@jeremyphoward@lexfridman Indeed; I haven't used it yet but will soon try it. I think it can resolve the RNN issues you mentioned. Nim is a delight of programming paradigms and compiles to super fast C bins. Swift is sort of a tough sell I think. Maybe the community will expand post-swift+TF though.
I'm done with supervised learning. Our deep learning models are largely mathematically brute-forced #AI parlor tricks. My future efforts will be learning and developing at @Numenta and for @neuralink, when available.
@RFrei46 @nocanundustan@voxdotcom haha nice try except we know trump outed epstine's pedophilia, never went to the island, wasn't "friends" with epstine by any stretch of the imagination, and has put more pedos in prison than any other president.
Putting in my vote for #NetflixZnation | I'm constantly streaming while I work. All day, everyday. Point is I've seen a lot of series and Z Nation is easily in my all-time top 5. Circling back to Season 5 and realized no 6. Please @Netflix, it feeds my soul.
Amazing #MachineLearning / #neuroscience interview! | Deepmind's DNC for #AI was practical for now but the cortex is the future | Also thought of Gun / @marknadal along the way when he described their model's composite data structure. | https://t.co/zYCcQnAKrU
@marknadal I'm so ashamed of so many of my so-called peers in Silicon Valley. Idk wtf the cool lot of them are thinking. They've lost their damn minds.