Two examples of automation
Google recently demonstrated their virtual assistant called Duplex. Complete with ums and ahs, the system was able to make a stylist appointment and a restaurant reservation, convincingly human in mannerism and ability to navigate the twists and turns of the unstructured conversation.
This automated warehouse by Ocado has countless battery-powered shuttles scurrying around a grid to collect and deposit groceries in bins.
Whereas Google Duplex is conceived to work with humans, the Ocado warehouse is basically a gigantic vending machine where humans are never to tread. In the Duplex video, the system does its best to behave like a human, whereas the Ocado warehouse is built from the ground up around the strengths and limitations of clever but relatively simple machines.
Jobs at risk
The grocery warehouse looks like a marvel of efficiency and is sure to eliminate the need for many workers that would otherwise be hired to work in a more conventional warehouse. This is offset by the massive human effort devoted to the warehouse’s planning and construction (not to mention the labour that goes into building the materials and robots offsite). It is safe to assume that, in order for such installations to be economically feasible, the labour up front will be outweighed by the labour that would otherwise be required to operate a warehouse for several years down the road.
One’s immediate thought when watching the Duplex demonstration is that anybody whose job is to answer straightforward queries on the phone is in peril. While the demonstration shows a personal assistant artificial intelligence calling a service provider to make a booking, it is easy to imagine the reverse happening as well, provided the AI has some awareness of the businesses’ service availability and scheduling.
What is striking about this example is that before this demonstration I would have assumed that a hair salon or a restaurant that wanted to avoid taking bookings on the phone would have drummed up some sort of online form or app – the development and/or implementation of which would once have been lucrative, high-skill work.
Of course Duplex is being developed by highly skilled Google employees, but once developed, it’s general purpose nature lets it be deployed far and wide at very little cost. It may be that each “conversation” held by Duplex takes a lot of computing resources somewhere in Google’s sprawling server infrastructure, but Moore’s law will soon make that insignificant, as well.
The peril and the promise
The Ocado automation scheme creates a simplified environment to accommodate unintelligent machines. There may be some machine learning used to optimize the routing of the robots, but these machines lack the sensors and flexibility needed to navigate an existing warehouse built with human pickers in mind.
Duplex, on the other hand, will wade into roles previously held by humans without needing an altered landscape.
This highlights the rapid change and widespread affect we can expect from AI. Once code develops to handle a task previously exclusively performed by humans, the marginal cost of using that same code to displace all people who hold that job is almost zero.
Robots vs Artificial Intelligence
We are accustomed to the idea of hardware taking on tasks that had been done by humans – the tractor replacing the hoe, the robot replacing the factory worker, the automated teller machine replacing the bank worker.
These tools were less expensive to operate than the cost of equivalent human labour, but not immediately, and not infinitely so. Each tool had a marginal cost that encompassed human labour. The end result was gradually enhanced productivity that enabled a gradual improvement in the standard of living of those who developed the new skills needed to produce the new tools.
Physical or mundane
Robots were first built for repetitive, sometimes dangerous tasks. Now AI is about to step in to take on the mundane tasks – phone reception, driving – and even tasks that are high skill but also repetitive – deriving insights from large data sets, analyzing medical data to predict prognoses, or predicting what products will be best sellers.
The urgent opportunity
Directly or indirectly, our new technologies will reduce the cost of goods by drastically reducing the labour input costs. This promises great abundance to those who can afford the goods produced. It also means that we as individuals have enormous leverage to produce new goods and services faster and more cheaply than before these new technologies existed.
Have a product in mind? Use artificial intelligence to assess potential demand. Need a prototype? Use 3D modelling, 3D printing and virtual/augmented reality to validate the product without large tooling expenses. Need to get funding to launch your idea? Try a super brand like Kickstarter get your idea in front of prospective investors.
The new technologies that eliminate the mundane invite – and require – your creativity.
