OpenAI just gave artists access to Sora and proved the AI video tool is weirder and more powerful than we thought

A man with a balloon for a head is somehow not the weirdest thing you'll see today thanks to a series of experimental video clips made by seven artists using OpenAI's Sora generative video creation platform.

Unlike OpenAI's ChatGPT AI chatbot and the DALL-E image generation platform, the company's text-to-video tool still isn't publicly available. However, on Monday, OpenAI revealed it had given Sora access to “visual artists, designers, creative directors, and filmmakers” and revealed their efforts in a “first impressions” blog post.

While all of the films ranging in length from 20 seconds to a minute-and-a-half are visually stunning, most are what you might describe as abstract. OpenAI's Artist In Residence Alex Reben's 20-second film is an exploration of what could very well be some of his sculptures (or at least concepts for them), and creative director Josephine Miller's video depicts models melded with what looks like translucent stained glass.

Not all the videos are so esoteric.

OpenAI Sora AI-generated video image by Don Allen Stevenson III

OpenAI Sora AI-generated video image by Don Allen Stevenson III (Image credit: OpenAI sora / Don Allen Stevenson III)

If we had to give out an award for most entertaining, it might be multimedia production company shy kids' “Air Head”. It's an on-the-nose short film about a man whose head is a hot-air-filled yellow balloon. It might remind you of an AI-twisted version of the classic film, The Red Balloon, although only if you expected the boy to grow up and marry the red balloon and…never mind.

Sora's ability to convincingly merge the fantastical balloon head with what looks like a human body and a realistic environment is stunning. As shy kids' Walter Woodman noted, “As great as Sora is at generating things that appear real, what excites us is its ability to make things that are totally surreal.” And yes, it's a funny and extremely surreal little movie.

But wait, it gets stranger.

The other video that will have you waking up in the middle of the night is digital artist Don Allen Stevenson III's “Beyond Our Reality,” which is like a twisted National Geographic nature film depicting never-before-seen animal mergings like the Girafflamingo, flying pigs, and the Eel Cat. Each one looks as if a mad scientist grabbed disparate animals, carved them up, and then perfectly melded them to create these new chimeras.

OpenAI and the artists never detail the prompts used to generate the videos, nor the effort it took to get from the idea to the final video. Did they all simply type in a paragraph describing the scene, style, and level of reality and hit enter, or was this an iterative process that somehow got them to the point where the man's balloon head somehow perfectly met his shoulders or the Bunny Armadillo transformed from grotesque to the final, cute product?

That OpenAI has invited creatives to take Sora for a test run is not surprising. It's their livelihoods in art, film, and animation that are most at risk from Sora's already impressive capabilities. Most seem convinced it's a tool that can help them more quickly develop finished commercial products.

“The ability to rapidly conceptualize at such a high level of quality is not only challenging my creative process but also helping me evolve in storytelling. It's enabling me to translate my imagination with fewer technical constraints,” said Josephine Miller in the blog post.

Go watch the clips but don't blame us if you wake up in the middle of the night screaming.

You might also like

TechRadar – All the latest technology news

Read More

Microsoft just gave Windows Copilot a ChaGPT-4 boost and the ability to explain screenshots

Microsoft came out hot with its Windows Copilot 365 and Copilot for Windows announcements last year, and presented Copilot as a general virtual assistant to help with your whole digital life. So far, we’re yet to see Copilot reach its potential, but it looks like we’re one step closer: Microsoft is reportedly gearing up to add a powerful new feature that will allow users to take a screenshot, submit it to Copilot, and ask Copilot to explain what’s in the screenshot. 

As far as we know, an “add a screenshot” button is rolling out to the general public – meaning you may already be able to try it. If not, you should be able to very soon. This button should appear in the Copilot panel, prompting you to select a part of the screen, confirming that you’ve captured what you want to discuss with Copilot, and then upload it to the Copilot or Bing right-side panel. When I tried it, I did also need to submit some text to go along with it, such as a question or additional context. 

A screenshot in a Microsoft Edge window with a Copilot panel open on the right, with arrow pointing to new

(Image credit: Future)

Once the screenshot is uploaded, you can talk about it and ask about anything within it or relating to it with Bing Chat or Copilot on Windows 11

As you can see in the screenshot, the new button sits next to the existing image upload button. You can  try the new feature at Microsoft’s Copilot website, or over at Bing Chat

Windows Latest has their own demonstration of this new feature, and I had a go of my own. First, I took a screenshot selection which showed a full description of video creation platform HeyGen's YouTube channel and asked it to tell me two things: whose channel it is and what it’s about.

Copilot returned this: 

This is the YouTube channel of HeyGen. It is a next-gen video creation platform that turns text into professional spokesperson videos in minutes. They offer premium avatars speaking in multiple languages and professional video templates for various use cases including marketing, e-learning, and corporate communication. You can find more information about HeyGen on their website.

This is pretty accurate, and reminded me of one feature in particular that I really like about Bing Chat and Copilot – they readily and very visibly provide sources and websites that you can visit to double check the information. 

Once you make a selection of your screen, you can make markings on it and draw on it. You can also add specific instructional visuals to help Copilot understand your query, and you can move your selection window around to a different part of the screen altogether. 

According to Windows Latest, Bing Chat recently got a ChatGPT-4 boost granting it a new level of functionality and this is likely making its way into Copilot as well. Apparently access is currently only granted to select users, and this development enables Copilot to engage in conversations about emotions. Currently, there is a limited pool of users who can try this for themselves and access is seemingly given at random, and it will be available to all who access Windows Copilot and Bing Chat very shortly.

Microsoft Bing logo on a white smartphone screen

(Image credit: Shutterstock / Primakov)

Microsoft charts a course ahead with Copilot

Microsoft has been pretty definitive in its messaging that Copilot is a big deal for the company, and will be a central feature in several products like Microsoft 365 and Windows, but not just those. 

In a pretty major (yet not terribly surprising) development, Microsoft is planning to add an actual physical Copilot button into the hardware of newly manufactured products as early as 2024. Microsoft is doing this in its continuing effort to make computing, especially AI-powered computing, simpler and more seamless for users. This was detailed and confirmed in a recent Windows Experience Blog post written by Yusuf Mehdi, Executive Vice President and Consumer Chief Marketing Officer at Microsoft.

For the rest of us not ready to throw our older Windows devices out quite yet for this new button, you can bring up Windows Copilot with the shortcut Win+C (if you have updated your Windows 11 version to one that has Windows Copilot included). 

According to Microsoft itself, the introduction of the Copilot key will be the most notable upgrade to the Windows keyboard in almost thirty years. It likens this future introduction to the addition of the Windows Start key, which is putting a lot of faith in Copilot itself so I imagine we’ll continue to see major developments to Copilot throughout this year. I think especially with Copilot’s development, Microsoft is one of the most exciting companies to watch this year. 

YOU MIGHT ALSO LIKE

TechRadar – All the latest technology news

Read More

Brave, DuckDuckGo just gave you another way to flip Google the middle finger

Brave has announced that its web browser will now allow users to bypass AMP pages hosted by Google, which it claims are harmful to both privacy and the state of the web.

The new De-AMP feature will instead funnel web users to content hosted directly on the publisher’s website, minimizing the opportunity for additional tracking and meddling to take place.

Not to be outdone, rival privacy software company DuckDuckGo rushed to Twitter to reveal that its apps and extensions now offer similar functionality, but the specifics of the implementation are not yet clear.

Google’s AMP troubles

Rolled out in 2015, AMP (short for accelerated mobile pages) is a system whereby stripped-back versions of trending web pages are preloaded and served up via Google servers.

When AMP was first announced, Google said it beleived the system would help ensure rich web content such as video and animation would load rapidly and behave consistently across all platforms, thereby improving the web experience.

However, the scheme has come under criticism from publishers and privacy advocates alike, who say AMP gives Google yet more signals to gobble up in support of its digital advertising business, creates confusion as to the source of information and forces publishers to build their websites to Google’s desired spec.

“AMP harms users’ privacy, security and internet experience, and just as bad, AMP helps Google further monopolize and control the direction of the web,” wrote Brave, in a blog post.

And in a Twitter thread, DuckDuckGo presented a similar justification for its decision to move against the initiative.

“AMP technology is bad for privacy because it enables Google to track users even more,” said the firm. “And Google uses AMP to further entrench its monopoly, forcing the technology on publishers by prioritizing AMP links in search and favoring Google ads on AMP pages.”

Since the launch of AMP, a number of publishers (including Future plc., parent to TechRadar Pro) have abandoned the system. And now, browser vendors like Brave and DuckDuckGo are coming out with their own tools to help web users bypass AMP altogether.

“Where possible, De-AMP will rewrite links and URLs to prevent users from visiting AMP pages altogether,” explained Brave. “And in cases where that is not possible, Brave will watch as pages are being fetched and redirect users away from AMP pages before the page is even rendered, preventing AMP/Google code from being loaded and executed.”

Brave’s De-AMP feature is now available in both Nightly and Beta versions of its browser and will be enabled by default in the next full public release. TechRadar Pro is awaiting further specifics about DuckDuckGo’s efforts.

TechRadar – All the latest technology news

Read More

Brave, DuckDuckGo just gave you another way to flip Google the middle finger

Brave has announced that its web browser will now allow users to bypass AMP pages hosted by Google, which it claims are harmful to both privacy and the state of the web.

The new De-AMP feature will instead funnel web users to content hosted directly on the publisher’s website, minimizing the opportunity for additional tracking and meddling to take place.

Not to be outdone, rival privacy software company DuckDuckGo rushed to Twitter to reveal that its apps and extensions now offer similar functionality, but the specifics of the implementation are not yet clear.

Google’s AMP troubles

Rolled out in 2015, AMP (short for accelerated mobile pages) is a system whereby stripped-back versions of trending web pages are preloaded and served up via Google servers.

When AMP was first announced, Google said it beleived the system would help ensure rich web content such as video and animation would load rapidly and behave consistently across all platforms, thereby improving the web experience.

However, the scheme has come under criticism from publishers and privacy advocates alike, who say AMP gives Google yet more signals to gobble up in support of its digital advertising business, creates confusion as to the source of information and forces publishers to build their websites to Google’s desired spec.

“AMP harms users’ privacy, security and internet experience, and just as bad, AMP helps Google further monopolize and control the direction of the web,” wrote Brave, in a blog post.

And in a Twitter thread, DuckDuckGo presented a similar justification for its decision to move against the initiative.

“AMP technology is bad for privacy because it enables Google to track users even more,” said the firm. “And Google uses AMP to further entrench its monopoly, forcing the technology on publishers by prioritizing AMP links in search and favoring Google ads on AMP pages.”

Since the launch of AMP, a number of publishers (including Future plc., parent to TechRadar Pro) have abandoned the system. And now, browser vendors like Brave and DuckDuckGo are coming out with their own tools to help web users bypass AMP altogether.

“Where possible, De-AMP will rewrite links and URLs to prevent users from visiting AMP pages altogether,” explained Brave. “And in cases where that is not possible, Brave will watch as pages are being fetched and redirect users away from AMP pages before the page is even rendered, preventing AMP/Google code from being loaded and executed.”

Brave’s De-AMP feature is now available in both Nightly and Beta versions of its browser and will be enabled by default in the next full public release. TechRadar Pro is awaiting further specifics about DuckDuckGo’s efforts.

TechRadar – All the latest technology news

Read More

He gave us the GIF: Stephen Wilhite has died at 74

 Internet technology pioneer Stephen Wilhite passed away on March 14 from COVID-related complications. He was 74 and he leaves behind an incredible legacy, the GIF, a game-changer for the blossoming World Wide Web in the 1990s.

Wilhite, who was interested in compression technology, created the GIF at his home in 1987. “I saw the format in my head and then started programming,” he told the New York Times. He then brought the technology to his job at CompuServe, the first major Internet service provider in the US, where he made finalizing tweaks.

In addition to his passion for technology, post-retirement Wilhite was an avid outdoorsman and enjoyed building model trains in his basement.

GIFs made the Internet

The GIF or Graphics Interchange Format has been a massive component of the Internet since its inception. TechRadar’s US Editor-in-Chief, Lance Ulanoff noted in 2016 that “for webmasters in the 1990’s, GIFs were as crucial to the site-building process as HTML….if HTML was the skeleton of our websites, then GIFs were the skin and blood.”

Per his obituary, Wilhite received a Webby Lifetime Achievement Award in New York in 2013 for his invention, during which he reiterated the correct pronunciation of GIF via, naturally, a GIF that simply stated: “it’s pronounced ‘JIF’, not ‘GIF’”. The crowd roared with excitement in response as Wilhite walked wordlessly off the stage.

In further response to the never-ending debate about the pronunciation, a somewhat annoyed Wilhite told the New York Times “The Oxford Dictionary accepts both pronunciations. They are wrong. It is a soft ‘G,’ pronounced ‘JIF’. End of story.”

While today you may associate GIFs with short animations that you see in memes or send in group chats to your friends, the early days of the format were much smaller in scale, usually consisting of just a few low-resolution frames at a time or even single-pixel spacers to help prop up complicated HTML designs.

From humble beginnings to humor

Being the building blocks of the early Internet, it’s almost surprising to observe how they're used today for comedy purposes, displaying fan-made excerpts of classic shows like Friends, or even in some cases, severely compressed, yet full episodes of Spongebob.

GIFs like those blew up on Tumblr and Reddit in their early days despite being regarded as jokes before the heydays of the sites; “No serious web developer or artist would use GIFs,” noted Lance.

These days, we all look upon GIF endearingly; they have withstood the test of time and prospered despite new technologies sprouting around them. Platforms like Giphy were created by people who love the format, and others like Tenor have followed in its wake.

GIFs are a critical component of the Internet that will likely never be replaced, so may Wilhite rest in peace knowing that the Internet will always celebrate his invention.

While the correct pronunciation may be dismaying for some, a commenter on The GIF Pronunciation Page says that the pronunciation may be an homage to peanut butter being “one of the principal three programmer foods”, with the other two being Doritos and Pepsi.

TechRadar – All the latest technology news

Read More

He gave us the GIF: Stephen Wilhite has died at 74

 Internet technology pioneer Stephen Wilhite passed away on March 14 from COVID-related complications. He was 74 and he leaves behind an incredible legacy, the GIF, a game-changer for the blossoming World Wide Web in the 1990s.

Wilhite, who was interested in compression technology, created the GIF at his home in 1987. “I saw the format in my head and then started programming,” he told the New York Times. He then brought the technology to his job at CompuServe, the first major Internet service provider in the US, where he made finalizing tweaks.

In addition to his passion for technology, post-retirement Wilhite was an avid outdoorsman and enjoyed building model trains in his basement.

GIFs made the Internet

The GIF or Graphics Interchange Format has been a massive component of the Internet since its inception. TechRadar’s US Editor-in-Chief, Lance Ulanoff noted in 2016 that “for webmasters in the 1990’s, GIFs were as crucial to the site-building process as HTML….if HTML was the skeleton of our websites, then GIFs were the skin and blood.”

Per his obituary, Wilhite received a Webby Lifetime Achievement Award in New York in 2013 for his invention, during which he reiterated the correct pronunciation of GIF via, naturally, a GIF that simply stated: “it’s pronounced ‘JIF’, not ‘GIF’”. The crowd roared with excitement in response as Wilhite walked wordlessly off the stage.

In further response to the never-ending debate about the pronunciation, a somewhat annoyed Wilhite told the New York Times “The Oxford Dictionary accepts both pronunciations. They are wrong. It is a soft ‘G,’ pronounced ‘JIF’. End of story.”

While today you may associate GIFs with short animations that you see in memes or send in group chats to your friends, the early days of the format were much smaller in scale, usually consisting of just a few low-resolution frames at a time or even single-pixel spacers to help prop up complicated HTML designs.

From humble beginnings to humor

Being the building blocks of the early Internet, it’s almost surprising to observe how they're used today for comedy purposes, displaying fan-made excerpts of classic shows like Friends, or even in some cases, severely compressed, yet full episodes of Spongebob.

GIFs like those blew up on Tumblr and Reddit in their early days despite being regarded as jokes before the heydays of the sites; “No serious web developer or artist would use GIFs,” noted Lance.

These days, we all look upon GIF endearingly; they have withstood the test of time and prospered despite new technologies sprouting around them. Platforms like Giphy were created by people who love the format, and others like Tenor have followed in its wake.

GIFs are a critical component of the Internet that will likely never be replaced, so may Wilhite rest in peace knowing that the Internet will always celebrate his invention.

While the correct pronunciation may be dismaying for some, a commenter on The GIF Pronunciation Page says that the pronunciation may be an homage to peanut butter being “one of the principal three programmer foods”, with the other two being Doritos and Pepsi.

TechRadar – All the latest technology news

Read More