AI, Kya Bolti Tu? | Part Sechs

My experiments with AI stuffses

This is a series of posts to share my experiments with generative AI.

Making AI describe a news report in order to then create an image that could fit well for that report

I tasked Bard with summarizing a news report to assess its summarizing skills.

**Used The Onion because real news is mostly doom and gloom. 🥲

This is the summary that Bard came up with:

A group of friends gathered around their alcoholic acquaintance Jason Peck to have a frank but necessary talk about buying the next round of drinks. They were concerned that their buzzes would start wearing off if Peck didn't get the bartender's attention soon. They took turns describing their sincere concerns, but Peck just pretended not to hear them. The friends eventually gave up and went to the bar themselves, leaving Peck to contemplate his drinking problem.

The article is a humorous take on the challenges of dealing with an alcoholic friend. It highlights the frustration that friends can feel when their alcoholic friend's drinking starts to affect their own lives. The article also suggests that sometimes the best way to deal with an alcoholic friend is to simply walk away.

Here are some of the key points of the article:

  1. A group of friends are concerned about their alcoholic friend Jason Peck's drinking.

  2. They gather around him to have a frank talk about buying the next round of drinks.

  3. They are worried that their buzzes will start wearing off if Peck doesn't get the bartender's attention soon.

  4. They take turns describing their sincere concerns, but Peck just pretends not to hear them.

  5. The friends eventually give up and go to the bar themselves, leaving Peck to contemplate his drinking problem.


I then asked Bard for prompts that would describe an image that fit this description. I wanted to see how different styles would affect the results, so I tried a few different ones. I ended up selecting a few that I really liked.

✒️High Renaissance Painting: A High Renaissance painting of a group of friends confronting an alcoholic friend in a bar. The scene is set in a dark and smoky atmosphere, and the friends are all looking concerned. The alcoholic friend is slumped over the table, with a drink in his hand. He has a glazed look in his eyes, and he is ignoring his friends. The friends are all talking to him in hushed voices, and the overall mood of the painting is one of sadness and concern. The painting is rendered in the style of High Renaissance masters such as Leonardo da Vinci and Raphael, with attention to detail, realistic proportions, and idealized beauty.

✒️1970s Comedy Movie Style: A 1970s-style comedy image of a group of friends confronting a drunk guy in a bar. The scene is set in a smoky and seedy bar, and the friends are all dressed in casual clothes. The drunk guy is slumped over the table, with a drink in his hand. He has a glazed look in his eyes, and he is ignoring his friends. The friends are all talking to him in hushed voices, but he is too drunk to even hear them. The overall mood of the image is one of humor and absurdity.

✒️Graphic Novel: A graphic novel panel of a group of friends confronting a drunk guy in a bar. The scene is set in a smoky and seedy bar, and the friends are all dressed in casual clothes. The drunk guy is slumped over the table, with a drink in his hand. He has a glazed look in his eyes, and he is ignoring his friends. The friends are all talking to him in hushed voices, but he is too drunk to even hear them. The panel is drawn in a realistic style, with a strong focus on the facial expressions and body language of the characters. The overall mood of the panel is one of tension and drama.

Observations

  1. Bard's ability to understand the gist of a short and simple article suggests that it is capable of understanding complex concepts and generating images that are relevant to the text.

  2. The use of generative AI images to convey the points of an article is a real-world use case that could potentially replace stock photos.

  3. Stable Diffusion XL's ability to generate high-quality images from both short and long prompts suggests that it is a powerful model with a wide range of applications.

  4. The rapid development of generative AI in recent months is indicative of the field's potential to revolutionize the way we create and interact with images.

Write a comment ...

Ishan Manjrekar

Designs Games | Makes Podcasts | Clicks Pictures | Travels | Curious