3 ms·

Look at this majestic beast GPT-4 just drew! https://imgur.com/a/ogZ6qI0 https://imgur.com/a/ogZ6qI0 Interestingly, I think you often get better results with C

Look at this majestic beast GPT-4 just drew! https://imgur.com/a/ogZ6qI0 https://imgur.com/a/ogZ6qI0 Interestingly, I think you often get better results with ChatGPT web interface than the API (the OP used a system message), but for a funny reason. Because ChatGPT typically first responds by insisting it can't actually draw, it ends up accidentally chain-of-thought prompting itself into drawing better. "I can't draw, but I can describe a detailed drawing for you in words." So instead of winging it with the SVG code and getting confused, it goes into it with a plan. This is how it described the unicorn I linked: __User__ Draw a unicorn! __ChatGPT__ As an AI language model, I am unable to create visual images directly. However, I can describe a detailed image of a unicorn for you in words. Picture a magnificent unicorn, standing tall and proud in a serene setting. The unicorn's coat is a shimmering, pure white, and it glistens under the warm sunlight. Its body is strong yet elegant, with a graceful arched neck and powerful, slender legs. The hooves are cloven and emit a faint, otherworldly glow. The most striking feature of the unicorn is its majestic spiral horn, which emerges from the center of its forehead. The horn is long, slender, and gently tapers to a sharp point. It is iridescent, with colors swirling together like a mesmerizing opal. The unicorn's mane and tail flow behind it like silken strands, shimmering with an ethereal pastel rainbow of colors. Its eyes are large, expressive, and full of wisdom, with a depth that speaks of untold secrets and ancient magic. The unicorn's ears are pointed and alert, picking up the slightest rustle of leaves or the gentlest of whispers. The unicorn's expression is one of serenity, exuding an air of peace and tranquility that surrounds it like a gentle embrace. Though it is a creature of great power, the unicorn carries itself with a sense of humility and grace that inspires awe and respect in all who behold it. __User__ What's the SVG for the image you just described? __User__ continue (to finish the file) Some more examples here, but with many iterations of "Improve this" over and over: https://twitter.com/jonathanfly/status/1639657098605740035 https://twitter.com/jonathanfly/status/1639657098605740035

int_19h 3y ago
Here's a somewhat more convoluted example: https://imgur.com/a/yhIGvho https://imgur.com/a/yhIGvho The first image is what came out of the initial prompt. Then I tried to get it to improve things and to add animations; the second image is the end result (and yes, the sun is animated as described). Here's the complete chat log: https://gist.github.com/int19h/18bb9d96b4d8418da2531513970d8006 https://gist.github.com/int19h/18bb9d96b4d8418da2531513970d8... It's obviously far from perfect, but the fact that it can do this much already - meaning that it understands broadly how the objects would relate to each other spatially for something like this - is already extremely impressive IMO, given that its training and inputs are entirely textual.
- JonathanFly 3y ago
  Totally agree. The samples in the OpenAI paper didn't seem that impressive and there were so few of them, I assumed it must not be very good. But even the one-shot version of the scenery you linked is great. Very efficient use of SVG features.
user_named 3y ago
ChatGPT has the InstructGPT model and additional prompt templates on top if it. I'm not sure if the GPT api includes those, I would guess not as that would interfere with some of the use cases you may have. So that's probably why output would differ.

int_19h 3y ago

Here's a somewhat more convoluted example: https://imgur.com/a/yhIGvho https://imgur.com/a/yhIGvho The first image is what came out of the initial prompt. Then I tried to get it to improve things and to add animations; the second image is the end result (and yes, the sun is animated as described). Here's the complete chat log: https://gist.github.com/int19h/18bb9d96b4d8418da2531513970d8006 https://gist.github.com/int19h/18bb9d96b4d8418da2531513970d8... It's obviously far from perfect, but the fact that it can do this much already - meaning that it understands broadly how the objects would relate to each other spatially for something like this - is already extremely impressive IMO, given that its training and inputs are entirely textual.

JonathanFly 3y ago
Totally agree. The samples in the OpenAI paper didn't seem that impressive and there were so few of them, I assumed it must not be very good. But even the one-shot version of the scenery you linked is great. Very efficient use of SVG features.

user_named 3y ago

ChatGPT has the InstructGPT model and additional prompt templates on top if it. I'm not sure if the GPT api includes those, I would guess not as that would interfere with some of the use cases you may have. So that's probably why output would differ.