Uber’s Not-So-Sharing, Automated Vision is Expensive
The news is in that Uber’s second financial quarter saw losses of more than $400 million due to its ongoing investment in self-driving cars. Apart from adding to the evidence of Uber’s hubris, this brings a conflict between decentralization and automation to the fore. Relevant to the world of blockchain in several ways, the notions of “points of failure” and removal of human agency seem foremost.
Uber is the most important recent Silicon Valley wunderkind in financial terms, perfect as a next chapter after the emergence of the search and social media colossi.
Having decimated the traditional taxi industry with the combination of cell phone GPS, smart contracts and gamified scaled micro-incentives for drivers, the company sought on the back of rapid success to become the unicorn that grew a second horn. Not only would it break apart traditional hired transport, but the nature of the automotive industry itself. Cars that drive themselves, guided by its maps and satellite data — in fact the idea had decades of precedent, with much of it funded in the U.S. by DARPA, the US Army and the U.S. Navy. DARPA being the bright sparks that incubated military drone tech and the Internet.
Other entities were also interested in self-driving vehicles and Uber wanted to use its logistical power to be the first to implement a workable prototype on a large scale. Over the past 18 months they dropped between $125 and $200 million on trying to make it work, which amounts to a variable of between 15% and 30% of quarterly losses.
Since its early accomplishments a number of setbacks have scuffed the company’s veneer: the failure to crack China, which cost $1 billion a year for three years in a row, conflicts with local authorities resulting in bans from cities and countries, and the departure of the founder Travis Kalanick from the CEO role after various “inappropriate behavior” accusations. A death in Arizona from a rather stupid sounding software bug earlier this year put trust and regulation for Uber’s self-driving vehicles and likely the whole concept back towards the horizon — where many people still feel it belongs. For one thing, with a human being, as error prone as they are, our ethics and legal systems can at least seek to find a truly accountable actor after a disaster. Not so with the self driving cars or Tesla’s autopilot — corporations can point to disclaimers or scapegoat individuals, displacing blame and litigating their way out of trouble as they do.
It seems that investors have even advised the “sharing” app Uber to sell its experimental self-driving car section. Hundreds of millions of dollars a year for the past several years have not produced a workable prototype with a clear path ahead for commercialization, reports the Information website.
One consideration here is to what extent we should want or allow “algorithmic” or smart contract management, in combination with AI and physical automation, to govern our lives.
Defenders of increased automation will point to human error with the same habit of argument as gun rights activists: the machine is never to blame. Wired magazine was pleased to side with the robots in the aftermath of the fatal crash in Arizona. If the question of safety on the roads is always one of human error, might it be better to relinquish human drivers altogether?
In the case of the terrible Air France disaster of 2009, it was arguably the interaction of autopilot and pilot at a stressful moment that caused the crash. The fatal moment was when the pilot overrode the autopilot. This leads to the conclusion that one should reduce or remove one party in the combination. Overall, by taking pressure from pilots and adding massively useful aerodynamic data to engineer’s storehouses, autopilot has undoubtedly made flying safer — it is unlikely to go away. Writing more barriers against human control into airplanes’ protocols could mitigate against hijacking, too.
This is exactly the way many people see the future of the roads — even if human drivers are allowed to remain, their vehicles will all be entered into a tracking system, along with the self-driving cars. The system will simply stop any two vehicles or any vehicle about to crash into an object.
Yet, the question, ported over here from the blockchain world, remains: if I create an automated system, won’t I have to relinquish decentralization to increase (certain kinds of) security and speed? And If I centralize control, do I concentrate points of failure too much, putting too much power into the hands of a company or the predictive or reactive power of an algorithm? The robo-cars might be unavoidable, but will they, unlike the Arizona car, be able to avoid knocking us into the roads?
About Genaro Network (GNX)
The Genaro Network is the first smart data ecosystem with a Dual-Strata Architecture, integrating a public blockchain with decentralized storage. Genaro pioneered the combination of SPoR (Sentinel Proof of Retrievability) with PoS (Proof of Stake) to form a new consensus mechanism, ensuring stronger performance, better security and a more sustainable blockchain infrastructure. Genaro provides developers with a one-stop platform to deploy smart contracts and store the data needed by DAPPs simultaneously. Genaro Network’s mission is to ensure the secure migration of the core Internet infrastructure to the blockchain.
Official Telegram Community: https://t.me/GenaroNetworkOfficial
Telegram Community (Russian): https://t.me/GenaroNetworkOfficial_Rus
Genaro Telegram Journal: https://t.me/genaronetworknews
