Very interesting article! I also use NLM for quick research. When I discover a new company, I upload the annual reports and let NLM create a mind map to give me a rough idea of the business and its strategy. From there, I then delve deeper into specific areas. NLM feels like a shortcut to reading and understanding a company, but I think that if you really want to understand a company, you have to make the effort to work through the annual reports yourself and get an idea what drives the company.
This is an incredibly practical guide for investors. The ability to cross-reference multiple years of reports and track management promises against actual results is a game-changer. I especially appreciate the structured prompt for creating analyst reports - it forces the AI to be quantitaive rather than just qualitative. The quiz feature is a clever way to ensure retention too!
Great article! AI is a great investing tool if it's used as an analyst instead of a source of truth. The biggest performance boost comes from letting AI brainstorm different scenarios so that the human can build on top of that with second-order thinking.
This article comes at the perfect time, and I especially appreciate the point about 'No hallucinations. No guesses.' because it addresses a truely fundamental trust issue with current LLM-s, making NotebookLM a powerful tool for serious analysis.
Actually this was my first thought after opening this article, does it hallucinate since these are important data points, one hallucination and you’re taking in the wrong information…
Haven't used notebookLM yet, I'll be giving it a try. I've been doing what you describe by creating a project in ChatGPT with some high level instructions and documenting expected outputs/questions I want answer and then one thread for each stock. In the thread I dump the latest earnings presentations etc (I often use perplexity to fetch them).
What would notebookLM do better in this case? I've got a few other use cases to try with it
Very interesting article! I also use NLM for quick research. When I discover a new company, I upload the annual reports and let NLM create a mind map to give me a rough idea of the business and its strategy. From there, I then delve deeper into specific areas. NLM feels like a shortcut to reading and understanding a company, but I think that if you really want to understand a company, you have to make the effort to work through the annual reports yourself and get an idea what drives the company.
What is NLM
NotebookLM
Going to give it a try
NotebookLM is my private tutor.
❤️🙏
This is an incredibly practical guide for investors. The ability to cross-reference multiple years of reports and track management promises against actual results is a game-changer. I especially appreciate the structured prompt for creating analyst reports - it forces the AI to be quantitaive rather than just qualitative. The quiz feature is a clever way to ensure retention too!
Great Content
Great article! AI is a great investing tool if it's used as an analyst instead of a source of truth. The biggest performance boost comes from letting AI brainstorm different scenarios so that the human can build on top of that with second-order thinking.
How do you know it's not hallucinating?
NotebookLM hallucinates less for 2 reasons:
-it only uses the sources you give it.
-before answering, links every line back to the exact passage it used as a source.
but at the end of the day, it’s still an AI, so you should always verify.. but nothing comparable to other AIs
This article comes at the perfect time, and I especially appreciate the point about 'No hallucinations. No guesses.' because it addresses a truely fundamental trust issue with current LLM-s, making NotebookLM a powerful tool for serious analysis.
Actually this was my first thought after opening this article, does it hallucinate since these are important data points, one hallucination and you’re taking in the wrong information…
Ive been using Notebook a lot. Its a fantastic resource.
Very interesting use case
Haven't used notebookLM yet, I'll be giving it a try. I've been doing what you describe by creating a project in ChatGPT with some high level instructions and documenting expected outputs/questions I want answer and then one thread for each stock. In the thread I dump the latest earnings presentations etc (I often use perplexity to fetch them).
What would notebookLM do better in this case? I've got a few other use cases to try with it
++ Good Post, Also, start here 100+ Most Asked ML System Design Case Studies and LLM System Design
https://open.substack.com/pub/naina0405/p/bookmark-most-asked-ml-system-design?r=14q3sp&utm_campaign=post&utm_medium=web&showWelcomeOnShare=false