About the use of generative AI

Intro

The increasing use of generative AI has flooded the internet with technical articles and infographics which weren’t created by humans.

This personal page contains technical writings and (sadly) I need to make clear about my position on using generative AI here. So, here we are.

Goals

  • Explain that I don’t use AI for generating content for this website
  • Present personal thoughts about using generative AI
  • Present how I use AI

Non-goals

  • Rant about AI
  • Create a guideline of using AI
  • Tell what’s good or bad about AI

Do I use AI? And what about the evil “algorithms”?

Well…

Let’s make it clear something: we’ve been using AI for decades. We’ve been using algorithms for millenia. If it somehow surprises you, go check what artificial intelligence and algorithms are. If you’re online, probably you use AI and doesn’t even notice it. Everybody uses algorithms.

Do I use generative AI?

Sometimes.

Sometimes I need to name something and it can suggesting something good. Sometimes I need to check the grammar and the syntax of a text that I wrote. Sometimes I use it to generate a really simple and well-known code snippet (e.g. a simple HTTP request in Python).

But in all those cases I check everything. I don’t accept blindly the suggestions. I even prefer to type the content instead of copying it. I am the author and I need to do it, the AI generated stuff is only a suggestion.

The LLMs have their own way to write (sometimes called “AI accent”). It is soulless, too much perfect, generic, sterile and without personality. Texts, even the technical ones, are not only the about words or what they are saying, but how they are saying and why they are saying.

You may not agree with me, and that’s ok. I’m not saying I’m right. On the contrary, I think it’s too early to know what’s right about using LLMs. I also believe this is the case where the technology evolved so fast that we don’t exactly know how to use it. I only think that this is a mid-term approach, which tries to leverage some benefits of AI without losing the authorship. If you don’t agree with me, I’ll b happy to hear your opinion (since it’s not generated by AI :-)).

Was any content in this website generated by AI?

NO. Again: NO. There’s no reason for me to generate a text using generative AI and post it here.

Why? If the content can be generated by an AI, you can ask your favorite LLM to do so. I won’t spend my time doing that. I’m not a “content creator”, I don’t post here because of money, I don’t have any pressure to do it regularly. Of course, I’m happy when I see people liking something that I posted here (e.g., when I wrote about Ruby and Matz shared it), but if no one reads, it’s ok. But I don’t know how can someone be happy being recognized for something that they didn’t do.

I only write something here if I have something to tell.

Do I hate generative AI?

Like any technology, hating it is pointless.

Do I generate code with AI?

Currently, in extremely rare cases, e.g. the aforementioned snippet code generation that I would copy from documentation, StackOverflow and so on. Mostly for things that I don’t care if they are bad code and that I am the only user. I never shared a source code written with AI on GitHub or similar.

However, I’m not against it by itself. If my job is to make something work and AI-generated code helps me do that, I think would be ok to use. But we need to be responsible for what we’re doing. If I can’t explain exactly what the code is doing, I wouldn’t deliver it.

Do I use generative AI in my personal projects?

Currently no, but I have some ideas of side projects integrating LLMs through their APIs, but their code will be 100% written by me. I work on side projects because I want to learn and because I want to have them fun with them, and I can’t learn something or have fun asking an LLM to do the dirty work.