Welcome to Incremental Social! Learn more about this project here!
Check out lemmyverse to find more communities to join from here!

eveninghere

@eveninghere@beehaw.org

This profile is from a federated server and may be incomplete. Browse more on the original instance.

eveninghere , (edited )

So... there will be organizations that will train devs for this, and that will outright ban LLMs. I know which mine will be. Time to reconsider my job, and possibly my place of work..

eveninghere ,

Well, I can agree with many points you made. I also think that FG is overreacting.

The only thing I want to say is that Fediverse by design lets administrators choose which instances they connect to. It's rather unfair because there's virtually no free alternative choice to fediverse, but as we choose fediverse, disconnection is a thing. The only question is whether FG is going too far this time. And if it is, we either convince FG to retreat or build an alternative to FG.

eveninghere ,

The rules for being listed on fedi.garden will require blocking instances cited in human rights reports on genocide.

And this is their announcement on this oddly specific rule.

I mean, the wording "cited in multiple reports of massacres or genocides" is strange enough. An organization can be "cited" for doing anything. Can I write up two BS reports so that we ban any instance I don't like? Sounds like a teenager mimicking a TV politician speak, tbh.

Fediverse should be based on a mature protocol imo.

Backdoor found in widely used Linux utility breaks encrypted SSH connections | Ars Technica (arstechnica.com)

TL;DR there was a backdoor found in the XZ program. All major distros have been updated but it is recommended that you do a fresh install on systems that are exposed to the internet and that had the bad version of the program. Only upstream distros were affected.

eveninghere ,

Ars Technica sounds like a weirdo to me these days. Loves to attack big techs (although understandable), now adds this to their description of Arch.

eveninghere , (edited )

viewers of select YouTube videos

So, now a policeman who doesn't like whoever YT user can unseal who that is... For a far-right policeman that can be anybody whose comment looks "woke".

I think I shouldn't use YT anymore. Actually, any SNS that uses my email...

Edit: they collected user identities on who watched videos on fucking "drones".

eveninghere ,

I think it's unrealistic for now.

Peertube works for organizations like Blender because they completely retain the copyright. Other than that, video hosting is too expensive and gains little audience in the first place.

eveninghere ,

The problem with this narrative is that Musk lost money doing so.

eveninghere ,

It must become illegal for a video game not to integrate this system

eveninghere ,

No way they can form a proper response to you on GDPR without citing GDPR. This is either utter incompetence or a lie. Wondering if one could sue them just for this reply message.

eveninghere ,

As an academic I love this. On Wikipedia there's actually fights among different expert disciplines going on. It is better to allow different instances operated by different discipline summarize knowledge from their own perspective.

eveninghere ,

I can respect your comment. The problem with Wikipedia's scholarly articlesI wanted to raise was that some group of researchers (or businesses) wash away others' views. In other times, mathematicians try to satisfy everyone from different disciplines, and write a very abstract article that covers everyone's view yet is too academic and hardly readable to most readers who actually need Wikipedia.

eveninghere ,

If you're correct, to me the usefulness of Wikipedia is actually different from that of encyclopedia, and the pattern I'm arguing goes against that.

eveninghere ,

Or they don't want the tool of specifically the CCP.

eveninghere ,

The current internet search is becoming obsolete. People are able to tell apart BS, though. This means, there's a possibility for a smarter filter. Hard to tell whether we will see one in the near-future.

eveninghere ,

you're right. I should've written some people

eveninghere ,

Why is arstechnica so cynical about specifically big tech?

eveninghere ,

I know, but arstechnica could also write other kinds of articles.

eveninghere ,

Agreed. If you vote for Biden anyway, he doesn't have the incentive to care.

Do they wait for the next 4 years to do this again? I like the intention but honestly don't understand the point.

eveninghere ,

That's why Uncommitted is pointless. Thanks for confirming.

eveninghere ,

Nah, you can do peaceful protests that would give incentives to Biden.

So, no, not only. Also not that I think that way.

eveninghere ,

During a closed-door meeting last week, Professor Ester Fuchs, who is one of the chairs of the task force, invoked a Supreme Court justice’s famous line about pornography: “I know it when I see it.”

Lmao

Apple will require notarization for apps from third party app stores, and will disable updates for apps installed via third party app stores if staying outside EU (support.apple.com)

As far as I can tell this basically means that all apps must be approved by Apple to follow their "platform policies for security and privacy" even if publishing on a third party app store. They will also disable updating apps from third party app stores if you stay outside the EU for too long (even if you are a citizen of an EU...

eveninghere ,

What boggles my mind is that the level of sandboxing displayed in Apple's App Store is not really interpretable to me.

I also see something like "the developers indicated they do not collect sensitive information." Yeah, but why would they indicate otherwise if they were malicious parties?

Probably, the only way to get sort of assurance is to choose an open source project, but App Store doesn't guarantee that the code on Github matches the app in the Store.

eveninghere ,

There's no guarantee on third party tools continuing to work with Lemmy. Something as critical as deleting images, which can cause problems like revenge porn and such, must be given priority by the official project.

eveninghere ,

Maybe encryption can help? Instances only copy the encrypted image, and only the original instance provides the encryption key to client apps. This way, the original instance can de-facto delete the image copies by simply refusing to issue the key.

eveninghere , (edited )

Lemmy providing an open API does not mean that third parties maintain compatibility forever. (Edit: for example, what happens if the third party app gets taken away by a malicious maintainer? Or becomes buggy, or un-maintainable, project dies, etc.)

I don't want to upset you, but I think I have to say this for the sake of the community. The attitude like "we provide an API, so third parties will follow," is what is causing instance admins' distrust in the first place.

eveninghere ,

And that's fine. If one instance has to be able to delete one revenge child porn from all the instances, it has to be what you call a DRM. Also, don't cache the key because that's exactly the opposite of what this is meant for.

eveninghere , (edited )

I think it's a dead lock. The replies show that they can't even understand the concern.

That's a typical death to a project. For, there will never be a moment for the team to address the concern. Whatever you try, the team won't move an inch.

I don't know what instance admins are thinking, but there's no point complaining at this point.

eveninghere ,

I'd say, if you're concerned, better safe than sorry. Nobody can take the legal responsibility here.

eveninghere ,

But why not issue trackers?

eveninghere ,

Kbin website implements this.

eveninghere ,

Well, to be precise, kbin groups posts with the same URL.

eveninghere ,

Yeah, those granpas asking me to fix their machines are like, "and, what do we do with my OneDrive which I don't understand what it is?"

eveninghere ,

Here we go again

eveninghere ,

This is actually an unfair experiment. This behavior is not specific to questions about MS. Copilot is simply incapable of this type of discussion.

Copilot tends to just paraphrase text it read, and when I challenge the content, it ends the conversation like this, instead of engaging in a meaningful dialogue.

Tumblr and Wordpress to Sell Users’ Data to Train AI Tools (www.404media.co)

this could not be timed worse for Tumblr which is in huge hot water with its userbase already for its CEO breaking his sabbatical to ban a prominent trans user for allegedly threatening him (in a cartoonish manner), and then spending a week personally justifying it increasingly wildly across several platforms. the rumors had...

eveninghere ,

It should be illegal for the company to own user-generated contents. They should at least pay the users.

eveninghere ,

You must be kidding. You surely haven't heard about Fediverse.

eveninghere ,

What I meant is that the data generally belong to the user on Fediverse, and your original comment ignored that.

eveninghere , (edited )

I mean, even if that's tue, I don't count it as "ownership" if they change the monetization scheme for what I wrote, without giving me a good chance to say what I get in return. Reddit even allegedly put back comments which users deleted.

It's near-impossible to delete all my own comments on Reddit, for example.

eveninghere ,

Hard disagree.

  • Companies don't give you a good chance to prevent them from eroding your control on the data further.
  • Even if they technically disown the data, they're doing even worse. They are de-facto doing anything they want without taking responsibility
  • Open source / CC licenses don't apply to whatever user data people are talking about here
eveninghere ,

And it doesn't change what I wrote.

eveninghere , (edited )

Trust me, fb ads are way worse than it used to be. Mine is flooded with celebrities advertising scams in the ads, in my region.

"Hi, I'm [some celebrity name]. When I was young, I didn't have money, but investing on stocks changed my life! ..."

These scammers use ChatGPT to generate millions of these and fb is now chaos because admins can't do anything, I tried reporting and tuning ad preferences, but there's no solution because they are literally flooding the platform. These even aren't copy pastes because the AI generates a huuuge variety of them. The fact that the ad posts mimics actual posts (by fb's deliberate design) doesn't help.

And I hate celebrities, so seeing their face already makes my day a bad one.

eveninghere ,

How would people on Mars even claim their rights? I never heard a country recognize their own jurisdiction on Mars.

If Musk enslaves these people, not return them to earth, etc., how can they stop that from Mars?

eveninghere ,

What's funny, we complain about the terminology use of AI, but nobody can actually define the intelligence.

eveninghere ,

The problem with these definitions is that they are verbal. Some could argue ChatGPT is capable of understanding, while others could do the opposite. I don't even believe it is capable of abstraction.

The Turing test was novel in that we could test the intelligence of AIs without actually defining intelligence. And it's still useful because researchers probably can't agree on a rigorous definition of intelligence.

eveninghere , (edited )

I'm a scientist entering the industry and can't agree more. Too many lies. There are handful of companies that do deliver, but, generally speaking, many businesses seem to bet on the naivety of investors. Some even do it unintentionally because the bar is simply too low.

But here's the thing. I noticed in my life that there are naive people following weak narratives no matter what. These people can't doubt your arguments. And if you do business, these people are apparently the most solid support.

Edit: it's perhaps also true that this majority of investors forces companies to lie by investing almost solely on LLMs

eveninghere , (edited )

The dilemma is, ChatGPT can write better reports than most graduate school students in my country. For what these problematic vast majority of students do is remember, not analyze.

Specifically for this context, students are not trained to analyze what they are asked (input query for ChatGPT). When I ask a unique question in their assignment, they can't even form a response. They just write a generic text that doesn't try to answer my question.

They seem to copy and paste what's in their brain. And when it comes to copy and pasting, i.e. mimicking what people do, ChatGPT is the champion in some sense. Hell, OpenAI even tuned it to generate balanced stance, and that's also what students can't do.

Finally, 90% of the population perform actually worse than these graduate students.

eveninghere ,

Until like 3 months ago, I felt the ChatGPT revolution was going on. Every 10-year plan my colleagues in AI research was actually completed in a few weeks(!) by a completely different research team at the opposite side of the planet. The hype was so high that every expert had same plans resulting in surreal competitions.

After that, the LLM businesses entered the b2b space, with all potential customers asking ChatGPT to search information in their pile of documents. That was the next big thing.

We haven't heard back from the pile of garbage so far...

jlou , to Technology

Apple Is Trying to Kill the Open Internet!

https://youtu.be/up-zUEFNMww

@technology

eveninghere ,

Oof. Can they just let me run my own app on my phone?

  • All
  • Subscribed
  • Moderated
  • Favorites
  • •
  • random
  • incremental_games
  • meta
  • All magazines
    •