I’m undertaking a 1000-day reinvention project, blogging here daily to track my progress. In Thursday Thinker, I share a smart idea or theory.
Shortly, I will see the last of my three kids leave the maternal payroll. My youngest, completing a master’s degree in two weeks, has a job lined up for after graduation. I’m so proud of how hard she has worked to get to this point. And I’m pleased that she has chosen a field that while, not initially high-paying, represents a direction that won’t be disrupted by artificial intelligence: social work, and one that represents a real contribution to the world.
I say “not initially high paying” because once she achieves her LCSW (Licensed Clinical Social Worker) designation, she has many career paths open to her, including opening a private practice. This can be a flexible and reasonably lucrative pathway, allowing a clinician to work from home doing online counseling, working less than full-time, and having great control over her work life.
The changing workforce
I agree with Molly Kinder, who writes about the intersection of AI and work, that we are very far from a world where artificial intelligence and robotics completely automates away human labor. She calls this Reality 3, where Reality 1 is today, “AI is everywhere, the displacement headlines are everywhere, but the aggregate labor market data is mostly fine.”
She writes of Reality 3:
[In] thinking through a timeline for Reality 3, I envisioned a single, concrete environment: an airport. The pilots, the flight attendants, the gate agents managing irate passengers when a flight gets cancelled, the people loading bags, the people cleaning the terminals, the people cooking and serving the food, the security screeners and customs officials, the wheelchair attendants, the maintenance crews, the people unclogging a toilet at 2am in Concourse B. How many of those jobs realistically go away in the next five years? Ten? Some of them, sure. Most of them, probably not. The physical world is sticky. Edge cases dominate. Humans demand humans, especially when something has gone wrong.
It’s not just that humans demand humans but that robotics to replace discretionary human action in unstructured environments do not exist right now. Many jobs in the real-world economy — nursing, in-person education, food preparation and serving, construction, transportation — are unlikely to be replaced by so-called physical AI within the next ten or twenty years.
What comes next?
Kinder writes:
Here’s what I think is much more likely, and much sooner. It isn’t a “job apocalypse” that destroys every job in the economy. Instead, it is a hard, messy period of concentrated pain with job losses clustered in specific, desirable jobs.
Here’s why. The work that AI is most directly capable of replacing is cognitive work performed at a computer — in offices and professional sectors. These are precisely the jobs that have grown the most as a share of the American labor force over the last fifty years. They are also the best paid and most coveted jobs in the economy.
She calls this Reality 2, the “messy middle.”
Recent layoffs at tech companies such as at Meta might make you think this disruption of cognitive work is happening at a radically accelerated pace. But I don’t believe it is. While current generative AI models can assist and augment experts in jobs dominated by cognitive work, it cannot replace the human judgement that is required to accomplish work of value such as designing a new software system, crafting a legal strategy, or designing a course syllabus.
Rather, tech companies overhired in the post-pandemic era and now that they need to invest in expensive infrastructure to survive they must cut where they can, and that’s in their bloated workforces.
Do AI agents replace people?
The near-term end and even disruption of the cognitive workforce is far overstated. AI agents aren’t as amazing as you might have imagined; they are just “internal automations with a language model bolted on.” Some examples:
Telehealth founder. Wanted “an autonomous AI receptionist that handles everything.” After an hour on a call I told her she needed a workflow that reads intake forms and routes them to the right clinician. We shipped it in six weeks. Saves her clinicians four hours a day. She paid me again last month.
Fintech client. Wanted a “fully agentic finance copilot.” What they needed was a script that reconciles ACH discrepancies before they hit the dispute queue. One model call, the rest plain code. Saved them a full ops hire.
Medspa chain. Wanted “AI marketing automation.” What they needed was a job that watches their booking system for no-show patterns and triggers a personal recovery message. Three steps. No agent. Booked 14% more revenue last quarter.
All of these software solutions were possible before the advent of generative AI models. Does AI possibly make it faster to build, because it can write the code and help a person decide on the system architecture? Yes. But it doesn’t make humans obsolete, even if they’re only doing thinking tasks.
I don’t disagree with Kinder that a messy middle might be coming, but like physical robotics that can replace nurses and food service workers, it is probably quite a bit further off in the future than pundits are claiming.