neural networks - AI Weirdness

Left: the original hotel room. Right: the hotel room as recreated by DALL-E3 from Bard's description, full of skulls and skel

The spookiest Halloween scenes

By Janelle Shane On October 30, 2023 - 5 min read

Google Bard has the ability to describe images. But it turns out what you get depends a lot on how you ask. I gave Bard this image and the prompt "Please describe this spooky Halloween scene". On the right is the image I got when I took the

Botober 2023

By Janelle Shane On September 30, 2023 - 5 min read

Since 2019 I've generated October drawing prompts using the year's most state-of-the-art text-generating models. Every year the challenges are different, but this was one of the hardest years yet. Large language models like chatgpt, GPT-4, Bing Chat, and Bard, are all tweaked to produce generic, predictable

Drawing of a walrus with "imagine a tiny little walrus. you call them a 'snowbonk' and put them in the fridge so they chill"

Trolling chatbots with made-up memes

By Janelle Shane On September 24, 2023 - 2 min read

ChatGPT, Bard, GPT-4, and the like are often pitched as ways to retrieve information. The problem is they'll "retrieve" whatever you ask for, whether or not it exists. Tumblr user @indigofoxpaws sent me a few screenshots where they'd asked ChatGPT for an explanation of

Image is of a brown baby giraffe with no spots. AI gives long descriptions but fails to report that it has no spots.

AI vs a giraffe with no spots

By Janelle Shane On August 28, 2023 - 4 min read

On July 31, 2023, a giraffe with no spots was born at Brights Zoo in Tennessee. She's a uniform brown with pretty white highlights around her face and belly, like a Jersey cow or a white-tailed deer. Image recognition algorithms are trained on a variety of images from

Baby one-piece bodysuits with designs of Galaxies on Ice, Dandelion Underwater, and Fart Whale

Baby onesie designs

By Janelle Shane On August 04, 2023 - 5 min read

A reader wrote in a while ago with a suggestion: they were about to have a baby and wondered if I could use AI to come up with some new ideas for baby onesies. I can't find the letter any more, and I don't remember how

Non-native speaker essays misclassified as AI-written 48% up to 75% of the time vs 0%-12% for native speakers

Don't use AI detectors for anything important

By Janelle Shane On June 30, 2023 - 4 min read

I've noted before that because AI detectors produce false positives, it's unethical to use them to detect cheating. Now there's a new study that shows it's even worse. Not only do AI detectors falsely flag human-written text as AI-written, the way in

Emoji include some kind of baby seal-horse and a sun-lion-eagle griffin.

The emoji of the future

By Janelle Shane On June 17, 2023 - 2 min read

Some of the recent image-generating models have this thing where they can fill in the blank parts of images. It's handy when you want to show them exactly how to give you more of the same. Like these animal emoji. See if you can tell which ones I

ChatGPT confidently losing at tic-tac-toe

Optimum tic-tac-toe

By Janelle Shane On May 26, 2023 - 2 min read

ChatGPT text can sound very knowledgeable until the topic is something you know well. Like tic-tac-toe. Once I heard that ChatGPT can play tic-tac-toe I played several games against it and it confidently lost every single one. Part of the problem seemed to be that it couldn't keep

ChatGPT describes code that draws a pink pony but actually produces a pink pig face

Chatbot, draw!

By Janelle Shane On May 05, 2023 - 3 min read

I'm interested in cases where it's obvious that chatbots are bluffing. For example, when Bard claims its ASCII unicorn art has clearly visible horn and legs but it looks like this: or when ChatGPT claims its ASCII art says "Lies" when it clearly says

Chatgpt: Here's the ASCII art of the word "lies" (generates block letters that clearly read "sip")

What does this say?

By Janelle Shane On April 15, 2023 - 3 min read

Large language models like ChatGPT, GPT-4, and Bard are trained to generate answers that merely sound correct, and perhaps nowhere is that more evident than when they rate their own ASCII art. I previously had them rate their ASCII drawings, but it's true that representational art can be

Tag: neural networks

The spookiest Halloween scenes

Botober 2023

Trolling chatbots with made-up memes

AI vs a giraffe with no spots

Baby onesie designs

Don't use AI detectors for anything important

The emoji of the future

Optimum tic-tac-toe

Chatbot, draw!

What does this say?

Writing

Subscribe

Recent Posts

Botober 2024

Bonus: "Ignore all previous instructions" gets weirder

An exercise in frustration

Bonus: A unicorn goes downhill

Hidden 3D Pictures

Hidden sheep

Follow

Botober 2024

Bonus: "Ignore all previous instructions" gets weirder

An exercise in frustration