New hires at this AI startup have to earn their AI privileges
Executive Take
Giving new hires unlimited AI access before they know the job creates workers who can't catch AI's mistakes. Valon's fix shows the real cost of AI overuse isn't just spend, it's judgment that never develops.
Executive Summary
Mortgage-software AI startup Valon now requires most new hires to learn their jobs without AI, until managers confirm they can spot AI mistakes. CEO Andrew Wang says overuse of costly AI models was creating "AI slop" and stunting learning. The policy is projected to cut annualized token spending from $15-20 million to $4-5 million. Engineers are exempt due to peer code review.
Why It Matters
HR and AI leaders should care because onboarding built around unrestricted AI access may be quietly eroding the skills new hires need to catch AI errors. This is a live template for gating AI access by demonstrated competence, not tenure.
Bizquad Perspective
Most companies chase AI adoption metrics while ignoring that unrestricted access can prevent the judgment needed to actually supervise AI, which Valon's 75% cost cut proves was pure waste.