Follow

"I worked for a company that used [...] a machine learning algorithm that combined the personality test results with data from people's resumes to determine which candidates were the best fit [...] The system saw we had lots of managers named Michael, so it kept recommending managers named Michael"

@trickster Obviously the name Michael means "good manager" in ancient Hebrew

@DeeUnderscore Legally changing your name to "Michael Goodmanager" as a career-hack.

@trickster Breaking into the company's servers to steal their machine learning model and using it to generate the most hireable name, so that you can rename yourself to it

@DeeUnderscore I could go depressing with it (you messed up, and now you have to hack your way to a lifeless job), or stealthy (this is a hook part of a larger story to infiltrate a megacorp).

But we're already full on bizzare/depressing hooks from running Paranoia.

@trickster the company later renamed itself Michael, Michael and Michael .co.

@trickster Yes, people confuse fit to the (company) culture you have, and the one you wish you had...

@trickster love too include names in my machine learning algo as if there's any connection between name and ability

@trwnh "well, how are you supposed to separate the men from the women then??"

@padraic_padraic "You have to keep the names in, otherwise, how are you gonna separate the MALES from the WOMZ??"

@trickster
sorry but that (if it's not supposed to be a joke) sounds like total BS. I don't even see how you would translate names into features for ML classifiers.

@Maltimore What I like to believe is that the initial poster had a hazy at best understanding of what ML meant, or what their system did.

@trickster the sad part here is always the conclusion:
"The system kept recommending managers named Micheal. So machine learning algorithm are useless."

Instead of, you know, admitting the input parameters were just garbage, And that's not even the hidden biases that machine learning algorithms pick up and get people confused, because they're supposed to be "100% objective".

Sign in to participate in the conversation
Mastodon for Tech Folks

The social network of the future: No ads, no corporate surveillance, ethical design, and decentralization! Own your data with Mastodon!