The shape of the primordial fluctuation spectrum is probed by cosmic microwave background fluctuations which measure density fluctuations at z similar to 1000 on scales of hundreds of Mpc and from galaxy redshift surveys, which measure structure at low redshift out to several hundred Mpc. The currently acceptable library of cosmological models is inadequate to account for the current data, and more exotic models must be sought. New data sets such as SDSS and 2DF are urgently needed to verify whether the shape discrepancies in P(k) will persist.