That time a data scientist told me to trust the algorithm over my gut

I was helping a buddy set up a hiring filter for his small business in Denver, and this guy from a local tech meetup kept saying just let the algorithm sort resumes by keywords. I went with it for three weeks and we ended up interviewing three people who all lied about their experience on paper. The algorithm couldn't catch the bullsh*t because it just counted buzzwords like 'Python' and 'agile' without any context. Has anyone else found that these systems miss the obvious red flags a human would notice in five seconds?

3 comments

3 Comments

jessica7075d ago

People lean on these tools like they're magic when really they just do pattern matching. It's the same thing with those automated chat bots on company websites that can't understand a basic question if you don't say it exactly right. The algorithm only sees what it's been trained to see, not the stuff that's missing or doesn't add up. That's why you'll see stores use automated inventory systems that never catch a broken shelf or a messy aisle until a human walks by. We traded simple common sense for speed and now everyone acts surprised when the machine misses the obvious.

logan6585d ago

And yeah, that store inventory thing hits close to home. I watched a self-checkout kiosk refuse to let me buy a bag of oranges because it couldn't recognize the shape, but it happily scanned a can of soup I accidentally set down twice. People forget these tools are just fast guessers, not thinkers, so they miss the obvious stuff every single time.

charles_young925d ago

Fast guessers" is a good way to put it, but they're really just fancy calculators.