I keep checking bard vs Claude.ai vs GPT and bard got significantly better during recent weeks. It excels in image recognition, getting jokes hidden in images and has a very informed opinion on current news at least in my examples. It was really bad a couple of months ago but that seems to shift 

  Quote

How AI reduces the world to stereotypes

https://restofworld.org/2023/ai-image-stereotypes/

just nuke it from orbit..

  On 10/11/2023 at 7:27 AM, iococoi said:

https://restofworld.org/2023/ai-image-stereotypes/

just nuke it from orbit..

"An American Person" didn't generate any fat people, so it's not all stereotypes.

Also, this one is pretty cool and horrifying.
 

  On 10/11/2023 at 7:27 AM, iococoi said:

https://restofworld.org/2023/ai-image-stereotypes/

just nuke it from orbit..

Stereotypes are the natural consequence of having a system that tries to model the world using categories and course grained approximations, it's impossible to avoid this if you're trying to model a complex domain, humans are not above this.

ha!

https://huggingface.co/ostris/ikea-instructions-lora-sdxl

  On 10/15/2023 at 6:38 PM, iococoi said:

ha!

https://huggingface.co/ostris/ikea-instructions-lora-sdxl

Expand  

lol

Some video stuff on TikTok Ive seen is very impressive. Like 1000 times better than cgi. 

idiocrazy as we type ...

https://www.technologyreview.com/2023/09/19/1079832/chinese-ecommerce-deepfakes-livestream-influencers-ai/

  On 10/19/2023 at 9:02 AM, iococoi said:

idiocrazy as we type ...

https://www.technologyreview.com/2023/09/19/1079832/chinese-ecommerce-deepfakes-livestream-influencers-ai/

 

that's nuts. i'm sure some of them are making $$$$ and becoming famous

Was I gaslighted by an ai fanny? Will Sam Alternativeman give me compo???

  Quote

The tool, called Nightshade, messes up training data in ways that could cause serious damage to image-generating AI models. 

https://www.technologyreview.com/2023/10/23/1082189/data-poisoning-artists-fight-generative-ai/

I really love that thats what ChatGPT really looks like: 

https://goodinternet.substack.com/p/ai-is-a-shoggoth

  On 10/25/2023 at 11:35 AM, o00o said:

I really love that thats what ChatGPT really looks like: 

 

https://goodinternet.substack.com/p/ai-is-a-shoggoth

Expand  

Yeah I like the shoggoth way of viewing these things I reckon it's pretty accurate. I highly recommend this article of you're interested in this sort of thing: https://www.lesswrong.com/posts/vJFdjigzmcXMhNTsx/simulators basically viewing these models as simulators, since they are trained to predict the next token, they must learn how to model the systems that produced the tokens. Because they have been trained on pretty much the whole Internet it means they are in theory able to simulate (to some degree of resolution) a wide range of types of people and situations.

agi is not a real concept

just cuz u can make sentence doesnt mean it is

  On 10/26/2023 at 6:03 AM, zlemflolia said:

agi is not a real concept

just cuz u can make sentence doesnt mean it is

Care to explain further? I feel like this depends on the definition, if you define it as some software running on a computer that is capable of all the mental capabilities of a human(in a functional sense, being able to learn, complete tasks etc, not something like consciousness) then I don't see why this isn't a real concept or possibility.

Maybe you mean that the concept itself contains a contradiction that makes it impossible? Like the term "communist utopia".

  On 10/26/2023 at 6:03 AM, zlemflolia said:

agi is not a real concept

just cuz u can make sentence doesnt mean it is

Assuming Artificial means taking place on a manmade non-biological substrate and General Intelligence means the ability to solve arbitrary problems with equal or greater competency than humans why is AGI not a real concept?

  On 10/26/2023 at 11:15 AM, vkxwz said:

Care to explain further? I feel like this depends on the definition, if you define it as some software running on a computer that is capable of all the mental capabilities of a human(in a functional sense, being able to learn, complete tasks etc, not something like consciousness) then I don't see why this isn't a real concept or possibility.

Maybe you mean that the concept itself contains a contradiction that makes it impossible? Like the term "communist utopia".

ill pretend i didnt read that awful last sentence

anyway, it necessarily is not going to be a good thing.  intelligence is not something we even understand.  there are extremely intelligent people actively working on awful projects that hurt humanity horribly.  "AGI" is such an ill defined goal that nobody even know what it is, that it basically just means "super weapons".  it will not serve humanity's needs, by definition, since it will be owned by the owning class, and it serves the needs of its owner only

  On 10/26/2023 at 12:37 PM, Real Human Bean said:

Assuming Artificial means taking place on a manmade non-biological substrate and General Intelligence means the ability to solve arbitrary problems with equal or greater competency than humans why is AGI not a real concept?

solve arbitrary problems? what does that mean then?  i doubt you can say, because nobody knows,

I find the facial expressions poorly executed for what they could do. They should use some monitor showing a waveform 

  Quote

AI companies have all kinds of arguments against paying for copyrighted content

https://www.theverge.com/2023/11/4/23946353/generative-ai-copyright-training-data-openai-microsoft-google-meta-stabilityai

I haven't watch it but they played a report in Portuguese news about some scams that have been happening... someone did some fake news using the image of a well known diabetes doctor claiming that he had a cure...

  On 11/6/2023 at 11:10 AM, cruising for burgers said:

I haven't watch it but they played a report in Portuguese news about some scams that have been happening... someone did some fake news using the image of a well known diabetes doctor claiming that he had a cure...

Expand  

Been aware of this issue for years now. I knew it since I saw first faceswaps apps. The video summarizes the potential of this new technology, involving this philosophical question about what's real and whats not. It forced me to be skeptical about news and all major corporate media stuff output recently.

deep fakes hardly matter, you already have meatfakes spewing propaganda on tv