The Download: gig workers training humanoids, and better AI benchmarks
The Download: Gig Workers Training Humanoids, and Better AI Benchmarks
Gig Workers Training Humanoids
In a world where robots are becoming increasingly intelligent and autonomous, a new group of workers is emerging to help train them. Meet the gig workers who are recording themselves doing chores, cooking, and even playing video games to help train humanoid robots. These workers, mostly based in developing countries, are hired by companies like Micro1 to record themselves performing various tasks, which are then used to train robots to perform similar tasks.
Zeus, a medical student in Nigeria, is one such worker. He records himself doing chores, cooking, and even playing video games to help train robots. Zeus is one of thousands of workers hired by Micro1 to record themselves performing various tasks, which are then used to train robots to perform similar tasks.
The work can be challenging and weird, but it pays well locally. However, it raises thorny questions around privacy and informed consent. The workers are often asked to record themselves performing intimate tasks, such as cooking or even personal hygiene routines, which can be uncomfortable and invasive.
Better AI Benchmarks
For decades, AI has been evaluated based on whether it can outperform humans on isolated problems. However, this approach is flawed, as AI is seldom used in the real world in the same way. In reality, AI operates in messy, complex, multi-person environments over time, which makes it difficult to assess its performance.
Angela Aristidou, a professor at University College London and faculty fellow at the Stanford Digital Economy Lab and the Stanford Human-Centered AI Institute, proposes a new approach to benchmarking AI. She suggests that AI should be evaluated based on its performance over longer horizons within human teams, workflows, and organizations.
Aristidou's proposal, called Human-AI, Context-Specific Evaluation, aims to assess AI's performance in real-world scenarios, rather than in isolation. This approach would require AI systems to be designed and trained to work in collaboration with humans, rather than simply trying to outperform them.
Quantum Computing for Health
In a laboratory on the outskirts of Oxford, a quantum computer built from atoms and light awaits its moment. The device is small but powerful, and its abilities will be put to the test in a competition to solve real health care problems that "classical" computers cannot.
The prize will go to the quantum computer that can solve real health care problems that "classical" computers cannot. However, there can be only one big winner, and the competition is fierce.
Must-Reads
- OpenAI just closed the biggest funding round in Silicon Valley history, raising $122 billion ahead of its blockbuster IPO.
- Iran has threatened to attack 18 US tech companies, eyeing their operations in the Middle East.
- Artemis II is about to fly humans to the Moon, with experiments that will set the stage for future explorers.
- Putin is trying to take full control of Russia's internet, with new outages and blockages cutting the country off from the world.
- A robotaxi outage in China left passengers stranded on highways, with police blaming a "system failure."
Quote of the Day
"From now on, for every assassination, an American company will be destroyed."
— Iran's Islamic Revolutionary Guard Corps (IRGC) threatens US tech firms in an affiliated Telegram, per CNBC.
One More Thing
ACKERMAN + GRUBER
How one mine could unlock billions in EV subsidies
In a pine farm north of the tiny town of Tamarack, Minnesota, Talon Metals has uncovered one of America's densest nickel deposits. Now it wants to begin mining the ore. Products made from the nickel could net more than $26 billion in subsidies through the Inflation Reduction Act (IRA), which is starting to transform the US economy.
Deep Dive
The Download
The Download: AI-enhanced cybercrime, and secure AI assistants
Plus: Instagram's CEO Adam Mosseri has denied claims that social media is "clinically addictive"
By Rhiannon Williamsarchive page
The Download: 10 things that matter in AI, plus Anthropic's plan to sue the Pentagon
Plus: The US DoD has been secretly testing OpenAI models for years
By Thomas Macaulayarchive page
The Download: Quantum computing for health, and why the world doesn't recycle more nuclear waste
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. A $5 million prize awaits proof that quantum computers can solve health care problems In a laboratory on the outskirts of Oxford, a quantum computer built from atoms and light awaits…
By Thomas Macaulayarchive page
The Download: protesting AI, and what's floating in space
Plus: The US government wanted to use Anthropic's AI to analyze bulk data collected from Americans
By Rhiannon Williamsarchive page
Stay connected
Illustration by Rose Wong
Get the latest updates from MIT Technology Review
Discover special offers, top stories, upcoming events, and more.
Enter your email
Privacy PolicyThank you for submitting your email!
Explore more newsletters
It looks like something went wrong.
We’re having trouble saving your preferences.
Try refreshing this page and updating them one
more time. If you continue to get this message,
reach out to us at
customer-service@technologyreview.com with a list of newsletters you’d like to receive.
