This week's issue is all about machine learning, or deep learning that lets computers explore real life data and use it to track, create, sort, and filter the world around us with an almost human-like ability. Of course, there are cracks here and there, which can as amusing as it is relieving.
This one is somewhat math-y but absolutely fascinating. Skim to the examples of what it produces--impressive, realistic Shakespeare, Linux code, Paul Graham essays.
This is absolutely hilarious. Clickotron is worth a gander--the results are as amusing as they are sublime.
A friend and I played around with this, with some pretty hilarious results. It's particularly fun when the auto-reply will generate responses to its own responses, like: "That's weird." "No, it isn't."
I like this article. I could spend hours upon hours exploring abandoned Minecraft servers, obscure areas of Second Life, the detritus of digital worlds.
OK, so this has nothing to do with machine learning, but it's great.
Deaf Girl's Glitchy Music Recs
Nico / Deaf Girl puts together glitch-inspired music track recommendations. He's awesome, and his picks are great. Annotations are his!
Nico! These tracks are seriously banging! Thank you!
"One small step for selfies, one giant leap for cheap deep-learning autonomous video-surveillance drones."
Interesting aesthetic contemplation on different depictions and perceptions of robots in popular culture.
Yeah, this isn't that surprising.
I am a data scientist, and this is why Minority Report will never happen."