The Hire Is Almost a Side Effect. The Real Value Is Everything Else the Challenge Produces OpenAI just launched "Parameter Golf." A public challenge where anyone can build the most efficient AI model under strict constraints. No resume. No application. Prove you can do the work. Top performers get recruited into $500K+ roles. Wendy's ran a version of this play (I wrote about it a few weeks ago). Their "Chief Tasting Officer" contest got thousands of people filming themselves eating Wendy's and posting to TikTok. Cost? One $100K contract. Value? Thousands of branded videos they never had to produce. The friction is the filter. And the hire is almost a side effect. OpenAI reviews hundreds of novel approaches to model efficiency from every submission. Free R&D they never commissioned. Wendy's got free content and a news cycle that hit Forbes. One mechanism, four outputs: a hire, free work product, earned media, and a public signal that your bar is high. This works outside tech. You need a maintenance director for 2,000 apartments. Publish an anonymized 90-day maintenance log and challenge candidates to diagnose what's broken. The top submissions are free consulting reports on your actual operation. Need a forensic accountant? Publish fictitious financials with embedded irregularities. Give candidates 48 hours. The person who flags something you didn't even plant? That's your hire. The obvious objection: won't people just dump the challenge into AI? Some will. You'll know immediately. I see this constantly. People send me AI-generated analysis on deals, and the surface looks polished. Structure is clean. Vocabulary is right. But the insights are generic. AI misses what only an insider catches: the vendor relationship that explains inflated costs, the seasonal pattern that signals a capital problem, the lease comp that doesn't fit the submarket. Unless AI has been trained on your domain with your data, it produces confident work that an expert spots as hollow in 30 seconds. That's the point. You need the domain expert reviewing submissions. Someone uses AI to organize their thinking and still delivers real insight? Great, they proved they can use the tools. AI did the thinking for them? You'll know. I'd know in my field. You'd know in yours. Resumes were always a proxy for capability. Proxies work until they're cheap to counterfeit. A well-designed challenge reviewed by someone who knows the work is the one filter AI can't fake its way through. What challenge could you design for your next open role? https://lnkd.in/ejk3TuNg