I just got a private walkthrough of

Google's new Gemini 3.0 launch from

Logan Kilpatrick. And I'm not

exaggerating when I say this is the most

powerful vibe coding tool I have ever

seen. And he takes us through how to get

the most out of it, how to create 3D

fully featured games, how to create a

startup from scratch. Today's episode,

if you stick to the end, is an unfair

advantage with how to get the most out

of Google's new powerful Gemini 3.0.

Enjoy.

>> All right, we got Logan on the pod. By

the end of this episode, Logan, what are

we going to learn? we're going to see

Gemini 3 in action building apps uh and

hopefully also showcase just like the

breadth of what this model is possible

uh and what what it's capable of. So

hopefully folks feel the magic uh as we

look through some of these demos

>> and by the end of the episode will they

have like an unfair advantage around how

to get the most out of this thing?

>> Yeah, I mean you're going to be able to

build apps using Gemini 3 Pro for free

in AI Studio. Um, and hopefully you'll

see all the sort of like pro features

and stuff like that as well. So, uh,

this is the moment to really ramp up

building things with with AI and, uh, I

think we'll deliver on that.

>> Let's let's get into it.

>> Yeah, I love it. So, the sort of like

top level, obviously, Gemini 3 is here,

uh, which is super exciting. Lots of

lots of hype over the last few months.

Uh, it's our new state-of-the-art model.

Um, sort of worldclass intelligence. Um,

and I think this like mission statement

for Gemini 3 really being this idea of

helping you bring anything to life. So

if you have a business idea, you have

sort of a artifact, a picture, a

textbook, a science paper, a receipt,

whatever it is, like Gemini 3 has this

capability to like bring whatever the

the thing is that you have to life,

which is really exciting. So we're

looking at it in AI Studio right now.

It's available across a bunch of Google

services like the Gemini app if you're

sort of looking for an assistant or uh

the API for developers and enterprises

if you want to build this experience uh

or build Gemini 3 into your own

products. Um but we're looking at it

right now inside of AI Studios vibe

coding experience. So if you want to go

and build an AI app uh or just build an

app in general and use Gemini 3 Pro to

either power that app or to build that

app for you um we can go ahead and do

that. So there's a huge breath. Um, and

Greg you I'll sort of click on some

examples and we can talk through them.

But, uh, yeah, the TLDDR is like just

the overall design aesthetic is

incredible. So these are a bunch of the

examples of, uh, you know, new landing

p. Not that the world needs more landing

pages.

But, you know, the aesthetics are

beautiful, which I think is the highle

uh, overview. Um, and these are like one

shot using Gemini 3 Pro, which is great.

So if you if any of these speak to you,

you can click into them and sort of

remix and change and deploy them and all

that stuff. Um the immersive games in 3D

worlds is this really exciting bit and I

think there's so many people and this is

like not super

business practical. Um maybe it is and

you can build like a small games

business, but I just I know so many

people who like got excited about coding

and building software and businesses and

stuff like that because they're really

excited about games. Um, and it's just

hard to make games if you've ever tried

before. Uh, like maybe just as hard as a

bunch of other maybe even harder than a

lot of other like traditional businesses

uh because of the complexity. And I

think with Gemini 3 Pro, you can just

vibe code games um in real time, which

is really really awesome. So I don't

know if any of these games speak to you,

but we can go in and click through one

of them really quick and then um talk

more about like very practical ones.

Well, so they all speak to me and I

wanted to actually speak to the point

around I don't know if games could, you

know, be helpful in your business and my

my take is that vibe coding games as a

marketing asset is so like you haven't

done that before. companies have not

done that before, but if you want to

create like um you know Gemini Racer for

whatever brand that you're working on,

let's say you're Lululemon or something

and it's just like a fun little

Christmas theme Lululemon runner, like I

don't know, but the fact that the fact

is like if you can like in a few prompts

build a game that could get traction on

on X on Tik Tok, on Instagram on you

know it might it's it's now possible.

>> Yeah, I love that. I think games are

also like a just a proxy which is I

think the other like very practical

point which is games are just a proxy

for like the model's overall capability.

So like hey maybe you're like I don't

actually care about building games. Um I

think as you see some of these examples

it like speaks to the level of

sophistication that the model has an

understanding of. Um and we can just

I'll click into one of these examples

really quick. Um and we can do this sky

metropolis one and like go and build a

bunch of uh stuff. I have not played

this game before, so I don't really

know. Uh, I assume the point is build a

great city. Um,

but all of this vibe coded inside of AI

Studio, which is really, really

exciting. And then the there's like some

AI integration as well, which is like

the city and the humans and the sort of

all the stuff that's happening is built

around um is powered by our models

behind the scenes. Um, so it's a fun

example of sort of what's possible. Um,

and I think this again, this is meant,

and Greg, you and I were talking off

camera about this. This is meant to

showcase like, you know, single prompt

generations. Uh, maybe like a couple of

multi-turn in some case, but like this

isn't something that like we paid a team

of like 10 engineers 5 months to come up

with and then put it into our demo. Like

this is literally a vibecoded example

that the team was able to build pretty

quickly. Um, so I think it speaks to

like how accessible it is to build a

bunch of this stuff. I feel like the

games, you know, people are going to

look at vibecoded games like right now

and they're they're going to think it's

a game, it's a toy. But I actually think

that people are the the people who

understand how important gaming is. I

mean, game games are content, right? And

you know, content is very important.

Content makes businesses. So, I think

games should be taken equally as

serious. So, I know I'm going to be

playing building games uh and shipping

games and seeing what happens. how, you

know, as a way to like spread whatever

it is I'm working on.

>> 100%. So, this is that like VR. I don't

know if this is actually going to break

my laptop because I'm in Inception right

here.

What?

Are you kidding me?

>> You You get the point. I I wanted to

pull out of that experience because

we're my I'm not hooked up to multiple

displays uh audio inputs or outputs. U

so I think this was potentially going to

break the recording, but uh that is all

vibe coded just using the camera out of

the box, which is a ton of fun. And so

you can go in and um the cool thing is

this is the tempo strike game that you

should be able to see. Um you should be

able to like remix this experience. So I

think this is the really fun part which

is like hey you're like hey this app is

actually weird in some way. I want to

add these features. You can go into the

gallery directly remix whatever it is.

Um and again this was vibecoded with

Gemini 3.0. Another one that I love is

sort of if you've seen this like famous

ball bouncing example, Greg, I don't

know if you see these on the internet

where people are like make the this is

like the quintessential AI test where

it's like put the ball in the box and

see if it bounces as it rotates in a

circle. Um so I came up with this idea

for how do you take that example to the

extreme? Um and so we have like 20

different simulators running at the same

time. This was oneshot vibecoded as

well. Um, and you can sort of mess with

the simulation speed, what happens when

you change gravity, um, what happens

when you change the rotation speed of

all these apps. Uh, and things get

things get a little crazy. Uh, you can

change the bounciness of the balls, all

this stuff. And I think it's just like

the gist of this one is you sort of take

these quintessential tests that people

were doing with AI models and I was like

how do you make this like the extreme

version um and this is very much an

extreme version of the ball bouncing

example and it just works out of the box

again literally singleshot vi vibe coded

um which is crazy like which I think

shows the power of this model and I

think is actually an underlying reminder

that you have to be ambitious with this

new model like what it's what it's

capable of is like probably beyond what

you're thinking. Um, so like ask for

more. Uh, ask push the model like ask it

to build more. Try to try to come up

with some of these ambitious ideas. Um,

and I think it brings a lot of them to

life which is awesome.

>> It's not just landing pages.

>> It's not just landing pages. Like please

please do more than just build landing

pages with Gemini 3 Pro. It really is

super powerful. And again, obviously

this is a gimmick example, but I think

it's meant to sort of like show you the

art of the possible. If you take sort of

what you were trying to do before and

bring those new ideas to life, but um if

we go back uh and potentially look at

some of these I'm feeling lucky ones.

Let's see if there's any good ones. Nano

Banana

or maybe Greg, what's the business idea

of the day for idea browser? Maybe we

pop that in and we see what we and see

what we get. This could be a good one.

>> Let's check it out. Let's check it out.

Open it up.

>> Ideabrowser.com.

>> Matching platform for AI generational

gap talent matching platform for AI

development teams.

This is actually interesting. Let me

what's the best Oh, and we have there's

prompts and everything. This is this is

perfect actually. What's the what's the

best way of bringing it over?

>> Are we just creating a landing page or

are we creating the app?

>> We should build a product. We should go.

We should go for the ambitious version

of this.

>> Of course, that's the type of guys we

are right?

>> I'm curious what happens if I just take

this top.

>> Yeah, the idea and and sort of like if I

just

>> You can try it.

>> Yeah, copy all of this. Let me see. Is

there I'm just trying to see is there

like other details about

maybe this stuff down here. I'll maybe

I'll just take in the raw literally just

command A and drop this in there and see

what happens.

This could be an interesting idea.

Build me an app based on what I saw from

ideabrowser.com today. I copy and pasted

the whole website below for the business

idea

idea.

Bring this

idea to life.

Make it fully functional.

go deep

and help me make this up.

So, what we're going to So, what we're

trying is we took a an idea, a free

idea, and we're just going to pl paste

it into

into Gemini, and we're going to see what

comes out.

>> Got the idea, copied it over, lots of

random context. uh didn't do any major

formatting updates which you ideally

would want to do but let's give it a

shot. Um and we'll see what happens. And

again for folks who haven't seen this is

the AI Studio build experience. You can

sort of put in whatever you want from a

prompt perspective. You can put in

multimodal inputs. You see on the lefth

hand side the prompt that I put in. The

model is thinking um and it's coming up

with what the planet should look like.

There's a bunch of like AI suggested um

sort of uh different additions to the

app that I can make in this context

about making a better matching

algorithm. Um there's also just

suggestions about the overall product

experience. So you can take your code,

you can deploy it, you can share it,

etc., etc. Um and then actually

afterwards we can directly look at the

code, manually edit it, uh preview on

different devices, all those kinds of

things. Um, and already so the model

came up with a plan only had to think

for 18 seconds which uh is definitely

longer than I thought for um and now

we're sort of going and generating the

actual website. And one of my favorite

things is um and we've been very

thoughtful about this from a vibe coding

perspective like if you've never written

software before one of my qualms with a

lot of products out there is that you

sort of see too much detail. Um, and

what I really like is like maybe you're

like, "Hey, I am sort of curious about

what's actually happening behind the

scenes, but I'm not a deep engineer and

I don't really know what's going on."

You can just hover over some of these

files and it'll actually tell you what's

happening. So like mock data.ts

uh is the mock data for users and market

stats. You can go to this Gemini service

config, which is Gemini service for

enabling compatibility between profiles.

Um, so you can get sort of like a quick

intuition behind what's actually

happening in this app. And then if you

have like, oh, I have some idea or I

need to tell someone else, I actually do

work with the developer and they're sort

of going to like take this idea I have

and then go and implement it. You can

sort of point them to some of these

files. Um, or just ask the model

directly to be like, hey, you know,

update X, Y, and Z file, or hey, update

X, Y, and Z setting. Um, and it should

be able to do all of that, which is

awesome. So, the moment of truth. Uh,

did we get a working app? Uh, let's go

full screen and we'll see. And just to

take us back to the idea really quick

was generational gap talent matching

platform for AI development teams. Um,

pairs junior developers with senior

industry veterans to create balanced AI

project teams. Junior developers bring

technical knowledge of emerging tools

but lack practical experience while

seasoned professionals contribute deep

expertise but need access to cutting

edge approaches. Um so it sounds like

it's a people matching platform uh based

on sort of like talent and background

and expertise. Um and there's actually a

business model around this and and sort

of memberships and subscriptions and

enterprise access which is super

interesting. So let's see how close we

got. Um,

I don't know what your reaction is,

Greg, but I feel like this looks like a

reasonable

looks like a fine landing page. Um,

nothing crazy.

>> Yeah, super. I mean, it's clean. I like

the interactions on like the AI talent

mark and is booming. Um, I would

personally like probably reprompt it to

I I'm just more of like a I like light

websites right now, like you know. Um,

so if you don't mind,

um, a light website like with with like

pops of color, which is,

you know, every designer hates that

feedback. Make it pop more.

But this isn't this isn't a designer.

This is uh, Gemini 3. So hopefully

>> Well, G Gemini 3's design capability is

really good. So thankfully, it won't

hate the feedback that we should make it

um that we should make it light. And I

actually agree with you. I I prefer

light websites. Let's see if there's

actually any functionality. Oh,

interesting. So, I clicked on I'll go

back really quick. I clicked on find

your match and it sort of takes me to um

a bunch of people and then let's see if

I click

analyze fit and then it's sort of

analyzing based on something. I don't

know what's happening behind the scenes

right now, but we'll see. I assume the

model the AI model is actually sort of

like analyzing. Oh, interesting. AI

compat. The pair exhibits expect

exceptionally high cap compatibility for

generational talent fusion. I love that.

Generational talent fusion. Um,

this is super cool.

Yeah. I I mean

I mean this is one prompt and it feels

like just this is my gut reaction. This

is the first time I'm seeing this. This

just feels like this is kind of like an

elevated version of any vibe coding

platform I've seen. like it just feels

like it's going a it's getting a little

more horsepower.

>> Yeah, 100%. Yeah. And so in the

enterprise team builder, there's like an

available talent pool and then I can put

these folks in and then I can let's

analyze the balance of the of the team

and the roster. Um and again behind the

scenes for folks who haven't played

around this experience before, we have

sort of all the Gemini API like all of

Google's AI model infrastructure hooked

up behind the scenes. So when you ask

for AI features, uh the model is like

really well primed and has all the right

context in order to go in and sort of

make all those uh make all those

features possible. And that's where

we're seeing like on all these random

pages like there's an AI team balancer

uh that's sort of analyzing the

different profiles of the current roster

of folks. And then if I sort of were to

remove two of these people and reanalyze

um you know it should give me a

different output based on Sarah and fake

Robert's uh profile specifically. Um

which is really cool. Um yeah and I

think the thing to to your point oh and

this is interesting dashboard your

personal command center. All right. So,

we got the new the new version. Um,

>> hell yeah.

>> This looks better.

>> This just looks like clean. You know,

this if I didn't know better, I wouldn't

know this is vibe coded.

>> Yeah, I love it. Well, it's purple.

That's how you know. That's how you know

that it's purple.

>> That's true. Actually, the purple.

>> Purple.

>> We should change the purple. I know

that. You know.

>> Let's do it. What do you uh I'll tell

I'll say I don't like the purple. Um,

what's your favorite color? Should we

>> green? Let's do green.

like purple vibe.

So my favorite color, Greg,

>> the the M dash is to AI writing, which

is what what purple is to vibe coding.

>> Exactly.

Green. And while we're having the model

do some some work for us, anything else?

Uh maybe we I'll just say and throw in

five new AI features as well.

You know, while the model's doing the

work, we might as well uh we might as

well get it to do some more stuff. So,

we'll switch the colors to green. We'll

add some more features. Um I like this.

I I think the design is also better. I

asked it to I forgot what was my prompt

before. Um I said spruce it up, add some

social proof. Um, and if we go back to I

think it added this like industry is

talking section which I like. It's like

a common sort of pattern uh from a

landing page perspective. Um, so I like

that the it the design also like looks

good. I feel like this is like a common

uh sort of approach of how I would see

that that sort of design show up which

is great. And how how expensive

you know are we talking like you know in

terms of you know tokens and and what

you know how how how does how much does

it cost?

>> Yeah that's a great question. Um so the

fun part is that uh the model is

available especially through this

experience of like you want to build an

AI app or build an app in general. It's

completely free. Um so you can go to

ai.studio/build studio/build access this

like start building your product or

experience. It doesn't cost anything. Um

which I do think is like part of this

headline story of making this this

capability accessible. There are limits.

Um and after you hit the limits we uh we

sort of prompt you to switch over to a

to a paid API key and then you can sort

of keep going as far as you want and use

it as much as you want. Um the actual

model is $2 per million input tokens and

$12 per million output tokens. Um, so

it's priced below some of the sort of

comparable competitive models like uh um

GBT 5.1 Pro as an example and Claude 4.5

uh is is like lower priced than both of

those models and that's for uh input

tokens less than 200,000 and then over

200,000 the the price I think it's like

doubles for both. So, it's $4 per

million in um and I think 24 per million

out though. I'd have to fact check that.

Um so, similar sim similar pricing

pattern to what we've done in the past

for all these pro models. Uh which is

great.

>> It's adding a bunch of features which is

exciting. I feel like I'm going to go

build this uh this company after this

and help

>> interview simulator generates a role

specific interview question to help you

practice. Skill gap analysis comparing

your skills to market trends. smart

icebreaker. Like that's where there's

opportunity in using like this product

is, you know, we're going from web apps

to smart web apps, right?

>> Yeah. And actually what I love too,

Greg, is like I sometimes have this like

writer, maybe not writer block, but like

vibe coding block where I'm like I don't

know the ideas like I don't know what

the best ideas are and I routinely use

this this sort of prompt which I have

here which is uh just like add in five

additional features and the model sort

of takes in all this context and all

these ideas that I wouldn't have thought

of and like even if they're crappy ideas

I can always just remove those features

but like the one feature that shows up

here that's like actually really Right.

Um it it's just like it feels like that

magic moment. So I feel like that loop

of just like add five additional

features, let the model figure out what

they are, and then play around with them

works really really well. Um but we have

your green website now. What do you

think?

>> I mean, dude, this looks this looks

fresh.

Like I'm not even saying this. This is

like a fresh, you know, it looks great.

>> So now we can connect with people. And

then I think it's sending it's like

drafting some sort of like draft email I

guess for reaching out to Sarah perhaps.

Uh, hey Sarah, Alex Chen here. Your

assistant design work is impressive. I'd

love to connect and learn from your

perspective. Thanks. Send it. I don't

know what it's going to do when I press

send. And obviously this button does not

look good, but um

hopefully

doesn't work. Uh this is a great

actually example of um this annotate

feature which I don't know if folks have

played around with but I love this. Uh

and I can just be like box in this

little section and

>> I can add a text box and say

this send button doesn't work

and seems oddly placed.

>> Oh wow. I didn't know that that existed.

>> Yeah. And you can just add it to the

chat.

it goes down here and then already pres

applied the edits shown in the

screenshot. Um, which is great. So now

we'll fix that which is wonderful. And

then we have our enterprise team roster

team configuration. This is great. Um, I

feel like we should I feel like we

should try some other ideas or play

around with some other stuff. But I feel

like we have the sort of like bones of

of bringing this idea to life. And then

I'll sort of just hit two more points

really quick, which is um, you know,

let's say this idea, you know, let's say

this was the fully working version. and

you're like, I'm ready to go. I'm ready

to start my sort of talent matching

platform. Um, you can go and deploy the

app and and actually start sharing with

people externally. Um, you can also

share the app with a friend um or like a

small group of people and um so you and

me could keep iterating on this or if

you want to sort of just get the

feedback from like two or three people

that you know and be like, hey, does

this idea make sense? Does this UI make

sense? Um you can actually share that

and continue to iterate together, which

is really cool. or move the code to

GitHub, download it, do all those types

of things that you want to do. Um, so

it's uh hopefully folks will be able to

actually build some of these ideas and

and share them with folks.

>> That's really cool. We had an offsite

recently and we had like a vibe coder in

the offsite and they were just like as

we're spewing ideas, they're just like

building in the meeting and I feel like

that that's going to be the future,

right?

>> That is the future. I was literally I

was on a customer I was on a call with a

clothing brand company and they were

they were telling me about how they're

like we want to you know we've always

wanted virtual tryon and we've always

wanted like this like AI enabled mood

boarding so that we can sort of try out

different things and I'm just like I'm

like yeah okay keep telling me I'm just

like spitting and I'm like okay and

here's the app for that idea and here's

the app and this is like live on the

actual customer call. So, I think

there's for folks who are like building

businesses, especially if you're like

trying to get your initial customers or

like trying to win certain customers,

like pe people have not seen that

flywheel. If you go to sort of the long

tale of like random businesses in

America and like meet the business

owner, they have not played with these

tools. Um, and if you can sort of build

something in real time, like I think

that blows the mind of so many people

and also being able to like customize it

based on their experience. Um, and al

maybe we can actually do an example of

that because I think this is really

powerful, which is um, and this is maybe

too too meta, Greg, so you can tell me

we shouldn't do this instead, but I want

to go to ideabous.com and I want to

screenshot this website and I want to

rebuild ideab browser in AI Studio um,

and have AI actually generate all this

stuff for us. Um, just to showcase like

how good it is at sort of getting close

to this design idea. Um, but I'll wait

for your consent because I don't want to

I don't want to I don't want to steal

the idea of browser thunder from you.

>> No, absolutely.

>> All right. So, I'll I'll zoom zoom out a

little bit. Uh, I'm sharing this tab.

And I'm just going to zoom out so that I

could try to capture as much of this

content

as possible.

And I'll scroll down,

capture the screenshot.

>> I don't know if you'll be able to do it,

but you know. Yeah, we'll give we'll

give it a shot. We'll give it a shot. Um

I we can't I'm sure your team spent a

lot of I think there's there I'll I'll

make the comment which is a lot of the

magic is in the nuance of this stuff.

Like I think your team and others like

have great perspectives. So I don't mean

to I don't mean to say that it's easy to

copy any of these ideas. Um, but I

literally I'm going to say clone this

UI. Exactly.

make uh make me a button to generate

tomorrow's idea based on the daily

business trends

through Google search

through grounding with Google search

make it all work

be interactive

and PowerPoint.

All right, we'll give this shot close

UI.

Exactly.

Um and and like the practicality of this

and like other than I don't I'm not uh

sort of endorsing going to uh idea

browser and copying the website. Um I

think what what like in practice

actually like the real use case for this

that our team benefits a lot from is we

do this with AI studio ourselves. So

like we our team is building a product,

we want to sort of iterate continuously

on what that new product can look like.

We literally take the AI studio UI, we

screenshot it, we say, "Hey, what would

it look like if you sort of changed this

configuration or added this feature here

and there?" And that flywheel really

works well. And um actually inside of

Google, we've seen like a huge amount of

traction of teams going and using AI

Studio to like rapidly iterate on their

product experience. Um, and I think it's

like even if it's not building the like

full production version. So like if

you're if you're watching this and

you're sort of like past the stage of

just like complete vibe coding where

you're like building something net new

from scratch, even if you have an

existing company, taking the the

product, screenshotting it, iterating on

the UI, um, works really well. And I

think it goes back to my other features,

my suggestion of um which I do all the

time is I screenshot the A2D UI, I'm

like clone it and add five new features

and the model just like comes up within

like some of the things are really dumb

and don't make any sense. Um but like

every once in a while, like every every

three times I get something where I'm

like, crap, we need to build that.

That's such a good idea. Um and I didn't

have that before. So I feel like we

should hopefully do this for idea

browser, too. And maybe we'll come up

with the next feature for for the

website, which would be really cool.

>> Totally. And I will say, you know, you

can also screenshot an app and then just

find a new niche or find a new

geography, you know what I mean? Like

>> I was I was reading about Wealthfront

today, which is a Robin Hood of Canada.

And like, you know, they're doing, I

don't know, 3 400 million ARR or

something like that. And it's like, it's

just Robin Hood for Canada. It's just

the same concept, but focused on a

different geo. So there's opportunities

to build businesses that are just new

geos or just vertical, you know, highly

verticalized.

>> Yeah, I love that. Um, and if folks are

watching on the lefth hand side, I

generated the app. It ran for 95

seconds. No preview uh was displayed. So

I told the model, I don't see anything.

The screen is white and blank. Uh, so

hopefully we'll This is the This is the

experience of vibe coding, honestly.

Like sometimes things don't work. Um,

and uh, you have to take another shot at

it. But, uh, we'll see if it we'll see

if it gets it on the second shot. It's

usually just if there's nuance context

that folks use AI Studio a bunch. And if

you run into this, there's like all of

this like special formatting we have to

do in AI Studio to like make the app

compatible with this rendering

environment that you see on the right

hand side which is usually what trips

the model up because it's used to just

like building software fully

unconstrained and then it comes into AI

studio. We have like a bunch of very

specific instructions of like making it

work with this uh rendering environment.

So, uh, it is a common a common failure

mode. So, if this happens to you,

literally just tell the model you don't

see anything. Um, and it should it

should fix it automatically. And we're

doing a bunch of stuff to hopefully

mitigate that in the future as well.

>> Mhm. Don't give up, basically.

>> Don't give up. Yeah. Yeah. I mean,

you're like one prompt away from it

working all the time. Uh it's it's like

that meme of like the the person in the

cave with the pickaxe and like the

diamonds are right behind and it's like

you're that little gap is usually like

literally one the one prompt um is

potentially that gap which is really

cool. All right, let's see. It realized

it was missing the index.html file. It

was incomplete. Uh which makes sense. So

hopefully it'll fix it very quickly. Um

and I think for folks again who have

used 2.5 pro, I think 3.0 O Pro is like

roughly on par from a latency

standpoint. So there's no um I think in

some cases it's actually it's a little

bit higher latency, but obviously the

model is much smarter. Um and yeah, so

it shouldn't be like a a worldly

different experience from like how long

you wait to actually do some of these

experiences. Uh which is helpful. But I

am excited to see this idea browser

version.

>> Is this the best vibe coding product on

the market?

>> I think it is hopefully. Um I I think

there's I'll make the comment. Let's

look at this and then I'll answer your

question. I think there is a we we have

a bunch of gaps. There's a bunch of

things that I'm really excited about.

Like it's free. It's powered by 3.0 Pro.

It has all these AI features built in.

Like there's a bunch of things that are

unique that I think we're like best in

class at. And then there's some very

specific things that I think we're still

behind on. Um but this is it. This is my

idea browser clone. Uh it looks pretty

good compared to

>> to what you all had before. And let me

just pan back to what it was on the

other uh the original idea browser one.

And I'll So this is what it looked like

before. And I don't know if it's

possible to show two tabs at once or we

can do it in post or something. The the

magic of post production. Um and this is

what it looks like the version that I

just vibe coded. Um

>> and the thing that I'm most excited

about is like generating future ideas.

So I'm super curious actually if I press

this if this feature works out of the

box. Um, but let's see. So, it removed

some of the stuff.

>> It's loading. Something's happening.

>> And we're using I asked it to use Google

search. So, I don't know. This this

could potentially be this is somewhat

unconstrained. Um, so it's supposedly

like pulling in the news of the day and

coming up with um uh coming up with

something potentially interesting. So,

what is it? Rapid AI. Uh Ding oversight.

Businesses are grappling with how to

scale because of the EU AI act, etc.,

etc. Yep. Ethos AI is a central control

plane where businesses define

parameters, ethical guidelines, and

operational boundaries for their

deployed AI agents. Interesting. Um, and

I wonder I wonder where it's sort of

grounding from or actually it shows sort

of where we're grounding from in this

context. Cool. Um,

>> and I think some and like AI governance

is the is the sort of key word that it's

that it's tracking which is really

interesting. Um,

>> yeah, governance readiness checklist,

ethos starter kit. This is cool. Um,

nice.

>> I I would never I don't think I've ever

seen anything this good. Like

screenshot, put it in, and it looks this

good and it works

on any other product.

>> I love it. This is This is I mean, we

did a little bit of there's a little bit

of AI magic on the AI Studio side, but

like the really cool story is like this

really is the Gemini 3 Pro model out of

the box. So like if you want to use AI

Studio to do this stuff, awesome. If you

want to sort of build your own version

or use some other tool that's powered by

this model, awesome. Um, the model is uh

it's available across the entire

ecosystem. It's available in our API.

It's available in our first party

product. So like you can use this uh

hopefully like wherever it is that you

want to you want to build your ideas.

Um,

>> right.

>> And yeah, it it just feels like magic

when when it works. And this again, this

is my favorite example of sort of like

continuing to iterate. And I don't know,

Greg, if we do maybe one more generation

of this of like if you ever had there

was some feature that you wish uh idea

browser had that hadn't that you all

haven't built yet and maybe we could try

adding that in.

>> I do. So um you know every day thousands

of people go to ideabrowser.com for

ideas and I think that you know people

are looking for co-founders.

M

>> so let's say I was into this e ethos AI

uh you know building this I wish I can

press a button that says like match me

with a co-founder to build this.

>> That is a great idea. I love that and I

feel like that is very true. Like it's

um

sometimes sometimes you just need

another person. I feel this way all the

time. I'm like, I've got lots of great

ideas, but if I have another person

who's in the trenches with me pushing on

that idea, like I feel like I can

actually make it happen versus sometimes

I feel uh when I'm pushing on these

ideas by myself, it's it's not the same.

Um, so I love that idea. And let's

actually use I don't want I'm done

typing, so let me use text to speech and

we'll see if we can uh see if we can

make this idea come to life. All right.

I love where I browser is at right now,

but um I was just talking to my friend

Greg and he suggested we should add the

ability to search for co-founders as

part of this process as well. So, can

you sort of mock out and sketch up the

experience? Uh make it as functional as

humanly possible where like different

people can come in and match um you know

with your potential co-founder to bring

this idea to life. Uh it should be like

front and center in this experience. As

soon as you have the idea uh being

presented to you, you should also have

the option to like go and find a

co-founder. When you click that, it

should sort of take you into the

experience where you can see different

people's backgrounds, their profiles,

etc., etc. Um, and then maybe while

you're doing that, uh, also try to add

in three additional ideas to this

website. Um, and we'll we'll sort of see

where that takes us. This and so we have

this text to speech feature which also

sort of cleans up a little bit uh the

idea. So, I'm just going to assume this

is right and not even read it. Uh,

because we're vibe coding. Um, and it

sort of removes a bunch of the ums and

a's and other random stuff that's not

relevant for the idea. Um, so we'll see

where this we'll see where this takes

us. And let's actually just quickly go

back and check in on what happened with

our co-founder

feature.

>> Near build the idea. Find a co-founder.

>> Yeah, let's do it. Find a co-founder.

Wow. I mean that's pretty that's pretty

dope.

>> Let's connect send some requests to

them.

I can upgrade premium get curated.

>> Wow.

>> That's really smart.

>> All these different filtering. Nice.

Let's do Sarah.

>> And it's on brand, right? Like this

looks like video browser brand.

>> The subtle the subtle details. I love

it. Um, and I actually I wonder what

else it added. Uh,

>> let me know. Uh, comment in the you in

YouTube if I should if we should add

finding co-founders on ideabrowser.com.

>> You should. I'm going to go comment as

soon as the video go live goes live. Um,

so this was the different features. Add

a co-founder button. Implement a list of

three additional predefined ideas at the

bottom of the page to allow quick

navigation. Let's see what did it add

before. Um, oh cool. So this is actually

coming up with this actually I think is

similar to what you already have in idea

browser which is like yesterday's ideas

and I think uh a bunch of other stuff

like that. So this is coming up with

like different opportunities and sort of

going back um so I think it it took it

took my suggestion to add new ideas or

features in a in a different direction

than I was anticipating but I feel like

it it works actually.

>> Yep. For sure.

>> This is great. Uh, so I'll send this to

you and you sort of um can go and

replace the version of idea browser with

the version that we vibe coded. Uh, we

should be off to the races.

>> Awesome. Anything else you wanted to

show?

>> I think that sort of is most of the

stuff again. Um, we're early uh in the

experience of like letting you build

really really cool powerful apps. Um, so

send us feedback uh send us feedback on

the model. Again, the model is available

across the ecosystem on on Google

products, on thirdparty products. It's

available in the APIs. Um, so you can

sort of build with it wherever you want

to. Um, and if you want to vibe code

apps for free in AI Studio,

a.studio/build,

uh, and you can sort of bring your ideas

to life, which is exciting.

>> Totally. If you need ideas, just go to

ideabrowser.com,

copy, put it in here,

>> screenshot the app, build your own

version.

>> Totally. Go crazy. have fun. Um, no,

this has been really uh this has been

really fun. Thank you for coming on the

show and and sharing it with the

audience and um I'll include links to to

get started with uh Gemini um 3.0 uh in

the show notes and uh let me know what

you built everyone. Logan, thanks again

for coming on and I'll see you next

time.

>> Yeah, this is a ton of fun. Thank you,

Greg. And uh yeah, hopefully we can

build some more stuff together.

>> I I hope so, too. See you.