It has been argued (Rudolph and Sanders 2001 Phys. Rev. Lett. 87 077903) that continuous-variable quantum teleportation at optical frequencies has not been achieved because the source used (a laser) was not ‘truly coherent’. Van Enk and Fuchs (2002 Phys. Rev. Lett. 88 027902), while arguing against Rudolph and Sanders, also accept that an ‘absolute phase’ is achievable, even if it has not been achieved yet. I will argue to the contrary that ‘true coherence’ or ‘absolute phase’ is always illusory, as the concept of absolute time (at least for frequencies beyond direct human experience) is meaningless. All we can ever do is to use an agreed time standard. In this context, a laser beam is fundamentally as good a ‘clock’ as any other. I explain in detail why this claim is true, and defend my argument against various objections. In the process I discuss super-selection rules, quantum channels, and the ultimate limits to the performance of a laser as a clock. For this last topic I use some earlier work by myself (1999 Phys. Rev. A 60 4083) and Berry and myself (2002 Phys. Rev. A 65 043803) to show that a Heisenberg-limited laser with a mean photon number μ can synchronize M independent clocks each with a mean-square error of .