The last jobs on earth

Hi, and happy Tuesday.

Friday night, Cape Verde - a chain of small islands most of us have never heard of - took the defending World Cup champions - Messi's Argentina - to the brink, before losing to a late goal.

Quite frankly, a machine would have known Cape Verde’s task was impossible.  The humans didn't - so they went after it anyway.

And this is what motivates so many people to watch and adore sports.

Our instinct - to want the human version of something, with flaws included - is work that survives when everything else gets automated.

But what other types of work also survive? I tested Claude Fable on this question.

The first thing I discovered is that jobs showcasing authentic human origin - while having durable value - are almost nonexistent.

If you take professional athletes as a form of this job, for example, there are roughly 19,000 in the entire US economy. Pay for most is closer to $25,000/year than the top-end splashed in the media.

If your kid wants to be the next Messi: godspeed. The odds are about those of the island nation.

(We use the US throughout because it’s an advanced economy that arguably is ahead of job-loss trends through outsourcing, offshoring etc)

But there are three other kinds of work that need uniquely human characteristics - characteristics that no amount of automation can substitute.

Embodiment. These are jobs where a body has to be in the room doing something fiddly. Home health aides are now the single largest occupation in the US - nobody wants a robot bathing their mother. 

Accountability. Someone has to be answerable when a decision is expensive or risky. Nurse practitioners are the fastest-growing job in healthcare, up 40% over the last four decades. Also in this group are doctors, pilots, and managers and leaders - such as you - whose name is on the decisions made.

Durable relationship. Someone you actually trust. Therapists and counselors - mental-health counseling is up 17% - as well as the teacher your kid will still talk about at 40.

But my view is that AI is largely the latest form of digital transformation - building on shifts such as the desktop computer, Internet and mobile - though it may accelerate this ongoing transformation. 

As such, we should already see, through automation from the previous waves of digital transformation, a shift to these durable categories of human jobs.

To test this hypothesis,  I tasked Fable with taking data on all the jobs in the US since 1980 and categorizing them into the above 4 categories, plus an “everything else” bucket.

No one has categorized jobs in this way before.

In the past, without AI, this would have taken a small army of analysts crunching through data for weeks.

This is what it came back with:

A few key takeaways:

• Since 1980, filing, data entry, and other kinds of routine work - the “everything else” category in the chart - fell from 46% to 38% of all jobs. 

• Two human moats absorbed almost all of that displaced work: accountability climbs from 15% to 19%, and relationships from 7% to 11%. 

• Embodiment has been a flat ~31% for four and a half decades - the biggest (and most boring) slice on the chart.

• Authentic human origin sits at a rounding error the whole way, around half a percent.

Looking back over 45 years, we see that the automation apocalypse has already been happening, quietly.

This analysis, of course, has a few problems.

• Is the AI correctly interpreting what the numbers in the data mean?

• Are we being naive, not seeing something in how to think about this problem?

• What would people who have been studying these problems for decades think?

Which speaks directly to how we’re seeing the work that we - at Prescouter - do shift.

Clients can do these types of analysis using AI themselves, but are the results trustworthy enough to make decisions with?

Clients often even wonder if they are asking the right questions - given innovation often comes from the fringes, where they are not the experts.

Our in-house experts - as well as those from our 10,000+ extended network - are often brought in to not just weigh in those questions, but build more robust analyses and dashboards that build confidence and trust - sometimes starting from their scratch AI work (such as this jobs example).

AI and automation is essentially shifting Prescouter’s work further along the accountability and durable relationships dimensions. I expect this is true of many knowledge work-orientated jobs.

Underlying all this is the durable, unglamorous truth: humans will always have problems.

And as long as there are problems, someone gets paid to solve them. Automation just keeps narrowing that to the four things a machine can't do - show up in a body, be answerable, be trusted, and be beautifully, watchably human.

Which is why, when I think about what to tell the next generation, the answer that keeps rising isn't a coding bootcamp or a specific credential. It's emotional intelligence. Two of the four human moats - again accountability and durable relationships - are about working well with other people and being the one others trust with the decision.

The full dataset, the classification behind every job, and the chart you can poke at are here:

https://auxee.com/the-four-moats-where-work-survives-the-machines/

If this topic interests you, let me know. If there is enough interest, we can give it the full Prescouter treatment.

In the meanwhile, let’s see how another overperforming underdog - Egypt - do with their turn at the defending champions, later today.

Best,

Dino