Search results

Facts and experiments are demonstrating that it isn't more nuanced than that. The training phase is effectively using gradient descent to find the perfect compression algorithm for the provided training set.

But I read it on a random website so it must be true.

You're missing the point. This has nothing to do with arguments about human creativity or derivative art. The training process literally compresses and stores copies of all the training data, aka stolen content in the case of copywritten content, using what is effectively computing's funkiest...

Because this place is insane, I don't know if you're being serious, but in case you are: We know for an absolute fact that any given model is a fixed string of numbers. The fixed string of numbers requires a massive amount of time, power, and compute to generate from the training set. That...

It's not that it's "blocked" from accessing the live web, it's just not how it works. These systems do not incorporate new data in realtime, the model is a fixed string of numbers. It... fuck it, here's a gif

Transformer neural network models do not go out and search the live web for content to incorporate into their output. The data has to be inside the training set. Bing Chat has a layer of Microsoft voodoo that allows it to incorporate web references into final the output, but that's a...

Everything that guy said is bullshit. "This is precisely how the human brain works. You have fragments of “facts” and you back fill with filler words. This fact is lost on most AI researchers." -some guy in a tweet Anyone claiming AI models and human brains work remotely similar knows nothing...

Not sure why @BoozeCube negged me. We know that Alliance is slightly overpopped on Lone Wolf (because it is still intermittently locked) yet that bar shows it massively underpopped. It does not line up with reality.

The soulless sameness of it all doesn't fire off red flags for you? Or the fact that the more 'aligned' it gets, the more everything just looks the same?

I don't see how those numbers can be remotely accurate.

5275

I literally work on platforms for Conversational and Generative AI for use in call center applications, using the very latest models and APIs. I understand how these technologies work. The 'video version' is just a bunch of still frames at slightly different angles. It did not 'improve'...

That's because it's just copying them directly from the images in the training set. It's not generating anything, it's just smoothing. The god emperor truly has no clothes here. They're just computationally expensive plagiarism engines.

Using an absolute shitton of compute to generate a bunch of disjointed, lifeless filler scenes that subtly shift perspective for 3 seconds is only impressive to drooling retards.

From the inventor of the web:

SoFi

SoFI

Do you remember any of the character names in Chronicles of Riddick besides "Riddick" and "Necromonger guy?"

I like Chronicles of Riddick and Jupiter Ascending. I like them because they are bad but also good. Then there's movies like Demolition Man, which are also bad but also very good. This just seems like it's bad.

I got trolled for saying this after only playing for a couple nights and the trolls were right. Honestly, Lone Wolf alliance has a pretty good community.

Where do I watch this?

In an ideal world, every consenting content producer gets their work trained into these models, contributions above a certain percentile threshold get attributed, and everyone gets compensated anytime their content is used. In the world we're heading for, Woke OpenAI owns the entire internet...

That's just your lack of reading comprehension, I already covered this in that same post: The href shows where the content came from, and links to where you can go find it. It therefore serves as both a reference and an attribution. The img src is a reference to the cache on some CDN...

That's all anyone is asking, but it's not technically possible the way the technology is built. If you remove it from the training data it does not remove it from the model, aka the current version of the product. They would have to re-run the training and produce an entirely new model anytime...

https://www.reuters.com/legal/transactional/ny-times-sues-openai-microsoft-infringing-copyrighted-work-2023-12-27/ Full text of the complaint:

Yes. Many people, repeatedly, including many ongoing lawsuits. It is not technically feasible. They cannot remove the content from the model. Think of the model as a vast network of interconnected numbers. No one, not even the creators, understands how any given piece of content or any...

Yes, thanks for losing the argument. Further, for all of these copywritten books, Google displays the copyright page of the books. It attributes the content to the original content producer. LLMs and Diffusion models do none of these things. They attempt to pass the copied content, stored...

You're 100% demonstrably wrong. Our own forum threads do this all the time. Because of the storage limits on this website, much of the content of any of the picture and gif threads are linked from elsewhere, either to Imgur's backend, or Discord's CDNs, etc, or many other websites. That is...

There is no valid legal argument that compressing the original content itself and storing it inside the model, aka OpenAI's product, qualifies as either Fair Use or Transformative Use.

You are very obviously way out of your element here on a technical level, and your argument sucks on top of it. When you do a Google Image Search and are presented with the indexed images, you get this: So the image you're presented in the index is in fact pulled from the original source...

Except theirs, of course.

You don't seem to get how this works. It is not browsing the web and grabbing a screencap from Dune for you. That screencap from Dune is stored inside the model from the training run; that math is part of the model's neural network and model weights, its knowledgebase. Therefore, they have...

You have walked into my trap card. Google hyperlinks to material. Google does not present those images as their own product. Further, for 25+ years, robots.txt has allowed you to control how your content is indexed by search engines. Additionally, Google gives massive amounts of control to...

This is going to fuck up all my pre-greench 99 parses.

One writer would:

What it demonstrates is that it contains the copyrighted images inside the model.

"AI" Literally just Napster.