This paper derives confidence intervals (CI) and time-uniform confidence
sequences (CS) for the classical problem of estimating an unknown mean from
bounded observations. We present a general approach for deriving concentration
bounds, that can be seen as a generalization and improveme