Concise Is Nice: The Perfect AI Text Length Every Time

I ran the numbers so you don’t have to

Bernard Builds
Generative AI
Published in
4 min readApr 29, 2024

Man measuring lines of text on digital screen

Do you remember your high school days?

You had to write an 800-word long essay…

And if your essay only had 600 words, you’d fail. If it had 799 words, you’d fail.

The only way to pass was to submit an essay with at least 800 words.

I wish AI would work like this, but it doesn’t. For some reason, it doesn’t care about the size (ba dum tss).

You can say how many words it should have. How many characters to include, but the result is still a couple words short.

Many people get frustrated about this problem, so I’m trying to crack it once and for all.

Here goes nothing…

How to Control Length of AI-Generated Text

Tools like ChatGPT, Claude, or Gemini don’t understand the numbers. And that’s where the problem is.

You ask for a Facebook post that has 400 words, but AI responds only with 200 words. You ask to make it longer and it responds with 204 words. You ask again and it responds with 150 words.

I hope this is going to change in the future. But for now, we need to try something hacky.

Create an account to read the full story.

The author made this story available to Medium members only.
If you’re new to Medium, create a new account to read this story on us.

Or, continue in mobile web

Already have an account? Sign in

Published in Generative AI

All the latest news and updates on the rapidly evolving field of Generative AI space. From cutting-edge research and developments in LLMs, text-to-image generators, to real-world applications, and the impact of generative AI on various industries.

Responses (4)

What are your thoughts?

For me, “brief” and “concise” mean the same thing.

Oh, I love experiments like this, and the results are genuinely quite intriguing. Same as yourself, I would have thought brief and concise would be remarkably similar in the output, but seemingly not.

Thank you so much for running this experiment…...

--

Would be very interesting to also get the standard deviation for each terms, this would help to understand the precision of the adjective. I would expect some adjective to have small variation, and some to have bigger one.

--