Artificial Intelligence Developments (ChatGPT etc)

[weltweit]

Todays announcement is pretty amazing

Link? what announcement?

 

[UnderOpenSky]

Hi UOS, I know how to prompt ChatGPT to write things based on asking it increasingly detailed questions, but how would I get it to improve the grammar of a piece that I had written?

Type improve the grammar or check this readability and paste what you want it to check underneath.

 

[pbsmooth]

Link? what announcement?

maybe this? quite hard, predictably, for an idiot like me to understand exactly what this latest AI billion-dollar company actually does...

Amazon steps up AI race with up to $4 billion deal to invest in Anthropic

Amazon's employees and cloud customers will gain early access to technology from Anthropic as part of the deal.

www.reuters.com

 

[Kid_Eternity]

Well worth a read, big update rolling out:

ChatGPT can now see, hear, and speak

We are beginning to roll out new voice and image capabilities in ChatGPT. They offer a new, more intuitive type of interface by allowing you to have a voice conversation or show ChatGPT what you’re talking about.

openai.com

 

[Signal 11]

British government quietly sacks entire board of independent AI advisers

The British government has quietly sacked an independent advisory board of eight experts that had once been poised to hold public sector bodies to account for how they used artificial intelligence technologies and algorithms to carry out official functions.

It comes as Prime Minister Rishi Sunak drives forward with a much-publicized commitment to make the United Kingdom a world leader in AI governance, and ahead of a global AI Safety Summit being arranged for November in Bletchley Park.

 

[Signal 11]

These 183,000 Books Are Fueling the Biggest Fight in Publishing and Tech

This summer, I acquired a data set of more than 191,000 books that were used without permission to train generative-AI systems by Meta, Bloomberg, and others. I wrote in The Atlantic about how the data set, known as “Books3,” was based on a collection of pirated ebooks, most of them published in the past 20 years. Since then, I’ve done a deep analysis of what’s actually in the data set, which is now at the center of several lawsuits brought against Meta by writers such as Sarah Silverman, Michael Chabon, and Paul Tremblay, who claim that its use in training generative AI amounts to copyright infringement.

 

[Signal 11]

Well worth a read, big update rolling out:

ChatGPT can now see, hear, and speak

We are beginning to roll out new voice and image capabilities in ChatGPT. They offer a new, more intuitive type of interface by allowing you to have a voice conversation or show ChatGPT what you’re talking about.

openai.com

loooool this is their "Head of Safety Systems"

Spoiler: tweet text

Lilian Weng (@lilianweng)

Just had a quite emotional, personal conversation w/ChatGPT in voice mode, tlaking about stress, work-life balance. Interestingly I felt heard & warm. Never tried therapy before but this is probably it? Try it especially if you usually just us it as a productivity tool.

Quote tweeting Sam Altman (@/sama) saying "voice mode and vision for chatgpt! really worth a try. openai.com"

 

[DotCommunist]

its still a reverse turing test then.

 

[Karl Masks]

How do i get Skynet or whatever it's called to do some art for my music so I don't have to keep using free clip art.

 

[kabbes]

For the record, therapy is most definitely not a process of pretending to listen to someone while not actually making sense of a single thing that they say.

 

[Shippou-Sensei]

How do i get Skynet or whatever it's called to do some art for my music so I don't have to keep using free clip art.

Midjourney Quick Start Guide

Learn how to use the text-to-image service, Midjourney on Discord or the web to create custom images from simple text prompts.

docs.midjourney.com

 

[Lazy Llama]

Interestingly, now that OpenAI have said ChatGPT is able to browse the internet for current information, we see this in the visiting robots info:

 

[platinumsage]

I don't have real-time browsing capabilities or specific up-to-date information on particular threads or conversations about the PCS conference on the internet or any social media platforms. However, I can provide some general information.
PCS (Picture Coding Symposium) is a conference that gathers researchers and experts in the field of visual data coding to discuss, exchange ideas, and present new advances in the field. Topics often include image and video compression, processing, transmission, and more.
To find specific threads or conversations about the PCS conference, I would recommend:

1. Searching on social media platforms like Twitter, LinkedIn, or Facebook using the hashtag #PCSconference or similar keywords related to the event.
2. Visiting forums and discussion boards related to image and video coding or processing.
3. Checking the official website of the PCS conference for announcements, updates, and forums.
4. Performing a Google search with keywords like "PCS conference 2023 discussion" or "PCS conference forum."

Always ensure to respect the privacy and terms of use of any online forums or discussion boards you participate in.

 

[Lazy Llama]

You can’t expect it to make much of the contents of urban75, especially without further disambiguation on what PCS might be.

 

[UnderOpenSky]

How do i get Skynet or whatever it's called to do some art for my music so I don't have to keep using free clip art.

Midjourney Quick Start Guide

Learn how to use the text-to-image service, Midjourney on Discord or the web to create custom images from simple text prompts.

docs.midjourney.com

If you just want to mess around and not use Discord then Stable Diffusion is quicker to get into. I've seen more impressive stuff from Mid Journey though.

Stable Diffusion Online

Stable Diffusio Online Demo. FREE forever. Create beautiful art using stable diffusion ONLINE

stablediffusionweb.com

 

[Lazy Llama]

Beth Frey who did the cover of Roisin Murphy’s latest album has some spectacularly strange and disturbing AI-generated videos on her insta

Beth Frey (@sentientmuppetfactory) • Instagram photos and videos

111K Followers, 1,012 Following, 526 Posts - See Instagram photos and videos from Beth Frey (@sentientmuppetfactory)

instagram.com

 

[Karl Masks]

are there any AI that dont' want money? I'm not prepared to cough up right now

 

[Buddy Bradley]

Interestingly, now that OpenAI have said ChatGPT is able to browse the internet for current information, we see this in the visiting robots info:
View attachment 393263

Do they respect robots.txt?

 

[Shippou-Sensei]

Used word dic

are there any AI that dont' want money? I'm not prepared to cough up right now

DALL·E mini by craiyon.com on Hugging Face

Discover amazing ML apps made by the community

huggingface.co

But you get what you pay for. Also most places give you a bunch of credits for signing up

 

[Karl Masks]

Used word dic

DALL·E mini by craiyon.com on Hugging Face

Discover amazing ML apps made by the community

huggingface.co

But you get what you pay for. Also most places give you a bunch of credits for signing up

That didn't happen with midwonder or whatever it's called. I might come back to it later. I don't mind them charging, I just can't justify it right now

 

[Shippou-Sensei]

But you get what you pay for. Also most places give you a bunch of credits for signing up

With the promt "People in a place called urban75 arguing about what comes first "cheese or beans" it is difficult to tell if the argument is serious or not."

Free


Premium


As you can tell it's a bit of a mixed bag.

 

[danski]

With the promt "People in a place called urban75 arguing about what comes first "cheese or beans" it is difficult to tell if the argument is serious or not."

Free
View attachment 393301

Premium
View attachment 393303

As you can tell it's a bit of a mixed bag.

Then free one is bang on the money.

This is exactly what I visualise when people argue about the beans/cheese sequence of assembly.

 

[Shippou-Sensei]

Then free one is bang on the money.

This is exactly what I visualise when people argue about the beans/cheese sequence of assembly.
View attachment 393306

I don't know what that is but I'm certain ET mastubates to it.

 

[Karl Masks]

how do thes images get generated? Are they stolen from existing works found online? Am i breaching copyright by using them myself?

 

[xenon]

Interestingly, now that OpenAI have said ChatGPT is able to browse the internet for current information, we see this in the visiting robots info:
View attachment 393263

So soon we can. ‘Chat GPT, write a reply to this post in the style of butchersapron’

 

[xenon]

How do i get Skynet or whatever it's called to do some art for my music so I don't have to keep using free clip art.

There are a few services that do this. I us night café. and then adjust the size in Microsoft paint. if I could find one, an accessible one that would make large images that would be good.

 

[NoXion]

how do thes images get generated? Are they stolen from existing works found online? Am i breaching copyright by using them myself?

Some artists have made that argument. Personally I think they're talking bollocks. Transformative works are allowed under fair use, AI-generated images are more transformative than a collage, and nobody is trying to sue people making collages. While you can use prompt engineering to generate images similar to the AI training material, including watermarks, I'm not convinced that should a problem for the AI creators, rather than the people are using the AIs to recreate copywrited material. That would be like suing Adobe because of people using their Photoshop program to rip off artwork.

I don't think a legal precedent has been set either way yet.

 

[Lazy Llama]

So soon we can. ‘Chat GPT, write a reply to this post in the style of butchersapron’

Write a forum post suitable for urban75.net in the style of "editor"
-=-=-


Subject: The Ever-Changing Urban Landscape: A Reflection on Gentrification
Author: UrbanEnthusiast
Greetings fellow urban explorers,
As the editor of urban75.net, I wanted to take a moment to reflect on a topic that has been a constant undercurrent in our discussions over the years: gentrification. Our beloved cities are ever-evolving, but with change often comes a complex mix of emotions and concerns.
Gentrification, in many ways, embodies the dual nature of urban development. On one hand, it can bring revitalization to neglected neighborhoods, breathing new life into once-decaying areas. This can lead to improved amenities, safer streets, and a sense of community pride that is genuinely heartwarming to witness.
However, on the other hand, gentrification can also bring about displacement, rising living costs, and the loss of the unique character that made these neighborhoods so appealing in the first place. It's a challenging issue that requires careful consideration and open dialogue.
I encourage all of you to share your thoughts and experiences with gentrification in your own cities and neighborhoods. What are the positive aspects you've seen? What are the drawbacks? Have you witnessed any successful efforts to mitigate the negative impacts of gentrification?
Additionally, let's remember that our community is diverse, and our experiences with gentrification can vary greatly. It's essential to approach this discussion with empathy and an open mind, respecting the diverse perspectives that exist here on urban75.net.
In the spirit of urban exploration, let's continue to engage in constructive conversations about the ever-changing urban landscape. Together, we can better understand the forces at play and work towards more inclusive, vibrant, and sustainable cities for all.
Looking forward to reading your insights and stories.
Warm regards,
[Your Name]Editor, urban75.net
-=-=-

I remain unconvinced

 

[Shippou-Sensei]

I don't think a legal precedent has been set either way yet.

Some stuff has.

As the work was not made by a human it cannot have copyright.

A lot of the precedent come from the monkey selfie case.

 

[NoXion]

Some stuff has.

As the work was not made by a human it cannot have copyright.

A lot of the precedent come from the monkey selfie case.

Now that I don't understand. AI image models don't do anything without a human entering a prompt. They're not autonomous agents like animals are. So what's the difference between me creating an image in Photoshop, versus using an AI model to do it? In both cases my input was needed for the image to ever exist in the first place.

Sounds like a dumb ruling made by someone who doesn't understand that AI image models aren't actually sentient.