That would be awesome, but I'm not sure that sufficient computing power to manage this exists in something easily portable (I'm not certain it doesn't either, I simply don't know, but I do know that most neural network "deep learning" uses far larger and more complex hardware). OTOH, if we can keep something like Moore's law going for another decade or so, I expect that we'll definitely be able to have such a thing then.
Re: Well ...
Date: 2019-12-05 10:40 am (UTC)