Our understanding of the unification of jetted AGN has evolved greatly as jet samples have increased in size. Here, based on the largest-ever sample of over 2000 well-sampled jet spectral energy distributions, we examine the synchrotron peak frequency - peak luminosity plane, and find little evidence for the anticorrelation known as the blazar sequence. Instead, we find strong evidence for a dichotomy in jets, between those associated with efficient or 'quasar- mode' accretion (strong/type II jets) and those associated with inefficient accretion (weak/type I jets). Type II jets include those hosted by high-excitation radio galaxies, flat-spectrum radio quasars (FSRQ), and most low-frequency-peaked BL Lac objects. Type I jets include those hosted by low- excitation radio galaxies and blazars with synchrotron peak frequency above 1015 Hz (nearly all BL Lac objects). We have derived estimates of the total jet power for over 1000 of our sources from low-frequency radio observations, and find that the jet dichotomy does not correspond to a division in jet power. Rather, type II jets are produced at all observed jet powers, down to the lowest levels in our sample, while type I jets range from very low to moderately high jet powers, with a clear upper bound at L_300MHz_ ~ 10^43^ erg/s. The range of jet power in each class matches exactly what is expected for efficient (i.e. a few to 100 % Eddington) or inefficient (< 0.5% Eddington) accretion on to black holes ranging in mass from 10^7^M_{sun}_ to 10^9.5^M_{sun}_.