Elvintree

And then the gorgeous purple Decepticon one: 

also, it’s not red carpet but I refuse to ignore these beauties:

Geraldine Pilgrim - Dreams of a Winter Night

hey! i’m kicking off the #ADHDInvasion hashtag for adhd awareness month with a comic about CONSEQUENCES, and how my lifelong failure to react to them has been a huge source of frustration for me and for the ppl in my life who rely on me. even worse, it makes me seem apathetic or careless to others, when in reality it weighs really heavily on me - i just struggle to show it.

the best way i’ve found to deal with this is accepting that shame-based motivation DOES NOT WORK AT ALL for me (which is hard to do, when it’s all that’s been modeled for you by parents/educators/bosses) and try to replace fear/shame based consequences with positive outcomes: i.e. instead of “everyone will be so disappointed in me if i don’t do this,” sometimes it’s more helpful to think “if i do this on time, i’ll feel so relieved, and everyone will be happy that i pulled through.”

not writing

writing

people reading their stories

Humans have pushed against the perfections of electronics for years. One of the ways they’ve done this is by designing effects pedals meant to distort the sounds of music. The pedals usually have evocative names (“Morley Emerald Echo”, “Happy Snake Lucifuzz Fuzz”, “Lovepedal Purple Pickle Vibe”). I had never really heard of these until Will from 608device scraped 669 pedal names from harmony-central.com (with input from the stompbox fans at ilovefuzz.com). Could a neural network learn to imitate them?

If I trained a neural network from scratch, there probably wouldn’t be enough data for it to do more than memorize the short list of names and spit them back at me. So, I turned to a pretrained neural net called GPT-2, which I’ve used before. From the time it spent looking at huge sections of the internet, this neural net starts with a knowledge of which letters and words tend to go together in English. Using a colaboratory notebook set up by Max Woolf, I started with the pretrained GPT-2 neural net and then trained it for just a little bit on the list of effects pedals. 

It is? kind of? spookily good at this?

A Tale Of Two Reverb Systems A Room Like No Other Best Damn Bass Chuck Collins TinyTone Octifer Meat Loaf Real Sweet Reverb Boiling Point Coffee Deep Dish Fuzz Cloudthunder Screaming Bird Tremolo Damn Fine Aluminum Finger Wah Amounts Vibe The Horribly Overgrown CIA Chorus  Release the Buffalo FLYBOARD Glamour Box Booooooom Blood Orange Overdrive Beelzebub Wet Screamer Ear Flanger Bigfoot Audio Radium Nitrate Distortion Boogie Woogie Nonesuch Tones Embryo Chorus Mr. Bubblegum Powered By God Tube Distortion

Although some of its names were a bit weirder.

Bat Face Down Under AshesToYourFingerBox K.S.Aji Sea Slug Chorus Big Chocolate + Yellow Magic Delay Big Cheese Big Cheese Box of Peas Bite Outside The Bacon Battery Me And A Big Bit of Cheese Dentures Distortion Monster Dish Blondie Moose Ant Farm Chorus/Vibrato Anneliese Gillenamy The Nightlight Feminizer Achtung Baby TinyMuffin Boost Akahashi Harp of Confusion Barton Junior Finger Pan Gelato Espresso Machine Bare Knuckle Biochemistry Professor, Neuroscientist And Librarian

And these are… frankly a little unsettling?

Metalliv Robot Zookeeper Machines Are People Too Dangerous But Not Unbearably So Beneath The Planet The Human Genome May Be Hacked Am I really that Transhuman Love and Sex Are A Mercy Clause Atomic Brain Check Thank God We’re All Flying Again Disastrously Varied Mental Model Doomed To Think Or Predict in Chord Form Dazzling So Beautiful Yet So Terrifying Electric Sheep Fuzz Empowered By the Power Of Mind Control Misanthropic Fondue From the Oasis of Luxury Krautrock Parlor A Shockingly Simple Method To Turn The Phrase, “I want this!,” Into Something Still Not Quite Envelope-Less Wheezy And Dreadfully Close To Normal