Sidebar Sidebar
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.

X Training Grok AI On Tweets Without Notifying Users

The unregulated AI sector will Hoover up all sorts of trash in tweets and spit it out at you elsewhere.

Classic garbage in garbage out

Except it will be garbage in, even stinkier garbage out.

And the investors are starting to get impatient with AI profits. They threw their money at the latest craze without much thought or care and now are impatient about seeing returns.

It sure looks like much of the AI industry will be a stinker and only a select few that are cautious and sensible get the best out of it where it’s appropriate.

Some of the AI features are truly terrible… take a photo and it’s not quite right… take another? Oh no, just use AI to fake a perfect one with what you got.

This is just one aspect of the AI stink. It sounds harmless… but it’s not real.

Look back in ten years and you’ll be wondering which photos are real and which are doctored.
 
  • Like
Reactions: uacd
Everyone who needs data and can feed data to AI is doing this. Without training data, AI will be just a search engine, video and text and picture editing tool. Which by the way is what you will get with 10 or 12 or possibly 14 GB RAM iphone. I am concerned about that RAM AI hype I sometimes see in discussions. This is still good marketing move and nothing else.
The only concern with AI here I have, is that there might not be enough data arround us to train AI that will be able to teach us something new. For example to teach us on how to train next generation AI without having enough training data. Even Go or Chess will in the end have limited number of possible moves. Limited energy and cooling is only delaying factor.
 
Last edited:
  • Like
Reactions: vjl323
So glad I don't use it, I only used it for a week, years ago when it was twitter, I certainly would not use it now that Musk has his hands in it,.
Also, how much longer are we going to get told that X, was formally Twitter? I think the majority of people who want to know, now know.
 
In a very short time, the only way to rid yourself of AI will be to rid yourself of the entire internet. I am prepared to do just that.
 
"X Training Grok AI On Tweets Without Notifying Users"

Like all other AI-Chatbots before, without users even knowing they are existent.
🥱
 
Why train AI on something as garbage as random tweets by random people, when you could train it on – I dunno – books, scientific papers or literally anything that has at least some value? Also I can see why people would be upset if AI was trained using their writing, art or likeness, but their tweets? Oh no, AI is going to be able to automatically generate endless sentences that may vaguely be inspired by the endless sentences I wrote that have no actual content.
 
  • Like
Reactions: Someyoungguy
I’m not “for” this, but I think it goes without saying and has been common knowledge for DECADES that content you publish to the web is basically openly available for others to do with as they please, much less, the owners of the platform hosting it. Anyone shocked at this needs to revevaluate their internet usage. Don’t want it used in AI? Don’t post it online.
 
  • Like
Reactions: Analog Kid
Musk is doing everything within his power to create a xenophobic, homophonic, racist chatbot. He has all the right tools within his arsenal of companies.
 
If you ask most AIs about something recent, like an election, they haven't got a clue because their training data is too old. Maybe this will differentiate Grok from the others

The little summaries of the latest news stories that Grok already provides are quite well done - it's already summarising things that happened this morning
 
Last edited:
  • Like
Reactions: gleepskip
shadowboi said:
Wait, which copyright are you talking about?

Twitter has always been a cesspool of useless 140-280 character “posts”. How can someone even copyright 💩posting.
Click to expand...
Copyright isn't something that someone has to do. It applies automatically as soon as you publish an original creative work which, yes, includes 140-280 character posts. This is even more true when you're talking about still images, or music, or video that's posted.

I haven't read Twitter's ToS lately and it's entirely possible that the license (of your copyright) that you grant them to display the content you submit to the service already included an escape hatch that would cover something like training Grok. But speaking more broadly about AI training, a company hoovering up information from public websites is generally not going to have a problem with privacy laws in any way (not that the US really has much in that regard) but the copyright violations involved are going to be their undoing.
 
  • Like
Reactions: uacd
Analog Kid said:
I find it hard to get up in arms about this... I don't use Twitter, but if I did I'd just assume the company used my data for all manner of things.
Click to expand...
And not just Twitter the company. How many “news stories” are out there that boil down to “here is a flimsy story based on some embedded tweets from people we perceive to be influencers”? If it’s on Twitter, it’s open to the whole world, and that’s the way it’s always been (and Reddit is pretty similar in that regard).
 
  • Like
Reactions: Analog Kid
In the meetings at X I guarantee you the strategy was to intentionally not tell users and to hide the setting to turn it off in the least used platform for X. This type of **** is hilariously obvious.
 
You must log in or register to reply here.
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.
Top Bottom
Back