Transcriptional output scales genome-wide with cell-size and growth rates. As a result, cells of different size and physiological states contain different numbers of mRNAs and proteins. How such remarkably coordinated regulation is achieved, remains largely mysterious. Moreover, little is known about the impact of such global changes in molecules numbers on regulatory networks. RNA and protein numbers exhibit cell-to-cell variation due to the stochastic nature of gene expression. Levels of noise in gene expression are linked to molecule numbers, and therefore are connected to growth rate and cell size. Yet, whether and how these basic features of the cell affect gene expression noise remains vastly unexplored. We have used single-molecule counting of mRNA to explore the relationship of noise in gene expression with cell-size and growth. We find that cell-size is often the major source of correlation between expression of gene pairs. Surprisingly, we also uncovered non trivial regulation of major regulators of transcription. Mathematical modelling of single-cell data confirms earlier finding that transcription rates adjust to cell size. Finally, we have sampled gene expression in single proliferating cells by RNA-seq, providing a glimpse into the molecular heterogeneity of fission yeast cells as a function of growth.