It’s not just you. The messages we’re being delivered these days about AI are conflicted.
The real impact is taking longer than expected. It’s already here. It’s hype. You can’t ignore it. It’s going to take jobs away. It’s going to create more jobs than we’ve ever seen.
When history is written, this may all work itself out into one cohesive narrative. But right now, those of us trying to make sense of the change are left turning over every stone for signs of the truth.
As a longtime media obsessive and even-longer-time word lover, I believe that subtle shifts in headlines can say a lot.
Two recent ones particularly stood out to me. Both were about Amazon, but to this observer, they carry a global point.
“Amazon to make massive $20B+ investment in Pennsylvania AI jobs — with data centers in tow,” Technical.ly wrote on June 9.
Big dollars! But for me, it was the phrase “AI jobs” that jumped out.
As a former Technical.ly editor, I have plenty of firsthand evidence that a Technical.ly article typically talks about tech jobs, so the use of “AI jobs” is a significant shift.
But more to the point, what is an AI job?
Read the article — you should always read the article — and it’s clear that the word choice in the headline was prescient. These AI jobs are much different than “tech jobs” as we’ve conceived them over the last decade.
In Amazon’s Pennsylvania investment, AI jobs will be tasked with operating nuclear power plants and building data centers. This is “the backbone for America’s AI infrastructure,” according to an AWS executive, and it’s kickstarting a massive investment in workforce development.
That spine will need innovation to power its brain, but this is clearly a departure from the last two decades of tech. Instead of focusing on software and connected devices that promised to vault us into a new generation of boundaryless knowledge work, AI is harkening back to a more broad-shouldered category of industry that was once synonymous with Pennsylvania.
Complex models require massive computational capacity to process data, which in turn requires lots and lots of energy. So we have to build more physical infrastructure to deliver it, operate those plants and figure out how to make them more productive. The former steel mill site in Bucks County that will house one data center may have simply been chosen because the space was available, but it’s dripping with symbolism nonetheless. Coal delivered the massive energy gains we needed for the industrial revolution, just as nuclear power promises to do in the age of AI. Plentiful steel helped us create marvels of engineering that reshaped how we moved and organized ourselves, just as data pledges to do for AI today.
There was good-paying work in all of it during the 20th century, just as Amazon’s announcement promises for the 21st.
The 100-year arcs may write themselves. But in the more immediate timeline of years and decades, it’s worth thinking about the shift that this is bringing to the landscape of technology and careers today.
That brings us to a second headline about jobs and Amazon in recent weeks, where the New York Times wrote that “Some Coders Say Their Jobs Have Begun to Resemble Warehouse Work.”
It was doubtless lost on the editor that Amazon redefined warehouse work for the Internet generation with its fulfillment centers, where humans and robots worked side-by-side to collapse the time between click and delivery.
Over the last decade, we also saw how Amazon shaped local economies. Throughout post-industrial cities, the appearance of a fulfillment center provided both a brand name and an instant job creator, checking off a couple of boxes on the wishlist of an economic development professional.
Now it is the coders who hold tech jobs that are doing something like warehouse work, but they won’t be in next-generation warehouses.
After all, the coders were never in fulfillment centers. They were largely in Seattle, and then more came to DC metro, and then we learned they were distributed in a few other locales, as well.
As Amazon’s pre-pandemic HQ2 competition showed, attracting tech jobs required a vastly different playbook that was very much of its time. HQ2 became a litmus test for how the leaders of midsize cities from the Rust Belt to the Sun Belt to the South assembled the pieces of an innovation ecosystem that would attract and retain software developers for years into the future — and the land to accommodate a massive new mixed-use property.
Most had some combination of university and corporate anchors to grow tech talent, investors and incubators to seed startups, and tax and quality-of-living policy changes to improve the city’s brand among millennials in the sought-after developer cohort.
The tech push produced new companies and reinvigorated urban life, but we’ll not soon forget how the era was equally characterized by a supernova of a software engineering job market that eventually crashed into some balance of inflation and automation.
Last month’s announcement in Pennsylvania has a flavor of “America’s Next Top Model”-style competition, as Amazon held out the promise that it may bring data centers to additional communities.
But HQ2 it is not.
In fact, if Amazon were to launch a bid for HQ3 today, the ingredients would likely be different. The company’s announcement in Pennsylvania still states universities and innovation hubs are needed, but in quotations supporting the initiative, politicians also cite water and construction, working families and clean energy. Success may yet again be measured in wattage and square footage, not app downloads.
Tech is about where we’re heading, and the shift to AI jobs is underway. It will reshape where engineers and construction workers seek work, how they build careers, where they move their families, and what motivates them. They will work on AI and use AI to get the job done. This all may or may not be inevitable. But somewhere between tech giants and local leaders, steel mills and headlines, the signs of change are there.
How will you read them?
