Framework for Thinking About Automation
Automation has been top of mind for a lot us recently. And if it is not it ought to be.
People all over the world are wondering if their work could be automated, how secure is their job, and if new technologies are sympathetic to what they do?
A recent documentary in the New York Times caught my attention. It is an unfortunate story of factory closing down and moving to Mexico. The individuals in this documentary didn’t make their criticism loud, but made it nonetheless. Taking aim at the new offshore employees, their language skills and much lower wages.
I invite you to consider where would such criticism go when training a robot to replace you
However, some of them did see how circular and out of their control such criticisms are. And realized that cost reduction, creating abundance and generating value are forces as silent and powerful as the flow of water.
We can subjectively argue about moral responsibility (which exists) but in the interest of simplifying the complexity of this economic equation I will try and stick to objective material.
The Machines Are Already Here
It is no secret that anything that can save money for a business, will happen.
& anything that can be broken into sets of instructions can be done by a machine
& machines are becoming exponentially more capable in narrow manipulation of intellectual and physical property
It follows then that if what you do can be broken into sets of instructions and will save money (with no value lost) your job will be automated.
Let’s expand on this a little further.
Making is becoming cheaper to the point of becoming free.
“The fact that a unit of wealth is created today with much fewer workers compared with 10 or 15 years ago is possible because digital businesses have marginal costs that tend towards zero. ”
—Klaus Schwab, The Fourth Industrial Revolution
“During the past century, we have used raw materials, new machines, and innovative business models to create unprecedented abundance and democratize luxury with physical goods. Now, as abundance markets open up in the digital economy, there are dozens of areas across knowledge-based industries in which the machine will do almost everything. But at the same time, we’re seeing the impact of digital abundance in the physical world, as well … A good example of this is Narayana Health, which is using the new machine to bring a form of abundance to those in need of heart surgery.”
– Malcolm Frank, Paul Roehrig & Ben Pring, What to Do When Machines Do Everything
The following case study is the work of Narayana Health where automation of everything but the operation itself (logistics, tests et al) the cost of heart surgery down from ~ $100,000 to $12,000. Even with accounting for difference in wages this is a phenomenal use of technology.
- making is becoming cheaper
repetitive making (factories, production) is easily transferable to a machine
companies will start moving such making to a machine if it makes financial sense
The Human Advantage
So if we wanted to hack this situation, and keep our work relevant for longer we should first start by operating on multiple domains.
Machines can only operate on a single domains. Your industrial bearing machine can’t order your lunch, or do your taxes.
As Oren Etizoni (from the Allen Institute for Artificial Intelligence) puts it in a recent talk even recent wins in the game Go by Google’s algorithms need to be put in context. The game played by a machines is highly structured, with discrete sets of moves, binary evaluation (win/lose) and “infinite” labelled data.
We can think about machine automation like infinite sets of tracks, with a train moving on it. The train will go exponentially faster, and might even it make it to supersonic speed, but will never leave that logical domain which is the tracks.
So the first thing we ought to do (as humans) is to consider what tasks (note the difference between tasks and jobs) are repetitive and require low levels of judgment. You should take for granted that these tasks will automate.
Now try and think about the connections you might make between different “train tracks”, your area of expertise, your subjective (and ill-strcutured) skills.
Lean into those, develop them and think about what would unlimited computation can do to excel your output.
The truth is that this level of computation is on its way to help you.
The big question is if you’re going to be around long enough to be the person linking different tracks, or the one running out of breath chasing a supersonic train.