I thought of this way of saving notebook limits. Basically, i would upload all the sources in one notebook and generate podcasts/videos of each module by just selecting the appropiate sources. Ideally this would save notebooks as i would not generate one for each unit, but at the same time i'm worried that it might lead to a worse performance (for whatever reason).
If anyone has tried this or it's just a dumb question let me know

Why do two accounts (both are pro) produce drastically different results with the exact same prompt and source?

I tested the exact same setup on two different accounts, using the same source material and the same prompt, and the difference in output quality is honestly huge.

For example, when generating a slide deck:

- Account #1 can generate around 15 slides with good quality, structure, and accuracy.

- Account #2 only generates around 10 slides, even when I explicitly tell it to create 15 slides.

- The quality and accuracy of the second account's output are significantly worse overall.

What makes this confusing is that I understand that even within the same account, running the exact same prompt again can produce a different result. That's not really my issue. In Account #1, even when the output varies from run to run, the results are consistently within an acceptable/good quality range.

With Account #2, however, the problem seems fundamentally different: the output is consistently much weaker, less accurate, and less capable, despite using exactly the same prompt and source.

Has anyone experienced something similar?

I'd really like to understand what's happening here, because the difference between the two accounts is too significant to feel like normal randomness.

​

Hi everyone! I’m trying to improve and structure my study method for Law, and I’d like to use AI as a study aid throughout the different stages of exam preparation.

I’d like to create a structured study process, using a different prompt for each stage:

1. Initial explanation: upload pages from a textbook or study material and get a complete, clear, and detailed explanation, without leaving out important concepts, so that I can also use it to take notes and write down the explanation.

2. In-depth analysis: analyze the more difficult topics, connect different concepts, highlight definitions, principles, exceptions, and connections with other areas of the subject.

3. Testing: use open-ended questions, quizzes, and questions similar to those I might get in an actual exam, ideally with gradually increasing difficulty.

4. Feedback and correction: receive detailed feedback on my answers, understand my mistakes, and identify which concepts I need to review.

5. Learning and memorization: use techniques such as active recall, spaced repetition, targeted questions, and structured summaries to retain the material more effectively.

6. Exam simulation: finally, have a realistic mock exam with unpredictable questions and detailed evaluation of my answers, as if I were actually being examined by a professor.

The goal is to be as well prepared as possible for the exam—not just to understand the material, but also to remember it, make connections between concepts, and explain everything clearly and accurately.

Have any of you used AI prompts or a similar method to study Law?

I’d really appreciate it if you could share your best prompts, especially prompts designed for the different stages of the study process.

Thanks a lot to anyone willing to share their experience or suggestions!

Let me explain, i'm logged into my university website reading a pdf file... being the website something like http : / / 123456 /pluginfile.php/74860/mod_resource/content/2/El%20proceso%20ense%C3%B1anza.pdf

But when I tried to load that PDF into my notebook, i don´t get any kind of error but when you look into the file loaded it's something like "Login : username / password "

being the only way to load the file to download it to my pc and then upload it again.

Getting the same error with extensions like NotebookLM importer.

is there anyway to do that? Or even better, any extension that first download the file and automatically upload it to my notebook?

I tried generating a slide deck from my notebook on the ios app this morning and it keeps failing. I then went to the desktop/web version and the 'slide deck' option is removed there. What's going on? That was such a useful feature.

Does anyone has intel on this? Or better alternatives? I like that it generated images as well.

failing \"slide deck\" artifact generation

missing \"slide deck\" option on the website

I am on mobile is it a paid feature? Because every time it says generation failed please try again if it is paid they don’t make they very explicit on the app but if not what’s happening?

I use NotebookLM to study and generate podcasts that I listen to on the bus. I find it really useful, but I’m curious about what people who use it regularly still feel is missing.

Would you want it to remember what you already understand, create a more structured learning path, adapt quizzes or practice based on your mistakes, connect knowledge between notebooks, or something completely different?

What feature would make the biggest difference for you?

I just wanted to ask as a medical student can it be useful i think notebook lm can be but obsidian i am not quite sure or am i wrong just wanted to ask fellow peeps about it

I'm personally into the AI landscape but somehow haven't tried this tool yet.
For those who have some experience with it, what tips do you have for a new user?

I know the questions are broad and maybe unclear, and that's why any tip or advice is welcome.

Is it just me, or is NotebookLM playing the podcast with sound even if the phone audio is muted? And not even that, it's always playing the same volume. It cannot be controlled anymore.

The "plays sound when muted" behavior only happens with the native phone speakers. When using the car Bluetooth, the volume is fixed and cannot be made louder or quieter. When using Bluetooth headphones on the smartphone, I can adjust the volume. BUT if I use the volume knob on the smartphone, not the Bluetooth sound gets adjusted, but the "subtitle reader"(?) volume? I will try to add a picture in the comments. And this volume bar has no effect. Only if I adjust the Bluetooth volume manually.

Android 17 Pixel 8 NotebookLM version: 1.54.2.964195800

For me, by far the bigger issue is that it test you based on the whole notes intead of only regsrding the sections you had studied. So it's useful when your already pretty far in your study but not for starting imo. Am I the only one? How do you guys solve this?

It seems powerful but I'm not sure I need it instead of Notebook, or along with Notebook. Any informed advice welcome.

​

I've been using NotebookLM heavily for research and studying, and ended up writing down everything I learned along the way. A few things that took me way too long to figure out on my own:

Audio Overviews aren't just a novelty — they're genuinely useful for reviewing dense material passively (commute, chores, etc.), but they work best when your source notebook is tightly focused on one topic rather than a mixed bag of unrelated documents.

How you phrase questions matters a lot. Vague prompts get vague summaries. Asking something specific like "what are the counterarguments to X in these sources" gets far more useful answers than "summarize this."

Organizing notebooks by project, not by source type, made a huge difference — one notebook per research question/project rather than dumping everything into one giant notebook, so the AI's context stays focused.

I turned my notes into a full guide since I couldn't find one that covered the whole workflow start to finish — happy to share it if anyone's interested, and also happy to just answer NotebookLM questions here if that's more useful.

Notebook lm is great with extensive capabilities of artifact generation and resource ingestion.I am just curious to know for which tasks do you use notebooklm research, content creation,learning,automation,resource gathering and crawling or something other.How do you add more power to it which problem do u people face and how do you overcome it.

Hey guys, my first time here.
I used my notebook as a grammar checker for my movie subtitle translations.
Normally it works pretty fine, but somewhere like 2-3 weeks ago until now, it keeps randomly cutting off some parts of the subtitles, each part includes 1 minute of dialogues while the rest can still be generated. While I used the same prompt, could it be that the notebook's model is lowering the amount of texts generated or I should change something to make it back to normal?

I used longer length option already so I think the problem only comes from the prompt, am I right?

i remember Claude’s website saying it could handle PDFs up to around 1k pages, but a few days ago i tried uploading a PDF that was 2k+ pages just to see what would happen.

surprisingly, it handled it really well.

I was asking questions about pretty specific things buried in the document, and not only did it find the answers, it also gave me the line numbers for where the information came from when I asked. That part is actually super useful for me because I often need to go back to the original PDF and verify the source rather than just trusting the AI answer.

the only thing i really wish Claude had now is a better built-in PDF viewer.

Like imagine if I could click the citation / line number in Claude and it would open the PDF right there, jump to the exact sentence it used, and highlight it. For someone doing finance or research or working with documents where references actually matter, that would make the whole experience sooo much better.