AI Can Make You Fast Before It Makes You Good
AI can make a new employee dramatically faster before they have developed the judgment to know whether the output is actually good. Andrew Wang, CEO and co-founder of Valon, has taken a deliberately cautious approach: new hires first learn the job, then earn more freedom to delegate work to AI once they can recognize bad outputs, catch mistakes, and know when human judgment is still required. The framework is simple: LEARN Understand the job, what good looks like, where mistakes happen, and what the important exceptions are. EARN Show that you can judge the output, catch errors, and know when to stop and escalate. LEVERAGE Then use AI aggressively for research, comparison, summarizing, drafting, automation, and other work where it can multiply your effectiveness. This is not an anti-AI argument. It is a management argument. AI should multiply competence, not replace the need to develop it. There is also a business case: powerful AI models can be expensive, and using them indiscriminately for basic work can waste both money and compute. Learn the work. Earn the judgment. Then leverage the machine. Sonny Saggar, MD Professional Problem Solver NewPhysician.org
Episode notes
AI can make a new employee dramatically faster before they have developed the judgment to know whether the output is actually good. Andrew Wang, CEO and co-founder of Valon, has taken a deliberately cautious approach: new hires first learn the job, then earn more freedom to delegate work to AI once they can recognize bad outputs, catch mistakes, and know when human judgment is still required. The framework is simple: LEARN Understand the job, what good looks like, where mistakes happen, and what the important exceptions are. EARN Show that you can judge the output, catch errors, and know when to stop and escalate. LEVERAGE Then use AI aggressively for research, comparison, summarizing, drafting, automation, and other work where it can multiply your effectiveness. This is not an anti-AI argument. It is a management argument. AI should multiply competence, not replace the need to develop it. There is also a business case: powerful AI models can be expensive, and using them indiscriminately for basic work can waste both money and compute. Learn the work. Earn the judgment. Then leverage the machine. Sonny Saggar, MD Professional Problem Solver NewPhysician.org