The Cannon is a data-driven method for determining stellar parameters and chemical abundances from stellar spectra in the context of large spectroscopic surveys. It holds exciting promise for bringing qualitatively different stellar surveys onto a consistent physical scale, essential in the large-survey era. The Cannon uses no physical stellar models, is very fast, and achieves comparable accuracy to existing survey pipelines using significantly lower SNR spectra; it requires only a set of objects observed in common between the surveys. In this talk, I will introduce the methodology and present recent applications to (1) inferring labels from APOGEE spectra, (2) spectroscopic determination of mass and age, and (3) bringing APOGEE and LAMOST onto the same physical scale, enabling the first measurement of alpha-enhancement for 450,000 LAMOST giants.