SoftBank & Boston Dynamics – Homecoming.

Boston Dynamics finds a credible home.

• Boston Dynamics may have finally found a credible owner spelling the end of click-bait robots that lose money and instead focusing on bringing its intriguing technology properly to market.
• Boston Dynamics is the company that produces the well-known (at least on YouTube) robots with legs that are capable of negotiating and dealing with obstacles and rough terrain.
• It was first spun out from the Massachusetts Institute of Technology to be purchased by Google in 2013 probably for a lot less than $1bn.
• (When a public company makes an acquisition of a private company it does not have to disclose details if the sum is not significant relative to the size of the acquirer)
• This is why I estimate that both Google and SoftBank which bought Boston Dynamics from Google in 2017 probably both paid a lot less than $1bn.
• Hence if the mooted sale to Hyundai goes ahead at the $1bn price tag, SoftBank will have made a good return on the investment.
• The speculation is that Boston Dynamics will be sold to Hyundai Motor Company (part of the Hyundai Chaebol) which does not make a lot of sense.
• Instead, I suspect that if this sale goes ahead, it will be Hyundai Heavy Industries that purchases it as this is where the Hyundai Robotics activities are held.
• I focus on Boston Dynamics because despite being kicked around like a football it is an important company.
• This is because it is one of the only companies that I know that is actively trying to address Moravec’s Paradox which is one of the most perplexing problems with modern artificial intelligence.
• Moravec’s Paradox is the observation that tasks that are hard for humans (e.g learning to play Chess to a grandmaster standard) are easy for machines while tasks that are easy for humans (walking) are almost impossible for machines.
• This is why almost all robots that move around do so on wheels as teaching them to walk on legs is fiendishly difficult.
• To me, Moravec’s Paradox results from the techniques being used to create artificial intelligence being insufficient to confer real intelligence upon machines.
• Almost all AI around today involves machines doing what they are instructed to do or making inferences based on statistical characteristics and patterns within data or some combination of the two.
• Boston Dynamics has had some success in addressing this issue which means if it can find a credible way to market and demand for its technology then it will have a sustainable competitive advantage.
• This is why the acquisition by Hyundai Heavy Industries is interesting because its subsidiary Hyundai Robotics is a major manufacturer of industrial robots.
• If Boston Dynamics technology is integrated into Hyundai’s robots in a way that improves its functionality materially, then Hyundai stands to gain a lot of market share.
• Furthermore, because this is a difficult area to solve, it will be a while before anyone catches up with Hyundai.
• So far, I am not convinced that anyone has really tried very hard because while Boston Dynamics has been great at generating YouTube views it has not been good at making money.
• Hence, the rewards for making robots that walk on legs have yet to be determined explaining the lack of interest.
• This is a bold move by Hyundai but whether it can make anything of this quirky company and its technology remains to be seen.