¶ Lyria 3
Josh: Just in the last week, Google unveiled some pretty unbelievable new AI software. Josh: One of my favorites being this new thing called Lyria 3. Josh: So on screen here, I have a music generator that I actually am going to prompt to make a song for us. Josh: So what I've loaded up here is make me a 2010s Kanye style stadium anthem about Josh: the Limitless podcast hosted by us.
Josh: And what I kind of want to do is, is have it do this in real time while we discuss Josh: the product while we discuss the software. Josh: So Lyrae 3, what you noticed on screen there, which was super cool, Josh: is you choose the genre you want, you feed a prompt to the model, Josh: and after you fed the prompt to the model, it will go and generate you lyrics, the full production, Josh: and a full 30-second song about whatever it is you want.
Ejaaz: Yeah, I saw a hilarious example of a lady asking her husband to do the dishes, Ejaaz: but instead she wrote it as a message and created a song and sent it to Dan, Ejaaz: her husband and it played just like hilarious music all jokes aside i think Ejaaz: this is more than just a gimmick or a novel you're about to see the quality Ejaaz: of this thing through the jingle it's just generated but this could replace
Ejaaz: a music series entirely at least for for 30 second excerpts for tiktok or whatever that might be are Josh: You ready to listen to it we have our output ready. Ejaaz: Let's go Josh: Whoa, even an outro too. Wow, okay, that's pretty good. Ejaaz: Just to be clear, that was generated in about under two minutes. Josh: I'd say less than one, fast. And the production is good. Like the lyrics sound good.
Josh: It's like a proper rap song. And I think that's one of the novel breakthroughs Josh: with this technology is that a lot of these tools that Google in particular Josh: has been releasing recently are really enabling creators to have this powerhouse Josh: suite of tools that let you do things like create music. Josh: If we want to have an intro song, we can actually use that or you can kind of Josh: workshop it and change a few of the lyrics and change the style.
Josh: But I think what's really fun is that you can actually choose the types of songs Josh: that you want to create and you can embed them into whatever content you're making. Josh: So there's a second feature to Lyria 3, which is around generating photos or Josh: generating songs through photos.
Josh: So what I have here is a photo of cold and snowy New York. We just got a blizzard Josh: last night, and I'm going to say, generate me a song that reflects the mood Josh: in this very snowy picture. Josh: And yeah, I mean, we were under a blizzard warning lockdown last night. Josh: Things are really bad here in New York, and maybe it will. Josh: Yeah, well, that's right. EJ's went to Florida, but maybe what we can do here
Josh: is kind of synthesize that feeling through music. So we'll see what it does.
¶ Multimodal Capabilities
Ejaaz: This is something that Google is really well known for, which is multimodality when it comes to AI. Ejaaz: It's not just a great LLM or chatbot. Ejaaz: It understands visuals such as videos and images. It understands sounds. Ejaaz: Remember, this isn't the only, I guess, sound model that they've created before. Ejaaz: They've had live translation with their AI that translates any kind of, Ejaaz: I think, 40 plus languages into whatever language that you want.
Ejaaz: They've had experience in this. And the fact that their models now have this Ejaaz: intuition and this judgment, this capability to be able to see a snowy day and Ejaaz: say, hmm, maybe I'll generate a vibey song for that is pretty impressive. Ejaaz: It looks like we have it already. Wow, that was under 30 seconds. Josh: Yeah, you ready to give it a listen? Ejaaz: Yeah, play it. Ejaaz: Are your spirits lifted?
Josh: Sonically, it sounds nice. I feel like I'm listening to a like kind of knockoff Josh: Demi Lovato, like main character energy type thing. And I like it. Josh: And to your point, I think this is something unique to Google in the sense that Josh: because they have these world models, they have that understanding of what they're Josh: seeing in a way that I think is more intuitive than a standard language, Josh: text-based language model would do.
¶ Watermarking
Josh: And another interesting thing that I learned while kind of Josh: playing around with this is they have this tool called synth id and Josh: what it is is it's a digital watermarking service that's baked Josh: into these files that we're listening to so that Josh: if you have an algorithm that can detect it it can actually know with high levels Josh: of certainty that ai was used to generate the music which is good for copyright
Josh: and for kind of trademark issues and the way it works is it injects this unhearable Josh: waveform within the music that we just heard that we can't detect as human beings Josh: but should you feed it to an AI, Josh: it'll be very easy for it to decrypt that and let you know that this was AI generated. Josh: So pretty cool stuff. I really enjoyed Lyria 3. And I think a lot of people Josh: would have fun playing with this. It's free to use, available in Gemini.
Josh: And yeah, super fun for people who, Josh: wanted to be producers or just want to send their friends a funny birthday song.
Ejaaz: Well listen my take on this is that it's actually Ejaaz: i'm going to go back on what i said earlier it's probably not going to take over music labels Ejaaz: anytime soon but for like v1 or v3 you can imagine what this probably looks Ejaaz: like in about six months time and if it's taken if it improves at any of the Ejaaz: rate that any of the llms have like this is going to be a pretty insane thing Ejaaz: to do um on the synth id thing josh the reason why i was wondering why when
Ejaaz: you were explaining it why it sounds familiar to me Ejaaz: they do this with AI images as well where they kind of like mix up a few certain Ejaaz: pixels in an image that is generated to tell you that this is AI generated. Ejaaz: They do this with text as well where they subtly kind of like change certain Ejaaz: word choices to do the same thing. Ejaaz: My question to you is or rather to both of us is can you create models now that just don't do that?
Ejaaz: Like this is a baked in watermark that Google kind of put up, Ejaaz: right? But presumably you could create a model that just Ejaaz: I guess, illegally copyrights a bunch of this stuff? Josh: Well, my assumption is that if you can detect it, then you can reverse what Josh: it's detecting by using the same tool. Josh: So like if an AI can be aware of the trademark, it can probably reverse engineer Josh: those few pixels or that waveform that exists and remove it entirely.
Josh: So I suspect it probably goes both ways, but it is nice to know and nice to have. Josh: And I think it's funny because we are following this trend of, Josh: again, reverse CAPTCHAs, where we're building things for AIs to recognize and Josh: not humans to recognize and this is another step in that direction.
Ejaaz: Well why i like it as well is um if Ejaaz: you are a music creator that's listening to this or even like an artist Ejaaz: that creates images you might be thinking well they might be stealing my work Ejaaz: and i get nothing for it with watermarks like this it'll be recognizable to Ejaaz: your attributes or brands and maybe you end up getting paid out for this something Ejaaz: like in some kind of future system that doesn't really exist right now and it
Ejaaz: definitely beats any kind of archaic royalty system that existed before so So I don't know, Ejaaz: I just like the technology. I think it's more than just a watermark. It's pretty cool.
¶ Pomelli
Josh: Yes, and that is not the only Google Labs cool new thing because there was another Josh: really fun creation that they had that they published to the Google Labs team Josh: called Pomelli. Do you want to walk us through what this post says? Ejaaz: Yeah, so I mean, I hate to say it, but it's true.
Ejaaz: From a single image of your product, you can now generate several thousand dollar Ejaaz: worth photo shoot, which includes people that feature your product, Ejaaz: different types of lightings and textures, Ejaaz: backgrounds, all in a matter of seconds. Ejaaz: And the reason why that's a crazy thing to say is Ejaaz: there's a lot of photographers that spend a lot of their time and expertise Ejaaz: creating these photo shoots that now are pretty much out of a job.
Ejaaz: And I'd also take that a further step and say that there are a lot of fashion Ejaaz: models that also may not be able to benefit from this as well. Ejaaz: But I have to say the feature is so cool. The fact that you can go from a single Ejaaz: person sitting at home that's ordered or generated their own product to a full-on Ejaaz: advertised kind of website that features these slick different images. It's just so cool to me.
Josh: Yeah. And we have some examples of this. I mean, one of them I used for the Josh: bankless website that we use and posted about it. Josh: And it got like a million views because it's really awesome.
Josh: What it does is, is you feed it a URL. So if you have a personal brand, Josh: if you have a website, if you have any sort of content on you, Josh: you feed it the URL and then it populates this business DNA sheet where it shows Josh: you your logo, the fonts, the colors, your tagline, all the values. Josh: And it It extracts this kind of identity that you can then use to create this Josh: marketing material with.
Josh: And another fun example that we generated with the Bankless stuff was a hat. Josh: We had a hat in the merch store and it's just this very basic bland hat. Josh: But what I did is I fed it through Pomelli and I asked it to do a product shoot. Josh: And what it generated on the other side was really impressive. Josh: It shows, yeah, we could see here it has a model wearing the hat that looks Josh: like it was shot professionally. The lighting is very cool.
Josh: The everything looks real. I can't really like if if you were to show me this Josh: in a magazine or just scrolling on x I would think that's zooming.
Ejaaz: In here josh this Josh: Like there's really incredibly real indications and Josh: this took 30 seconds to generate all by feeding him Josh: the hat and then it created this really epic product shot that was Josh: just this fun kind of hero image showcasing the hat Josh: showcasing the logo in the middle of it and I think to Josh: your point I mean this is something that a lot of creators would see as a huge
Josh: barrier where it it costs a lot of money to get these professional looking shots Josh: to hire models to have the lighting and the photo shoots and the reality is Josh: is that it's really not that difficult to do if you use this tool so what i figured we could do, Josh: is we can try to use Pomelli here live and do a demo for Limitless. Josh: So I actually have it loaded up right here. Josh: We can run through exactly how it works and create our own business DNA.
Josh: So when you click let's go, it asks you to use the website. We kind of have Josh: a website, limitless.bankless.com if anyone wants to go check it out. Josh: And I'll feed it in there. And what you'll see is it takes a few minutes. Josh: It'll analyze everything. It clicks through and it starts to understand what Josh: colors you use, what images you use, what type of copy you use.
Josh: And while that's thinking, maybe we could kind of describe a little bit more Josh: about how it works and what it's used for. Ejaaz: Yeah. So the rough kind of steps would be you can enter your website URL and Ejaaz: Pomelli basically just scans your entire business and extracts, Ejaaz: as you mentioned earlier, something that would be referred to as, I guess, a business DNA. Ejaaz: So it kind of judges your vibe. It gets your vibe. And it's a theme of all these
Ejaaz: different products. Google's models are very good at kind of gauging sentiment. Ejaaz: So it'll understand kind of like the tone of your voice, the color palette, Ejaaz: the fonts, the brand identity, stuff like that. Ejaaz: And it generates a completely kind of novel marketing campaign that hasn't existed Ejaaz: for any other product before or is based on the type of vibe that you generally like.
Ejaaz: And with all these things, I think that, well, Ejaaz: For example, like in previous jobs that I've worked, Josh, I've worked with Ejaaz: massive marketing departments and a large chunk of their time and money is spent Ejaaz: on creating these types of photo shoots, gauging the sentiment. Ejaaz: Now, what I will say, OK, I have to argue the other side here, right?
Ejaaz: I've seen a bunch of demos of these product shoots. And after a while, Ejaaz: you kind of see a kind of similar sense and style. Ejaaz: Some of the product shoots and angles are kind of repeated. So if you want to Ejaaz: get something unique, you do need to be very descriptive and kind of know and see what you want. Ejaaz: And that argues in favor of the product shoot photographers, Ejaaz: which already have the ideas and experience, kind of like a movie director.
Ejaaz: They're not going out of job because of video models. Ejaaz: They kind of know the artistic shots and they just need to kind of upgrade their toolkit per se. Ejaaz: So I think we're going to see something similar like that rolling out. Ejaaz: I don't think it's at its final form just yet.
¶ Product Shoots
Josh: I think you're right in the sense that like it can get you 80% of the way there, Josh: but it's not going to give you that professional look so for the Josh: people who don't have these budgets this is awesome for the people who Josh: do i mean there's clearly a a gap still that Josh: exists but you have to assume that gap is going to be filled fairly quickly Josh: so it generated our dna and you could see it's already working on generating
Josh: these automatic um kind of images that we could use perhaps to post on twitter Josh: to get people excited i mean it shows master the llm power shift it understands Josh: the topics that we're talking about the trends um opus the deep dive context mastery This just. Ejaaz: Jogged my memory. This reminds me of the live demo we went through with Nano Banana. Ejaaz: And I believe this is using Nano Banana on the back end, right? Ejaaz: Google's image generation model.
Josh: I believe so. So we're using Nano Banana Pro. And we're going to test that by Josh: using it for a photo shoot. So what I did is I stole a little hat. Josh: And I'm going to drop the hat in here.
Josh: And we can have it do a photo shoot in real Josh: time that allows us to generate our Josh: own limitless branded imagery so Josh: we'll load that up we'll change the format to Josh: be let's say square and then we could choose the photo shoot template so maybe Josh: we'll choose flat lay in use and then we'll do two of these product shots here Josh: and we'll say it looks good and generate and now while that generates we can
Josh: talk about yeah maybe the things that are powering this so one is nano banana Josh: pro but two is i mean this is using gemini 3.1 pro as well which is the new model, Josh: brand new just came out.
Ejaaz: Last week a lot of a lot of new launches um Ejaaz: i was also looking at the costing for a typical product Ejaaz: photo shoot josh it's between 500 bucks to Ejaaz: 5 000 bucks on the high-end scale and a lot of the images Ejaaz: that we're seeing here are on that high-end scale it takes Ejaaz: a while to kind of like edit in post-production and make Ejaaz: things like really crystal clear so that's super impressive um
Ejaaz: what i'm also impressed by is this is like available to quite a few uh different Ejaaz: regions so typically when google launches their ai products it's just for the Ejaaz: u.s only but this is available in the u.s canada australia new zealand so think Ejaaz: of like the entire swathes of people that now get access to a tool that cost Ejaaz: them cents instead of these thousands of dollars it's just so cool yeah
Josh: It's um it's amazing and i think it's a testament to the direction that they're Josh: going towards which is just kind of complete and total value creation for anyone out there. Josh: Like they're working on, they have their coding agent, they have their IDE, Josh: they have their video generation with VO3, they have audio generation now, they have brands and...
Josh: Brand DNA and I guess marketing and understanding of that. And there really Josh: is no industry that's safe from these tools. Josh: And as these tools continue to get better, they're just going to become way Josh: more proficient and way more optimal as it relates to kind of not replacing Josh: these jobs, but enhancing and augmenting people's abilities that work in these spaces.
¶ The Evolution of AI Models
Ejaaz: But let me ask you this question, Josh. Do you feel that way about all these products in unison Ejaaz: today because my take is it's great individually but i want an entire suite Ejaaz: that can manage all of this for me and they're working towards that right with google ai suite Josh: Yeah there's going to be a single comprehensive tool Josh: that has all of these things and it's exciting Josh: so we have the outputs here which um i didn't load the
Josh: logo onto and i didn't ask to load the logo onto so perhaps Josh: that's why our logo isn't actually on there but i mean Josh: to your point you can see the bankless one that showed you earlier looks very Josh: similar to this one yep and the model actually looks Josh: like kind of similar to the last one and you can see there are Josh: some i guess kind of restraints and limitations Josh: as it relates to the quality of the outputs but to have something this good
Josh: this quickly this easily it's really impressive and it's a fun tool to use um Josh: and like i mentioned earlier under the hood gemini 3.1 it just came out last Josh: week um there's a lot of good and bad news about it so maybe we should talk about that next. Ejaaz: Yes okay let's get back to being suited and booted uh gentlemen we have to get Ejaaz: into the llms um google dropped a brand new model and on paper
Ejaaz: it's pretty damn impressive. What I've got in showing on the screen right here Ejaaz: is the ARC AGI2 benchmark. Ejaaz: Now, just for context here, there was an ARC AGI1, but the models became so Ejaaz: good that they needed to create a brand new test to make sure that these models Ejaaz: were actually getting more intelligent. Ejaaz: So ARC AGI2 is a benchmark that was set. Ejaaz: And I just want to focus on a little difference here. And by little, I mean a lot.
Josh: This is more than a double. Ejaaz: That's more than a double. It's a 46 percentage point increase from Gemini 3 Pro. Ejaaz: And I want to point out that this is a 0.1 version update. Ejaaz: We've gone from 3 to 3.1. This isn't even Gemini 4 yet. And we've seen such Ejaaz: a major leap in intelligence and reasoning. And the reason... Josh: I just looked this up, actually, just real quick. It came out in November 18th of 2025. Ejaaz: So this is Josh: Less than four months.
Ejaaz: The iteration cycles for these model updates are getting astounding, to be honest.
¶ Advances in Gemini 3.1
Ejaaz: We've spoken about Claude Opus 4.6 and ChatGPT 5.3 Codex. Ejaaz: These are the new coding models from Anthropic and OpenAI, respectively. Ejaaz: Those released within a week, sorry, within an hour of each other. Ejaaz: And their previous model version updates were three weeks prior to that. Ejaaz: So the fact that these models are improving at a rapid rate doesn't surprise me.
Ejaaz: The secret source behind that is these models are most likely working on themselves, Ejaaz: i.e. they're reviewing their own code and updating themselves, Ejaaz: which is a scary topic, which we'll get into another time. But back to Gemini 3.1 Pro. Ejaaz: It excels in two major leaps. Number one, what you're seeing on the screen here, AGI 2, reasoning.
Ejaaz: So it understands what you're saying. It goes deep into the weeds of your prompt Ejaaz: and gives you a really well thought out answer. Ejaaz: Again, on paper. The other thing that it's really good at is coding. Ejaaz: It's insanely good at generating visuals specifically using code and understanding Ejaaz: the physics of simulations that you create. Ejaaz: But I have to point out that there's been a bunch of bad feedback from people Ejaaz: actually testing these things.
Ejaaz: So what I've just explained to you has been on paper, but in practice, Ejaaz: people observe that the reasoning actually kind of defeats its own self. Ejaaz: It starts getting into these thinking loops where it starts doubting itself, Ejaaz: questioning its own answers, and you end up waiting 10 minutes for what would Ejaaz: otherwise be a very simple answer.
Ejaaz: And then other people's experience is when they use the web app version to access Ejaaz: Gemini 3.1, it's actually a much more reduced model that just agrees with them. Ejaaz: So it kind of sounds a lot like GPT-4.0, which is known for being very agreeable, Ejaaz: which OpenAir actually decommissioned last week. So there's a lot of mixed reviews for this. Ejaaz: There are a few cool examples from the big man himself, Jeff Dean.
Ejaaz: He's showing us an example here of Gemini 3 Pro on the left and Gemini 3.1 Pro on the right. Ejaaz: As you can see, the SVG generations are just a stark difference, right? Ejaaz: The physics is super cool. The animations, look at the backgrounds on these things. It's so cool. Ejaaz: This is a really cool demo of someone using it to do urban planning. Ejaaz: I'm going to mute it so I can show you the start over here.
Ejaaz: Um, it's just super cool how we can map track and figure out spatial awareness. Ejaaz: It's a really spatially astute model, which is just super cool for creating Ejaaz: kind of games or any kind of simulated demos. Ejaaz: But yeah, the, the actual practical feedback from people hasn't been that great.
Josh: It's funny to see the evolution of these models in between major releases, Josh: because I mean, what happens during these incremental releases, Josh: they're not doing this huge pre-training loop where they're, Josh: they're building an entirely new model. Josh: They're taking the existing core model and they're kind of building on top of it. Josh: And what you find is that these incremental models become spiky in some different areas.
Josh: So some are much better at some things, but the trade-off is that it gets worse. Josh: I mean, a good example we had is GPT 4.5, I think it was, which was trained Josh: to be an excellent writer, but it was actually so bad at everything else and Josh: so expensive that they had to depreciate it.
Josh: And I think what you see throughout these kind of incremental evolutions, Josh: 3.1 being one of them, is that they're trying to improve them and they will Josh: improve them in some areas, but there are some unexpected areas in which the Josh: net quality actually declines. Josh: So each one has its own personality. It has its own skill set.
Josh: Net net they're better because they've been refined and they've been distilled Josh: down into just a highly refined model but there are going to be those like downsides Josh: that we see and i think that's just the natural part of these incremental releases Josh: is kind of seeing how these these models evolve over time from that huge core base model my.
Ejaaz: Major lesson from this uh i think which is one that we've learned a while ago Ejaaz: is just don't trust the benchmarks just trust your own experience and obviously Ejaaz: different people use these models for different things um do i think this is the end of Google? Ejaaz: No, but they did go on a gargantuan winning streak. Ejaaz: And this may have subdued them just slightly. Ejaaz: The good news is I hear that Gemini 4 is cooking up an absolute beast.
Ejaaz: And what I like is that they're integrating. Yeah, I'm integrating. Ejaaz: What I like is that Google's integrating a lot of these models into a singular Ejaaz: interface. I mentioned this earlier. Ejaaz: The one thing that annoys me is I need to go one place for Nano Banana. Ejaaz: I need to go another place for VO3, their video generation model. Ejaaz: And then I need to go another place to access Gemini 3.1.
Ejaaz: Sometimes you can access it through the API, sometimes it's through the web Ejaaz: app, sometimes randomly it's for free through Google search. Ejaaz: And so I'm like, I just want one place to go to. I'll give you my money. Ejaaz: Just let me combine all these things into one experience. Ejaaz: And that's all I want for like a single price, a single Netflix subscription. Ejaaz: And they're moving towards that phase, which I think is great.
Ejaaz: And so my guess is maybe they kind of dropped the ball here because they're Ejaaz: focusing on all that other stuff. Ejaaz: The other part I want to mention is Google's distribution is just insane.
¶ Google's Distribution Power
Ejaaz: Just through the other two features, Liria and Pomelli, they've been able to Ejaaz: distribute these to tens of millions of people on day one. Ejaaz: And that feedback loop cannot be understated. Ejaaz: So whereas Google, you know, may have this initial reaction, Ejaaz: they have been the quickest to bounce back. Ejaaz: Don't forget where Google was two years ago when they had the worst AI model.
Ejaaz: And now they're like fighting for the throne pretty much. So I wouldn't count them out. It's cool. Josh: Yeah, it feels like the Adobe Creative Cloud Suite on steroids for Google and Josh: for creators. And I think that's something... Josh: I can get really excited about. And what you saw with these demos today is it's Josh: real technology that's here today and it's improving very, very quickly.
Josh: So I'm excited for 3.2, 3.3, and then 4.0 whenever that comes, Josh: which I'm sure is going to shock the world. Josh: But that wraps everything up for our fun little demo day. Josh: I would encourage you to go and try these things. I mean, a lot of them are Josh: just available for free, which is pretty cool. Josh: You can just go and test them out, create a song, send it to your friend, Josh: roast them, whatever it may be. Josh: It's a fun way to kind of help creators, help marketers.
¶ Make Stuff
Josh: These tools are awesome and they're coming so quickly so if Josh: you enjoyed please don't forget share it with a friend subscribe Josh: to our newsletter it comes out twice a week every single week last week we had Josh: our biggest week ever we never had more views than last week so the mission Josh: is working and it's thanks to you guys for sharing it with your friends and Josh: liking it and commenting rating five stars on your favorite podcast platform
Josh: each has any final notes before we sign off here i. Ejaaz: Have one small piece of feedback sorry feedback i have a small piece of homework Ejaaz: for those of you who are up to it um i'm desperate to hear your music creations Ejaaz: um or even your product photo shoots so try and find josh and i uh on twitter Ejaaz: we'll link to our handles or uh profile pages below Ejaaz: and dm us i want to hear some of your creations because i'm genuinely uh interested
Ejaaz: in how creative people can get um and maybe it's a breakup text or maybe it's Ejaaz: telling your girlfriend to uh i don't know make something for you by the time Ejaaz: you get home i don't know i'm gonna i'm gonna try it out later and see what Ejaaz: it's like. But yeah, let us know your questions. I want to see what you guys are up to. Josh: Awesome. Well, thanks so much for watching. And yeah, we'll see you guys in the next one.
