c
https://huggingface.co/Dragonstalker/Nudify_Utilities/tree/main
https://huggingface.co/Mdp01/Lora/blob/main/nudify_v12.safetensors
https://huggingface.co/spaces/outsourceit2day/free-nudify-ai-v1.0
Also, doesn't LAION5b still have a bunch of bad stuff in it?
https://huggingface.co/Kernel/sd-nsfw
What you’re asking about shouldn't even be a question.
The HuggingFace employee did nothing that violated the rules of this site, nor break the law, by saving posts.
We have come to Bluesky expecting a different set of people, than what we left behind.
That’s how the auditors put it to me when I reviewed with them at Salesforce in ~2021
that's a number that's generally considered safe for *operational* data
it's not a 30-day free for all if the usage is abusive, not at all
I highly doubt the courts consider research over public web data “abusive”.
Is CommonCrawl in violation of GDPR?
it was all very conditional, and rests on the usage not being e.g. blatantly scraping clear PD without prior interaction
https://reddragdiva.dreamwidth.org/606812.html
Social programs need to be developed to get rid of AI anxiety imo
When we label data for model training for instance I know we go through contracts of the labelling company to make sure labellers are fairly paid.
Many people at Hugging Face are immensely kind and talented and helpful, you can just go to forum/GH/website and see the attitude.
Large companies are training models and gating them and making you pay per use with no transparency on biases, fairness of the model or societal/environmental impact.
https://bsky.app/profile/chrisoffner3d.bsky.social/post/3lbwhh6qmks2u
But HF brand is undamaged among people who know what it does.
I went to my boss @julien-c.hf.co asked that day if I could use company's compute and he said "have whatever you need".
https://hf.co/blog/using-ml-for-disasters
this is not fair.