I’ve borrowed and modified this title, of course, from the great Raymond Carver short story collection, What We Talk About When We Talk About Love. Like love, human employment is an emotionally-laden topic. Freud famously stated, “Love and work are the cornerstones of our humanness.” People need meaningful connections with other humans—also somewhat at risk these days—and they need to undertake productive activity.

Work often sucks, of course, and all of us have specific work tasks that we wouldn’t mind handing over to an intelligent machine. But humans often derive both satisfaction and sustenance from work, and 30% of global workers in an ADP survey are concerned that AI might take their jobs. Although it’s not entirely clear that AI is at fault, there is plenty of speculation that it’s already causing a decline in hiring, particularly for entry-level workers.

The Vendor Conspiracy of Silence

For years, however, AI vendors of all types have largely been silent on the topic of job loss. They may acknowledge that some jobs will be lost, but quickly point out that “many new jobs will be created.” The glass is always half full. ServiceNow, for example, in announcing new agentic AI capabilities (“the next evolution of great experiences”, said this:

We want to help make people more productive so they can do work that’s meaningful…Using GenAI and related technologies, we can deploy AI assistants that accomplish manual, repetitive tasks, working in partnership with people who supervise their work. This frees people to focus on more strategic and meaningful work.

The ubiquitous Sam Altman is even more optimistic in a late 2024 blog post about AI agents and superintelligence:

We believe that, in 2025, we may see the first AI agents “join the workforce” and materially change the output of companies. We continue to believe that iteratively putting great tools in the hands of people leads to great, broadly-distributed outcomes…We are beginning to turn our aim beyond that, to superintelligence in the true sense of the word. We love our current products, but we are here for the glorious future. With superintelligence, we can do anything else. Superintelligent tools could massively accelerate scientific discovery and innovation well beyond what we are capable of doing on our own, and in turn massively increase abundance and prosperity.

It’s great that AI is going to let all of us humans achieve our best selves and be highly prosperous. However, I’m not sure there is evidence for this assertion. And while vendors have largely toed the highly optimistic line about job loss, some are beginning to discuss less positive futures explicitly, for better or worse.

Some Vendors Break the Silence

Some vendors are becoming much clearer about the possibilities for large-scale job loss. Dario Amodei, the CEO of AI vendor Anthropic, told Axios in an interview that:

AI could wipe out half of all entry-level white-collar jobs — and spike unemployment to 10-20% in the next one to five years…Amodei said AI companies and government need to stop “sugar-coating” what’s coming: the possible mass elimination of jobs across technology, finance, law, consulting and other white-collar professions, especially entry-level gigs.

Amodei at least seems to regret this possibility. Other vendors are almost gleeful. At Mechanize, an agentic AI startup, the founders seem to relish putting people out of work. From a New York Times article about the company:

“Our goal is to fully automate work,” said Tamay Besiroglu, 29, one of Mechanize’s founders. “We want to get to a fully automated economy, and make that happen as fast as possible.”

So the cat’s out of the bag. If these companies are successful, many of those reading this article will be unemployed. How do you feel about that? Does it make you want to buy their software for your own business? Was it always the case that many business leaders wanted this level of automation, but it just wasn’t available? And if it becomes available, will we soon be seeing “hordes of citizens of zero economic value,” as Michael Malone and Bill Davidow put it in a 2014 Harvard Business Review article? If so, what will those hordes do?

What If It Does Happen?

People who admit there will be job loss from AI often say that it will turn out fine in the end, as it did for the Industrial Revolution and the Agricultural Revolution. They say that workers who become unemployed because of AI will find new jobs. Occasionally they admit that there will be some “dislocation” while workers get retrained and look for new jobs that didn’t exist before AI.

But it’s likely to be much worse than minor dislocation. In the Industrial Revolution, for example, life for those who lost their jobs was unpleasant for many years. In 1770, for example, 20% of women and children in the UK were employed in the hand spinning industry. By 1780, automated spinning machines began to create unemployment for these workers. According to one historical source, the effects on employment persisted until the 1830s. For those workers—and I suspect those affected by AI—there wasn’t rapid relief. One need only read Dickens’ Oliver Twist or Great Expectations to understand what “dislocations” mean in real human lives.

I hope that the doomsayers about AI and employment tomorrow are wrong, but that possibility is not sufficient justification for doing nothing today. I don’t have high hopes for the current U.S. Congress providing relief for unemployed workers—just take a look at the Medicaid requirements in the “One Big Beautiful Bill Act,” which is only beautiful if you are rich. But retraining and supporting workers who lose their jobs to AI should certainly be a campaign plank for Democratic candidates, and for Republicans who want to retain their populist popularity.

The Yale economist Robert Shiller—a Nobel Prize winner in 2013—once said at about that same time that, ”Advancing machine intelligence is the most important problem facing the world today.” I thought it was a bit of an exaggeration, although I have always admired Shiller’s work as an economist. Now I think he was simply prescient. He also argues that we should create a new insurance product that protects people against loss of income from AI. That sounds like an excellent idea; perhaps for a while many potentially-displaced workers can still afford the premiums.