Increasingly, front runner patients and practitioners want to use state-of-the-art science for rapid lifestyle based cure of diseases of affluence. However, the number of new health studies per year (>500.000) is overwhelming. How to quickly assess state-of-the-art and use new opportunities for rapid patient DIY (Do-It-Yourself) health improvement? In order to develop a health literature hybrid AI to aid DIY rapid health improvement, we analyze user side functional requirements. A cross case design analysis is conducted for hypertension and T2D (Type 2 Diabetes), two major cardiometabolic conditions in our society. Our analysis shows that current DIY health support is ‘watered down’ advise, prone to medicalizing rather than empowering patients. We propose hybrid AI user requirements and discuss how a 2030 hybrid AI health support system can stimulate new ways of working in health and cure.