NotebookLM -- A breakout GenAI Tool for enhanced communications.

[Kingsley Idehen]

Hi Everyone,

I personally believe that enhanced communication is the ultimate benefit of AI. NotebookLM from Google exemplifies this worldview.

As someone who has studied (and admired) John F. Sowa’s work for many years, it felt natural to use one of his presentations to demonstrate why I like NotebookLM so much.

[1] https://notebooklm.google.com/notebook/a33f7df2-e03c-4795-bfe5-b8b0690510f2 — notebook shared publicly.

[2] https://www.openlinksw.com/data/pdf/Language_Logic_and_Cognitive_AI.pdf -- rendition of ESWC presentation

[3] https://www.openlinksw.com/data/infographics/Language_Logic_and_Cognitive_AI.png -- Infograhic

Comments and thoughts are welcome.

Enjoy

-- 
Wishing Everyone a Happy, Healthy, Peaceful, and Prosperous New Year!

Kingsley Idehen	      
Founder & CEO 
OpenLink Software   
Home Page: http://www.openlinksw.com
Community Support: https://community.openlinksw.com

Social Media:
LinkedIn: http://www.linkedin.com/in/kidehen
Twitter : https://twitter.com/kidehen

[Mark Underwood]

A useful demonstration.  Hope to see this on LinkedIn if it's not already there.

Mark

--
All contributions to this forum are covered by an open-source license.
For information about the wiki, the license, and how to subscribe or
unsubscribe to the forum, see http://ontologforum.org/info
---
You received this message because you are subscribed to the Google Groups "ontolog-forum" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ontolog-foru...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/ontolog-forum/bdf6389c-bc74-46dd-a69d-802048376fc5%40openlinksw.com.

--

Mark Underwood 
Linkedin  

[Jans Aasman]

Thanks for sharing Kingsley. Pretty impressive condensation of 69 pages. What did you ask the NotebookLM to get to this result? 

Happy New Year to all of you too, cheers, Jans

[John F Sowa]

Jans and Kingsley,

I had an offline discussion with Kingsley about those slides.  They look pretty, and some of them are quite good.  But others have mistakes or miss some important issues.

One of the worst errors, which most people would not notice, is slide 7 of Language, Logic, and Cognitive AI.  It shows  three examples:  From left to right, Peirce (1897, existential graph). Predicate Calculus;; Common Logic.

The second and third are correctly labeled.  But the first is by Frege, not Peirce..  Literally, that graph would be translated to English as "It is false that for every x and for every y if x is a cat and if y is a mat it is false that x is on y"  It shows that Frege's version leaves much to be desired.  The graph by Peirce, which is not shown in that slide is very simple, elegant, and readable:   Cat--On--Mat.

That is a terrible mistake, which is confusing for anybody who goes to the detailed slides for more information.  But other slides are quite good.

For example, slide 11 about sharks and dolphins is beautiful and instructive.  The other slides are nicely done, but I would like to edit the text to clarify some points that are not totally wrong, but not exactly what I would prefer to say. 

Furthermore, it has only 12 slides.  I'll like to add 3 more to clarify some issues and state a better conclusion.

Summary:  It should be an interactive tool rather than a take it or leave it production.

John

 

---

From: "Jans Aasman' via ontolog-forum" <ontolo...@googlegroups.com>
Sent: 1/1/26 1:57 PM
To: ontolo...@googlegroups.com
Subject: Re: [ontolog-forum] NotebookLM -- A breakout GenAI Tool for enhanced communications.

[Kingsley Idehen]

Hi Jans,

I simply provided the PDFs url and asked NotebookLM to generate a deck. Note, this first example includes a little hallucination that reminds us of the: never trust, always verify dictum for anything generated by an LLM.

Hallucination caught by John:

"The first diagram, which is labeled Peirce (1897), is actually by Frege.  The equivalent diagram by Peirce is very simple:   Cat--On--Mat."

A revised deck has been added to the notebook denoted by the same url. It too was generated by simply clicking on the "slide deck" generator control.

[1] https://www.openlinksw.com/data/pdf/Common_Logic_and_Cognitive_AI_2.pdf -- deck 2 PDF

Kingsley

To view this discussion visit https://groups.google.com/d/msgid/ontolog-forum/CABd59dpCXEnbczCeShXsnwKhAFEby9oZVta4SSj76Z-ymWi52w%40mail.gmail.com.

-- 
Regards,

Kingsley Idehen	      
Founder & CEO 
OpenLink Software   
Home Page: http://www.openlinksw.com

[Kingsley Idehen]

Hi John,

Tool calibration is handled via notes. You can use notes to guide how the system generates decks and other related artifacts.

NotebookLM is quite sophisticated, but some experimentation is required to understand how best to control it.

My experiments were done to spark discussions about what’s possible with LLM-powered tooling aimed at improving communication and related workflows.

Kingsley

--

All contributions to this forum are covered by an open-source license.
For information about the wiki, the license, and how to subscribe or
unsubscribe to the forum, see http://ontologforum.org/info
---
You received this message because you are subscribed to the Google Groups "ontolog-forum" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ontolog-foru...@googlegroups.com.

To view this discussion visit https://groups.google.com/d/msgid/ontolog-forum/a006e8ff08e54604957220b2c8272f10%40f051fef026d24f7b9daca45db2270171.

-- 
Regards,

Kingsley Idehen	      
Founder & CEO 
OpenLink Software   
Home Page: http://www.openlinksw.com

[Jans Aasman]

  And so I learn something new every day. Btw: I use NotebookLM regularly and it is really great for research. However, I recently did a project with NotebookLM that needed cross-referencing of information from multiple documents in a very systematic way, and that completely and utterly failed. I read up on NotebookLM and it is almost certain that it stores documents in chunks in a vector store, but it clearly doesn't have a good knowledge graph behind it for some decent GraphRAG :-) The same task worked great in Gemini CLI 3.0  

To view this discussion visit https://groups.google.com/d/msgid/ontolog-forum/543074f3-ea06-4901-a1c6-4d3bedef53b4%40openlinksw.com.