Will robots eat your lunch?

I’ve been re-reading Physics of the Future by Michio Kaku, and found some of his predictions for the next few decades are a bit unsettling. The good news is that technology is improving quickly, and the rate of improvement is accelerating. In the coming decades, we can look forward to ongoing diagnosis that prevents potentially lethal diseases, frictionless transportation, and advanced AI assistants.

However, there is a flip side. In the chapter on "winners and losers" near the end of the book, Dr. Kaku notes that for many occupations, robots or artificial intelligence may soon replace us outright, eliminating our ability to create value and thus our means of making an income.

The above has significant implications for the future of jobs. It might be impossible to reliably see 30-40 years into the future, but foreseeing how the next decade will develop technologically isn't too difficult. This is worthwhile, if only to help us avoid careers that may be obsolete soon.

The first step to identifying the threats and opportunities presented by computers is to understand their limitations.

Perhaps the most important insight is that computers today have limited parallel capabilities. We often talk about multi-core CPUs, which can perform a few calculations at a time much more quickly than humans can. Conversely, a human brain has 100 million neurons that all operate in parallel, though at speeds much slower than computers.

It's clear that computers and humans are categorically different. Though today's computers are superior in many ways to humans, they have at least two glaring deficiencies:

1. Limitations in pattern recognition
2. No "common sense"

A robot or computer can be programmed to be very good at solving a specific problem, but would be inefficient (and sometimes unusable) when faced with problems that aren’t yet well-defined. Currently, and for the near future, the best applications of computers and robots will be when smart humans design automated solutions to specific problems.

Such solutions will replace some jobs but not others. Let’s first look at which jobs may be at risk, and then at jobs that will likely be safe for at least another couple of decades.

Repetitive work will be automated

Occupations that solve a very specific problem, where the solution can be made much more efficient through automation, will soon be replaced by robots doing the same. Some examples, in ascending order of approximate value to society created:

• Retail industry: cashiers.
• Hospitality industry: receptionists, maids.
• Food and beverage industry: servers, line chefs, dishwashers.
• Transportation industry: drivers, pilots.
• Interpreters.
• Securities traders.
• Teachers.

There will be fewer middlemen

Technology will gradually shorten the value chain. For example, distribution of any product/service that does not require complex sales (eg. most things less expensive than a car) could likely be automated.

We first saw this in retail many years ago with the introduction of the vending machine. Today, all of physical retail is being replaced by automated processes that seek to replicate the basic experience of the venerable vending machine. Soon, we will be able to use a mobile app to order and pay for most items, which will then be automatically selected and transported from a warehouse via a delivery robot directly to our doorstep. Warehouse workers, cashiers, and delivery drivers will no longer be needed.

Meanwhile, the middlemen who survive will be those who can add value through common sense or adaptability. For example, most CEOs will probably never buy other companies online - it's just too important of a purchase. Mergers and acquisitions involve complex sales, where relationships with investment bankers based on trust and customized advice are usually built first before transactions can happen.

Jobs that will still exist

Any occupation that requires adaptable problem-solving, communication or an emotional level with humans, or imagination/creativity will be more long-lived.

• Law enforcement. People may never feel truly safe entrusting security and justice to robots. In the United States, it's even written into the constitution that citizens must be tried by their peers.
• Entertainers and creators. As Dr. Kaku notes, “Computers are not good at determining what makes us cry or laugh, since they cannot cry or laugh on their own, or understand what is funny or sad.”
• Communicators such as negotiators or public relations experts, who often rely on empathy and understanding their counterparties' emotions to achieve results.
• Truly good customer service and/or capable assistants. I'm reminded of a scenario in The Devil Wears Prada, where the main character was tasked with obtaining copies of a Harry Potter novel that had not been released yet. This complex task would certainly have been a tall order for a robot assistant to achieve.
• Leaders, who almost by definition create value through adaptability, imagination, communication, and common sense.

The lists above are by no means exhaustive, and are likely to need updates before the end of this decade. Rather than focus on the details, I think it's worthwhile for us to follow and learn about nascent technological advancements and how they might impact various industries and careers going forward. My advice in the meantime is to learn continuously, allowing us to adapt to new developments as they appear.