How I use LLMs to write

What does this mean?

This means that the content is 100% human written. With all the possible imperfections like grammatical errors and typos.

The content is 100% written by me, but a LLM made the proof reading, removing typos and grammatical errors. But the form is not modified from AI.

The content is a mix of sentences written by me and written by AI. During the proof reading, the AI likely changed the form of some sentences or made entire paragraphs.

The content is completely generated by AI. I made only proof reading and prompt the topic.

Introduction

Transparency is really important for me.

I think that the large use of LLMs (or AI if you prefer) to generate content created a sort of sadness among the written content online. Or at least this is what I feel.

I was very very Hyped at the beginning, when OpenAI released the version 3 of ChatGPT. And I used it also to generate completely the content of my blog.

I started immediately to have a bad feeling. Yes, I was asking for specific topics and I was reviewing the text, but... it sounds fake.

I don't know about you, but after some time playing with LLMs and reading a lot of AI generated stuff online, I can quite easily spot when a text is generated or not.

Then I thought, that like every new stuff that seems disruptive, it brings also some good things. It is easy to think that human written content will desappear, due to ChatGPT and other LLMs, but what is true is the opposite.

Now, more than ever, human written content is special I got a really nice feeling now reading a human content instead of a generated one, and I give a lot more value to something that comes from a real brain.

The Idea

So I stopped using LLMs to generate content, but I started to use it only for proof reading and grammatical corrections (I am not english mothertongue). And I thought that I would like to know, if I am reading some human content or AI generated one.

So the idea: why don't I say it to the reader? Also if there are some AI generated parts, it would be much better at least to know it.

You know how much AI touched the content you are reading and how. So you can better value what you are reading. And maybe, knowing that a little or no AI is involved, you can value more some content, that before could sound not so interesting.

Take the good part of innovation!

A lot of time, when something disruptive like LLMs comes around, the panic to loose what we have kicks in. But in fact, there is always something that we can use as advantage. Look for example the possibility for the people to buy high end cameras. Everybody is now a photographer, and everybody panic that photography was over. The opposite! In the sea of images, know only the real good ones comes up, this means that this new thing brought a boost of the quality of the best pictures.

For LLMS is the same. Human content is more valuable now, and that's the reason of this little information on my blog posts.