4 ms·
Sigh… It’s interesting that they blame the ML models, on the one hand it seems like a perfect example of ML gone wrong, on the other hand losses at this scale (
by d136o 5y ago
Sigh… It’s interesting that they blame the ML models, on the one hand it seems like a perfect example of ML gone wrong, on the other hand losses at this scale (hundreds of millions) don’t just signal that some data scientist’s model was not accurate enough… who’s running the ship over there? The ML model?
Also I’m really curious about the geographic distribution of their inventory… let’s do some Data Science on this blow up!
- nradov 5y agoIt's not really ML gone wrong, more like a misapplication of the technology. No one who understands ML ever claimed it could produce accurate forecasts of chaotic systems.
- SomewhatLikely 5y agoWeren't there headlines recently about a Google deep learning model producing the most accurate weather forecasts? Weather after all is the system from which chaos theory was first inspired.