Anil DashVerified account

@anildash that's not fair. plenty of companies A/B test any number of scenarios to find upper/lower boundaries. ad load for example.

@hunterwalk deliberate software crashes? Come on now.

@anildash it's like testing latency, no? We all know that no latency is best but *zero* latency has lots of costs, I'm sure *most* large..

@anildash ..internet companies have control groups at different points that *slow down* in addition to *speed up* their services?

@hunterwalk sure, I know the ways different performance vectors are used for testing. This is different, in obvious ways.

@anildash i think you need to give engineers LOTS of freedom in % testing to get data for different hypotheses.

To be fair, I think Facebook has gotten a little more respectful towards users in the years since that test, but… well. Facts is facts.

“We had to make more frustrating & confusing to talk to your loved ones because one of our managers needed more data.” #imaginethemeeting

@anildash A depressing fact: even learning of this very deception still won't make users quit. There is no "choice".

@mrgan @anildash reminds me of that time Comcast intentionally degraded my internet connection so oh wait

@anildash Anil ... ... Do you really think this happened? Not doubting org's disdain, but this stinks.

@tnofuentes I find it quite believable, yes.

@anildash Where was that published?

@gilbert the link in the tweet.

@anildash isn't this the other way around: this is what people think about facebook? :) @FrankPasquale

@ilparone @anildash FB engaging in human subjects experimentation without consent or even notice http://www.newcriticals.com/facebooks-model-users/print … ==> disrespect