Marc Raiabert, the founder and chairman of Boston Dynamics, gave the world a menagerie of two-legged and four-legged machines that are capable of stunning park tour, infectious dance routines and diligent plank stacking.
Riabert now wants to lead a revolution in robot intelligence and acrobatics. And he says that recent developments in Machine Learning have accelerated the ability of his robots to learn how to perform difficult movements without human help. “The hope is that we can produce a lot of behavior without having to make everything that robots are doing,” Raiabert recently told me.
Boston Dynamics may have developed pioneer robots, but it is now part of a busy suit of companies that offer roboths and humanoids. Only this week did a startup called figure showed a new humanoid called Helix, who apparently can release groceries. Another company, X1, showed off with a muscular humanoid called Neo Gamma that did odd jobs around the house. A third, Apptronik, said it is planning to scale up the production of his humanoid, called Apollo. However, demos can be misleading. Also, few companies also reveal how much their humanoids cost, and it is unclear how many of them they really expect to sell as home helpers.
The real test for these robots will be how much they can do independently of human programming and direct control. And that depends on the progress as Riabert's that is praising. Last November I wrote about efforts to make completely new types of models for driving robots. If that work starts to bear fruit, we may see humanoids and four -legged friends ahead faster.
Boston Dynamics sells a four -legged robot called Spot used on oil brigs, construction sites and other places where wheels struggle with the site. The company also makes a humanoid called Atlas for research. Riabert says that Boston Dynamics has used an artificial intelligence technique called reinforcement learning to upgrade the ability of spot to run, so that it moves three times faster. Atlas also helps the same method to walk more confident, says Rialle.
