Can't wait to read about BERT on the front-page!

"temperature=0 & sampler=deterministic is all you need".

Deterministic predictions are sought by those looking to offload their decision responsibility to AI, whether to lower perceived legal risk or otherwise. This is fake risk reduction, i.e. "risk theater".

Unfortunately, a deterministic prediction utterly fails to yield an uncertainty measurement which is critical to have in actual risk reduction. If you want the variance in measurement, it is vital to obtain multiple measurements. This also gives a confidence interval.