
Imagine not just reading a blog like lotharschulz.info but engaging in a dynamic chat with its content. This is exactly what I explored using Google’s NotebookLM.
Among the various options available, I tested uploading PDF exports of my blog posts and simply providing the URI of my blog to NotebookLM.


Contextual Differences
One of the most significant differences I noticed was in the context when asking for top items from a set of blog posts. Inspired by chaining prompts, I adapted this idea to the set of blog posts on lotharschulz.info:
Prompt
Given the set of relevant blog post, please compose answers to these questions:
What is the oldest post? Context about the question: Use the publication date as indicator e.g. https://www.lotharschulz.info/2025/01/06/transformative-ai-solutions-for-supplier-due-diligence-save-time-ensure-precision/ was published January 6, 2025, where as https://www.lotharschulz.info/2024/12/22/troubleshooting-a-wordpress-database-error-updating-failed-could-not-insert-post-into-the-database/ was published December 22, 2024.
As a result https://www.lotharschulz.info/2024/12/22/troubleshooting-a-wordpress-database-error-updating-failed-could-not-insert-post-into-the-database/ is older than What is the oldest post? Use the publication date as indicator e.g. https://www.lotharschulz.info/2025/01/06/transformative-ai-solutions-for-supplier-due-diligence-save-time-ensure-precision/. Please find the oldest post from all posts on https://www.lotharschulz.info/.
What is the newest post? Context about the question: Similar to above, but in reverse order. The last post in the list of “oldest” posts is the newest post.
What is the post with the most comments? Context about the question: The post https://www.lotharschulz.info/2023/12/28/the-semicolons-little-ballet-in-happy-rust-programming/ contains comments that are listed in https://www.lotharschulz.info/2023/12/28/the-semicolons-little-ballet-in-happy-rust-programming/#mh-comments. In the set of all posts, find the number of comments per post and order the resultset by number of comments per post.
What is the most successful post? Context about the question: Please refer to the context of the question below.
What is the most popular post? Context about the question: In https://www.lotharschulz.info/ in the right navigation pane, there is a section: Popular. This contains a list of popular posts and the top most post of the list is the most popular post.
Ensure that the answers are accurate, have a friendly tone, and sound helpful.
Answer:

I asked for the same prompt based on a different Notebook that is powered by uploaded blog posts as PDFs:
Answer:
I’m happy to answer your questions about the blog posts. Here are the answers:
- Oldest post: The oldest post is “Estimate multi module stories” from March 24, 2014.
- Latest post: The latest post is “Denouement: 5-3-1 Narrative Rule” from January 20, 2024.
- Post with the most comments: The posts “Repository Archiving & Unarchiving with GitHub GRAPHQL API”and “How to break GitHub GraphQL Pagination”have two comments each.
- Top performing post: To determine the top performing post, I would need additional information about what you define as “success.” Are we considering views, time on post, comments, social media shares, or other metrics? Please give me more context so I can answer this question.
- Most popular post: The most popular post is “Transformative AI Solutions for Supplier Due Diligence: Save Time & Ensure Precision”.
While these answers are not fully accurate, they are an improvement over earlier versions. Also, note that there is a maximum capacity of items (in my case PDFs) one can upload:

Capabilities of NotebookLM
NotebookLM offers a variety of capabilities that enhance the user experience, including:
Answers with Links
Similar to Perplexity, NotebookLM can provide answers with relevant hyperlinks.


Audio Summary
Convert text content into an audio summary for easy listening.

Briefing Note Summary
Create a concise summary of the content as a briefing note.

Timeline Note
Generate a timeline of events or topics covered in the blog.

FAQ Note
Produce a FAQ-style note summarizing common questions and answers.

Prompt Engineering
From various sources on the internet, I chose the two prompt engineering approaches below that fit the content of my blog:
Generated Knowledge Prompting
Inspired by Liu et al. 2022 I asked the following question:
Prompt
What’s the safest programming languages used in lotharschulz.info?
Knowledge: use the knowledge from all articles on lotharschulz.info as well as the general knowledge you have about safe programming languages.
Answer
The sources provided do not provide any direct information about which programming language is considered the safest on lotharschulz.info However, the articles cover various aspects of software development and technology that might be relevant to security. Here is a summary of the relevant points:
- Security-related topics: The sources deal with various security aspects, such as offline password generation, secure authentication at GitHub with SSH keys, validating supplier data with AIand troubleshooting software problems.
- Technologies used: The articles mention technologies such as WordPress, Rust, Java, Kotlin, and AI, all of which have their own security features. There is a mention of “Copilot” for test development, but no further execution on the underlying programming.
- Meaning of Unicode: An article describes that Unicode characters can lead to database errors in WordPress, indicating a potential security issue if input is not validated correctly.
Conclusion
There is no direct information in the sources about which programming language is considered the safest on lotharschulz.info. However, one can deduce from the information that the author is interested in security-related topics, uses different technologies, and is aware of the importance of software security. Rust is briefly mentioned and has a reputation for being a safe language, but this does not indicate that it is the main language of the site.

The answer is ok. The selection of information related to security is ok and the answer preamble provides the necessary background.
Retrieval Augmented Generation (RAG)
While reading about similar approaches to NotebookLM, I found Retrieval Augmented Generation Streamlining the Creation of Intelligent Natural Language Processing Models, which appears to be the technical foundation for NotebookLM.
Inspired by the paper I asked:
Prompt
What is the main sentiment of technology related articles on lotharschulz.info?
Answer
The technology-related articles on lotharschulz.info convey a predominantly solution-oriented and practical feeling , which is characterized by the following aspects:
- Focus on problem solving: Many articles describe concrete problems and their solutions in the technical fieldExamples include troubleshooting WordPress database errors, fixing a bug in a (X)ubuntu upgrade or dealing with authentication issues at GitHub.
- Using technology to increase efficiency: There are articles describing how technologies like AI can be used to optimize processes, such as automating supplier auditing with AI using Copilot to support software development.
- Security awareness: The articles also cover security topics, which can be seen in articles on offline password generation and secure GitHub authentication shows.
- Diversity of topics: The articles cover a wide range of technical topics, from software development to system administration to artificial intelligence.
- Concrete instructions and tips: Most articles offer practical instructions and tips that readers can implement directly in their daily work, e.g. instructions on how to fix a specific bug or how to improve software development processes.
Overall, the articles give the impression that the author is pursuing a pragmatic and application-oriented approach. The articles not only describe problems, but also always offer concrete solutions and instructions, which makes them very helpful for other technology enthusiasts. The consistently positive attitude signals trust in technology as a tool for overcoming challenges and improving processes.

Pretty accurate – well done, NotebookLM 😉
UX – Glitch
NotebookLM offers the ability to save answers as Notes. However, I encountered a UI glitch:
The glitch is triggered by Chrome’s translation feature. Although I asked the prompt questions in English, my Google settings produce answers in German. For the sake of this post, I used Chrome’s translation to convert the answers back to English. When scrolling through notes with active translation, the issue occurs. Without browser translation, the bug is not reproducible. This suggests that the actual bug might be in the browser/translation combination and could be independent of NotebookLM.
Conclusion
Exploring the capabilities of NotebookLM with my blog, lotharschulz.info, has been an enlightening experience. The tool offers a range of features that enhance how we interact with blog content, from providing answers with links and audio summaries to generating briefing notes, timeline notes, and FAQs. The experiment with prompt engineering demonstrated the potential for generating meaningful insights and summaries from blog posts, although some limitations and contextual misunderstandings were noted.
As we move forward, tools like NotebookLM will undoubtedly play a role in transforming how we consume and interact with digital content. Whether you are a content creator or a reader, these advancements promise a more interactive, efficient, and user-friendly experience.
Thank you for joining me on this exploration. Stay tuned for more insights and innovations on lotharschulz.info!
Leave a Reply