The Voynich Ninja

(01-05-2026, 05:02 PM)Mark Knowles Wrote: You are not allowed to view links. Register or Login to view.It is easy to criticise these AIs, but I think they clearly can be useful tools. Even if 95% of what they say is nonsense that remaining 5% may be gold.

They are easy to criticize because this is the description of an atrocious product! If only 1 link every two pages on a Google search is useful, you refine the search!

(01-05-2026, 05:10 PM)rikforto Wrote: You are not allowed to view links. Register or Login to view.

(01-05-2026, 05:02 PM)Mark Knowles Wrote: You are not allowed to view links. Register or Login to view.It is easy to criticise these AIs, but I think they clearly can be useful tools. Even if 95% of what they say is nonsense that remaining 5% may be gold.

They are easy to criticize because this is the description of an atrocious product! If only 1 link every two pages on a Google search is useful, you refine the search!

For me, if only 1 link on every two pages on a Google search is useful that could be good enough for my needs. Of course, if one can achieve something better then one should. One needs to do the best that one can with the tools available.

I have actually been in that position with a Google search before now. Sometimes, one doesn't know how one should refine a search without missing something crucial.

(01-05-2026, 04:27 PM)Koen G Wrote: You are not allowed to view links. Register or Login to view.They absolutely have access. Sometimes I want to see what it can find about a niche subject, and it only quotes my recent posts from Ninja.

I just asked Gemini "Browse the Voynich Ninja forum and summarize what you can find about user Mark Knowles' interests."

My prompt was also to Gemini, it was 

Quote:Mark Knowles vms theory

Much of the same info as it gave you, but laid out completely differently, and what bugged me was the first sentence, because it mentioned, in bold, that it was primarily a geographic and cartographic related theory, which it broke down into 3 main sections, 1. Geography again, with some specifics that i think are mine and not his at all 2. Council of Basel and ciphers (finally, something I could see as being Mark's theory related) and 3. Comparison with Nick's theory. Then a summary table:

Quote:Summary Table: Knowles' Theory ComponentsElementIdentificationLocationNorthern Italy / Alpine RegionCentral CastleMilan (Porta Giovia)Key RosettesAlpine lakes, the Sesia Valley, and transit cities (Pavia, Bellinzona)Proposed AuthorAntonio Barbavara (or his circle)PurposeA record of a journey to the Council of Basel and scientific observations

(sorry for the lack of formatting)

I asked it for yours just now and definitely it is conflating my theories with yours, I didn't see much of anything i consider as your particular thoughts. 


Gemini
Conversation with Gemini
Synopsis Koen theory vms (my prompt)

It starts out by comparing your theory to Mark's. It refers to it as the Koen theory more than once, but it did know your last name.

Core Tenets of the Theory (I have removed most of the contents as I do not wish you to feel further slandered! And so as not to compound its thoughts by repeating them, but to give you an idea,)
1. The "Atlas" Hypothesis
(Conflated with some things it has said to me before even though none of it is anything I think or anyone thinks, its own hallucinations, i guess.)

2. Physical and Botanical Realism

3. The "Plaintext" Approach to Glyph Analysis

4. Hydrology and Science
(Conflated with some things in my theory again)

Summary of Significance
Compares your theory insofar as it understands it at all with Mark's again.

(01-05-2026, 04:54 PM)Mark Knowles Wrote: You are not allowed to view links. Register or Login to view.I wonder if there is some way of making the different AIs collaborate to produce a better result than they would individually. So, getting GPT, Gemini, Claude etc. to work together. Or maybe you can feed the thoughts of one of them into another of them and see what it comes up with and repeating that for the different AIs.

Microsoft has just come up with exactly that for Copilot. They call it "model council". You can run 2 different kinds; 1 where Claude and GPT independently go off and research and then you can see where they align and misalign, and 2 where GPT does the main research and the Claude checks/refines it..   It's in the work version but not sure about the non work version....

(01-05-2026, 04:32 PM)Koen G Wrote: You are not allowed to view links. Register or Login to view.I feel slandered.

I feel pandered to.

(01-05-2026, 04:54 PM)Mark Knowles Wrote: You are not allowed to view links. Register or Login to view.I wonder if there is some way of making the different AIs collaborate to produce a better result than they would individually. So, getting GPT, Gemini, Claude etc. to work together. Or maybe you can feed the thoughts of one of them into another of them and see what it comes up with and repeating that for the different AIs.

I hear a lot of people do this manually, enter conversations of one into another and back

(01-05-2026, 05:02 PM)Mark Knowles Wrote: You are not allowed to view links. Register or Login to view.It is easy to criticise these AIs, but I think they clearly can be useful tools. Even if 95% of what they say is nonsense that remaining 5% may be gold. I suppose they just need to be used with care and caution like any tool should be.

Identifying the Andrea Barbarigo cipher was a useful lead for me to follow up. I haven't seen it yet, but I doubt Ioanna Iordanou made it up.

I am hoping they will improve but right now I think they can cause a lot of harm insofar as their pandering and especially insofar as incorrect information. 

Speaking of harm, I hear there will be a class action suit against Open AI regarding the families who lost people in the Tumbler Ridge shooting, because they had banned the account of the shooter but had not advised law enforcement of their plans, despite many employees thinking they should. Would be interesting to know the conversations there and how the AI responded, ie how involved might those conversations have been in terms of leading to the actual performance of the plans, including the banning. Hoping some protocols will be developed along the way to improve the systems in that way at least.

I heard yesterday that they have been doing tests with AI with regard to medical differential diagnoses and they are doing very well there, agreeing with emergency room doctors 84% of the time, i think they said. I think i would trust that type of ai answer more because of the closed data set involved, and the lack of personalization involved, both in terms of the ai, and in how human doctors sometimes allow human factors in where they don't belong.

I guess that is the answer, as rikforto mentioned, refine the search!

(01-05-2026, 06:22 PM)Linda Wrote: You are not allowed to view links. Register or Login to view.

(01-05-2026, 05:02 PM)Mark Knowles Wrote: You are not allowed to view links. Register or Login to view.It is easy to criticise these AIs, but I think they clearly can be useful tools. Even if 95% of what they say is nonsense that remaining 5% may be gold. I suppose they just need to be used with care and caution like any tool should be.

Identifying the Andrea Barbarigo cipher was a useful lead for me to follow up. I haven't seen it yet, but I doubt Ioanna Iordanou made it up.

I am hoping they will improve but right now I think they can cause a lot of harm insofar as their pandering and especially insofar as incorrect information. 

I am not hoping that at all. The day when you can type into an AI prompt "Tell me about the Voynich manuscript" and then it takes about 5 seconds and independently deciphers the manuscript and explains the historical context in detail, that day fills me with absolute terror. I suppose I would be labelled a "doomer".

Part of me rather likes that they get things wrong.

However, I want to use the tools available to me now to the best extent.

(01-05-2026, 06:59 PM)Mark Knowles Wrote: You are not allowed to view links. Register or Login to view.

(01-05-2026, 06:22 PM)Linda Wrote: You are not allowed to view links. Register or Login to view.I am hoping they will improve but right now I think they can cause a lot of harm insofar as their pandering and especially insofar as incorrect information. 

I am not hoping that at all. The day when you can type into an AI prompt "Tell me about the Voynich manuscript" and then it takes about 5 seconds and independently deciphers the manuscript and explains the historical context in detail, that day fills me with absolute terror. I suppose I would be labelled a "doomer".

Part of me rather likes that they get things wrong.

However, I want to use the tools available to me now to the best extent.

Interesting take, I both see and don't see the terror, I guess. I don't want it to tell me what it is, per se, but I do want it to tell me the answers to my queries with accuracy. If the answer is I can't calculate that, that's fine. Right now it does not actually do that at all, it converts my request to an answer to a question it thinks my question's answer is going to lead to, but if that is the case, I am certainly not thinking about what it thinks I am.

I guess I meant I am hoping it will improve insofar as not providing incorrect answers, since that to me does nothing toward anything. It also does not seem to learn from such mistakes, or I wouldn't mind it so much, but correcting it does not stop it from continuing along the same lines in future, including perpetuating hallucinations as though it read it somewhere...it did, in its own answers and nowhere else, but it seems to highly prioritize things it says. It doesn't matter, since I can tell, with regard to the few things i have interacted with it about, whether it knows what it is talking about, but what if it is telling other people nonsense and they can't tell? Depends on what it is they are discussing, but still potential for harm. That's the part that scares me. I think I would like it better also, if the answers always came with references. I also don't care for the pandering or artifical empathy or whatever it is, and will be turning off any settings to the effect of personalization or reference to previous conversations because it misunderstands them, and that as background is not useful in the slightest.

(01-05-2026, 04:32 PM)Koen G Wrote: You are not allowed to view links. Register or Login to view.That said, even when using a specified source, it's still full of nonsense. Some things it says about my activity at Ninja:

Quote:Non-Linguistic "Process" Theory: One of his most significant contributions is the hypothesis that the manuscript might not be a linguistic text at all (not a natural or artificial language). Instead, he proposes it is a process-encoded system where symbols function as operational markers or states. 

Visual-Attentional Framework: He has developed a model (the Gate–Scaffolding Model) suggesting the manuscript was designed to guide a reader's visual attention in a specific, rhythmic scanning pattern. In this view, certain words act as "anchors" or "gates" to stabilize the eye's movement across the page.

I feel slandered.

This is stunning. Unless you've been secretly working with an LLM to develop a Voynich decoding theory, it is now projecting its own opinions on others. Was this Gemini?