I was honored to be among the first who have signed The AI Manifesto. The title of this article is currently just my own bold prediction, but come back in a year or so and we'll see.
I was honored to be among the first to sign The AI Manifesto. The title of this article is currently just my own bold prediction, but come back in a year or so and we'll see.
Its author, German-based developer Christopher H. Stappert, put together five short, unsurprising, yet deeply inspiring principles:
- Never let an LLM speak for you.
- Never let an LLM think for you.
- Never let an LLM own your work.
- Never let an LLM replace your curiosity.
- Never let an LLM discourage someone else.
Despite sounding almost trivial, those short rules really do sum up the topic better than I ever could. They are like guidelines, a beacon in the ever-changing stormy AI sea. New tools keep emerging every day. Entirely new ways of working are being discovered (hello prompt engineering, hello vibe coding, hello context engineering, hello “whatever waits around the corner yet to be described”). How do we avoid getting lost out there in the wild?
Some of us are eagerly early-adopting everything. Some (like me) are curiously scratching the surface. Some are trying to ignore. Some are scared...
I am a techno-optimist. I believe AI can ultimately change our lives for the better. But only if we maintain control and our natural will to explore and improve. We must not give up our minds to the machines and brain-rot into mental oblivion. And we should never forget to share the fruits of our wisdom with others - this is the last, and perhaps the most important, principle.
You can read the story behind and the (brief) explanation of each rule at https://ai-manifesto.dev/.
And if those five amendments resonate with you as they do in me, you can also digitally sign them easily via your GitHub or LinkedIn account. Or spread the word as this simply deserves to be well known.
I checked this article with ChatGPT for correctness - with the AI Manifesto in mind
Cover was generated by Sora in two iterations (then it kept failing to fix the typo)