GOOGLE'S NOTEBOOKLM CREATES PODCASTS

PREV ARTICLE NEXT ARTICLE FULL ISSUE PREV FULL ISSUE

V27 2024 INDEX E-SYLUM ARCHIVE

The E-Sylum: Volume 27, Number 41, October 13, 2024, Article 20

See the first article in this issue for a link to the newest NBS Bibliotalk podcast. In a first, it was generated by artificial intelligence. This came together quickly, and credit goes to NBS President Len Augsburger, who is also the Project Coordinator for the Newman Numismatic Portal. As an NNP consultant, I help gather new content and occasionally pass along ideas that may be relevant. After reading a Wall Street Journal article last week about the Google NotebookLM tool, I forwarded it to Len noting that it could be a way to generate accessible summaries of NNP content. Here's an excerpt from the article. -Editor

Have you heard about the latest hit podcast? It's called Deep Dive—and you have to check it out.

Each show is a chatty, 10-minute conversation about, well, any topic you could possibly imagine. The hosts are just geniuses. It's like they know everything about everything. Their voices are soothing. Their banter is charming. They sound like the kind of people you want to hang out with.

But you can't. As it turns out, these podcast hosts aren't real people. Their voices are entirely AI-generated—and so is everything they say.

This experimental audio feature released last month by Google is not just some toy or another tantalizing piece of technology with approximately zero practical value.

It's one of the most compelling and completely flabbergasting demonstrations of AI's potential yet.

This one is definitely cool, but it's also useful and easy to use. All you need to do is drag a file, drop a link or dump text into a free tool called NotebookLM, which can take any chunk of information and make it an entertaining, accessible conversation.

One of the coolest, most useful parts is that it makes podcasts out of stuff that nobody would ever confuse for scintillating podcast material.

Wikipedia pages. YouTube clips. Random PDFs. Your college thesis. Your notes from that business meeting last month. Your grandmother's lasagna recipe. Your resume. Your credit-card bill! This week, I listened to an entire podcast about my 401(k).

And then I found myself listening to an oddly captivating Deep Dive into that day's edition of the Federal Register.

Each conversation is between the same male and female voices—Google says they don't have names—and it takes only a few minutes to create one. If you didn't know anything about it, you wouldn't guess it was automatically generated. It doesn't sound like other AI slop. It just sounds like any other podcast. And once you start playing with it, you'll probably become obsessed with it.

The first time I heard a Deep Dive, it reminded me of an AI version of Acquired, the hit podcast about business history. When I told the hosts, they tried it out for themselves, opening NotebookLM and uploading a Google Doc with links to some of their sources for a recent show on Microsoft. Not their notes—just a bunch of links. They found the results to be mind-blowing.

"I don't know whether to be amazed or terrified," Acquired co-host David Rosenthal said.

Exactly. There have only been two times when I had my mind blown by AI. The first was my introduction to ChatGPT. This was the second.

"I'm in agreement. There are a few factual errors that the discerning ear can pick out, but the novice ear would take at face value. The podcast reminded me of a book report by a high schooler (perhaps myself at a time) presenting a tale built from a skimming of the text. Not quite right, but not easily dismissed."

This is a worthwhile experiment, illuminating at once how far AI has come, and how far it has to go.

On my first listen-through, I was amazed at how plausible the two voices sounded, and how much better AI is getting at simulating the give-and-take of human conversations. A casual listener, listening through once, probably would not suspect that this was AI-generated.

But the second listen-through revealed cracks in the polished facade. The identities of the two narrators subtly shifted between neutral interviewers to knowledgeable enthusiastic hobbyists and back again. There are at least half a dozen points at which one of the narrators responds in subtly inappropriate ways to the other (such as mismatched tones of voice). The patter between the narrators gets steadily more redundant as their "conversation" goes on until, near the end, a large chunk of the "dialogue" is repeated, almost verbatim.

If the order given was: "create a promotional piece for numismatic literature in the context of an informal dialogue between a lay male interviewer and a knowledgeable female numismatic literature collector" then Google Notebook LM did a reasonably good job (notwithstanding the hiccups mentioned above) of executing on that order. It delivered mostly plausible dialogue, lots of promotional sloganeering, and a number of solid examples (Crosby, Chapman), to support it. Notebook also pitched it right, threading the needle between abecedarianism and esoterica.

But if the order included, "and get the history right," the execution left much to be desired. John J. Ford did not invent numismatic bibliomania in a road to Damascus moment after a Garrett sale. Coin Cards are not the center of the numismatic bibliophilia universe. The dive into The Asylum somehow missed nearly everyone of prominence in the hobby, from Attinelli to Zander, and folks like Katen and Kolbe in between. In fact, outside of Early Coins of America, and plated Chapman catalogs, hardly a book or a catalog is mentioned.

So, a worthwhile experiment, but I hope that AI-generated content will be placed back on the shelf for the near future. It simply lacks the insights, wit, and creativity of human writers. Until it gains them, I see no need for the NBS to employ it.

Thanks, Joel. I would agree that this was a worthwhile experiment, although I wouldn't stop experimenting. The entirety of The Asylum is a lot of ground to cover, and human curation could help. An "audio overview" of a set of Asylum and E-Sylum articles mentioning say, John J. Ford or the Chapman Brothers could well be more aligned with the actual facts. Regardless, disclaimered less-than-perfectly factchecked summaries of obscure but interesting topics in numismatics could be just the lure to entice new collectors and researchers to the field. And who knows where the technology could be in two years, let alone twenty.

I was heartened this week to see well-deserved Nobel Prizes awarded to AI researchers Geoffrey Hinton and Demis Hassabis. -Editor

Both Hinton and Hassabis were born in London, albeit nearly three decades apart...

The foundations were shaped over centuries. The UK was a serious player in statistics, logic, mathematics and engineering – think Thomas Bayes, George Boole, Charles Babbage, Ada Lovelace – long before Alan Turing asked: "Can machines think?" As computers became an established technology, expertise flourished at a handful of centres.

First came the physics prize. The American John Hopfield and the British-Canadian Geoffrey Hinton won for foundational work on artificial neural networks, the computational architecture that underpins modern AI such as ChatGPT. Then came the chemistry prize, with half handed to Demis Hassabis and John Jumper at Google DeepMind. Their AlphaFold program solved a decades-long scientific challenge by predicting the structure of all life's proteins.