• D.C.
  • BXL
  • Lagos
  • Dubai
  • Beijing
  • SG
rotating globe
  • D.C.
  • BXL
  • Lagos
Semafor Logo
  • Dubai
  • Beijing
  • SG


In today’s edition, we look at how contests funded by the rich are driving AI advances that too weir͏‌  ͏‌  ͏‌  ͏‌  ͏‌  ͏‌ 
 
rotating globe
June 28, 2024
semafor

Technology

Technology
Sign up for our free newsletters
 
Reed Albergotti
Reed Albergotti

Hi, and welcome back to Semafor Tech.

The biggest stars in the tech industry often have unconventional trajectories. Many drop out of college and start companies. Security experts begin their careers as ethically questionable hackers or bug bounty hunters.

Katyanna has picked up on a new way young hot shots in tech are making a name for themselves: by entering AI coding competitions funded by wealthy donors eager to advance their pet interests.

I first wrote about one of these contests — the Vesuvius Challenge — last year. Entrepreneurs Nat Friedman and Daniel Gross were offering a grand prize of $700,000 to anyone who could use AI to read text on scrolls that were essentially burned and buried 2,000 years ago.

That project turned out to be a resounding success and it has since inspired others to try similar things. I won’t spoil some of the fun examples Katyanna found. You’ll have to read below for that.

I think we’ll see more of these contests pop up in coming years and they’ll be an increasingly common way to break into the industry. They combine a hackathon energy with something that resembles academic grants. But there’s no red tape or bureaucracy involved. Just put up some money and start a contest.

Plus, Semafor’s Marta Biino has a scoop on TikTok’s AI plans.

Move Fast/Break Things
Time Magazine

➚ MOVE FAST: Getting paid. Time, owned by Salesforce co-founder Marc Benioff, signed a licensing deal with OpenAI that gives the ChatGPT maker access to the magazine’s archives, which it can use to train AI models. Flush with cash, OpenAI seems willing to pay any major media company in similar arrangements that probably won’t be around for much longer.

➘ BREAK THINGS: Paying retainers. The Center for Investigative Reporting became the latest media company forgoing big paydays in favor of litigation. CIR accused OpenAI and Microsoft of using its proprietary information to train its AI models without asking. But payments are more useful than legal bills in the often cash-strapped media industry.

PostEmail
Artificial Flavor
Agility Robotics

Your SPANX may now arrive via a humanoid robot. Agility Robotics just signed a long-term deal with GXO Logistics to provide the robots in a facility that handles deliveries of the women’s shapewear and clothing.

The deal marks the first time a humanoid robot has been deployed commercially, Agility said in a statement. The company has been developing its robot, called “Digit,” since 2015. Originally just a pair of legs, Digit has gained more human-like parts and can now grab bins with its hands and walk around like a person.

Oregon-based Agility didn’t say how many robots it will be renting out to GXO. But the company is likely starting small and will expand as they take on more capabilities. What makes Digit interesting is that it can be used for many different purposes. AI labs all over the world are now experimenting with using generative models to create these robots. Digit uses a mixture of different methods.

Its deployment marks an important milestone. We’ve officially entered the era of general-purpose robotics. It will take some time before households have Jetsons-like robot maids, but that is the direction in which we’re headed.

PostEmail
Katyanna Quach

Inside the wild new world of AI prizes

THE SCENE

A British financier will pay millions to anyone who finds a way for humans to communicate with animals as part of a new generation of AI prizes that represent a fresh incentive to push technology forward.

Jeremy Coller, who made his money developing a way for a private equity fund’s backers to sell interests or assets to another investor, is offering $10 million to whoever can pass a kind of modified, interspecies Turing Test. The now archaic test examines whether machines can successfully imitate humans and trick them into believing that they’re chatting with another person.

The goal of the Coller Dolittle Challenge, instead, is to test whether humans can mimic animals and trick them into talking with humans. AI algorithms have helped analyze bat calls and whale songs, leading scientists to believe that communication with other species may be possible. Yossi Yovel, a professor at Tel Aviv University’s Sagol School of Neuroscience, who is chairing the competition, said it’ll be very difficult to show whether a human has successfully communicated with an animal, and that he’s still trying to finalize the competition’s rules.

“We want to decipher their communication system and we want to communicate with them using their communication system,” he told Semafor. “So training a dog to sit using human language does not count. If you can do it using barking, then that might be something else.”

The Coller Dolittle Challenge is one of a new set of efforts to drive developments in AI that are too weird for big companies, and too wild for academia. These types of contests are often driven by rich donors with niche, peculiar interests and a strong desire to change the world.

Yovel said Coller is “all for animal rights, which I think our competition is very relevant for because the more we understand animal cognition and communication, the better we will know how to treat them.” Singer Peter Gabriel, who was popular in the 1980s and ’90s, is also part of the effort as co-founder of the Interspecies Internet, a think tank dedicated to accelerating communication with animals and backed by Coller.

“When I played music with bonobo apes, I was stunned by their intelligence and their musicality,” Gabriel said on the Coller Dolittle Challenge website.

Coller Foundation

The winners of these AI contests often go from being unknown computer scientists to geeky rock stars. For example, the Vesuvius Challenge, launched by former GitHub CEO Nat Friedman and entrepreneur Daniel Gross, offered amateurs a path to fame and fortune in artificial intelligence.

Youssef Nader, a PhD student at Freie University, was part of the three-person team that won the $700,000 grand prize this year for deciphering the charred remains of ancient Greek scrolls. He is now famous among archaeologists.

“I met a lot of them, and all three of us sort of became celebrities in this field, of which we know nothing about,” he told Semafor. “[We] don’t know any Greek, but people assume we’re experts.”

He became captivated by the idea of downloading 2000-year-old data directly onto his laptop, and spent hours trying to figure out ways to piece together the characters in the scroll, sacrificing sleep.

These flashy million-dollar prizes offer an alternative way to solve quirky AI problems in new ways that push the technology forward outside the confines of academia and industry. Anyone with some technical chops can get involved.

“I think it’s a very interesting way of solving research problems that probably would otherwise be stalled,” Nader said.

Check out what AI contest Nader wants to enter next.  →

PostEmail
Mixed Signals

Hot off the debate stage, and into the podcast studio, Ben, Nayeema, and Max dig into who gets to decide who “won” the first presidential debate between Joe Biden and Donald Trump. Joining them, statistician Nate Silver, founder and former editor-in-chief of FiveThirtyEight, discusses election predictions and why people are so skeptical of polls these days. On Wednesday, Silver released his Silver Bulletin Election Model showing Trump with a 66% chance of heading back to the White House. Plus, Nayeema calls out Americans’ blindspots by taking us on a quick tour of elections around the world.

Catch up with the latest episode of Mixed Signals.

PostEmail
Semafor Stat

The number of users that Baidu’s Ernie AI model has, according to the company’s Chief Technology Officer Wang Haifeng. It also unveiled Ernie 4.0 Turbo as competition heats up in China’s AI landscape, and as OpenAI moves to block access there to its API amid rising geopolitical tensions. Baidu said its latest model will be available on the web and on mobile app interfaces, and developers will be able to build off of the technology.

PostEmail
Intel
United States Patent and Trademark Office/Screenshot

It looks like TikTok is upping its AI game. In May in the US, the viral video app filed to trademark the name “Genie” for an AI chatbot software that will, among other functions, simulate conversations, facilitate interaction and communication between humans and AI, and produce human-like speech and text, according to a US Patent and Trademark Office filing.

TikTok, which is owned by Chinese company ByteDance, is known for its impressive AI-powered algorithm, and the company has also been working on generative AI tools — in 2023, it appeared to be testing an AI recommendation chatbot called Tako, which could suggest videos based on user requests. The company was also using OpenAI’s code to develop its own large language models, The Verge reported. Earlier this month, TikTok unveiled new AI capabilities for its ad services, like digital avatars and AI dubbing.

But the recent trademark filing appears to go a step further, indicating that it’s working on a full-fledged chatbot that goes beyond search recommendations in the US.

The move comes as the company faces the prospect of divestment from its parent or a ban in early 2025, after the Biden administration determined the company’s Chinese ownership may pose a risk to national security. TikTok is challenging the measure in court, claiming it’s unconstitutional and violates First Amendment rights.

TikTok declined to comment.

— Marta Biino

PostEmail
What We’re Tracking

We wrote Wednesday about the translation company DeepL and its looming competition with Big Tech. So, of course Google announced the next day that it is adding 110 new languages to Google Translate, bringing its total to 243. The new additions, including Cantonese, came thanks to the use of generative AI models.

This just highlights the point of our story: Companies like Google can snap their fingers and get the ability to translate more than 100 languages and offer that technology for free.

DeepL may not be doomed because it’s carving out a specialty in serving businesses that don’t want to risk using free services to translate sensitive documents. But startups need to think long and hard about what kind of products will suddenly become free, including with the massive AI models being developed by Silicon Valley giants.

PostEmail
Hot on Semafor
  • The debate that Democrats feared.
  • How Biden lost the climate debate, too.
  • Kenya’s Ruto withdraws controversial tax bill as protests escalate.
PostEmail