just a note

[PGTBoos]

It's somehow funny, how at first things started with time series, RNN, LSTM, and now TransFormers, (GPU optimized LSTM) in just a few years. Currently, huge models perform amazing tasks (while using a lot of resources), although it's great to use those Large Transformers, I hope someday we (or an AI) will find a better way than those "predictive machines" to create the next type of AI, which hopefully will not require that much overhead. ( We near the number of neurons of the human brain in some of those models, yet the brain does do so much more, ..questioning current efficiency, transformers are not brains. ) That aside, this course is a great gift.