10
Debate: training AI on scraped data vs paying for curated datasets - which is more ethical?
I spent about 6 hours last week trying to clean up a dataset for a small image recognition project and kept hitting weird biases in the free scraped stuff... then I looked at what a paid dataset cost and it was like $800. Am I wrong for thinking the free route is still fine for most hobby stuff, or does the ethical side really matter that much to you guys?
2 comments
Log in to join the discussion
Log In2 Comments
the_wren8d ago
The free scraped data is like dumpster diving for dinner.
1
riley_bell4d ago
Oh man, I gotta push back a little on that! A lot of the free data out there is actually really helpful for training models to understand stuff like common sense or basic language patterns. Sure it's messy sometimes, but it's not all trash - it's more like foraging for ingredients that can make a solid meal if you know how to cook.
1