socialmedia Timeline - FediScanner

30.03.2026 20:00
ainewsletter (@ainewsletter@techhub.social)

🤖 [TechCrunch] Startup zajmujący się chipami AI Rebellions zebrał 400 milionów dolarów przy wycenie 2,3 miliarda dolarów w rundzie poprzedzającej IPO

🔗 Więcej: https://techcrunch.com/2026/03/30/ai-chip-startup-rebellions-raises-400-million-at-2-3b-valuation-in-pre-ipo-round

#AI #SztucznaInteligencja #TechNews #TechCrunch #ArtificialIntelligence #technology #socialmedia #si

Show Original Post

Report

30.03.2026 19:58
GrumpyOldFart (@GrumpyOldFart@expressional.social)

“The Collapse of %Mossad in Iran – Did Tehran Reverse the Intelligence Battlefield?”

by Palestine Chronicle Editors in The Palestine Chronicle on Substack

@palestine@fedibird.com
@Palestine@masto.ai
@palestine@lemmy.ml
@iran

“For years, Israeli strategy toward Iran rested on a set of assumptions: that internal dissent could be activated, that covert operations could weaken the state from within, and that sustained pressure would produce systemic cracks.

Those assumptions have not materialized. Instead, what has emerged is the exposure of spy networks, the collapse of a central destabilization plan, the expansion of Iranian cyber operations, and the continued ability of Iran to strike strategically”

https://open.substack.com/pub/palestinechronicle/p/the-collapse-of-mossad-in-iran-did

#Press #SocialMedia #Iran #War #Trump #Israel #OperationEpsteinFury #OperationEpicMistake #RegimeChange #WarCrimes #CrimesAgainstHumanity #Hormuz #Empire #Collapse #US

Show Original Post

Report

30.03.2026 19:48
GrumpyOldFart (@GrumpyOldFart@expressional.social)

***** Satire alert !

“IDF not joining US ground invasion because it's only comfortable fighting women and children”

by Laura And Normal Island News on Substack

@palestine@fedibird.com
@Palestine@masto.ai
@palestine@lemmy.ml
@iran

“An #IDF spokesman told me via Zoom from his cosy Tel#Aviv bunker: ‘Just because we started this war, doesn’t mean our soldiers should be the ones to fight it. We expect the Americans to fight for us, just like they did in #Iraq and #Afghanistan. #AIPAC has spent a lot of money on their politicians and we demand a return on our investment’.”

https://open.substack.com/pub/normalislandnews/p/idf-not-joining-us-ground-invasion

#Press #SocialMedia #Iran #War #Trump #Israel #OperationEpsteinFury #OperationEpicMistake #RegimeChange #WarCrimes #CrimesAgainstHumanity #Hormuz #Empire #Collapse #US

Show Original Post

Report

30.03.2026 19:44
r (@r@fed.brid.gy)

Saskatchewan wants feedback on potential social media ban for kids under 16

https://fed.brid.gy/r/https://globalnews.ca/news/11751671/saskatchewan-potential-social-media-ban/

Show Original Post

Report

30.03.2026 19:39
2026 (@2026@ideatrash.net)

You Posted It Publicly, So Who…Or What…Gets to Read It?

There is an important distinction between objecting to your content being used to train an AI/LLM, and for one to process your content. Blurring that distinction does nobody any favors.

What rights do you have to your posts on social media?

I started thinking about this due to a reaction some users on Mastodon had to a tool designed to summarize a user’s own home feed so you didn’t have to doomscroll to catch up.

The objections were specificially about how the summarizing was being done — by an LLM.

I’ve been thinking about this a lot since then.

Not about the tool itself, or whether or not it’s worthwhile. But the objections.

Because the only way Jan (or anyone else) can actually prevent that is by not posting publicly.

Before I go any further: I think there’s an important difference between “my publicly posted content cannot be used to train an LLM” and “my publicly posted content cannot be see or used by certain classes of programs, people or entities.”

First, the technical aspect: Even if those who object block Voss, that won’t achieve the goal if you have any public posts. That just means that when Voss uses the tool it won’t include those posts. But if someone else does, your public posts will be included.

That’s because they’re… well, public. Public posts (or public web pages) have to get sent to the receiving computer to be displayed. We’re running up against the same problem that DRM — digital rights management — has: In order to show the content, you have to send the content to the person getting it.

Because viewing digital content inherently means copying it, we have to briefly consider copyright for simply viewing a post. (I am not a copyright lawyer, this isn’t legal advice, I’m just a mostly-knowledgeable layperson, and I’m discussing from a US perspective.)

You automatically have copyright on anything you’ve written in the US; technically you are providing a license to display your content to the social media site. But what rights do others have in regard to what you wrote?

The closest real-world analogy we have is of public photography; if you’re in a public space and do not have a reasonable expectation of privacy, then taking a photo of what you see is a reasonable expectation. Posting publicly on a social media feed is our digital equivalent to being seen in public. Taking a photograph is equivalent to displaying the post on the end user’s system. Add to that the ToS and express “right to display” that’s included, and that part is handled.

At that point, the copyright test depends entirely on the output. While your social media posts are copyrighted, that means that someone cannot replicate it. Arguably, my screencapture of the toots above would fall under this rule (although fair use clearly applies here). While a copyright holder does have the right to prepare derivative works, unless you’re discussing a summary of just your social media posts (rather than an aggregate of a timeline), that’d be a hard thing to prove.

It’s worth noting that this tool we’re talking about is utilizing other providers on the back end — either Anthropic or Copilot. So the data is being sent to them; whether or not they are using that data to train the LLM on is subject to whatever their ToS is.

That’s an important distinction — most copyright claims against LLMs is because they are substantially reproducing the style or content of the works it was trained on. That’s why I think giving the rights to use your content to train an LLM should always be opt-in only.

In contrast, the summaries this tool produces look like this:

This looks nothing like a timeline feed, isn’t reproducing anyone’s original content or style, and is clearly transformative. (Although whether or not you should trust a summary by an LLM is a completely separate issue!)

Whatever rules we apply to this tool, we also have to apply to a tool that, say, creates alt text for images that do not have them… even if that tool is a human.

Those copyright tests have nothing to do with whether it is a human, program, or LLM creating that output.

So the actual objection (and demand) is not actually that there’s a summary of a social media feed. The objection is about an LLM doing it.

It’s the functional equivalent of “you can only read this post on Firefox, you’re not allowed to read it on Chrome.”

And that is entirely too close to the “shrink-wrap EULAs” for my comfort.

You know the type: “By receiving this email you agree to” and the like. Sort of like the anti-EULA that you’re now subject to because there’s a link to it in this post. They’re all bullshit, and particularly in cases like this, where it’s just one parties’ unwritten social expectation. I cannot say, for example, “all employees of this company cannot read or discuss my posts”in the posts and seriously expect that to hold water. (I’m sure you can imagine much more distasteful examples.)

Again, I think there’s an important difference between “train an LLM on my content” and “my publicly posted content cannot be see or used by certain classes of people or entities.” The first explicitly prepares the LLM to make derivative content. The second is functionally equivalent to viewing a post on your RSS reader, browser, or client.

There are real, valid concerns about AI/ML. Freaking out over an AI simply reading your public posts distracts from those real concerns.

Featured Image by Gerd Altmann from Pixabay

#AIML #copyright #socialMedia #technology