Uber and Alphabet subsidiary Waymo currently find themselves embroiled in a legal battle over whether or not the former stole the latter’s technology and used it to create its own autonomous vehicles. We are no wiser than the courts however, a set of documents leaked to Recode put an interesting light upon things.
As per the documents, Uber hadn’t exactly been having a gala time developing self-driving systems. We already know that the taxi aggregator was one of the staunchest supporters of the tech and had been clamoring to develop it as soon as possible. After all, Uber envisions a future when it can get rid of human drivers and all associated problems and move towards becoming a self-driven cab aggregator.
As things stand, the company is already testing its self-driving vehicles upon public roads. However, if this report is true, it raises serious questions upon the wisdom of allowing even that.
As per the report, Uber’s miles per intervention (which is the number of miles driven by a car before a human has to take over for any reason) stood at .9 miles in January. In February, the car had managed to cross the one mile landmark. However, it dropped right back down to .8 miles last week.
Speaking of critical interventions, which are instances where the driver had to take over in order to avoid accidents, in February, it stood at 125 miles. Over the next week or so, the number dropped down to just 50 before shotting back up to 160 miles. At present, it stands at a cool 196 miles.
As far as bad experiences are concerned, Uber’s self-driving cars managed to make it almost 4.5 miles before a bad experience took place, in January. The number dropped down to 2 miles in February and there the numbers remained while well into March.
I don’t know about you but to me, it doesn’t seem as if the company is ready to replace human drivers anytime soon. Meanwhile, it does appear to be really convenient that this data got leaked right around Uber’s ongoing legal battle with Waymo. While it will be up to judge and jury to decide, this particular piece of data may just help Uber. After all, it seems to imply that the cab aggregator’s technology is far from finished.
It also brings the decision of allowing the company to bring its vehicles on public road into question. Was this data perhaps not presented to the regulatory bodies that allowed Uber to bring its self-driving cars to the roads in Arizona and elsewhere? True, it is only by driving that the cars will improve. However, a bad experience every 2 miles does not look like very good odds to me.