Hacker News new | past | comments | ask | show | jobs | submit login

I've talked to colleagues about this before but I find it's mostly a cyclical nature fuelled by tunnel vision.

Every good couple of years researchers come up with a good advancement or fresh concept that reignites the community. However all that happens till the next breakthrough is basic tweaks. The amount of junk papers that slightly adjust the method that gets a few decimal place improvement then call it a snazzy name would fill a mountain. Drawing blood from a stone. People get so obsessed with specific methods thinking its the new great thing they don't stop and think it's probably not the only way.

Marketing and media is the worst though for general public perception. The amount of times they would warp ML into a magical pangea. "ML will make you skinny!" I mean it's just mathematical approximations, been around a while.

Don't get me wrong, I like this field. I'm fortunate that I get to apply it to a problem that helps people but at times I just want to shout from the rooftops that it's not a God it wont make all your dreams come true. Then they get annoyed it doesn't and we a bunch of stagnation articles like this.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: