The Memory Keeper - AI-generated podcast

 I don’t do podcasts. 

I was asked if I had tried to use Google’s NoteBookLM, specifically the podcast generator. I had not. I had experimented with NoteBookLM, but not the podcast feature. 

Knowing NoteBookLM would not support a TiddlyWiki HTML file I extracted high-level text from the Memory Keeper’s help content from the sooner-to-be-released (1.0.01) Wiki and created a Google Doc. Then, I proceeded to generate an AI-generated podcast from that content. 

Besides being a bit flattering and a bit humorist, it was also a bit misleading and in a couple cases downright wrong.

A day later I thought about what was reported in the podcast, thinking that a few of these wrongly reported features were good ideas I should implement.

The podcast reported that assertions (proofs) were made easy with templates. Huh, what templates? Great idea. Writing an assertion with the use of a template is not only easier, but now it enables me to be consistent. I wrote a couple templates and users can add their own. 

The podcast reported that the Memory Keeper went beyond typical family relationships, but this is misleading. The podcast states you can track neighbours, adversaries, and associates. Well, you can by associating individuals to an ancestor's event. In doing so you define the roles people had in the event. However, what if MK enabled the user to define relationships?

There are a few reasons this becomes useful:

1. Shirttail relationships. Those honourable relationships.  Honourary uncles and aunts.
2. I have blood relationships that are known between two people but I haven’t been able to fill in the gaps to naturally derive the relationship. Example: I know a person is a grandparent of another, but I don’t know this person’s children. So, I created a user-defined “grandparent” relationship type. Then in the person tiddler for the grandchild, I assign the “grandparent” relationship. Now in the tree, it can automatically assign the relationship in the tree diagram.
3. Relationships you don’t typically track, but in some odd cases you want to track it.  Example: fiancée. In most cases a fiancée becomes a spouse, leaving us not tracking a fiancée. My grandfather had a fiancée he didn’t marry. This information was obtained from his military record. Well, that was something I wanted to track; and I wanted it to appear on my tree. Why? Because I found a DNA match with a descendant of this fiancée. Aside: I also draw DNA matches on my tree diagrams. 

Here is the original AI-generated podcast.

Disclaimer: the examples in the podcast are not mine. Either the AI engine made them up or went out on the web and found them (which I highly doubt). Also, it is misleading in the podcast about having unlimited number of records. Limitations are based on the computer and disk performance. MK is not about importing tens of thousands of names and dates. If the number of people you are researching is in the tens of thousands, pick another app.