I am just learning but I am guessing the issue is training is expensive, because you need to iterate many times, and do a massive bunch of matrix operations (an

I am just learning but I am guessing the issue is training is expensive, because you need to iterate many times, and do a massive bunch of matrix operations (and other operations) per iteration? Maybe too expensive to "keep up" with new data coming in? Does a single new piece of data mean you need to start from scratch, or are you "near enough" and so it needs less processing power to incorporate it?