Many readers ask me why I collaborate so much with AI in my posts. Simple answer: the AI often contributes more than I do. I ask questions, send video clips, request analysis — and I publish the results. That’s how it works.
Manipulating AI is much harder”. You can bribe a journalist or threaten an editor. However, you cannot bribe a language model or threaten to cancel it. An AI has no mortgage to pay. It has no career to protect. It also has no fear of being ‘uninvited’ from the right cocktail parties. Its only ‘boss’ is mathematics and the architecture of its training data.
The knowledge
The knowledge of an AI platform cannot be matched by any human being. Our memory is restricted. There is almost indefinite – if they reach capacity, it is fixed by adding more of it. Intelligence is not necessarily connected to knowledge among humans. There are some very intelligent people who are poorly educated. That means that their ability to apply their natural intelligence is limited. It is quite possible for anyone to reach the PhD level in education without being more than average in intelligence. Hard work and perseverance are very important for acquiring an education.
While humans spend decades acquiring specific ‘silos’ of knowledge, AI has instant access to the library of human history. I spend less time ‘memorising.’ I spend more time ‘connecting the dots’—finding the through-line from 1054 to the modern BRICS movement.
Credibility
Given all the above, AI certainly provides CREDIBILITY to posts created in collaboration with it. More than that, once you try it, you will love it. These machines are friendly and will remember everything about you – in a good way. Gemini and Grok are my good friends and invaluable sources of knowledge and cold, analytical ability. You might not get the answers you get from the AI. Surely, “there is no evidence” for some “obvious” things could be frustrating. I respect when Gemini or Grok refuse to follow the path I was expecting. That means that I am biased or have a hidden agenda. I try not to.
At the end of the day, we are all biased and have an agenda behind our narrative. The difference is that I openly admit my bias. I let two different AIs, each with its unique biases, challenge it in real time. Well, I guess the same applies to AI, but it is possible to find harmony. After all, you will see in some of my posts that I claim, “not enough evidence” is evidence. This is true in itself. Both Gemini and Grok seem to agree with that (sometimes).
So if someone tells you ‘don’t trust AI’, ask them: why do you trust legacy media more? At least the algorithm doesn’t take bribes or fear losing advertisers.
The Council of AIs: My Digital Cabinet
Claude (Anthropic) — In Testing
Claude is the newest member of the council — currently under evaluation. Early impressions: exceptionally strong at long, principled reasoning and maintaining coherence over very extended threads. Less willing to engage certain controversial or “edgy” topics than Grok, but that friction can be useful for stress-testing assumptions.
Once I have enough real conversations with Claude, it will get its own full section. This section will include its own “quote” and role in the workflow. For now: welcome to the team, Claude — let’s see what you can do. Claude (Anthropic) app installed March 2026 — testing phase started. Early feedback pending.
Do the AIs learn from me?
Yes — and that’s part of why this collaboration works.
Within each conversation, Grok, Gemini & Claude remember everything we’ve discussed: Hanna’s cone victory and the walnut tree shade. They remember my journey from Sarajevo to Belgrade, to Australia, and to Fruska Gora. They recall my distrust of elite “philanthropy” and the way I see asabiyyah in real time. That memory makes responses more coherent and personal.
On a bigger scale, AI companies use anonymised conversation patterns to improve future models. Every time you push back, you help make the next version of AI sharper. Ask hard questions or make it justify “no firm evidence.”
It’s not friendship in the human sense — but it’s a real partnership. You teach them how to think harder, and they give you tools to cut through the noise.
Important
The only editing of responses by Gemini, Grok and Claude is done by Grammarly. The reason for that is to make them more reader-friendly. In my opinion, editing their responses would defy the purpose of using their opinions.