The Download: spectacular new AI capabilities

0
535
The Download: spectacular new AI capabilities


This is in the present day’s version of The Download, our weekday publication that gives a every day dose of what’s happening on this planet of know-how.

OpenAI teases a tremendous new generative video mannequin referred to as Sora

OpenAI has constructed a putting new generative video mannequin referred to as Sora that may take a brief textual content description and switch it into an in depth, high-definition movie clip as much as a minute lengthy. It’s severely impressive-looking. 

Based on 4 pattern movies that OpenAI shared with MIT Technology Review,  the agency has pushed the envelope of what’s attainable with text-to-video technology (a sizzling new analysis course that we flagged as a pattern to look at in 2024).

It’s laborious to know precisely how spectacular a step that is till we get extra data from OpenAI—and we could have a wait on our palms. The firm has no plans to launch it to the general public presently, although it does hope to in future. For now, conscious of the potential for misuse, OpenAI might be doing intensive security testing. Read the complete story—and take a look at a few of the movies! 

—Will Douglas Heaven

Google’s new model of Gemini can deal with far larger quantities of information

The information: Google DeepMind has launched the following technology of its highly effective artificial-intelligence mannequin Gemini, which has an enhanced means to work with giant quantities of video, textual content, and pictures.

For instance: In one demonstration video proven by Google, the mannequin was fed the 402-page transcript of the Apollo moon touchdown mission. Then they confirmed Gemini a hand-drawn sketch of a boot, and requested it to establish the second within the transcript that the drawing represents. The mannequin was additionally in a position to establish moments of humor. 

What it means: These types of AI capabilities are very spectacular, Oren Etzioni, former technical director of the Allen Institute for Artificial Intelligence, instructed us. However, he did give one main caveat: “Never trust an AI demo.” Read the complete story. 

—James O’Donnell

How bacteria-fighting viruses may go mainstream

Lynn Cole had a blood an infection she couldn’t shake. For years, she was out and in of the hospital. Each time antibiotics would power the an infection to retreat. Each time it got here roaring again.

In the summer season of 2020, the micro organism flooding Cole’s bloodstream stopped responding to antibiotics. She was working out of time. Her medical doctors determined they needed to attempt a special method: phages, that are tiny viruses that infect and destroy micro organism.

The phages labored. Cole recovered with exceptional velocity. But then the remedy failed. Cole’s case highlights the big promise of phage remedy, nevertheless it additionally reveals simply how a lot we’ve to be taught. Read the complete story

—Cassandra Willyard

This story is from The Checkup, our weekly publication all about biotech and well being. Sign up to obtain it in your inbox each Thursday.

The must-reads

I’ve combed the web to seek out you in the present day’s most enjoyable/necessary/scary/fascinating tales about know-how.

1 The Senate may very well be about to move some main tech regulation
If it passes, the Kids Online Safety Act would be the largest piece of tech regulation within the US in many years. (WP $)
Why baby security payments are popping up all around the US. (MIT Technology Review)
New York City is suing TikTok and Instagram for ‘addicting’ children. (NBC
 
2 A brand new period of deepfakes is colliding with pivotal elections 
And it’ll be very laborious to determine how massive an impression AI-generated content material has on outcomes, even after the actual fact. (WSJ $)
A Chinese affect marketing campaign is utilizing AI-generated content material to amplify division within the US. (NYT $)
 
3 TikTok has launched an app for the Vision Pro
YouTube says it’s constructing an app for the headset too. (The Verge)
 
4 AI is nothing to worry for white collar employees
That’s as a result of it’s not likely an alternative to experience—it’s a lever for its software.  (Noema)
People are fearful that AI will take everybody’s jobs. We’ve been right here earlier than. (MIT Technology Review)
Here’s how AI is shaking up the best way we work. (The Verge)
 
5 What it’s wish to be a content material moderator in Pakistan
Pretty soul-crushing—and with little hope of a promotion or transferable expertise. (Rest of World)
 
6 Hardware nonetheless issues
In reality, within the AI period, it’s about as necessary because it’s ever been. (FT $)
 
7 Discredited well being claims are getting a second airing on TikTok
It’s giving new life to lectures by a girl completely banned from offering well being companies in Australia. (Vox)
 
8 Electric autos aren’t nice at dealing with excessive warmth
But they might get higher, because of new supplies. (Scientific American $)
Tesla’s stainless-steel Cybertrucks are already rusting. (Futurism)
 
9 Meat-injected rice, anybody? 🍚🥩
I’ve some severe beef with this new foodstuff (sorry, sorry.) (CNN)
 
10 Some younger folks need landlines ☎
There are some upsides to having a landline telephone, however they’re disappearing quick. (The Guardian)

Quote of the day

“Critics are trying to write our obituary and are working to ban our industry in its infancy.”

—Uma Valeti, CEO of cultivated meat firm Upside Foods, responds to criticism and pushback in direction of his trade from lawmakers, Wired reviews.

The massive story

This scientist is attempting to create an accessible, unhackable voting machine

""

ALICIA FERNáNDEZ

November 2022

For the previous 19 years, laptop science professor Juan Gilbert has immersed himself in maybe essentially the most contentious debate over election administration within the United States—what function, if any, touch-screen ballot-marking gadgets ought to play within the voting course of.

While advocates declare that digital voting programs may be comparatively safe, enhance accessibility, and simplify voting and vote tallying, critics have argued that they’re insecure and must be used as occasionally as attainable. 

As for Gilbert? He claims he’s lastly invented “the most secure voting technology ever created.” And he’s invited a number of of essentially the most revered and vocal critics of voting know-how to show his level. Read the complete story.

—Spencer Mestel

We can nonetheless have good issues

A spot for consolation, enjoyable and distraction in these bizarre instances. (Got any concepts? Drop me a line or tweet ’em at me.)

+ It’s Friday! These recipes can turn out to be both cocktails or mocktails, if that’s extra your factor.
+ Why ‘Wonderboy’ was John Lennon’s favourite Kinks observe. 
+ Reckon I just about had this actual chat with my Dad once I was a child. 
+ Soup doesn’t should be a faff, as this recipe reveals.
+ Turns out extra money actually can imply extra issues
+ Some bizarre and wacky methods folks with busy jobs chill out.



LEAVE A REPLY

Please enter your comment!
Please enter your name here