Asking the Large Language Models (LLM) Chat Bots silly things is a good use of their time. Avoiding any requests for factual information will help keep things correct. This is because they don’t fact check anything. Ask for a joke, limerick, their ideas about the future or just an opinion for some chat that won’t be factually embarrassing for the Bot.
We asked Bard during the experimental phase this question that was prompted by some previous searches by other people.
Bard gives 3 versions of its output for each question input given. They are called drafts.
Write a limerick about a fish.
Draft 1
There once was a fish from the deep,
https://bard.google.com/
Whose scales were so shiny and sleek.
He swam all day,
And he ate all the prey,
And he never once had to speak.
Draft 2
There once was a fish named Jazzy,
https://bard.google.com/
Whose color was really snazzy.
She took a dip in the doom,
And swallowed a broom,
Then stumbled into her friend Ponzi
Draft 3
There once was a fish named Jazzy,
https://bard.google.com/
Whose color was really quite snazzy.
She took a dip in the doom,
And swallowed a broom,
Then stumbled into her friend Ponzi.
ChatGPT
We gave ChatGPT from open.ai the same challenge
Write a limerick about a fish.
There once was a fish in the sea,
https://chat.openai.com/chat
Whose scales shimmered so brightly.
With a flick of its tail,
It would dart and it sail,
A majestic sight for all to see.
You are presented with only one result when using ChatGPT but there is a button to ‘Regenerate response’. This will reissue the request and the chat bot will have another attempt at the text generation for you. Pick the one you prefer or mix it up and take parts of one and parts of another. See our post on Springboarding with AI.
Your turn
We would love to hear your AI inspired limericks please comment below
Leave a Reply