Self-Adaptive Learning through Teaching (SALT) is proposed to build an adaptive online social network to improve the learning process. SALT will utilize crowdsourcing, collaborative filtering and collective intelligence to design an adaptive system. It will monitor data through warehousing, analysis and mining to adaptively guide students to the most productive learning pathways. An agent-based simulation of SALT and its students will be used to study how the system converges over time and is affected by initial conditions.