shouldn't they be training the models to detect the poachers and not the animals to be poached?
I get that it is cool to be able to ooohcrazy train a model to select between different species but like all tools this is one that can be used by both sides-- and assuredly will be.
Training a model to identify poachers is A: way easier B: waaay more actionable C: way less sexy for a news report and fun for nature lovers, and NVIDIA's shareholders, unfortunately.
Unfortunately for the pangolins C will probably always trump. . .
> Training a model to identify poachers is A: way easier
Maybe, maybe not. What if the poachers do super difficult things like a) wear a cardboard box, or b) attach some twigs to themselves, or c) walk like a tree?
I get that it is cool to be able to ooohcrazy train a model to select between different species but like all tools this is one that can be used by both sides-- and assuredly will be.
Training a model to identify poachers is A: way easier B: waaay more actionable C: way less sexy for a news report and fun for nature lovers, and NVIDIA's shareholders, unfortunately.
Unfortunately for the pangolins C will probably always trump. . .
They're so cuutte I love them!!!!