Honestly, this kind of curve isn’t meaningful. It’s really misleading, actually. It assumes that the detected number of cases each day accurately represents the actual number of new cases each day, when it very definitely doesn’t.
Taiwan did very little testing for a long time followed by a bit more testing in concentrated areas for 3 days, and the observed curve is an artefact of that. Trying to extrapolate these 3 days of slightly increased testing out for multiple days just isn’t valid.