The AI discussion thread

[JEDI]

 

[Kaido]

Hi-res local rendering now available with the offline version of Stable Diffusion XL:

Stability AI releases Stable Diffusion XL, its next-gen image synthesis model

New SDXL 1.0 release allows hi-res AI image synthesis that can run on a local machine.

arstechnica.com

The user interface technology is starting to improve as well:

GitHub - AUTOMATIC1111/stable-diffusion-webui: Stable Diffusion web UI

Stable Diffusion web UI. Contribute to AUTOMATIC1111/stable-diffusion-webui development by creating an account on GitHub.

github.com

As an artist, this stuff is pretty disheartening. As a creative person, these tools are absolutely incredible! It's like going from a typewriter to using Siri to dictate your speech...you can do children's books illustrations, Powerpoint graphics, web design, billboards, magazine covers, anything you want!

 

[Kaido]

Very early test of A.I.-generated video:

https://twitter.com/x/status/1686923299392290816

 

[JEDI]

Very early test of A.I.-generated video:

https://twitter.com/x/status/1686923299392290816

it looks awful!

 

[Red Squirrel]

Honestly that's pretty good knowing that it is very open to refinement. Imagine just feeding a movie script into AI and it basically makes the whole movie. You could essentially program a movie like code if you want very specific things to be seen/happen.

 

[Kaido]

Honestly that's pretty good knowing that it is very open to refinement. Imagine just feeding a movie script into AI and it basically makes the whole movie. You could essentially program a movie like code if you want very specific things to be seen/happen.

They figured out hands in A.I. now, so they'll make progress on the video front too over time:

 

[Kaido]

Augie lets you do voice cloning:

Augie lets you make videos with an AI clone of your own voice

Never read a script again.

www.theverge.com

Aug X Labs | AI video editing

Aug X Labs is an AI-driven video technology and publishing startup, democratizing video creation by transforming words into compelling, relevant, and engaging video content. We help you convert your thoughts into compelling visual creations.

www.augxlabs.com

 

[Kaido]

So to recap NeRF technology, it's basically software that takes static 2D images & converts them into 3D scenes:

AI Artists with NVIDIA Instant NeRF

Compose beautiful scenes With new perspective.

www.nvidia.com

A more detailed explanation: (also read up on synthetic data)

Neural Radiance Field (NeRF): A Gentle Introduction

Understand how Neural Radiance Field (NeRF) works and how you can use it to generate novel images from a 3D scene, based on a partial set of images.

datagen.tech

I've been keeping an eye on this technology as it's been RAPIDLY advancing lately. Luma Labs has now released an A.I. flythrough tool that can takes pictures of your home or office or other space & make a video walkthrough:

Flythroughs by Luma AI

Create AI-generated cinematic videos that look like professional drone captures!

lumalabs.ai

SUPER incredible stuff! This is an A.I.-generated flythrough video based solely on 2D static photos. My wife does real estate photography; I do the video side of things, but don't go along on most jobs because of my IT work, so this is really promising technology for people who work in specifically in 2D photography! (iOS app is available FYI, but the output & pricing stuff appears to have some teething issues)

https://twitter.com/x/status/1692296771161170356

 

[Kaido]

Midjourney released inpainting, which allows you to selectively edit a portion of an image using generative A.I. This is an ENORMOUS step forward for end-users! You can do it by region with button options or enable via text prompts using the remix option. This is pretty cool because:

1. You can use AI to generate an image
2. You can then edit the image to your liking
3. The selective customization pulls on the AI dataset to generate whatever you want

Video demonstration:

https://twitter.com/x/status/1693730802670981166

So you can generate an image like this, but then selectively change a portion of the image, such as the front of the dress, so rather than just being stuck with the AI's "final" output, you now have selective creative control over the finished product:


Which means we can engage in creative iterative development to flesh out ideas:

https://twitter.com/x/status/1693743693402882317

 

[Kaido]

Nope, no thank you:

Using Generative AI to Resurrect the Dead Will Create a Burden for the Living

AI technologies promise more chatbots and replicas of people who have passed. But giving voice to the dead comes at a human cost.

www.wired.com

 

[Kaido]

Dang. RIP copywriters. "Write SEO-optimized and plagiarism-free content for blogs, ads, emails, and a website 10X faster." Uses GPT-4 technology. The toolset is incredible:

Writesonic - AI Content Writer, AI SEO Toolkit & AI Chatbots

Writesonic is the Generative AI platform for crafting factually-accurate, on-brand content with real-time data, optimizing for SEO, and building custom ChatGPT AI chatbots trained on your data.

writesonic.com

Sentence Expander:

Text Expander, Sentence Expander - Expand Sentences For Free

Writesonic’s AI text expander can help you to expand sentences in a click. Try expanding sentence for free!

writesonic.com

Article summarizer:

Text Summarizer | AI Text Summarizer

Summarize your text copy that provides the most important details, in a click with Writesonic’s Text Summarizer. Try for free today!

writesonic.com

Paraphrasing Tool:

"Rewrites That Are Better Than The Original: Writesonic’s paraphrasing tool lets you rewrite sentences, paragraphs, essays as well as entire articles with a single click. You'll receive 100% unique and plagiarism-free content:"

Free AI Paraphrasing Tool Online | Writesonic

Free Paraphrasing tool by Writesonic rewrites texts, emails, and articles instantly. Perfect your writing in moments. experience it for free online!

writesonic.com

 

[Ermitan]

I'm not SEO specialist, but had experience to apply for SEO and copywriting services abt 3 yrs ago. At that time they told me that they used AI tools for text writing. I don't remember what was its name, may be the writesonic, but it was not so popular and nobody told abt that. Perhaps, now it becomes more easy and lesstime to write text, but still good professional copywriter would probably get rarther better and faster results with the same AI tools, than newbie. AI is yet a tool and cannot write good SEO-optimized text without human mastered the tool. It's easy said than done.
At futurepedia now there is abt 5 000 AI services for different application and 180 + out of them are for copywriting. Although, I tried to write one article for blog with ChatGPT. Well, it was written by 20 sec (abt 700 words), but I spent abt 3 hours afterwards to re-write it until I got it "readable" for the readers.

 

[Kaido]

RIP customer service jobs

https://twitter.com/x/status/1680314562753490949

Well that was quick:

DoorDash launches AI-powered voice ordering technology for restaurants

 

[Kaido]

Well that was quick:

DoorDash launches AI-powered voice ordering technology for restaurants

Presto has already launched their A.I. drive-through chatbot at over 350 restaurant locations nationwide:

 

[lxskllr]

I want custom voice packs for my phone. I'd love to get navigational directions from Worf, with the trademark inflection and exasperation when I miss a turn or something.

 

[GodisanAtheist]

I want custom voice packs for my phone. I'd love to get navigational directions from Worf, with the trademark inflection and exasperation when I miss a turn or something.

- "But Captain, there is a faster route that also avoids tolls"

No, I want to go this way.

*SIGH* " Yes Captain, inputting navigation coordinates now"

 

[Kaido]

- "But Captain, there is a faster route that also avoids tolls"

No, I want to go this way.

*SIGH* " Yes Captain, inputting navigation coordinates now"

I'd pay for that lol

 

[Ermitan]

I'd pay for that lol

Me too LOL. Although, guess, the custom voice pack is quite real to get. With the AI tools, like Speechify, Resemable or Murf they can clone the voice. So, guess, very soon Presto would increase sales of the Peach Pie - once they start offering it with the voice of Marilyn Monroe Honestly, I’m also raving about these new horizons opened by AI, especially for making video right from a script, like @Red Squirrel said. Just the refinement and creating the right prompts is the matter and challenge yet. At least looks to me personally. Plus I think there is no yet a single self-sufficient AI tool to replace human in conversations or to make the movie as from a scratch till the end. I read recently abt reasonable ai challenges in video production and I understood that even professionals yet have to use a few AI tools in order to get fine-tuned video. There are so many ads, chat-bots, short promotional videos, explainers etc now made with AI, but still they make it with a few AI services applied. Refining prompts and mastering AI tools are looks to be the next genre of modern tech-art.

 

[Zeze]

These really freak me out.

If capitalism continues to be insanely top heavy while eliminating jobs, we will all be poor.

 

[Kaido]

Refining prompts and mastering AI tools are looks to be the next genre of modern tech-art.

"Prediction: AI prompting will be as important as an Excel certification in the future."

https://twitter.com/x/status/1696516063595290763

 

[pmv]

Augie lets you do voice cloning:

Augie lets you make videos with an AI clone of your own voice

Never read a script again.

www.theverge.com

Where does that leave "voice recognition" security systems?

Seems like a boon for scam artists who want to impersonate people on the phone, as well.

 

[Kaido]

Where does that leave "voice recognition" security systems?

Seems like a boon for scam artists who want to impersonate people on the phone, as well.

 

[Kaido]

AI gets text & logo support from Ideogram:

Ideogram

Helping people become more creative.

ideogram.ai

Pushbutton graphic design. Various fonts in various styles:









Also check out:

Civitai: The Home of Open-Source Generative AI

Explore thousands of high-quality Stable Diffusion & Flux models, share your AI-generated art, and engage with a vibrant community of creators

civitai.com

&

AI Image Generator - Create Art, Images & Video | Leonardo AI

Transform your projects with our AI image generator. Generate high-quality, AI generated images with unparalleled speed and style to elevate your creative vision

leonardo.ai

 

[GodisanAtheist]

AI gets text & logo support from Ideogram:

Ideogram

Helping people become more creative.

ideogram.ai

Pushbutton graphic design. Various fonts in various styles:

View attachment 85043

View attachment 85044

View attachment 85045

View attachment 85046

Also check out:

Civitai: The Home of Open-Source Generative AI

Explore thousands of high-quality Stable Diffusion & Flux models, share your AI-generated art, and engage with a vibrant community of creators

civitai.com

&

AI Image Generator - Create Art, Images & Video | Leonardo AI

Transform your projects with our AI image generator. Generate high-quality, AI generated images with unparalleled speed and style to elevate your creative vision

leonardo.ai

It has some trouble with words/sentences but generates beautiful images



If I ever decided to open a YouTube channel.for video game reviews:

 

[Kaido]

It has some trouble with words/sentences but generates beautiful images

View attachment 85048

I'm excited to see more integrations like spot-editing & generative fill. Blender is already experimenting with rendering-up 3D sketch type of stuff into AI-enhanced imagery. Imagine sketching with a digital tablet to get a baseline and then dragging & dropping in different styles for it to clone, rather than just using prompts. Art generation is getting a HUGE level-up thanks to A.I.!