This guy's research was cited in a Nobel Prize, and after 2,300 job applications he can't even get a call back from a company that sells frozen crab cakes. His name is Jack Weis. He got his chemical engineering PhD from Princeton in 2024, and the one paying gig he's landed since then was helping train AI. It lasted less than three months. Here's how good his research was. Back in 2022, while he was still a grad student, Jack co-authored a paper that used a neural network to simulate water turning into ice, modeling how hundreds of thousands of atoms interact as it freezes. Scientists had basically written that off as too complex to compute. Two years later, when Hopfield and Hinton won the physics Nobel for inventing the building blocks of neural networks, the committee picked his paper as one of a handful of examples of what the technology could now do for science. So he graduates and starts applying. He goes after research jobs at drug companies and postdocs at national labs, and most of them never write back. He opens the search up to any industry, anywhere in the country, and keeps getting mostly silence. Somewhere in there he applies for a food prep job at the crab cake company, and they ghost him too. In January he finally gets paid. A startup brings him on at $70 an hour to write chemistry problems and grade AI answers. The hours start around 20 a week and slide down to 4. Then whole weeks go by with no work at all, and in mid-March he finds out he's done through a group chat message. If you've never looked at how these AI training gigs work, this part will bother you. A lab that wants its model to get better at chemistry pays experts like Jack to write a few thousand hard questions with perfect answers. Once the model has learned from that batch, the lab has what it paid for, so the budget moves on to physics or law or whatever subject is next. You're getting paid well to hand over the one thing that made you hard to replace. When you zoom out it's kind of brutal. A Nobel committee looked at Jack's work and saw proof that AI could change science, and so far the main thing the AI industry has wanted from him is a couple months of answer keys. He's 30 now and living with his parents back in Salt Lake City. He's planning to take the LSAT in November, and in the meantime he's building a drone that uses AI to remove invasive weeds.