Hello, and welcome to TechScape. In this week’s edition: 23andMe files for bankruptcy, Nvidia forecasts a fusion of AI and robotics, and AI enables the creation of fiction at the pace of social media.
Genetic testing firm 23andMe filed for bankruptcy on Monday. The CEO and co-founder Anne Wojcicki has stepped down after several attempts at a buyout. Once valued as high as $5.8bn in 2021, the company’s financial failure is the finale to a long decline.
***
My colleague Julia Kollewe reports:
23andMe said late on Sunday that it had started voluntary Chapter 11 proceedings in the US Bankruptcy Court for the Eastern District of Missouri to “facilitate a sale process to maximise the value of its business”.
The loss-making company, which provides saliva-based test kits to customers to help them track their ancestry, added that it was operating as usual throughout the sale process. “There are no changes to the way the company stores, manages, or protects customer data,” it said.
I understand the urge to assure customers that there is “no change” to business as usual at 23andMe, but the company’s statement bears an unfortunate implication. In late 2023, the company disclosed that hackers had gained access to the personal data of 7 million customers, including their genomes. Not long after the incursion, hackers offered to sell the names, addresses and genetic heritage belonging to 1 million 23andMe customers with Ashkenazi Jewish heritage on a shadowy dark web forum. Though the hack did not only target Jewish customers, the proposed sale gave a grim example of what malicious denizens of net could do with 23andMe customers’ information.
One 23andMe participant, a man in Florida who discovered Ashkenazi Jewish heritage in his test, summed up the imbalance of the trade-off: “I didn’t know my family was going to potentially be a target. I may have put my family and myself in danger for something I did out of curiosity more than anything.”
The question for 23andMe customers is what will happen to the trove of genetic data that 23andMe has amassed in its years of collecting spit in tubes.
The ultimate promise of 23andMe – medicine personalized based on your unique genetic code – has not yet come to fruition. In the meantime, knowing the exact breakdown of your genetic ancestry is more novelty than medical necessity, and it’s not good business. Sometimes the test just reveals that you’re British, which left at least one Guardian writer nonplussed, and you find yourself having given your DNA away.
The question for 23andMe customers is what will happen to the trove of genetic data that 23andMe has amassed in its years of collecting spit in tubes. Over the weekend the California attorney general, Rob Bonta, urged the company’s users to ask it to “delete your data and destroy any samples of genetic material held by the company”, as is their right under the state’s law.
More on 23andMe
The stuff of science fiction
Jensen Huang, the Nvidia CEO, interacts with a small robot on stage during the keynote at the SAP Center in San Jose, California, on 18 March. Photograph: Brittany Hosea-Small/Reuters
A series of convergent developments in tech last week have my nose pointed at the future like a hunting dog.
Nvidia hosted its developer conference in San Jose, California, announcing new and more powerful chips that will offer greater computing capacity to artificial intelligence. The AI business, if Chinese model DeepSeek serves as a bellwether, is learning to maximize the results it draws from that computing power.
As in any good science fiction blockbuster, a lovable side character made an appearance at the conference. A Star Wars-inspired droid named Blue waddled onstage alongside Jensen Huang, the Nvidia CEO, during his keynote to say hello. Disney partnered with Nvidia to design and showcase the new bot, which holds in its brain software for modeling and processing the physics of its surroundings. Nvidia also announced an AI meant for robots, which likewise takes its name from a Disney franchise, Groot N1.
Nvidia’s announcements come as various AI companies make their first public forays into agentic models, which can take on tasks for you. Per early reviews, these products are not very adept yet.
But excitement i for the advent of artificial general intelligence (AGI) is growing, and real preparations are happening. More and more people who aren’t AI company CEOs anticipate the arrival of this powerful and versatile technology soon. Joe Biden’s top AI adviser, Ben Buchanan, gave an interview at the start of this month about how the US had planned for the widespread arrival of AGI under the previous administration. Soon after, a Times tech columnist wrote about why he’s come to believe the AGI hype.
An agentic AI with the capabilities of AGI plugged into the brain of a robot – baby, that’s a bona fide humanoid, and it’s a possibility that’s becoming easier to imagine even without the help of Isaac Asimov.
skip past newsletter promotion
Sign up to TechScape
A weekly dive in to how technology is shaping our lives
Privacy Notice: Newsletters may contain info about charities, online ads, and content funded by outside parties. For more information see our Privacy Policy. We use Google reCaptcha to protect our website and the Google Privacy Policy and Terms of Service apply.
after newsletter promotion
Tesla is struggling while Elon Musk mucks about in the White House
New entertainment: Using AI to illustrate short-form fiction on Instagram
Photograph: The Guardian
This week on my iPhone, I’m scrolling through the videos of @HolyFool36 on Instagram.
Created by a 26-year-old from Long Island named Dylan (he declined to give his last name), the account posts charming, retro and lightly spooky videos daily. They’re usually 90 seconds long. Dylan said he was inspired to create the videos by works of dark fantasy (Clark Ashton Smith), Elden Ring and other Dark Souls games, and analog horror videos on YouTube. Far from AI slop, the videos offer clever tidbits of the absurd and compelling stories in the form of occult instructions. The human touch is evident, though AI serves as the means of production. I enjoy them.
“I do the writing myself because I was born with the faculties to do that. I use AI to make the images because I don’t have those faculties. It’s a means to an end,” he said.
The account has amassed more than half a million followers since launching in the first half of 2024. It earns money via TikTok ads and merch sales, according to Dylan, but he’s kept his full-time job in tech.
“I went from a hobbyist to a niche internet micro celebrity!” he remarked. His fiancee has started an AI art page as well.
Dylan’s creative process involves multiple AI tools. He asks Dall-E to make the first draft of the picture in his head then runs the result through Midjourney to give it the retro video game sheen. If the story he’s writing requires animation, he uses Kling, though most of Holy Fool’s videos consist of collages of still images. All of his material features the same background music, a simple electronic synth melody, and the same narrating voice, which he generated and customized with ElevenLabs.
Artists across the US and UK have spoken out by the hundreds against the use of AI in the arts and what they see as theft by tech giants skirting intellectual property law. Their point is a fair one. Just this past week, the Atlantic created a way for authors to search LibGen, a database of pirated books, for their work. Many found their books there. Meta employees allegedly downloaded the database from peer-to-peer file sharing networks, a matter currently at issue in a copyright suit against the company over how it created its AI model Llama, specifically concerning whether the chatbot was trained on copyrighted material. The messages between Meta staff revealed in discovery about downloading the database are damning. Employees said licensing authors’ copyrighted work would be too expensive and slow, so they turned to more shadowy means of accessing mammoth amounts of texts and received Mark Zuckerberg’s personal approval to do so, according to court documents. Meta is worth $1.51tn, and Mark Zuckerberg’s personal fortune weighs in at $202bn.
However, none of these artists and authors are doing what the Holy Fool is. None are populating a fictional universe with daily short-form videos. Why would they? To create content at the speed that an Instagram feed without AI would be a full-time job. For many influencers, it is, but those video creators use their own faces and make videos about their real lives. Animating fictional videos so quickly and posting them for free is unlikely to give a worthwhile return on investment. All that is to say – there seems to be no labor lost in the creation of these kinds of videos, no artist who would otherwise be earning a living but has been replaced by AI. The creation of new kinds of serialized fiction seems like a positive use case for AI to me.
The wider TechScape
