The Download: generative AI for video, and detecting AI text
This is today’s edition of The Download, our weekday newsletter that provides a daily dose of what’s going on in the world of technology.
The original startup behind Stable Diffusion has launched a generative AI for video
What’s happened: Runway, the generative AI startup that co-created last year’s breakout text-to-image model Stable Diffusion, has released an AI model that can transform existing videos into new ones by applying styles from a text prompt or reference image.
What it does: In a demo reel posted on its website, Runway shows how the model, called Gen-1, can turn people on a street into claymation puppets, and books stacked on a table into a cityscape at night. Other recent text-to-video models can generate very short video clips from scratch, but because Gen-1adapts existing footage it can produce much longer videos.
Why it matters: Last year’s explosion in generative AI was fueled by the millions of people who got their hands on powerful creative tools for the first time and shared what they made, and Runway hopes Gen-1 will have a similar effect on generated videos. Read the full story.
—Will Douglas Heaven
Why detecting AI-generated text is so difficult (and what to do about it)
Last week, OpenAI unveiled a tool that can detect text produced by its AI system ChatGPT. But if you’re a teacher who fears the coming deluge of ChatGPT-generated essays, don’t get too excited.
The tool is still very much a work in progress, and it is woefully unreliable, only identifying 26% of AI-written text correctly as “likely AI-written” 26% of the time. Even while we should expect this number to improve, we’re extremely unlikely to ever get a tool that can spot AI-generated text with 100% certainty. Read the full story.
Melissa’s story is from The Algorithm, her weekly AI newsletter. Sign up to receive it in your inbox every Monday.
I’ve combed the internet to find you today’s most fun/important/scary/fascinating stories about technology.
1 Google has announced its own chatbot to rival ChatGPT
Bard won’t be available to the public for another few weeks. (The Verge)
+ It’s called Bard because it’s a storyteller, geddit? (NYT $)
+ AI is kindling the long-dormant search wars flame. (FT $)
+ Only ‘trusted testers’ have access to the system right now. (Wired $)
2 China’s ‘spy balloon’ furore has undermined Xi Jinping’s leadership
It certainly isn’t helping his ongoing efforts to stabilize tense relations with the US. (FT $)
+ The balloon was 200ft tall and carrying a huge load. (BBC)
3 We can’t really predict earthquakes
Even the best efforts aren’t able to provide much more than a few seconds’ warning.(WP $)
+ Rescuers in Turkey and Syria are working around the clock to find survivors. (FT $)
+ A deep-learning algorithm could detect earthquakes by filtering out city noise. (MIT Technology Review)
4 Andrew Tate groomed women into joining his webcam sex ring
Victims were approached on social media and dating apps. (Vice)
+ The “manosphere” is getting more toxic as angry men join the incels. (MIT Technology Review)
5 What it’s like to run an abortion hotline
Post-Roe, its volunteers are dealing with more calls than ever. (Vox) + The cognitive dissonance of watching the end of Roe unfold online. (MIT Technology Review)
6 This viral TikTok drug challenge never actually existed
But that hasn’t stopped panic from spreading across Mexico. (Rest of World)
+ The porcelain challenge didn’t need to be real to get views. (MIT Technology Review)
7 We’re still learning about the moon
NASA ShadowCam mission hopes to shed some light on its most mysterious craters. (The Atlantic $)
8 Austin has become a refuge for tech workers left jaded by Silicon Valley
The city’s cool, counter-culture vibe is a massive draw.(New Yorker $)
9 How EV batteries are made in America
But some components still need to be sourced from overseas. (WSJ $)+ How old batteries will help power tomorrow’s EVs. (MIT Technology Review)
10 The tricky ethics of de-aging actors with AI
Some critics argue it’s both demeaning and pointless. (The Guardian)
+ AI has learnt how to crush humans at Pokémon. (The Atlantic $)
Quote of the day
“People are afraid to have conversations.”
—Mary Jane Copps, a former journalist who coaches people on how to speak to others over the phone, tells Bloomberg why her customers are so reluctant to talk.
The big story
How big technology systems are slowing innovation
In 2005, years before Apple’s Siri and Amazon’s Alexa came on the scene, two startups—ScanSoft and Nuance Communications—merged to pursue a burgeoning opportunity in speech recognition. The new company developed powerful speech-processing software and grew rapidly for almost a decade. Then suddenly, around 2014, it stopped growing.
Nuance’s story is far from unique. In all major industries and technology domains, startups are facing unprecedented obstacles. They are growing much more slowly than comparable companies did in the past. And it will take not only strong antitrust enforcement to reverse the trend, but a fundamental loosening of restrictions like non-compete agreements and intellectual property rights. Read the full story.
We can still have nice things
A place for comfort, fun and distraction in these weird times. (Got any ideas? Drop me a line or tweet ’em at me.)
+ These paintings (and the odd photo) of London are magical.
+ Now this is a chatbot I can get onboard with—Cat GPT (thanks Charlotte! Top tip: meow at it)
+ If you’re worried about your attention span, you should feel safe in the knowledge that even ancient Roman philosophers struggled to concentrate sometimes.
+ Happy quarter-century to that sinister dancing baby.