The piece dives into the concept of aligned recommender systems, which are designed to prioritize what users would truly appreciate weeks later, rather than just maximizing immediate engagement. The author reflects on his own experience building a job recommendation system around twelve years ago. The algorithm used was based on collaborative filtering, which identified jobs that users with similar interests liked. Surprisingly simple, this approach worked effectively because it aligned the interests of job seekers and employers, both of whom wanted successful job matches. In contrast, the dynamics of dating app recommendations present a fundamental challenge. While the best outcome is for users to find lasting partners and leave the app, this directly undermines the app's revenue model. An app that succeeds greatly earns only a short-term subscription fee before losing its customer. This creates a dilemma where companies can either align with user satisfaction and sacrifice profits or focus on continual engagement, leading to user disappointment. The dating app industry has largely opted for the latter, prioritizing revenue over genuinely helping users find love. The author appreciates the Forethought design sketch for aligned recommender systems, which aims to optimize for long-term user approval instead of short-term interactions. However, he notes that the real difficulty lies not in the technology but in the economic structures that shape user experience. His past experience in job recommendations succeeded because everyone involved had a shared goal, unlike the conflicting incentives found in dating apps.