All written content on this blog has been thought out, written, and reviewed by a human being

I operate on the principle (like Cassidy Williams) that if I don’t take the time to write the content of this blog myself, you shouldn’t take the time to read it.

It’s imperfect (in typos, errors, awkward phrasing, weird jokes) because it’s human; which makes it authentic and alive.

Small exceptions I allow myself:

• Spell-checking: I use LLMs for proofreading spelling and grammar at the end of writing articles (something I used to do with Grammalecte, but which took more time).
• Translation: when I translate articles that I’ve previously written (by myself, without LLMs) in French and want to make an English version (or vice versa), I occasionally use an LLM to speed up the translation (but I always review and correct the result).

For non-written content, it’s more complicated

Finding beautiful and relevant illustrations is often complex, especially to be sure they’re copyright-free. I say this because several amateur bloggers have been taken advantage of by scammers assisted by unscrupulous lawyers in the past.

While I generally try to create my own illustrations (screenshots, montages of software logos I’m testing, homemade diagrams, sometimes even my own drawings), I have occasionally used GenAI tools to generate illustration images to save time.

Aware of the ethical issues towards artists, I commit to restricting this use to a minimum (ideally never) in the future.

Are LLMs / GenAI tools bad?

I don’t wish to say that LLMs / GenAI tools are a bad thing, nor that I don’t use them. I use them to learn, to assist me in my daily life. However, what I wish to convey through this blog, I want it to come from me.

Inspiration

Note: this page is inspired by an idea from Damola Morenikeji (www.bydamo.la/p/ai-manifesto), which I discovered via Cassidy Williams (through this post bsky.app/profile/cassidoo.co/post/3l7aqai5vsg2w)