A Princeton Ph.D. Applied to 2,300 Jobs.He Still Can＇t Get One.－By Lindsay Ellis,WSJ

JACK WEIS likes to joke that he’s just another unemployed guy living in his parents’ basement.

That leaves out a few key details: The Salt Lake City native has applied to more than 2,300 jobs since he graduated from Princeton in 2024 with his Ph.D. in chemical engineering. And he still doesn’t have a job.

He first pursued corporate science jobs and prestigious postdoctoral fellowships at national laboratories. Then he cast a wider net, applying to software engineering and other roles. Eventually, he was getting ghosted for food-preparation jobs, including at a company that sells frozen crab cakes.

For a time, Weis got a gig training AI on his expertise, including in engineering and fluid mechanics. That work dried up too.

I first interviewed Weis late last year, the day before his 30th birthday. Covering careers and the labor market, I endeavored to meet job seekers in all 50 states, hoping it would help me understand why many people felt so dim about their futures. Just 34% of Americans surveyed by Gallup in August said it was a good time to find a quality job.

The picture remains murky. The latest jobs report fell far short of expectations. The unemployment rate remains low, but the percentage of unemployed people out of work for 27 weeks or more was higher this summer than it’s been for more than a decade, with the exception of Covid-19 and its immediate aftermath.

The promise of AI—and unease around its application—also contributes to an environment that is harder than usual for young, college-educated workers to find work, while the opposite is true for those without degrees. More than 70% of U.S. respondents to a Pew Research Center survey released last month said they expected AI would lead to fewer jobs overall.

After dozens of interviews across all 50 states, Weis’s story struck me for the way it embodied the deep freeze that has paralyzed the current labor market in the past few years—what economists refer to as a low-hire, low-fire environment.

Many people have jobs, but they’re not leaving them, limiting opportunity for others. Those who don’t have a job are locked out of the growing economy as if it were a fortress, leaving them puzzled by the headlines of low unemployment and an accelerating GDP.

Employees laid off from stable jobs with skills heavily absorbed by AI are far less likely to find jobs than they were several years ago, according to new research from the Federal Reserve Bank of Richmond.

Part of the struggle to get a job is due to the way AI has infiltrated the job application process itself.

As employers are inundated with hundreds of applications, many made with AI or facilitated by it, companies are likewise employing AI tools to screen applicants.

Weis tried to circumvent the abyss of online job portals and one-way interviews with AI bots by trying to make connections with people behind the scenes. Princeton alumni would take his calls, for example, he said, but then send him to their companies’ jobs’ websites. He anticipated his résumé becoming stuck in the technological ether once again, saying: “You feel your heart sink.”

THE SON of two physicians, Weis graduated summa cum laude from Rice University and enrolled at Princeton for his Ph.D. in 2018. His research on ice formation was cited by the Nobel committee when it awarded the Nobel Prize in physics in 2024.

He started looking for a job in earnest at a time when companies pulled back on hiring graduates after a pandemic hiring spree.

Weis studied chemical principles and molecular interactions, and he aimed to work in science roles at pharmaceutical companies. The response to numerous applications was mostly silence, with a few rejections.

He soon widened his search geographically, and by industry. He sent hundreds of cover letters and résumés through job portals and often didn’t get an interview. He networked with Princeton and Rice alumni. He scoured classified ads in the Salt Lake Tribune. He read PitchBook updates to find startups that recently got influxes of cash, thinking it might result in job openings.

His mother, Dr. Holly Carveth, said she looked at her son’s applications, wondering if there was a red flag that made employers hesitate, but nothing stood out. They seemed to be well written. “It’s quite surprising,” she said. “I don’t know how to put it together.”

As the months ticked by, Weis found himself stumbling on dates when the topic of work came up. At first, he said, he would tell the person across the table that he had recently finished his Ph.D. A year later, “it feels a little odd to say that,” he said.

Weis and many job candidates I interviewed described job interviews that happened without humans—the applicants made a video recording of themselves responding to questions posed digitally on the call. Such a process is increasingly common in the application process: More than 60% of U.S. job seekers surveyed by Greenhouse said they had been interviewed by the technology in the last year.

Not every missed opportunity was a result of automated vetting processes. Some assessments had remarkably tight deadlines, which Weis believed was an effort to thin the applicant pool. At one point, he set out on an ill-timed backpacking trip as McKinsey sent an assessment over email for a consulting role. As part of the assessment, the company prompted him to play a computer game in which he had to organize an ecosystem of plants and animals in a virtual forest. He had 72 hours to do so. In 48 hours of that window, Weis was in the real-life forest and didn’t see the note.（The company said candidates can ask for extensions if they encounter scheduling challenges.）

He completed the test when he returned and got rejected.

IN JANUARY, Weis snagged a gig paying $70/hour through Mercor, a tech startup that pays experts to train AI in their various career fields.

The company, valued at $10 billion last year, has worked with OpenAI and Anthropic. It hires gig workers in fields as far-flung as astronomy, creative writing, law and investment banking.

For each shift with Mercor, Weis logged on to a program called the Studio. His job was to come up with questions about chemical engineering and input them into a large language model. The AI platform would generate answers to Weis’ prompts, which Weis would then evaluate and then try to improve. Sometimes he worked 20 hours a week, sometimes four.

Coming up with prompts could be easy some days—“Explain this at a sixth-grade level”— or more tedious when it came to teaching AI about, say, technical specifications. Sometimes when he struggled to brainstorm potential queries, he asked the AI to adopt strange personas: “You are the ghost of Blackbeard. Explain the principles of operation of a nuclear power plant in 300 words.” The software recorded his screen as he worked.

There were a few days when the client, whose identity wasn’t disclosed, didn’t have work for Weis. Those days became more frequent. Eventually, the days became weeks.

In mid-March, Mercor cut Weis’s contract. He learned through a message sent in a Slack group with 810 people: “If/when the project continues, we will likely reach out to y’all! .”

“Friendship with Mercor ended, unemployment is my best friend now,” Weis quipped in an email to me.

Weis was invited to a wedding in France. The couple was one of the last of his college friends to marry, and he thought it was important to go. He used the Mercor earnings to pay for his ticket. Around that time, he wrote to me, he was seeing a “suspiciously high” number of ads to donate sperm.

WEIS WAS THOUGHTFUL about the strange economic dichotomy many people in his position were experiencing. Since Weis defended his dissertation two years ago, the Dow Jones Industrial Average closed at record highs more than 50 times.

Staying in his childhood home, Weis has saved money and made some investments. He didn’t take on debt for his postgraduate degree—doctoral students are paid to do research rather than paying large sums in tuition.

“It’s cool that my stocks keep going up, I’m not mad about that,” he said earlier this year. “But it does make it a whole lot stranger.”

His parents compare notes with others whose adult children are struggling to find work. They remain shocked by the situation, and worry about their son’s ability to put down roots and one day have a family.

“We never dreamed he would have issues. He’s an Eagle Scout, he helps others, he’s environmentally conscious, he’s physically fit,” his father, Dr. John Weis, told me. “He’s smarter than I am, he’s more articulate than I am. He’s got an encyclopedic knowledge of history.”

Weis wonders if he should go to law school. He signed up for the LSAT in November. He has gotten involved in the Utah GOP and has sent petitions for rule making to federal agencies, including one urging the decoupling of employment and health insurance. (If fewer people stay in jobs for insurance, he reasons, there could be more openings.) He is also trying his hand at entrepreneurship. He used AI to design a drone with LED lasers to kill myrtle spurge, an invasive weed. “If I truly believe I have some value, I should go out and get it,” he says.

He still applies to new fellowships as they come up. In October, he unlocked his job-application account at the Los Alamos National Laboratory and realized that a role he applied for in 2024 had finally been filled.

For two years, the status had remained static: “Review in progress.”