Google's new AI podcast tool goes viral

there's a new hit podcast taking X in Tech World by storm it's called Google Deep dive it's taking on everything from the latest in science to classical literature the only catch the hosts aren't real and the producer is you this is all part of Google's new notebook LM project and dear daosa has more in today's Tech check Hi D Morgan I hope you're intrigued because instead of telling you about this new tool I'm going to show it to you now earlier in the show I heard that you guys we're talking about that JP Morgan iPhone

survey here's another version of that in podcast form with two AI generated podcasters have a listen get this we're diving deep into iPhone territory today but not just any iPhone chitchat right we're talking about why people are really shelling out for those new Apple phones luckily we've got some Clues to work with we've got this juicy JP Morgan survey yeah over 500 people oh gave their insights on smartphones and buying decisions that's a pretty good sample size actually yeah it gives us a good snapshot of what's really going on I mean I don't know about

you guys but this blew my mind this is all AI generated and it sounds pretty close to how someone might expect Carl and I or Carlin Morgan to sound on a podcast breaking down that JP Morgan survey right down to the ahun and the banter it's worth the full 10-minute listen which we'll post now analyst Notes One useful example for our audience but notebook LM is build more broadly as as a personal AI research assistant students are using it to make their textbooks into more engaging podcasts lawyers can upload legal briefs scientists with lab notes

and research papers this is powered by Google's most advanced AI model Gemini 1.5 Pro and judging by the reactions to it including my own and the tech check teams and the broader Tech Community on X this could be the next killer app in generative AI now the next big thing in gen has been multimodal capability or the ability to process and understand multiple types of input and output like text images video audio even sensory data audio in particular guys is proving a competitive Battleground with meta celebrity voice chat Bots open AI advanced voice mode that

was outlined at developer day earlier this week and apple and Amazon's promises for better voice assistant Siri and Alexa Google though may just have the viral moment here with this notebook LM multimodal capability plus a killer use case now its product lead posted this meme saying our tpus right now on X just a few days ago suggesting that all of this attention has strained Google's AI chip infrastructure and keep in mind guys that this is still in lab's AKA Google's beta or experimental mode as Google tends to do as we know though the space and

Technology it moves incredibly fast open AI billions of new capital that will make sure that it stays that way and they'll have a competitor you might just think finally that this is podcast and maybe gimmicky but that's like saying chat gbt was just a chatbot some of the biggest platform shifts are driven by consumer adoption and sets the stage for Enterprises to pick up and this podcast a good example of what gen AI can do I have so many questions I have so many and I I I I will definitely listen to the full thing

here um as a podcaster myself I'm really curious how this is going to shake out but the voices themselves I just want to go back to that because there's already been this sort of sticky issue of you know who owns the voices where the voices are coming from uh we know FTC is looking at stuff like this right now I just wonder how that's factoring in here right so you can sort of generate or build totally unique voices but if you want say a John Cena or Julie Dench like meta does you license it right

and so this whole licensing issue is becoming so interesting in generative AI in this case for Notebook LM you actually upload the documents your self you can upload text from a web page you can't go past pay walls um but you can upload up to I believe 50 different documents or audio pieces to give you something so it also says that they don't share it they don't use it to actually build further upon it but it's tricky Morgan um it does raise a lot of questions and the Really the industry is figuring that out now

how much do you license who do you license from how much do you pay that's all being negotiated in real time it does does seem like every day is some eye openening application that you can Envision 5 or 10 years from now maybe not even that long uh de fascinating thank you dear daosa