Hong Kong protestors consider all sorts of face masks 🤡

And how getting to Mars will need AI

Hi hi,

There was a ton of news and features this week! Debating capping how much I cover, but weeks like this would mean cutting things I find important. What do you think? Might try breaking up into sections (robotics/face rec) next week if there is as much, a suggestion from Neil!

Story of the week

Hong Kong banned people from wearing face masks in public last weekend. Protestors aren’t following the rules so far, who were using face coverings to prevent from being recognized by the city’s face recognition cameras (as well as turning to lasers to confuse the tech and straight-up cutting the cameras down).

One thing they aren’t doing? Using wearable face projections despite some tweets going around. The tool, shown above, is actually an art project created by students at Utrecht School of Arts in 2017.

The tech is likely not ready for prime protesting time. A story I did back in 2015 (so take it with a grain of salt!) said that making real-time face masks take months of work and are heavily pre-planned. But necessity is the mother of invention, and other hacks are already working. It’s just a matter of time.


African Americans are at a 10 percent higher risk of losing their jobs from automation than the general population, according to a McKinsey report.

Surprise! Amazon workers are watching and labeling footage from customers’ Cloud Cam footage to improve the algos.

Much like the internet at large, most deepfakes (96% by one count) are porn.

In other deepfake news, California just banned politically motivated deepfakes. Best of luck with that!

California also passed a law stopping police officers from adding facial recognition tech to body cameras.

John Deere, a company very famous for its software, wants you to known agriculture is a tricky problem for machine learning.

Rotterdam, Europe’s busiest port, is trying to run the whole operation with 10 to 15 people by relying on smart containers, autonomous cranes and software called Captain AI.

Not just anyone gets to write for the New Yorker—but they let a fancy NLP tool (GPT-2) try out.

Future lab rats could be lab robots.

Waymo is coming to LA.

The US is blacklisting a bunch of Chinese companies, including some of the AI heavy hitters, for developing tech that is being used against the country’s Muslim minority groups (also helpful in the trade war, coincidentally enough). Here is a tweet showing all the places that use the tech and aren’t impacted by the new U.S. sanctions.

NYC is considering passing a law that restrict how landlords and business owners can use face recognition tech.

You’ll have to submit to a face scan when you want a new mobile number in China. They already checked and collected information previously, but the face data is a level up.

Police robots don’t call human cops when their emergency button is pressed. Extremely useful.

Brexit isn’t the only issue for some Brits wanting to travel freely: The UK launched tech to look at passport photos that they knew didn’t work great on dark skin. (╯°□°)╯︵ ┻━┻

Gymnastics are looking into robot judges.

Facebook launched some new stuff for PyTorch, the company’s deep learning framework.

Datapoint of the week

100,000: How many elevators Thyssenkrupp connected to the cloud just in the U.S. in hopes of being able to predict needed maintenance before things break.

Prospecting Arxiv

We need AI to get to Mars (and beyond?), according to a delightfully nerdy and accessible paper put out by a NASA researcher. The tome, “Artificial Intelligence: Powering Human Exploration of the Moon and Mars,” outlines three autonomous mission operations capabilities where AI is needed.

Crew Autonomy gives astronauts tools to assist in the performance of each of these mission operations functions. Vehicle System Management uses AI techniques to turn the astronaut’s spacecraft into a robot, allowing it to operate when astronauts are not present, or to reduce astronaut workload. AI technology also enables Autonomous Robots as crew assistants or proxies when the crew are not present.

The paper dives into examples of how each of these instances would come into play and why it is necessary. It’s also easy to imagine these tools not only getting us to Mars, but being retrofitted to do cool things on Earth. After all, we already have things like artificial limbs and memory foam from previous research (and here a list of more tech we get to play around with on this planet that were originally developed for space.)


“Every time a driver wrests control from the computer to avoid an accident, it’s a potential teachable moment—a chance for the software to learn what not to do. It’s a calculated risk, and it’s one that federal regulators, used to monitoring for mechanical defects, may be ill-prepared to assess.”

“Tesla’s Autopilot Could Save the Lives of Millions, But It Will Kill Some People First” from Bloomberg

Until we all have our own portable face recognition technology deceivers,