Inspiration vs. mindset

Inspiration vs. mindset

How to avoid getting tangled up in your own thinking when working with AI.

Time and again, I find myself getting lost in my work with AI. Not in the technology itself, but in endless further thinking.

One thought leads to the next, one impulse opens up ten new spaces – and suddenly I find myself juggling concepts, ideas, quotes, agents and half-finished drafts all at once. What began as a focused question ends up as an open space for experimentation – even though I actually wanted to get something done.

Brainstorming vs. cognitive noise

The diversity of ideas triggered by AI is fascinating – but also exhausting. Generative AI expands our thinking, but – if we're not careful – it also increases cognitive noise.

For a long time, I tried to create order through roles: analyst, dramaturge, strategist, editor – digital colleagues who were supposed to support me. But the more roles I defined, the stronger the feeling became that I was merely the assistant director of my own thinking. When I combine several roles in a system prompt, it quickly becomes generic.

I think that in this case, too, the problem does not lie with AI. It lies in the unclear interconnection between my thinking and hers. That's why I'm now experimenting with a different approach:

Sequential thinking.

Instead of activating all roles at once, I consciously work in phases: exploration – structuring – consolidation – production – reflection. Each phase has its own logic, its own question, its own timing. Because I am convinced that AI – like us – needs context, rhythm and guidance.

When both intelligences work together, a different way of working emerges: slower, more conscious, but also clearer. This does not necessarily save time and resources, but it does lead to more thoughtful, better structured and more precisely formulated results. At least, that's how it is for me.

Perhaps this is the progress that is needed. Perhaps we don't need AGI and superintelligence – but rather a better understanding of the rules of collaboration between humans and machines. Not the perfection of AI, but the self-awareness of the thinker – and the ability to meaningfully combine human and machine thinking.

Those who understand how they think can better let AI co-think – with an emphasis on better, not faster or more.