[SciShow Intro plays]

Hank: The information age is changing nearly every aspect of life on Earth. I have a computer in my pocket that weighs less than 200 grams, and I can use it for everything from learning about quantum mechanics to watch cats be frightened by cucumbers. And last week, the White House hosted a summit to talk about a project to help information and technology do even more -- especially, to treat diseases. It’s called the Precision Medicine Initiative, and about 40 organizations are splitting more than $200 million to help make it happen.

So, what is precision medicine? It’s based on the idea that diseases work differently in different people, depending on their genetics, environment, and lifestyle. For example, if we know that some genes affect how diseases responds to treatment, then doctors can use that information to help pick the right treatment for a patient based on their genetic code. And doctors are already using this kind of approach -- cancer cells, for instance, usually have genetic mutations that allow them to grow and spread, and specific mutations can affect how the cancer will respond to certain treatments. So doctors can test for those mutations.

As genetic testing gets easier and cheaper, these tests are being done more often. But there’s a ton more information that we could be getting out of genetic testing -- we just don’t know about it yet. This initiative is hoping to change that. The main thing researchers need is data -- they need access to the genomes of lots and lots of people, so that they can look for associations between genetics, diseases, and treatments. So, some organizations involved in the initiative are developing plans to collect as much genetic data as possible, with patients’ consent. By 2019, they’re aiming to have a database with the sequenced genomes of at least a million Americans. And, hopefully, all of that new information will help researchers figure out which treatments work best for which people, and maybe even develop some new ones.

But, moving on, progress in the information age isn’t without its casualties. Sometimes, in order make things better and smarter... you just have to mess with a robot. Last week, robotics engineering firm Boston Dynamics posted a video of its newest humanoid Atlas robot. And it was awesome, and also kind of sad. In the video, Atlas does a lot of walking, slipping and sliding on uneven, snow-covered ground. They also show it moving around boxes -- but then, a human uses a hockey stick to knock the box out of it’s hands, or to push it away right as Atlas is trying to pick it up. Then pushing it away from the box with the hockey stick...

Eventually, Atlas gets attacked from behind and pushed onto its face, and has to get back up again. Robotics engineers...you must know what it’s like for people to push you around. It’s not nice. But, by pushing Atlas around, the research team is actually showing off capabilities that, for a robot, are a really big deal. Atlas is designed to be used for things like search-and-rescue missions -- which often involve a lot of uneven terrain and unpredictable obstacles. And, from the looks of it, improvements to the robot are making it better and better at getting around in that sort of environment.

For one thing, Atlas used to have a tether that would provide it with power and communications -- and sometimes act as a safety line to stop it from falling over. This version -- as you can see from the video -- doesn’t. Instead, it comes with a huge battery pack, uses wireless communication to get commands from its users, and relies on programming and hardware to keep from falling over, all on its own. And that programming seems to be working! Even though it’s walking up and down a snowy hill, Atlas manages to regain its balance before it falls. And when it eventually does get pushed over -- with what looks like a nice, strong shove -- the robot is able to get right back up again.

It also seems to be good at figuring out how to accomplish a task, even when circumstances change -- i.e., when someone’s being a jerk to it. That’s what’s happening when the box it’s trying to pick up gets moved around -- this problem-solving is an important skill for a robot that might be, say, searching through rubble. Those black-and-white patterns on the box, by the way, are basically QR codes, and the robot uses them to track the box’s location. For example, it might know that one pattern means “this is the top of the box,” and that to pick up the box, it needs to position its arms in a certain way relative to that pattern. If the box -- and therefore the pattern -- moves, Atlas can use its programming to figure out the particular way it needs to move to pick up the box.

So, all that pushing and shoving means that Atlas is improving, and fast. I, for one, can’t wait to see Atlas continue to improve. I just hope they give it, like, a nice fruit basket or something to apologize.

Thank you for watching this episode of SciShow News, and thank you especially to SR Foxley, our Patreon President of Space who helps make this show possible. If you want to help us keep making videos like this, you can go to patreon.com/scishow­. And don’t forget to go to youtube.com/scishow and subscribe!