You know that expression, “I could do that with one hand tied behind my back”? Well, OpenAI’s latest creation — a self-learning AI robot named Dactyl that has learned to solve a Rubik’s cube one-handed — doesn’t even have a second hand to be tied.
According to a recent article on The Verge, the robotics research division of OpenAI says its humanoid robot hand Dactyl has learned to solve a rubiks cube after being presented with virtual simulations before it was presented with a real cube.
Dactyl Won’t Beat Any Records… Yet
Even though it takes many minutes for the self-learning AI system to solve the puzzle, OpenAI says this accomplishment represents a leap forward towards their goal of developing a robot that can learn to complete tasks without having to train for many months or years and without needing to be specially programmed.
While many other robots can solve Rubik’s cubes very quickly, the Dactyl self-learning AI system differentiates itself through using the same approach it took to learn the Rubik’s cube in other self-learning situations. In short: it doesn’t aim to be a specific, narrow-use robot, it aims to be a general-purpose one, says The Verge.
More from The Verge’s report:
“That’s why OpenAI sees Dactyl’s newly acquired skill as equally important for both the advancement of robotic hardware and AI training. Even the most advanced robots in the world right, like the humanoid and dog-like bots developed by industry leader Boston Dynamics, cannot operate autonomously, and they require extensive task-specific programming and frequent human intervention to carry out even basic actions.”
This is One Step Closer to Sci-Fi
If a robotic hand that can learn new tasks — and, ostensibly, learn to perfect them over time — is capable of solving a Rubik’s cube one-handed, it isn’t too wild to think that it could learn other complicated tasks.
Who knows? Maybe some day, AV project managers will be looking to hire a robotic version of “Thing” from The Addams Family instead of rookie technicians. After all, the self-learning AI only needs one hand to accomplish what most human “cubers” need two to do.
Whether or not it is morally righteous to take those jobs away is a different story…