Like nearly everybody, we have been impressed by the power of NotebookLM to generate podcasts: Two digital folks holding a dialogue. You can provide it some hyperlinks, and it’ll generate a podcast primarily based on the hyperlinks. The podcasts have been fascinating and interesting. However additionally they had some limitations.
The issue with NotebookLM is that, when you can provide it a immediate, it largely does what it’s going to do. It generates a podcast with two voices—one male, one feminine—and provides you little management over the end result. There’s an elective immediate to customise the dialog, however that single immediate doesn’t let you do a lot. Particularly, you possibly can’t inform it which subjects to debate or in what order to debate them. You’ll be able to attempt, nevertheless it received’t hear. It additionally isn’t conversational, which is one thing of a shock now that we’ve all gotten used to chatting with AIs. You’ll be able to’t inform it to iterate by saying “That was good, however please generate a brand new model altering these particulars” like you possibly can with ChatGPT or Gemini.
Study quicker. Dig deeper. See farther.
Can we do higher? Can we combine our data of books and expertise with AI’s means to summarize? We’ve argued (and can proceed to argue) that merely studying tips on how to use AI isn’t sufficient; you should learn to do one thing with AI that’s higher than what the AI may do by itself. It’s good to combine synthetic intelligence with human intelligence. To see what that may appear to be in observe, we constructed our personal toolchain that provides us far more management over the outcomes. It’s a multistage pipeline:
We use AI to generate a abstract for every chapter of a ebook, ensuring that every one the essential subjects are lined.We use AI to assemble the chapter summaries right into a single abstract. This step primarily provides us an prolonged define.We use AI to generate a two-person dialogue that turns into the podcast script.We edit the script by hand, once more ensuring that the summaries cowl the appropriate subjects in the appropriate order. That is additionally a possibility to appropriate errors and hallucinations.We use Google’s speech-to-text multispeaker API (nonetheless in preview) to generate a abstract podcast with two members.
Why are we specializing in summaries? Summaries curiosity us for a number of causes. First, let’s face it: Having two nonexistent folks focus on one thing you wrote is fascinating—particularly since they sound genuinely and excited. Listening to the voices of nonexistent cyberpeople focus on your work makes you’re feeling such as you’re residing in a sci-fi fantasy. Extra virtually: Generative AI is certainly good at summarization. There are few errors and virtually no outright hallucinations. Lastly, our customers need summarization. On O’Reilly Solutions, our clients steadily ask for summaries: summarize this ebook, summarize this chapter. They need to discover the data they want. They need to discover out whether or not they actually need to learn the ebook—and in that case, what elements. A abstract helps them do this whereas saving time. It lets them uncover rapidly whether or not the ebook can be useful, and does so higher than the again cowl copy or a blurb on Amazon.
With that in thoughts, we needed to actually suppose by what probably the most helpful abstract can be for our members. Ought to there be a single speaker or two? When a single synthesized voice summarized the ebook, my eyes (ears?) glazed over rapidly. It was a lot simpler to take heed to a podcast-style abstract the place the digital members have been excited and enthusiastic, like those on NotebookLM, than to a lecture. The give and take of a dialogue, even when simulated, gave the podcasts vitality {that a} single speaker didn’t have.
How lengthy ought to the abstract be? That’s an essential query. In some unspecified time in the future, the listener loses curiosity. We may feed a ebook’s whole textual content right into a speech synthesis mannequin and get an audio model—we could but do this; it’s a product some folks need. However on the entire, we anticipate summaries to be minutes lengthy reasonably than hours. I’d hear for 10 minutes, perhaps 30 if it’s a subject or a speaker that I discover fascinating. However I’m notably impatient once I take heed to podcasts, and I don’t have a commute or different downtime for listening. Your preferences and your state of affairs could also be a lot totally different.
What precisely do listeners anticipate from these podcasts? Do customers anticipate to study, or do they solely need to discover out whether or not the ebook has what they’re on the lookout for? That is dependent upon the subject. I can’t see somebody studying Go from a abstract—perhaps extra to the purpose, I don’t see somebody who’s fluent in Go studying tips on how to program with AI. Summaries are helpful for presenting the important thing concepts introduced within the ebook: For instance, the summaries of Cloud Native Go gave a superb overview of how Go may very well be used to handle the problems confronted by folks writing software program that runs within the cloud. However actually studying this materials requires examples, writing code, and working towards—one thing that’s out of bounds in a medium that’s restricted to audio. I’ve heard AIs learn out supply code listings in Python; it’s terrible and ineffective. Studying is extra seemingly with a ebook like Facilitating Software program Structure, which is extra about ideas and concepts than code. Somebody may come away from the dialogue with some helpful concepts and presumably put them into observe. However once more, the podcast abstract is simply an outline. To get all the worth and element, you want the ebook. In a latest article, Ethan Mollick writes, “Asking for a abstract just isn’t the identical as studying for your self. Asking AI to unravel an issue for you just isn’t an efficient approach to study, even when it feels prefer it ought to be. To study one thing new, you’ll should do the studying and considering your self.”
One other distinction between the NotebookLM podcasts and ours could also be extra essential. The podcasts we generated from our toolchain are all about six minutes lengthy. The podcasts generated by NotebookLM are within the 10- to 25-minute vary. The longer size may enable the NotebookLM podcasts to be extra detailed, however in actuality that’s not what occurs. Slightly than discussing the ebook itself, NotebookLM tends to make use of the ebook as a leaping off level for a broader dialogue. The O’Reilly-generated podcasts are extra directed. They observe the ebook’s construction as a result of we supplied a plan, a top level view, for the AI to observe. The digital podcasters nonetheless specific enthusiasm, nonetheless usher in concepts from different sources, however they’re headed in a route. The longer NotebookLM podcasts, in distinction, can appear aimless, looping again round to select up concepts they’ve already lined. To me, at the least, that seems like an essential level. Granted, utilizing the ebook because the jumping-off level for a broader dialogue can be helpful, and there’s a steadiness that must be maintained. You don’t need it to really feel such as you’re listening to the desk of contents. However you additionally don’t need it to really feel unfocused. And if you would like a dialogue of a ebook, it’s best to get a dialogue of the ebook.
None of those AI-generated podcasts are with out limitations. An AI-generated abstract isn’t good at detecting and reflecting on nuances within the unique writing. With NotebookLM, that clearly wasn’t underneath our management. With our personal toolchain, we may definitely edit the script to replicate no matter we wished, however the voices themselves weren’t underneath our management and wouldn’t essentially observe the textual content’s lead. (It’s debatable that reflecting the nuances of a 250-page ebook in a six-minute podcast is a shedding proposition.) Bias—a type of implied nuance—is an even bigger subject. Our first experiments with NotebookLM tended to have the feminine voice asking the questions, with the male voice offering the solutions, although that appeared to enhance over time. Our toolchain gave us management, as a result of we supplied the script. We received’t declare that we have been unbiased—no person ought to make claims like that—however at the least we managed how our digital folks introduced themselves.
Our experiments are completed; it’s time to indicate you what we created. We’ve taken 5 books, generated quick podcasts summarizing every with each NotebookLM and our toolchain, and posted each units on oreilly.com and in our studying platform. We’ll be including extra books in 2025. Take heed to them—see what works for you. And please tell us what you suppose!
