This talk presents an update on some speech-synthesis research I have been involved in since leaving CSTR, and a few thoughts on an interesting retro-future application. I begin by briefly discussing recent work on controlling emotion and speaker characteristics in TTS. The main body of the presentation then describes and evaluates a novel setup – called "cyborg speech" – that generates segment-level controllable foreign-accented (synthetic) speech with native prosodic characteristics for speech-perception research. This work was previously presented at ICASSP this year. Finally, I outline an upcoming investigation of the perception of synthetic speech, intended to demonstrate the viability of modern synthesis paradigms as speech-science research tools.