It could happen by accident. They insert a new feature into the app which increases load times a lot. It A/B tests better. Not because the feature itself is good, but just because increased load times do better.

This seems especially devious if it's true. (I seriously doubt it is.)

What makes you doubt it?

We know these companies have people working to exploit psychological quirks/flaws to increase engagement. We also know that ethical concerns are somewhere below the bottom of the list of considerations when creating/implementing new "features".

Given the things we know, I'd be more surprised if this wasn't true.

Oh the cute naivete of your innocent optimistic doubts…

This sort of thing is well-known, so doubting this particular case is not a well-founded doubt.

How is this "well-known"? Do you have a source?

I don't mean this exact case, I mean stuff like intentional slowing down of results. People don't trust results that come too quickly. We like to feel that the computer worked hard for it. E.g. Airline prices or TurboTax results.

https://www.theatlantic.com/technology/archive/2017/02/why-s...

A more detailed article: https://www.1843magazine.com/features/the-scientists-who-mak...

I stopped reading the article at this point. This is a big claim to make without any proof.

I also stopped at this point. Seems like an absurd assertion with no evidence.