## Progressive Summary **Layer 3**: **Multi-year training pipelines postpone the only test that matters — do you like the actual job? — until exit options have already expired; front-load contact with the terminal role, not the training.** **Layer 2**: "...train for 14 years after college to become a surgeon only to realize... I don't like being a surgeon." — Dr. Arghavan Salles **Layer 1**: Viral TikTok self-report by surgeon Arghavan Salles (MD, PhD), captured Feb 2023 via a Twitter repost thread. Fourteen years of post-college training ended at 42 with the discovery she disliked the destination. ## Atomic Insight A long pipeline tests your ability to complete training, not your fit with the job it leads to. Trainee life and practitioner life are different jobs: different autonomy, different stakes, different daily texture. The preference signal — "do I actually want this?" — arrives only on exit, after sunk costs, identity-lock, and expired alternatives have made quitting maximally expensive. Two defenses: - **Sample the terminal state, not the pipeline**: shadow the post-training role; ask people five years past the finish line about their ordinary Tuesday. - **Pre-legitimized exit checkpoints**: schedule decision points where quitting is framed as a valid outcome, before each escalation of commitment. ## Cross-Domain Applications - **Product development**: waterfall delivers the market-fit test at the end; shipping early is the same fix. - **Careers**: [[Credentials Prove Instruction-Following — Skills Require Building]] — credentials measure pipeline completion, not destination fit. - **Decision design**: [[The Drift Trap - Default Paths vs Intentional Design]] — pipelines are powerful default paths; checkpoints reintroduce intention.