taps nose

This

Technically a grey area. How many humans do you see here posting walls of text and putting the attribution at the very top, if the attribution is different from their username of course.

ChadGPT

Maybe not a post joke world, but certainly one where Onion articles are more believable than real ones…

Exactly, I was trying to make that point, and specifically looked at Meta and the current “Torrenting is only Illegal if you seed” bs defence they’re using. Which was completely ignored by our friend over there.

Sure, cool…coolcoolcool…

So how about the ongoing lawsuit where there’s been hard evidence of Meta torrenting terabytes of data, and then claiming it’s not illegal because they didn’t seed?

And that’s literally only for people who posted on FB or Instagram or whatever other garbage data mines Meta are operating. The above is what I was talking about. The highly public trial that relates specifically to torrenting data illegally to train an LLM. That’s what I was talking about. Whoever got fucked by Meta’s T&C’s to train their AI are SoL, which is why I didn’t mention any training they were doing on their platforms specifically.

I personally believe there could be a place in a person’s creative workflow for the use of AI as a tool to enhance their own creative work…with caveats…

As for the ethics of using AI to copy and art style? It’s theft. End of.

These models are trained on stolen data. The artists/musicians/writers/intellectuals, or their estates, never gave permission for their works to be used to train these models. They never receive royalties, or payment of any kind, for the use of their works. And as we’re finding out, at the very least, Meta took that data…those creative works… illegally. People’s lives have been destroyed by laws put in place to protect IP. I personally feel those laws are fucked and should be fully scrapped in favour of something that actually protects the people creating these works. That doesn’t change the fact that when Joe Shmoe shares a torrent he could be hit with fines and possibly jail. Fines alone could essentially make a person’s life literal hell for however long they have left. The companies who have trained these models are likely going to get a “cost of doing business” slap on the wrist.

It’s ethically ambiguous if you look at it from the standpoint of “IP law shouldn’t exist” while totally ignoring that even if these companies get away with it common people nearly never will.

It can, but I’d like to see what it’s doing instead of just out and out blocking it. Still going to block it, just want to know who all it’s calling home to.

There an app I can run on android that’d log that?

That…is something I’d not tried yet…hmmmm. I’ll try it tonight and get back

because why else would you go to a whole other post to “prove a point” about downvoting?
It wasn’t you (you claim)

I do claim. I have an alt, didn’t downvote you there either. Was just pointing out that you were also making assumptions. And it’s all comments in the same thread, hardly me going to an entirely different post to prove a point.

We will not get the benefits of Generative AI if we don’t 1. deal with the problems that are coming from it, and 2. Stop trying to shoehorn it into everything. And that’s the discussion that’s happening here.

I agree. And while I personally feel like there’s already room for it in some people’s workflow, it is very clearly problematic in many ways. As I had pointed out in my first comment.

I’m not going to even try to justify to you what I said in this post or that one because I honestly don’t think you care.

I do actually! Might be hard to believe, but I reacted the way I did because I felt your first comment was reductive, and intentionally trying to invalidate and derail my comment without actually adding anything to the discussion. That made me angry because I want a discussion. Not because I want to be right, and fuck you for thinking differently.

If you’re willing to talk about your views and opinions, I’d be happy to continue talking. If you’re just going to assume I don’t care, and don’t want to hear what other people think…then just block me and move on. 👍

Nope. It only happens when connected to a mobile device. Tested it on all 3 of my android phones (current daily driver, back up phone, and an S4 that I use as a media player in my car,) my wife’s iPhone X, my Samsung tablet, and my work iPad. The S4 doesn’t have that behaviour, but isn’t really a viable option. The battery is basically only good enough to keep it powered for an hour or two, and doesn’t jive with my cell provider’s SIM it’s also running a Google free ROM… Both my more modern phones, my wife’s, and the tablets have the exact same behaviour. When I got them I had no intention of even installing the app. Wasn’t until I was trying to troubleshoot the audio drop out that I installed it. And through experimenting with permissions and services turned on and off decided to just not use them with my phone.
The settings provided by the B&O app aren’t in userland for android, and as far as I’ve been able to find, can’t be pulled up and implemented without the app.

and here you are, downvoting my valid point

Wasn’t me actually.

valid point

You weren’t really making a point in line with what I was saying.

regardless of whether we view it as a reliable information source, that’s what it is being marketed as and results like this harm both the population using it, and the people who have found good uses for it. And no, I don’t actually agree that it’s good for creative processes as assistance tools and a lot of that has to do with how you view the creative process and how I view it differently. Any other tool at the very least has a known quantity of what went into it and Generative AI does not have that benefit and therefore is problematic.

This is a really valid point, and if you had taken the time to actually write this out in your first comment, instead of “Tell that to the guy that was expecting factual information from a hallucination generator!” I wouldn’t have reacted the way I did. And we’d be having a constructive conversation right now. Instead you made a snide remark, seemingly (personal opinion here, I probably can’t read minds) intending it as an invalidation of what I was saying, and then being smug about my taking offence to you not contributing to the conversation and instead being kind of a dick.

I wasn’t able to use my WiFi ssid in home assistant automations without having location services enabled, for example.

I was recently(6 months ago) gifted a very nice pair of Bang & Olufsen BT headphones. They come with a 3.5mm hardline as well…but since my phone, and basically everyone else’s, no longer has a port for that…🤷
BT connection works amazing with both my laptops and my desktop. Zero issues. To use them with my android phone the B&O connection app has to be open, WiFi on, precise location approved. Because of my settings for precise location data not being available for apps in the background, the B&O app has to be focused. If WiFi is off, and precise location is denied the app refuses to connect to the headphones, despite my actual phone seeing and connecting to it with no issues. I’d just not use the app…but without the app there’s (absolutely intentional) audio drops. I know it’s on purpose because it happens in a pattern. 90 seconds of perfect audio, followed by four 2 seconds cuts 5 seconds apart, then another 90 seconds of perfect audio. So I just don’t use them with my phone. 🖕

Ok? If you read what I said, you’ll see that I’m not talking about using ChatGPT as an information source. I strongly believe that using LLMs as a search tool is incredibly stupid…for exactly reasons like it being so very confident when relaying inaccurate or completely fictional information.
What I was trying to say, and I get that I may not have communicated that very well, was that Generative Machine Learning Algorithms might find a niche as creative process assistant tools. Not as a way to search for publicly available information on your neighbour or boss or partner. Not as a way to search for case law while researching the defence of your client in a lawsuit. And it should never be relied on to give accurate information about what colour the sky is, or the best ways to make a custard using gasoline.

Does that clarify things a bit? Or do you want to carry on using an LLM in a way that has been shown to be unreliable, at best, as some sort of gotcha…when I wasn’t talking about that as a viable use case?

Oh, and it also hallucinates.

This is arguably a feature depending on how you use it. I’m absolutely not an AI acolyte. It’s highly problematic in every step. Resource usage. Training using illegally obtained information. This wouldn’t necessarily be an issue if people who aren’t tech broligarchs weren’t routinely getting their lives destroyed for this, and if the people creating the material being used for training also weren’t being fucked…just capitalism things I guess. Attempts by capitalists to cut workers out of the cost/profit equation.

If you’re using AI to make music, images or video… you’re depending on those hallucinations.
I run a Stable Diffusion model on my laptop. It’s kinda neat. I don’t make things for a profit, and now that I’ve played with it a bit I’ll likely delete it soon. I think there’s room for people to locally host their own models, preferably trained with legally acquired data, to be used as a tool to assist with the creative process. The current monetisation model for AI is fuckin criminal…

Yep!

See point #2. Hand written notes can be kept as is, one thing that’s not mentioned in that link is that Joplin can also use OCR to change handwriting to typed.

Joplin has apps for all UNIX-likes(Mac, Win and Lin,) iOS(phone and pad variants), and Android. It can import Evernote and OneNote. It’s built to E2EE notes to cloud services, currently supporting: Joplin Cloud, Nextcloud, S3, WebDAV, Dropbox, OneDrive or the local file system.
If you’d like to fully tell M$ to get fucked…I can recommend self hosting NextCloud.
Not incredibly difficult to set up, I’m fairly fluent in Linux systems…but by no means a “Wizard”, and my self hosting knowledge is extremely minimal.

There’s other options out there, but Joplin has all the apps…

20 years, would have been better if they’d stuck with that instead of cutting it to a one year and a day, with non-custodial supervision for another year…

Usually when the USPS cops get involved The. Hammer. Fuckin. Drops.
Shame to see the sentence reduced…

Would all of the Google spyware be stripped out of Google’s Android distribution?

No. Or, almost certainly no. If they were forced to sell Android, my guess would be that another tech giant, that makes most of their money from data harvesting, would be most likely to buy it. Meaning any spyware would remain, and just be pointed to different servers.