David Weinstock opened with a confession: one of his best-performing LinkedIn articles was written entirely by an LLM. He felt proud of it, until he sat down to write and couldn’t. He called it writer’s block. Really, he’d outsourced his thinking.
His argument: friction is where we learn and build judgment. AI can remove it. The question is when it should.
Chore or climb. Some work is a chairlift: take it and don’t look back. Scheduling interviews is a chore. Other work is a climb worth keeping. Explaining your numbers to a board or CFO is one, because you won’t have a prompt in that room.
New recruiters need the climb. Experts can hand sequencing to an AI sourcing agent because they know what good looks like. A new recruiter who never writes a sequence can’t judge one. (His question: if every new recruiter starts on autopilot, who’s the expert in 5 years?)
Prompt-and-paste has real costs. David asked ChatGPT for a recruiter posting in Stamford, CT. It left out the pay range Connecticut’s new pay transparency law requires as of October 1. He pointed to DaVita’s $298K fine in Colorado as the price of skipping that step.
The AI sandwich. Think first and add context. Let the model work. Then interpret, validate and finish it yourself.
He’s not against AI. He’s against AI slop. His challenge to the room: name 1 climb you’ll keep and 1 chairlift you’ll take.
