gpt-2 - AI Weirdness (Page 2)

AI recipes are bad (and a proposal for making them worse)

By Janelle Shane On January 31, 2020 - 3 min read

Pictured above is an abomination in the making, a lesson in why humans should never trust what a neural net says just because it’s based on math. It’s a neural net generated brownie recipe called Chocolate Baked and Serves [https://aiweirdness.com/post/159176882762/in-which-the-neural-network-gets-bored-halfway] , and its distinguishing

(Untitled)

By Janelle Shane On January 17, 2020 - 4 min read

lingthusiasm [https://lingthusiasm.com/post/190298658151/lingthusiasm-episode-40-making-machines-learn] : Lingthusiasm Episode 40: Making machines learn language - Interview with Janelle Shane > If you feed a computer enough ice cream flavours or pictures annotated with whether they contain giraffes, the hope is that the computer may eventually learn how to do these

AI Weirdness in 2019

By Janelle Shane On January 09, 2020 - 3 min read

Here’s a look back at a few of my favorite projects from 2019! 1. My book, You Look Like a Thing and I Love You I published my first book, a pop science look at how artificial intelligence works and why it’s making the world a weirder place.

Don't let an AI (even an advanced one) make you a cocktail

By Janelle Shane On December 31, 2019 - 3 min read

How far has AI come in a year? In 2018 I generated cocktails [https://aiweirdness.com/post/181110592232/dont-let-a-neural-net-mix-drinks] using textgenrnn [https://github.com/minimaxir/textgenrnn] and char-rnn [https://github.com/karpathy/char-rnn], two neural networks that learn to imitate text. I trained them on 1000 cocktails that Beth Skwarecki

The AI's Carol

By Janelle Shane On December 24, 2019 - 4 min read

In 2017 I decided to find out what would happen if I trained a neural net on 240 Christmas carols (collected by The Times of London and reader/neural net hobbyist Erik Svensson). The result? A neural net that was very confused about how Christmas works [https://aiweirdness.com/post/

Play AI Dungeon 2. Become a dragon. Eat the moon.

By Janelle Shane On December 06, 2019 - 6 min read

I wrote earlier [https://aiweirdness.com/post/187645450357/dungeon-crawling-or-lucid-dreaming] about a neural net-powered dungeon crawling text adventure game called GPT-2-Adventure [https://quicktotheratcave.tumblr.com/post/187432425523/shall-we-play-a-game-a-gpt-2-text-adventure] in which gameplay is incoherent and dreamlike, as you encounter slippery sign text, circular passages, and unexpected lozenge rooms [https://twitter.com/StrongBadActual/

Finest Pies

By Janelle Shane On November 26, 2019 - 4 min read

People wonder about what it would be like if a super-intelligent AI decided to place all of humanity in a realistic simulation. I wonder what it would be like if the simulation were built by today’s AI instead - whose computing power is somewhere around the level of an

How to begin a novel

By Janelle Shane On November 19, 2019 - 10 min read

Last year for National Novel Writing Month I trained a neural net called torch-rnn [https://github.com/learningtitans/torch-rnn/blob/valle-syllables/doc/flags.md#preprocessing] on 10,096 unique ways to begin a novel. It came up with some intriguing possibilities [https://aiweirdness.com/post/168051907512/the-first-line-of-a-novel-by-an-improved-neural] , my personal favorite

I trained a neural net on my own book

By Janelle Shane On November 05, 2019 - 2 min read

I spent the last two years writing a book on AI [https://www.janelleshane.com/book-you-look-like-a-thing]. It’s called You Look Like a Thing and I Love You, and today it’s finally out! And since training neural nets on things is a large percentage of my blog, of course

Halloween costumes by the neural net GPT-2

By Janelle Shane On October 14, 2019 - 4 min read

In my opinion, one of the best applications of neural networks is for generating Halloween costumes. Thanks to a dataset of over 7,100 costumes crowdsourced from readers of this blog, I’ve been able to generate Halloween costumes with progressively more powerful neural networks. In 2017, I used char-rnn

Tag: gpt-2

AI recipes are bad (and a proposal for making them worse)

(Untitled)

AI Weirdness in 2019

Don't let an AI (even an advanced one) make you a cocktail

The AI's Carol

Play AI Dungeon 2. Become a dragon. Eat the moon.

Finest Pies

How to begin a novel

I trained a neural net on my own book

Halloween costumes by the neural net GPT-2

Writing

Subscribe

Recent Posts

Minecraft with object impermanence

Bonus: In Which The Adventurer Attempts to Build a Website

Botober 2024

Bonus: "Ignore all previous instructions" gets weirder

An exercise in frustration

Bonus: A unicorn goes downhill

Follow

Minecraft with object impermanence

Bonus: In Which The Adventurer Attempts to Build a Website

Botober 2024