This talk first reflects on what genetics can offer the social sciences and vice versa, but also the advantages of working together. It delves into examples that have produced novel measures and better prediction of behavioural outcomes such as reproduction, education and status. The talk then turns to why genetic research can benefit from interaction with the social sciences, drawing attention to how heritability differs across birth cohorts and countries, that certain polygenic scores are highly correlated or vary by socioeconomic background or how a third predictor may be driving outcomes. The talk then stands back to question whether it is not only ethical but also technically plausible to use polygenic scores in policy interventions and beyond, including recent policy attempts. It then reflects on the impact of lack of diversity and representative samples in data (e.g., by ancestry, socioeconomic status, age, sex, country) on our findings. The talk concludes with several promising new large data collection efforts that contain novel measures and focus on hard to reach populations.