A billion-dollar question is whether precision medicine (aka personalized, "P4" or systems medicine) can substantially increase the utility of individualized disease prevention and population health 1,2. In this respect, the first results from the "Pioneer 100 Wellness Project (P100)" featured in last August's Nature Biotechnology issue, is a landmark 3,4. The study sheds light on an approach that has primarily existed as a vision and precedes the US National Institutes of Health's (NIH; Bethesda, MD) "All of Us Study", which will include a million participants in a similar scheme (http://www.allofus.nih.gov/). The researchers behind the study claim to demonstrate how measurement of personal data clouds over time can improve our understanding of health and disease and to identify "actionable possibilities", by which individuals can enhance health through preventive strategies. P100 is an exploratory study of associations in networks of biomarkers and risk factors established from large and dynamic data clouds of 108 participants. Over the course of a 9month period, participants underwent whole genome sequencing (yielding 127 polygenic scores for disease risks plus three copy number variations), three-times testing of metabolome (643 metabolites), proteome (262 proteins) and microbiome (4616 taxonomic units), 218 other clinical tests and measurements, as well as daily activity tracking via "quantified self" technologies. Associations between these biomarkers yielded a total of 3470 connections in a correlation network. The P100 study is a project designed to display the potentials for novel technologies and gather support for precision or P4 medicine (predictive, preventive, personalized and participatory), and it has gained prominent publicity in Nature 5. The P100 will also be scaled up to include 100,000 participants in the 100K Wellness Project 3. Against this background, we here discuss whether the P100 study actually supports the prospect of substantial benefits from datadriven disease prevention, and argue that it exposes severe challenges. The graphics of the correlational networks presented by Price et al. 3 do offer an interesting research potential for exploring connections in molecular networks and for identifying candidate biomarkers 3. However, as yet there are only a few examples where such This is a preprint version of a correspondence article published in Nature Biotechnology. Please see the following link for the final version: https://www.nature.com/articles/nbt.4210.