From the @ODELL told you so X files:
When X unveiled its newest terms of service, which go into effect on November 15, users quickly picked up on one change.
“By submitting, posting or displaying Content on or through the Services, you grant us a worldwide, non-exclusive, royalty-free license to make your Content available to the rest of the world,” the terms of service said, which includes the right to analyze any of that content “including, for example, for use with and training of our machine learning and artificial intelligence models, whether generative or another type.”
Basically, by continuing to use the platform, users will agree that X can use their data to train its AI models.
https://edition.cnn.com/2024/10/21/tech/x-twitter-terms-of-service/index.html
#data #privacy #technology #ai #artificialintelligence
Yes.
That said, AI models can run permissionless on any NOSTR relay as well, without consent or users knowing about it.
But that is the nature of the nostrich beast
True, but at least on Nostr you aren't also KYCd.
But is it legal, since you never agreed to it? If someone is caught doing it, can they be sued?
I mean it's public available data, can't they do that regardless?
Google scraping bots don't ask for permission either
Commercial versus non commercial usage.
If you haven't granted rights, that is an easy lawsuit if you can identify the bad actor.
Ok but still not clear, how does it differ between the scraping bots? Can i sue Google, pretty sure they commercially exploit the scraped data?
No difference. You'd need to be able to 1) identify the bad actor that infringed upon you and 2) be able to prove they did you harm of some sort.
Nobody would likely take that fight to Google because they'd just bury you in legal mumbo jumbo for years. You'd need some law office doing some sort of class action most likely.
Cool, we should get it going
Scraping bots should pay a sat per scrape. When will we get pay-to-reply on #nostr?
it’s being worked from multiple angles already. 🫂
Isn’t this generally already a rule on any social media today?
Allowing them to use your data for LLM models? I'd say very few have it in their TOS.
Theoretically if they have a long enough post history, they can develop a virtual persona bot. That's a new wrinkle but I was actually talking with my wife about this several years back.
It was the anniversary of her dad's passing and I said in the future some company would offer posthumous bots that have LLM'd email, social media, and other data sources so people could talk to their loved ones that have passed on. "Man, if I could only get advice from dad one more time..." type of shit.
Not LLM models specifically, just that anything you post they own. I thought Facebook had that for years in their TOS but I'm not sure about that.
Regarding the "dead loved ones bot", I remember someone made a service that offered such bots. I'm not sure if it was only a chatbot or with actual voice but that definitely exist already!
Nice, I was more thinking video as it would actually make the connection more real.
Dear god
What’s x? Is that the party drug?
That would be really nice
They are just making their activities legal now
This is something I wonder about all the casual use of ChatGPT out there. Aren't folks just feeding into a massive LLM which will undoubtedly be used against us all?
Sure are. That is why they're banned at many workplaces (like mine) because you could potentially slip PII, confidential information, etc... and they become part of their dataset.
“This was previously identified as an outcome.” - @ODELL
"Resistance is futile" -- Master from Doctor Who in 1976.
How did he know? LOL
"UP IN ARMS" AKA NOT GONNA DO SHIT ABOUT IT
Admitting to their crimes now doesn’t negate the crimes against humanity they already did.
Proprietary services are always about making money.
Oh no, the tweets you are posting publicly for everyone to see will be seen by ML. The actual bad part was shutting down API access to supposedly restrict 3rd party AI training. That completely killed the 3rd party clients, which is the only way I ever used it.
You shouldn't be angry about AI training on your speech. I hope all my notes are read by all the parties undoubtedly training LLMs on nostr. If you are posting your world views publicly for everyone to see, you should want AIs to see it, because it will incrementally bias their outputs to be closer to your opinions.
Every company is doing the same nowadays
That's how the sausage is made...
Oddbean new post about login | logout 
