StatsForecast: MSTL - confidence level gives same result as the point estimate - bug

[lken-stripe]

What happened + What you expected to happen

If you follow the example in the doc, the confidence intervals are all the same
https://nixtlaverse.nixtla.io/statsforecast/docs/models/multipleseasonaltrend.html

In the table where you see

MSTL | MSTL-lo-95 | MSTL-lo-80 | MSTL-hi-80 | MSTL-hi-95
-- | -- | -- | -- | --

all the values are the same

Versions / Dependencies

just statsforecast package 1.7.4 (latest)

Reproducible example

# see doc https://nixtlaverse.nixtla.io/statsforecast/docs/models/multipleseasonaltrend.html

Issue Severity

High: It blocks me from completing my task.