Recent evidence - in particular the hard X-ray spectra obtained by NuSTAR, and the large amplitude hard X-ray variability observed when ultraluminous X-ray sources (ULXs) show soft spectra - reveals that common ULX behaviour is inconsistent with known sub-Eddington accretion modes, as would be expected for an intermediate-mass black hole (IMBH). Instead, it appears that the majority of ULXs are powered by super-Eddington accretion onto stellar-mass black holes. Here, we will review work that delves deeper into ULX spectral-timing behaviour, demonstrating it remains consistent with the expectations of super-Eddington accretion. One critical missing piece from this picture is the direct detection of the massive, radiatively-driven winds expected from ULXs as atomic emission/absorption line features in ULX spectra; we will show it is very likely these have already been detected as residuals in the soft X-ray spectra of ULXs. Finally, we will discuss ULXs that do not appear to conform to the emerging ULX behaviour patterns. In particular we discuss the implications of the identification of a good IMBH candidate as a background QSO; and the confirmation of an IMBH/ULX candidate in the galaxy NGC 2276 via the radio/X-ray fundamental plane.