A/B testing, aka automatic remote gaslighting at scale
@foone
Yes but also:
There is something deeply democratic about A/B testing. Sadly there is so much capitalism plastered over it that it is hard to see.
@freemin7 @foone It's more scientific than democratic - most of the people in an A/B testing setup don't get to choose A and/or B.

@AT1ST
Democratic isn't about voting only.
It is about upward control and political equality.

In situations where following stated preferences leads to worse results then revealed preferences insisting on stated preferences is ... like insisting on a winner takes all rule.

@AT1ST i also deeply object to use of the word scientific here. I don't see how you could say it is scientific at all.

@freemin7 In the sense of double blind testing a lot of the time - neither side really knows at serving time who gets A or who gets B.

And the way A/B testing determines the results is more opaque than normal democratic means.

@AT1ST
A/B testing *is not* science. A/B testing is engineering.
Not research but search and optimization.

A/B testing is less opaque than the US voting system in terms of upward control. If you mean "less opaque" as in worse audibility this is a consequence of the employment of the method under capitalism not the method itself.

@AT1ST
A/B testing is not science because it doesn't form a model.
Experiments in A/B are not run based on a hypothesis constrained by past experiments which are invalidated or not based on an observation.

A/B tests give an ordering between alternatives on some metrics. A/B testing will never tell you that the dice are intransitive. Sciene will.

https://en.wikipedia.org/wiki/Intransitive_dice

Intransitive dice - Wikipedia