Originally published at: https://boingboing.net/2019/10/16/data-is-overvalued.html
…
1 Like
I’m auto-deleting right now.
3 Likes
i switched to duckduckgo instead. 4d chess! *taps head*
3 Likes
Their search results are generally better too! 
1 Like
did it mostly to tell people to duck it themselves
1 Like
That may be, but I don’t think it follows from this line of reasoning. If I build a model based on data, I can maybe then discard the data after 90 days without losing any value, but that doesn’t mean the model stops being valuable after 90 days. Assuming I continue to update that model with data I collect, use, and then discard, I also doesn’t follow that someone else can build an equally valuable model with only 90 days’ worth of data.
Trivial example: it would be very hard to predict my Christmas shopping needs based on my purchase history from July 2019 to October 2019, but if you knew that last November I bought toys for a two year old and a newborn, and built a model using that data, you can probably guess I’ll be buying gifts for my now-one-year-older niece and nephew this holiday season.
1 Like
This topic was automatically closed after 5 days. New replies are no longer allowed.
