This text is republished with permission from Surprise Instruments, a e-newsletter that helps you uncover essentially the most helpful websites and apps. Subscribe here.
Google’s NotebookLM is a brand new free service that allows you to apply AI to your personal notes and paperwork. You should utilize it to floor new concepts and discover contemporary connections in your ideas and analysis. Learn on for the way I’m utilizing it, what I like most about it, its limitations, and two fascinating options.
begin utilizing NotebookLM
- Enroll at notebooklm.google.com. It’s free for now, whereas in beta.
- Add as much as 50 paperwork into a set (or “pocket book”) of sources to discover.
- You possibly can create as much as 100 of those notebooks for numerous subjects.
- Every doc you add might be as much as 100MB and include as much as 500,000 phrases.
- You possibly can add PDFs, Google Docs, or pasted textual content.
- Inside every pocket book, there’s a “noteboard” area for saving AI responses and key passages out of your sources or to avoid wasting ideas as you work together with the AI.
- While you create your personal account, you’ll see just a few pattern notebooks to discover. Discover these notes and check out queries to get a really feel for the way it works.
Tip: Take into account separating your personal writings from analysis you’ve gathered from others. That can make it simpler to tell apart your materials sourcing later.
Instance: I’ve a set about entrepreneurial journalism, for instance, with a wide range of notes and supplies I’ve created over time. I’m constructing one other one targeted on the historical past of classical music, with articles, analysis, and notes I’m gathering from others. Having separate areas for these will assist the AI phase its evaluation, and also will assist me guarantee readability about sources.
What to make use of NotebookLM for
Analyze your personal materials
When you create a set, NotebookLM generates a abstract for every of the paperwork in it. It additionally provides topic tags primarily based on the content material. You possibly can then question a specific doc, a gaggle of paperwork, or the complete pocket book. That’s useful if you wish to analyze materials from a specific supply or a batch of paperwork on a selected topic.
Urged queries
NotebookLM will recommend just a few queries you can begin with, primarily based on its evaluation of your materials. For instance, in my assortment of entrepreneurial journalism notes, it proposed queries about greatest practices for presenting startup tasks.
It prompt that as a result of I had written extensively about presentation techniques within the paperwork I uploaded. After I clicked on that question, NotebookLM drew from numerous strands of supplies in my notes and created a abstract of a few of my key factors.
To assist me establish the place it was sourcing the fabric in that abstract, NotebookLM supplied an inventory of 10 citations from throughout the 9 paperwork I had uploaded. I may then click on on the citations to view these unique supply sections. I may additionally add that abstract to my noteboard as a brand new observe. That’s helpful when growing one thing new.
Ask your personal questions
NotebookLM additionally permits you to pose any query you need. For instance, I queried my very own entrepreneurial journalism notes about previous discussions of podcast income streams.
Responses from AI chat providers like Claude, ChatGPT, and Gemini might be arduous to hint to any explicit supply. (Perplexity, the most useful AI search service, does share useful citations.) With NotebookLM, against this, you profit from AI utilized particularly and narrowly to evaluation of your personal observe, with citations.
- Caveat: Typically I discover the choices NotebookLM makes about which sections of a supply to quote to be puzzling.
Constructing by yourself notes
After querying NotebookLM, it can save you its responses and construct on them to make new notes. These reside in your noteboard, together with different AI queries and new notes you make. For those who’re writing one thing new or getting ready a presentation, you should use NotebookLM to help you in exploring your supplies. You possibly can have a dialogue with your personal notes. That’s probably the most helpful facets of NotebookLM: It goes far past looking for key phrases.
Limitations
- As of Might 2024, NotebookLM is obtainable solely within the U.S. to folks 18 and older. And a few Google accounts might not have entry. I’m ready to make use of it with my private Gmail account, for instance, however not with my work account.
- There’s no cellular app, although the cellular Net model works properly for easy queries.
- The design nonetheless feels a bit clunky to me. The interface contains a number of panels—one on your supply paperwork and one other on your queries, alongside numerous notes you save or create. On condition that it is a new sort of instrument, it might take a while for a streamlined view to emerge.
- Google doesn’t prepare its fashions in your materials, which stays non-public to you. However in case you’re skeptical concerning the firm, it’s possible you’ll not need to add non-public materials.
Options
- Mem is a notes app with AI that trains itself in your notes. It surfaces associated notes routinely as you kind. You may also ask questions of your personal notes, very similar to NotebookLM. However Mem is a extra totally useful notes instrument. I wrote about why Mem is like a next-generation Evernote.You possibly can’t compartmentalize its AI evaluation for explicit notes or notebooks in the identical means you possibly can with NotebookLM. Mem works solely with Google authentication, so that you want a Google account to make use of it. Pricing: $100/yr or $15.
- AnythingLLM is a free software program instrument you possibly can obtain to run an AI mannequin by yourself laptop computer. You possibly can feed it any notes, paperwork, or recordsdata you need it to use AI to. You possibly can even run the mannequin fully offline, so that you don’t need to add the recordsdata to a cloud server. I examined this to use AI to a totally non-public set of recordsdata alone laptop computer. It’s surprisingly simple to arrange and use. I anticipated to spend an hour following directions, nevertheless it was up and operating in about quarter-hour on my 2021 MacBook Professional. I even ran it offline. AnythingLLM works on Mac, Linux, and Home windows desktops. I downloaded and ran the open-source Llama 3 mannequin, which requires about 4.7GB in area.
This text is republished with permission from Surprise Instruments, a e-newsletter that helps you uncover essentially the most helpful websites and apps. Subscribe here.