There are a few things that set Gary apart from other robots on the market today.
Some amazing things have been happening in the field of robotics lately. Every major technology company in the world, including Tesla, Samsung, and Toyota, appears to have announced its vision for a robotic future; Amazon has not only switched to fully automated warehouses, but it has also recently purchased iRobot - one of the leading robotics companies in the world. And if that isn’t enough, A Chinese game developer has recently appointed an artificial intelligence-powered virtual humanoid robot named “Ms. Tang Yu” as their new chief executive officer. And while there is no doubt that this is just a (pretty silly) publicity stunt, it says a lot about the public perception of robots.
The public is ready, and dare we say, Eager to start sharing their lives with autonomous robots. The 80’s dystopian vision of “the terminator” or “robocop” is seen as an amusing anecdote from the past, and people today are not fearing a future full of robots; they are actually welcoming it with open arms. Any self-respecting consumer electronic show has several booths dedicated to robotics technology, while Many of them have robots roaming around and interacting with the public or serving them drinks and snacks.
If you have ever visited one of these consumer electronic shows, you have probably seen unbelievably impressive robots showing off their skills as butlers; some of them can walk to a fridge, open it, take out a bottle of water, walk over to a table, pour some water in a glass and then serve it to the person running the demonstration. In fact, we have seen this done about a decade ago, so where are these robots? If they could do this all the way back in 2010, how come they are not walking around our houses, serving us coffee, and bringing us our slippers?
Well, the truth is that most of these demonstrations are very misleading; These companies are showing us that their robots are mechanically capable of performing these tasks while skipping over the “mental” part of the demonstration. What robots have always been good at is repeating a pre-programmed movement over and over again. That is why robots build cars - they can weld the same door to the same place again and again without ever making a mistake. What they can’t do is pick up a door if it is presented to them at a different angle. They have no idea what a door even is - all they know is to repeat the movement they were programmed to do.
The same thing can be said about these “Show bots.” while it looks like the robot is walking to the fridge to grab a bottle of water, what actually happens is that the robot takes an X amount of steps in a pre-programmed direction, then extends his arm in a fixed angle, makes a grabbing motion with his hand and so on. As long as the fridge is in the correct position, everything should go smoothly, and the robot should end up with a bottle in its hand. However - move the fridge a tiny bit to the left or right, and the whole thing crumbles like a tower of cards. The robot has no idea where the refrigerator is located, and it will reach out, grab nothing and proceed to pour a nice cup of nothing while not even realizing it has failed.
The reason we don’t have robot butlers in our homes is not mechanical; Disneyland has robots that can dance, sing and even swing like spiderman - Mechanically, robots can do amazing things. But robots have to be smart if you want them to help around the house or in a hospital or hotel environment. They have to be able to understand the world around them and be able to make decisions on the fly.
Suppose you ask a robot to bring you a bottle of water from the fridge. In that case, he needs to know what a refrigerator is, where it is located around the house, and how to navigate to it while avoiding obstacles. Next, he needs to know how to open any kind of fridge, identify the bottle, and differentiate it from everything else in the refrigerator. He then needs to calculate the exact way he needs to move his arm in order to grab the bottle and what force to exert on the bottle, so it doesn’t fall from his hand but also doesn't break. Then he needs to close the refrigerator behind him while holding the bottle and finally navigate back to the human and hand them their drink.
And THAT is what Gary does!
Gary isn’t just mechanically sophisticated; He is also incredibly smart. He understands simple prompts and can extrapolate all the different actions he has to do in order to fulfill the task he was given. So, for example, If you tell Gary to clean the dining-room table, he doesn't simply run the “Dining-room cleaning” script. Instead, he identifies the dining-room table, scans it, and then proceeds to pick up each item on it and place it in its correct place - dirty dishes in the dishwasher, milk carton in the refrigerator, half-eaten apple in the trash, and so on.
Gary uses computer vision, machine learning, and artificial intelligence to navigate the real world, Understand what he sees, and then plan and execute his actions accordingly. This makes him an actual All-purpose robot, one that can evolve and learn new skills. Gary can even learn from the life experience of other “Garys” worldwide, which makes his potential pretty much limitless.
Gary’s body is built to navigate and manipulate the world around him, but it is his brain that allows him to understand this world, plan his actions, grow, evolve, and become much more than just a robot that repeats a set of pre-programmed movements intended to make audiences go “ooooh.”
