This Week in Startups: AI demos: ChatGPT’s UI punch ups, Google’s generative image search, and more with Sunny Madra | E1829

This is DEF CON 1, Sundar, and anybody comes in and they're treating it like DEF CON 5 or 4 or 3,

take it serious, folks. And if you can't take it serious, Sundar, you have my permission,

out the door. Out the door. Nobody ever gets fired at Google. Now it's time to start firing people.

You come two weeks in a row without progress? Done.

This week in Startups is brought to you by, with Mercury Raise, startup founders no longer have

to navigate row blocks alone. Visit mercury.com-slash-raise to get access to a network, connections,

and advice. Squarespace. Turn your idea into a new website. Go to squarespace.com-slash-twist

for a free trial. When you're ready to launch, use offer code TWIST to save 10% off your first

purchase of a website or domain. And FitBod. Tired of doing the same workouts at the gym?

FitBod will build you personalized workouts that help you progress with every set. Get 25%

off your subscription or try out the app for free when you sign up now at fitbod.me-slash-twist.

All right, everybody. Welcome back to This Week in Startups. It's Monday. It's Monday,

which means it's Madra Monday. That's right. Sunny Madra is here for our This Week in AI demo.

Sunny, welcome back to the program. Good. Good to be back. Excited. Things keep moving quickly.

We've got a lot of stuff to do. Yeah, let's go to a demo. Let's just go right to a demo.

I know we've got a couple of new stories. I know Stack Overflow, there's a story,

there's a bunch of different stories, but I love when you demo stuff. Let's just get to it.

All right, awesome. Let's get into the first one. So,

you know, this is an area that we've all spent a bunch of time in, and so this product is called

Trove. Trove, G-R-O-V-E. Trove. Yeah, yeah, and their URL is GetTrove. So, G-E-T-T-R-O-V-E.co.

And what this allows you to do is create a survey, which we all do. We've used some products,

you know, some of our close friends. Survey monkey, type form, and then, of course,

Qualtrics. Those are all, you know, in that genre.

Exactly. And what this does is, you know, this kind of takes that to, like, a more modern workflow,

which I thought is pretty cool. And so, I've kind of pre-constructed one here. You can do one

alongside me if you want, Jake Hall as well. So, this one, I said, I wanted to create a form. The

main objective is to collect thoughts on where the next all-in summit should be.

Okay. So, it asks you for your main objective of the survey, and you put in a form to collect

thoughts on where the next all-in summit should be located. The title is All-in Summit Locations.

And it asks you for a tone of voice. Very nice. And you picked Whitty. What were the other choices?

Yeah, if I drop this down, it'll be professional, humorous, concise, sincere, sassy.

I almost went sassy, but I'll go Whitty.

Yeah, you know, it's trying to give you credit there. And then, you know,

sort of the welcome message when you jump in. So, it's going to create a chat-based survey,

which is really unique. And we'll see. Yeah. And so, what I did is, hey, you know, welcome,

we need your help by narrowing down where we should have the next all-in summit.

If there is one, yeah. Yeah. And we never know with this crew.

We never know, exactly. And then, I gave it just a few questions. What you'll see here is,

what is your recommended location? What do you prefer it? Is there a specific venue?

Do you have any ideas for after parties? How much would you like to pay for the event

and to keep the event accessible but yet exclusive? And so, you can see here, and I'll

just kind of go further into this in a second. But you can see here, on the right-hand side,

for those listening, I've got a chat going. So, it drops in with sort of that initial starting

prompt that we gave in. So, what's your recommended location? I said, I'm thinking NYC.

It says, ah, the Big Apple. Why do you think NYC is a perfect spot? I think NYC is incredible

start-up energy these days. NYC is definitely buzzing with start-up energy. Now,

if we were the whole all-in summit there, what would be a specific venue in mind?

Well, now wait, did you tell it what venue? Was that one of your sub-questions?

No, that was one of my sub-questions. One of your sub-questions. So,

instead of doing this as like, here's a four-question survey, it's just, hey,

I'm going to dump you to chat room, ask you the first question, then I'll follow up with those.

Yeah. That's really interesting. Yeah. I like that modality.

Me too.

Now, what if you ask something that's like off topic? So, if you said, who's going to speak at

the event? Yeah. You know, or how much are tickets? Yeah. Yeah. Okay. The event, okay.

I hope this thing is smart enough to say, who do you want to see speak?

And so, what it does is... It gets you back on track. Yeah. Yeah. Right?

Ah, okay. Before we dive into the after parties, let's focus on the speakers first. Who do you

think would be the perfect lineup of speakers in New York City? And after that, we can talk

about the epic after parties. Okay. So, if you go to the left in the interface here,

you had put the questions in what order? Let me see. What is your recommended location?

Why do you prefer this location as question two? Question three is, is there specific venue?

Great. Question four, any ideas for after parties? Question five,

how much would you like to pay to keep the event accessible? But yet, it'd be exclusive, right?

The big question everybody has about all-in-summit is, how has it become VIP? But how do we make it

bigger? Yeah. Yeah. Yeah. Folks, those are two different things. I love people who are upset at

the summit that other people got to come on a scholarship and that they had to pay a full

price ticket. It's like, okay, do you want to have like diversity at the event? Do you want

to have interesting people at the event? Like, and what do you care if you're rich? Like, just

relax. Anyway, I'm with you. And people are just... Sonny, you got Jason all hot and bothered with

this all-in-summit first example. You're trying to tweak me here. No, you know, for people who

are wondering, you know, surveys are great. I like to do things and I don't care what,

as an artist, and I do some things that are artistic. I don't want your feedback,

to be totally honest. I want to create an experience and have you enjoy it. Like, you know,

like if I go to a Scorsese movie, when I go see Killers of the Flower Moon, like I can't wait to

see it. Yeah. I don't want to tell Scorsese what I want in the movie. I want him to delight me and

surprise me. I have nothing to add to his art. So if you come to something like All-in, pay for

your ticket if you can afford it. Apply for a scholarship if you can't. If you get lucky enough

to get in on our scholarship and pay like a thousand bucks or something, a fraction of the

price of the VIP people, everybody just enjoy the show. And if you don't like it, you don't want to

take the risk, don't come. The end. That's it. We don't need your feedback, honestly. Like,

and if you don't like the after parties or whatever, you don't have to go. Go have a party

of your own. Go have dinner on your own. If you don't like the Barbie theme or the Blade Runner

theme or the 007 theme, that's okay. You don't have to like every party. You're like everything

every meal you've ordered. Do you like every restaurant you've ever been to? Of course not.

No. This is a nerve. I didn't think it would. I was just trying to...

No, I'm just, you know, it's some people over index on feedback. And I just want to like,

literally, if you have 1800 people come to something, there are one or two percent of

people who will complain about everything. Just like there are people, you ever go to a hotel,

Sonny, and you're like checking in or you're checking out or whatever. You're getting a

bottle of water from the front desk and there's somebody next to you. Oh, yeah. Oh, yeah. You

know the person next to you? Oh, yeah. They are just hot and bothered. Yeah.

Noise complaint. My room didn't get turned over. My room service came too late. It's like...

I had that exact thing happen to me last week. I was traveling and then I was down in LA and

basically, yeah, someone was checking in. I was checking in. Sorry. And the person beside me

was staying there and was just had the list of complaints. The list. They had their list. Yes.

They had their list. Yeah. I mean, there's an appropriate way to do this. Here's what I do.

I come with an outcome in mind and my outcome is never to berate anybody. My outcome is like,

I want an upgrade or I want a comp night or I want a fruit plate, whatever.

And I'll just say to them, hey, a lot of times I just want them to do better. So I'll say, hey,

listen, I don't know if you know this, but there is a huge construction site outside the window and

it starts at 7 a.m. and I got in late. Maybe if there's a construction site, you can let the

person know. I just want to let you know because it did detract from my experience. Do you want to

make sure if you have somebody coming in late. I think they want to know that stuff. They want to

know that. Yeah. And I had this happen one time. I'll tell you, this is like, I ordered a cappuccino

at the addition when I was in Miami and they charged me, and I am not kidding here,

$11 for a cappuccino. And I got the bill. It was like $16. It was a wrong bill.

And it was like, they put the service charge in, they did everything. It was like $16. I literally

had one cappuccino. And I said to the guy, I said, do you think if you charge $11 for a cappuccino,

you should tell the person? Because that's like triple a Starbucks cappuccino. I can afford it

or whatever, but it's just really weird. He's like, I tell this all the time, would you tell my manager

because it's really important for them to hear this? I said, sure, I'll tell your manager.

Manager comes over. I said, listen, I'm in an $800 a night room. I understand I'm at the addition.

I just feel abused by an $11 cappuccino. That makes no sense to any human being.

If I was at MetLife Stadium or something, I would be offended, but at least they tell you

you're paying $11 for a cappuccino or something. Yeah, you see it at the vendor stall, right?

You see it right up there. Yeah. And I said, if you're charging a great rate for the room,

you're doing great. The hotel's packed. You don't actually need to do this.

And it really is a turn off. And so I just want to let you know, because I'm in the service

business too, I'm going to consider myself in the service business. She said, I really appreciate

you saying that we've had this discussion in the staff meeting and I'm going to bring it up one

more time. And she goes, oh, by the way, it's comp. I said, no, I wasn't angling for comp,

you don't need to comp it. She said, I insist. I said, okay, and I just took a $20 bill and I

gave the guy a $20 tip. I said, I just approved you. This is not about the money. I'm giving the

waiter a $20 tip so you can charge me the $16 I give a $4 tip or you can just take the whole

$20. I don't care. I get up to my room. I kid you not, I get back to my room that night,

there's a giant fruit plate, not like a little mini cheap fruit plate, not like a skydating,

you know, filler fruit plate, like a literally jit in it with blackberries and blueberry.

It's not about a 1.0. It's like skydating 2.0 fruit plate. I get the really good fruit plate

and a bottle of champagne. I kid you not, a vocally called yellow label. That's like $1.25

and the fruit plate goes like that. And she just wrote me a really nice note. I really appreciate

like you giving us the note, whatever. Yeah. Mr. Callican. That's a different kind of feedback

than, you know, what we're talking about here. There's an approach to giving feedback with love

and kindness and with intent. Like I think the intent really matters. So I always wonder when

people, but anyway, I like this kind of survey format. I think this is easily replicable. I think

everybody will have this as an option, but I love the idea that the startup is thinking this way.

What else can it do? Is there anything else here because? Yeah. So yeah, let me get my

sharing back on this. This seems very 1.0 to me. I'll be honest. Like,

you know, somebody can whip this. I could have one of my startups in Founding University whip

this up in 72 hours. So where's this all going? There's got to be some destination here, right?

Well, yeah. Like so, I think that's a good point. Let's maybe double click on that for a second.

I think one of the things that we're seeing is an incredible ability to create, you know,

what would have been products before. Obviously, we've seen an entire business created on this

type of thing as just a feature. And this is, it's a good kind of segue into maybe our third

demo, which I'll jump to right away is that the ability to kind of create stuff quickly. I think

what this shows is a new user interface. I think we're entering the era of chat. It kind of aligns

up with what you're saying. You know, we didn't, we didn't finish this, but like, I think this

feedback form, you know, when aggregated and put together will be more valuable for the end,

the person looking for the feedback as well. Because they're kind of seeing a thought process

rather than people checking off boxes. Yeah. So why don't we do this? I came to the last event.

Let's just type this in for them. I came to the last event and I would rather have a sit-down

dinner than food stations. You know, I get that advice a lot. I like to do food stations that

are open the whole three hours and you can get food whenever you want and you can get, you know,

you can eat whatever order you want. I don't like sitting down and then you're late and then

you're three course deal with bad chicken. Got it. So for the next long summit,

you would prefer a sit-down dinner to keep things classy or are you open to exploring

other food options? And speaking of a specific, how much would you like to pay to keep the event

accessible? So it's keeping you on track. So it's capturing this for you as well,

which I think is great, which you wouldn't get if you just put four questions there that

all you were looking for. Yeah, I would just say I would pay five to 10k.

Let's see. So now I guess the results and how it organizes these roads, because we're giving

chat-based results. You know, one of the nice things about doing a form is that you can require,

you can force a number. So I would have asked what do you believe, what dollar amount is the

amount you want to pay for this and what's the most you would pay? I would ask you to get in two

forms there. But anyway, that's pretty good. And then I would build a chart, right? If I actually

cared about people's feedback. And you can give answer options and do all that. I see that. Yeah.

So I think this is really incredible. Obviously, people can replicate it quickly. But that said,

this team has come up with it. You should give it a try. And we'll go from there. I think it's...

And for startups, where you start is not where you finish. So what will happen here is maybe

they'll figure something out. I didn't see it suggesting new questions for us. But what it

should do is after the first... As people are answering this, it should suggest to you people

keep bringing up the dinners. And people keep bringing up sit-down dinners. Do you want to

ask a question about what you prefer? So if it's actually was suggesting questions,

that's where I think this could get very powerful. And sometimes machine learning

does that already, or they just add like, here are five common questions, like a wizard,

a wizard. But this would be really interesting if the... This is what I refer to as AI all the way

down because the... AI all the way down. Yeah, which is like the output of this goes into another

system, which then has AI looking at that, then saying, hey, we should modify the questions,

right? So your whole stack from top to bottom in this new world order should be AI kind of all the

way down. AI all the way down is a really good idea here. I agree. I will give Trove... The survey.

The survey. Give it a B. It's a solid B. Good start. Yeah. I mean, there's nothing to hate on it. It

works. Yeah, it works well. It's a single utility. I mean, it might be a C plus, actually, a B minus.

I'll give it a B minus. There's like a lot it can do, but it's a good start. It's clean interface.

I like it. Good start. It's not transformative yet. I don't think it's made the... I think

the best is yet to come for that app. And the best is yet to come is going to be

analyzing the results and telling me what I should be doing in the survey, as opposed to just

doing the survey in a chat format. Would you fund a company taking on Qualtrics or SurveyMonkey?

Of course, I have actually. We have one. We have one in the current law, yeah, in LA 29.

I haven't been on the show. Really? Yeah. So of course, I would take on anybody

who's got a big, large business is worth it because there's two or three possibilities,

you know, one you fail and then we know exactly what we lose. So just to give people an idea,

if it's an accelerator company, we know we've lost $100,000. But then, you know, and that's

kind of like your blinds. That's your big blind. That's your, see the flop bet. You see the flop

and you flop something nutty and something really juicy. You know, maybe we'll put another 250 in

and another 500K. And so we're in for 350 to 600K, but we typically own at that point,

call it 7, 8, 9, 10% of the company. So let's put it at 8% of the company.

Is that your target? My target is to get to 10% in the likely winners and 15% in the winners.

So in our fund structure, half the dollars, half of the 100 million

goes towards the top 30 to 40 companies out of the 400 in the portfolio. So if our portfolio

construction is 400 accelerator companies, founder university companies, early stage direct

invests, we tend to get to call it 8, 9, 10% in that early stage. And then we wait and see.

If something does really pop off, then we have another 50 million in the fund. If we raise

this $100 million fund, we're at 50 now, which I assume we'll glide into that. Then we could put

a million or 2 million into those winners. So when they're at a $20 million valuation, we can

buy another 5 or 10% of the company. If we buy 5% of the company, we get to 15% in the next

superhuman, calm, or Uber or Robinhood, which would be... Robinhood.

Yeah. Yum, yum. Skis. All right.

So that's really them. That's the betting strategy.

And we're going to see some in this era. We're going to see some in this era.

I mean, I'm more excited now than I've ever been. So for a company like Grove,

would I have them in our accelerator? Heck yeah.

We'll make that $100k bet. That's an easy one. And if they had $10k a month in revenue and it was

worth $5 to $10 million, I might make a $250k bet. Sure. If I like founders, if they're

builds or founders, two or three founders who know how to build product and they got good product

velocity, all that would be contribute to making an investment in that company.

Yes. And then...

And you know, the guy from Qualtrics owns the Jazz.

Yeah, he was just on the podcast. He was here for all sorts of stuff.

Yeah, Ryan. Oh, wow. That's amazing.

Ryan's a friend. Great.

I'm B-minus.

I'm B-plus. I think it, you know, and I haven't gone through the whole product.

Look, I'm a builder. That stuff's hard, you know. So we're coming at it from different

places. But like, look, I think it's a great first implementation. I've actually got to go

through the whole product, see what the reporting side of it looks like. And then that, but I do

think there's room to do that loop around helping me come up with the questions. But

I'm going to implement it. Actually, we're going to use it for a survey we're about to send out

to some customers. So yeah, do it. Let me know. Follow up with me.

All right.

Yeah. And shout out to the trustee. Congrats.

Being a founder is one of the most amazing journeys you could ever go on. I suggest you do it,

but you got to know it's going to be hard. And sometimes it's going to be a little bit lonely.

But with Mercury Raise, you don't have to go it alone. This is an amazing new program for Mercury.

I'm part of it. In fact, Mercury Raise is a founder success platform that's built

to remove the roadblocks at every step of the founder journey. And this is going to help you

with the number one thing that founders tell me they need, money. Yes, Mercury Raise gives you

access to investors and then they give you access to industry experts. And finally,

they connect you with fellow founders. What a great idea. It's like what I do in my

accelerator, except it's open to everybody, not just people we invest in. So here's how you use

it. And a lot of my founders have been using it and I'm super happy about this. If you're fundraising,

you submit your pitch and you get in front of hundreds of investors who are looking actively

to fund businesses like yours. And I'm in that database and I have found good companies in

there. If you're looking for guidance, you can tune in to unfiltered conversations with industry

experts. And if you're craving some community, which I know you are, you can meet fellow founders

and navigate similar challenges. So here's your call to action. Mercury Raise was created to help

founders navigate obstacles so more startups can become success stories. It's that simple,

they care. And you can take your startup to the next stage with Mercury Raise by visiting

mercury.com slash raise. That's mercury.com slash raise. So to the folks who built it,

keep grinding. Yeah, awesome. Great work. Okay, I'm going to jump forward a bit just in terms of

some of the ideas that we were talking about here, like how fast this is to create. So

earlier this week, we saw this, folks got wide access to chat, GPT with vision. We did one of

the first demos here. And someone did something pretty cool. They took just a screenshot of the

calculator app, provided a little prompt that said, Hey, your software development, make this for me.

And so I tried to basically recreate that because I was intrigued by it. So here in my window, I

took a screenshot of the calculator app, provided a similar prompt. And basically it created,

asked it to create the HTML. Jacob, you remember this from your early days, would have been a

pain to do this. It said, Hey, give me the JavaScript for the functionality. So it's not

only the UI for it. And it's also functionality. Then I said, Hey, give me a file that I can

download. And ultimately, and I'll just show some interesting uniques and tweaks. So it didn't do

it complete. And this is sort of my mistake. So I want to give a warning to folks that think it

can just be like a complete start from here and finish. Because if I go back and I'm just flipping

back and forth between some tabs here for the folks listening, the code it's given me is not

fully complete. And you can kind of look in here and it says you can continue creating rows of buttons

like the one above. So it gave me a zero, all clear, a plus percentage and a divide. But it didn't

give me the rest. And it obviously didn't give me the JavaScript. I wonder why. I wonder if it's

been asked to in the AI, when you ask it a very complex question, in order to not waste GPUs and

cycles, it says give the first 10% of the answer and let people ask for more. Like if I asked it

to write an entire novel that was 50,000 words, I wonder if it would actually do that. Or it would

say here's an overview, here's the first chapter and then make me kind of tease it out so it doesn't

you know, bog down the system with too much work. I had the exact same thought because after that

happened, I basically went back and said, can you provide me the complete HTML and JavaScript so

the calculator is really functioning? Yeah. And you know, I've been trying this all morning.

And it keeps and I think what's happening is something similar to what you were saying. I

think they're so bombarded. This is really heavy workload on them. You know, generating this much

code is a lot of tokens on the way out. I'm only paying them $20 a month. And so I think I'm getting

totally throttled from because, you know, I've tried this for even a few hours this morning,

and I couldn't get it to give me the complete, I keep getting this network there.

So you're being throttled for sure, which makes sense. I understand them doing that. They have to

preserve the user experience. It would be like if I said, on YouTube, hey, download the entire

channel, like there are these more Knopfler bootleg concert channels. And I found an app called

Downey, and I paid for it and then it didn't send me the code. But anyway, what it lets you do is

put in a playlist from YouTube or a channel and then you can just like pick what you want to

download. So there's like 20 concerts I really like. I like to download them. I can't find them.

I would pay for them too. I can't find them. Why don't you just download them into the YouTube app?

Because what happens is they get taken down these channels. And then I got to search for it. And

then when I'm skiing, you know, I do, when we're skiing, you and I ski sometimes together. I have a

nice Mark Knopfler concert on or a Diastrates concert or a Pink Floyd concert on. And then like

the concert I really loved is gone. And so I just wanted to download it, right? So, you know,

YouTube doesn't let you bulk download kind of situation. But there was a report earlier this

week to this point that and I don't know if we talked about it last week, but GitHub was losing

upwards of like $80 per month per user. Now, the CEO of GitHub says that's not true.

So someone was at the conference where they spoke about that and they said it's actually,

no, they're profitable on that. So there was kind of interesting. The report was completely

opposite to what they said on stage at the AI engineer conference. So once again, I just warn

people, journalists have typically somewhere between 10 and 30% of the story when they write a

story. Sometimes they'll get upwards of 50% of the story, even if you're, you know, even if you're

Walter Isaacson and you get to spend three years with Elon, you might get 60% or 70% of the story,

right? Like there's a percentage you can get reasonably, it's not a dig to journalists,

I was one, I'm telling you from my time on the other side of the table. And then obviously,

the person who's in the company who's the CEO, they probably have close to complete information,

80 or 90% of the information, even in their own company, they might not have 10 or 20% of

what's happening. You know, people might hide stuff from them, they might not be aware of something

that occurred. And in this case, I'm sure the CEO knows best. I'm a journalist. And what happened

was somebody leaked, hey, there are some users that are unprofitable. Of course, there are some

users that are unprofitable in any oil you can eat program, whether it's, you know, storage of photos

or using of an AI co-pilot. But overall, they know how much they're spending, they know how

much they're making, and they put those two numbers together and net, net, net, net, net,

at the end of the day, it's profitable, I'm sure. Also, it was Matt Freeman, the former CEO of Github.

The former CEO. So he probably has 70% of the information, you know, and the current person

has 90. And just on that, like just, you know, a quick aside here, but like, basically,

we're here for these sides. Yeah. Yeah. Yeah. But just what that really meant is, you know, they

came out and they set it on stage right here, that it's passed 100 million ARR. And basically,

that's $10 per month for co-pilot. That's 830 devs because they have about 9 million devs on

Github. So think about like the size. 10% are using it, yeah. 10% are using it, exactly. And

that's just co-pilot from Github. There are other options out there. So I'm shocked that only 10%

of Github developers are using this right now. I thought it would be 90%. But it might be,

they're talking about accounts. And those accounts are, some of those are dormant accounts, right?

So... Well, if you're paying the 10 bucks, you aren't dormant, right? Because you're...

No, no, no, I'm talking about the 90% that aren't, right?

Oh, I don't know. There's something like 50 million developers worldwide now, isn't it?

Yeah, they said Github has 100 million user base, but I don't know the active user base.

So I would want to know 830 of active. Let's say there's, let's say there were...

I think they said that was 10%. That's what I heard someone say.

Right, so maybe that if it's 10%, they have 9 million active, right? So they have 9 million

active. 10% makes more sense to me than 1%. But that's, I mean, who's not using it?

Yeah.

Who's not using it? It must have another solution maybe, yeah.

Well, you know, like people are still getting there, but like, what's crazy is,

you know, and we've seen this in our business and I see it in other places as well,

which is, you know, at this point right here, if your team is using this, this is like airline

Wi-Fi. Like it is, it could be $10 for the trip. It could be $50 for the trip. It could be $150.

Like you're going to pay for it.

For the top end user, they're going to pay it. Yeah. So what they said in the tweet is,

have heard from many enterprise AI leaders seeing meaningful productivity uplift, obviously,

which aligns with people saying they'd pay $100 per month for co-pilot.

That makes sense. I mean, the top enterprise software like Salesforce or something like that

is a couple of hundred dollars per month per person. And if you're spending, you know,

or Bloomberg terminals to that 20,000 a year, if you're a professional and you're getting paid

over 100,000 a year, your boss is willing to pay low one to 5% of your salary to make you 10%

more efficient. Easy. Easy. Easy.

So that's, and that's really how you should look at it. If there was a way, actually,

I could talk about producer Nick. Producer Nick has a $4,500 laptop.

Now, my laptop costs, I think I have the best MacBook Air and it costs like $1,800. No, he's

got, I'm the CEO and he's the producer. He's got a laptop, but he does video editing.

So for him to have a $4,500 laptop and trade it in every two years for whatever the state of the

art is, if it makes him 50% or 20% or 10% faster in editing, this week in startups and all in.

I have the numbers, actually, if you want them.

Yeah, let's hear it.

I'm going to draw an export, went from 18 minutes video and audio together to six minutes. So it's

three times faster. Yeah. So you save 10 minutes per episode on export. That's just one function.

Times 50 episodes, you know, 50 episodes, that's 500 minutes. 500 minutes is 10 hours a year.

10 hours a year. Boom. Yeah.

You know, over, and then you do it over the two years. That's one show. Yeah. I know you do the

other four episodes. So you start getting to, you know, maybe if he's 500 minutes times five

episodes, yeah, you draw by 2500 minutes a year. And then also on co-pilot, the

Wall Street Journal reported, again, take that for what you will, more than 1.5 million people

have used it so far and it's helping, and of the people that are using it, it's helping build

about half of the code for co-pilot users, about 50%. That makes sense to me, yeah.

Yeah. I mean, when I'm writing now and I get whatever software I'm using is giving me hints,

I would say one out of three times I take the hint, you know, if I write on the other,

and it's like hand, you know, I hit the tab button. Usually it's tab to accept it. And it's

like, yeah, I'll just tab, I'll take it. Sure. Yeah. What's your keyboard? What's your go-to

keyboard on your, do you use the iOS keyboard or do you use Google or Grammarly? I use Logitech

keyboard. It's typical, oh, on my phone. On your phone, yeah. I'm using the regular iOS one.

After I've written something, then I turn on Grammarly because the iOS one, when you're typing,

is most accurate, even with the ducking and whatever. Yeah. Yeah. The ducking mistakes and

things like that. You know the fix for that, right? No. No. In your like, you know, can you

do shortcuts? Uh-huh. Just turn ducking into, you know, ****. Yeah. Okay. Yeah. So tracking,

turning it to tracking. Yeah. So when I type tracking, it puts in ducking.

If your landing page is terrible, I'm out, right? Most consumers are. It's 2023.

You can't have an ugly website. Stop selling for okay or good and have great and great means

you're using Squarespace. It's out of the box. Beautiful. These websites have templates made

by the world's greatest designers that are going to engage your audience, let you sell anything.

And Squarespace over the past decade has just added feature after feature on top of the gorgeous

templates that are designed for mobile and the drag and drop web design with their fluid engine

is just perfect, easy to use. And you get built in analytics, marketing channel analysis, sales

data, all that stuff. You know, it goes beyond page views and site visits and time and all that.

And with Squarespace, you can create an online store or you can start a blog. Click of a button,

right? Easy peasy lemon squeezy. You can create a subscription business for members only content.

You're seeing a lot of that out there. It's simple. It's cost effective. It's gorgeous.

And they keep adding feature after feature after feature. That's when technology is at its best,

isn't it? When you pay one price, but the product gets better and better and better. You get that

with your Tesla, you get that with your iPhone, you get that with Squarespace. These are the

legendary brands of the internet of this era. Go to squarespace.com for a free trial. And when

you're ready to launch, I want you to go to squarespace.com slash twist. And they're going

to give you 10% off your first purchase of a website or domain. Go to squarespace.com

slash twist because they know we sent you. Okay. So the next one, this one, I couldn't convince

a team to do a demo this week, but maybe you guys can install this one next week,

but I want to talk about it because I think it's interesting. Okay. So this one's called

supermanage, supermanage.ai. Okay. And what it does is it goes through your slack

and figures out what everyone has done. And then it gives you the material to be ready for your

one-on-ones. Got it. I love this. I think this is really important. And look, it takes a little

bit of massaging to get this implemented with the team because it's like, hey, now I've got to be

more aware of it, but I really like sort of how they've talked about it here. I don't have the

demo, but maybe next week you guys install it and we can do it from your side. But the thing is,

the concept is really powerful because within your company, everything that's happening is

probably centered around Slack, hopefully. And she should be able to pull from Slack what the

conversations you need to have around contributions, challenges, engagement, and whatnot. And so

basically, you know, they have a free package, you add it in up to 15 team members with one

manager, sorry, and you go from there. Okay, so to be clear here, I install this on my Slack

and then it would go in, it would find every Slack message that producer Nick did.

And we actually do something called SOD and EOD. Every day I ask people as remote workers,

as part of the deal, you get to work remote and you get to do your own schedule, reasonably speaking.

All I ask is that you do a stand-up meeting in the morning with your team and that you do an SOD

and an EOD. SOD, five minutes, here's where I'm going to work on today. And you share that with

the team. So everybody knows what people are working on. So people are informed. And at the end

of the day, you say what you, you reply to that and say what you get accomplished. And sometimes

people put a little bit of filler in there and I have to like tell them like, don't put, I read my

email, don't, you know, like that's not an actual thing you got done. So I have to keep reminding

people it's not for like little like everyday tasks, it's for things you're accomplishing,

contributions that would change the fate of the company or in some way, right? So supermanage

looks for contributions and then it summarizes them and then it looks for challenges that people

are having and it summarizes them and then and then engagement as well. Right. Yeah. Yeah. And so

you like, look at this one based on the messages, Connie seems to be highly engaged and excited

about their work. Oh, yeah. I think it's really good. Yeah, this is I like this. And then what's

going to happen that's going to be really awkward is it's going to turn into a black mirror episode

where like producer Nick's going to know, oh, you know, I've got a show more, it's going to be like

Flair member in office space and they're like, you only want to wear three pieces of flair,

that's okay, but he's got 20 pieces of flair. So now Nick being like, oh my god, what a great day,

OMG, shout out to my peeps, you know, doing a great job and it's going to just like write

nonsense to get his engagement score up, you know? Yeah. So, but I mean, our team has good

engagement and not tell the team. That's the other way. Well, I do think that I do look at

things that the team doesn't know I look at, like I look at how many days people were on slack.

So I know who's 30 days out of 30 days and I've shared it a couple of times, but I can tell

who is super committed to the company based on, you know, if somebody, let's say we have 20 work

days and somebody's like, they were on slack for 20 of the work days, but not the other 10 days.

I start to worry like, is this like a nine to five type person now they have to drill down

because they could have had a week vacation and yeah, it's your right to not be on slack

on the week. But I'll be honest, like if you want to compete in venture, it's a 30 out of 30 day

job. Like I'm 30 of 30 every month, Nick's 29 or 30 out of 30. I mean, you could take a day off,

you have a child born, you're out of wedding, you went, you're like when I went to July 4th,

I was off the grid for a week. Yeah, but that's different. But the real deals in venture on

as EJ Cal are happening outside the nine to five hours. Of course they are. Yes. My experience

as an entrepreneur, very little of the real deals with investors and partners and you know,

all that is happening outside nine to five. Yeah, it's happening at night or the weekends,

right? They call you up, hey, let's go for a walk, hey, meet me a fill, let's get a coffee.

Let's wrap this up on the weekend. Let's, let's turn this out for Monday's meeting. I agree.

So Supermanage, I'm going to give this, we don't know if it works or we haven't used it,

but I'm going to give it, I'm going to evaluate it based on the idea. Based on the idea, it's a

B plus. It's a B plus. This is something that I think will, again, we would look at both these,

Trove and Supermanage. You can say to yourself, hey, this is going to be built in to these services.

We have another employee engagement company called 15.5. I remember I was there when they

got funded on stage. Yeah. So 15.5 is an amazing great founder and it's a great company. They're

going to need to incorporate some of this. So Supermanage will be going up against like say

15.5 or other services. There's a couple of other ones that manage the OKR process and employee

engagement or happiness. And what 15.5 does is it asks you a series of questions each week.

Those roll up to your manager, then your manager rolls up or their reports to

the CEO, let's say. And you can highlight like, hey, this is an interesting response.

And people really love using it. So again, some people don't like this stuff because they're

like, I'm a high performer. Why are you making me talk about nonsense? So there's a hardcore samurai

group that are just like, I did my job. There's blood on my sword. Go watch the tape if you want

to see me murder everybody who tried to attack the village at the end. Yeah. That's it. So there

are some people who find this kind of stuff offensive on its face. What I find it to be is

accountability to each other. And that's what I try to encourage people is like, hey,

if you're sharing this stuff, it's great. But sentiment is also nonsense. So the sentiment

stuff I'm not crazy about because again, I mean, I guess it could theoretically, but some of the

best people I've ever worked with are sarcastic, gnarly, contankerous. And those are literally

high performers sometimes are sardonic, sarcastic. Yeah. You're saying just the AI is going to put

them into this bucket and it's not necessarily that they're burning out or they're bad teammates,

they're just being themselves. Exactly. Usually the 10x engineers have all those qualities that

you're talking about. So like, you just have to be careful. Like sometimes people who are like

the popular person at the office at the water cooler, and they, you know, get everybody to

laugh, whatever, they're actually not getting GSD. They're not getting stuff done. I go even

further, Jason, I would say that negative sentiment is, I would say sometimes a positive

indicator. The most negative person I know consistently at launch is maybe the highest,

the two actually that are usually like, you know, the sky is falling, but are very concerned with

things going wrong are Ashley and Jackie. And they are probably your two highest performers at this

company objectively speaking. Ashley and Jackie are two managing directors. One's been with me for

10 years, one was with me nine years, and it does say something about me as a manager. I like people.

Again, all leadership is top down. I like people in my org who are like,

we could do better, or this person is good, but not great. You know, so they're kind of like

standard bearers. Ashley and Jackie, for me, shout out to managing directors, Ashley and Jackie,

because let's be like, you know what, I think this founder is insincere and, you know, they've

been negotiating this stuff, not in good faith. I think we maybe should pass on this investment

because the diligence process was, you know, like we don't, I didn't get the sense of honesty from

this founder. Like this happened like, you know, multiple times in the last decade,

where they correctly pointed out, and I, and I overrode them and made the investment.

And they were like, see, this founder was brutal during the diligence process and negotiating the

term sheet. And now look, they, the company, now most companies fail, so you don't get credit for

it failing, but we did, they did predict correctly downstream issues. And I said, listen, this is

a sins of omission business. I hear what you're saying. Of course, as a difficult founder, you

have some pink or red flags. I'm going to make the 100k, 250k bet anyway, in the case that this is

a person who is difficult to get along with, but a high performer. And I think in both cases that

happened, they were right. I was wrong. Or we are rated a nine out of 10 by a founder or by someone

who goes to an event. And instead of saying, great job, everybody, nine out of 10 is a great

score. They say, why wasn't it a 10? What was wrong with us? Why didn't we get a 10? How do we

see it? Why didn't we get the 10? Yeah, yeah. What could be better? Exactly. Yes. I think so.

You have to be careful with that. But okay, so great to the super managed people. I'm going to try

it actually. I want to try it. And I did see that Slack has made an announcement about an AI product

that will remind you of what happened since you've been gone. Oh, yeah. Yeah, you're right. But

there is something about how a bunch of these yeah. Yeah, recapping channels. Slack's AI tool

can recap channels and threads starting testing this winter. Slack, I'm reading from the Verge

here. Slack AI can do things like summarize channel discussions, which Zoom has been doing for a month

now. Threads and search for answers in relevant messages, files and channels. So I think that

those recaps are going to be really good. See, if you get to a channel and it's like, here's what's

been going on the past week, that could be very powerful. I agree with that. It's like a very

powerful, obvious tool. This is slightly different. And then it's going to be like

generative AI and Slack. Anybody who's got chat should have all the stuff built in. So the fact

that there's not a Dolly competitor built into Slack, and we're not asking for it to do stuff

inside of Slack, this could make Slack. This could breathe new life into Slack if they get AI.

It could make Slack really sticky. Or someone could finally take them out.

Or somebody could take them out. Right. If somebody built the AI first version of Slack,

I would fund that easily. If somebody wants to make an AI version of Slack, here's what it

needs to do. It needs to start me with pointing me to the discussions that are occurring. So when

I open up Slack, it's like, Nick and Matt and Jackie were discussing a new series for this week

in startup based on AI. And they're discussing potential advertisers, guests, and formats.

Click here to join the discussion. It would actually be summarizing their discussions in

real time. That would be incredibly powerful for a manager of a company with anything more than 20,

30, 40 employees. Because you can't keep up with all the discussions, like a top-level

thing. So you basically start from the top-level experience, and then you drill down. That would

be super powerful. Another great example of negative sentiment. Matt Kotler, what does he

always want? More inventory. We don't have enough inventory. I need to sell more. I need more. I

need more. Always. Exactly. Yes. I mean, listen, the people who one might interpret as irritable

or cantankerous, another person might interpret them as being a perfectionist or demanding

excellence. And I think the three people you described in our organization, they're pursuing

excellence. That's all. And for people who are mids, mids review pursuit of excellence as...

This person's a negative. This person's too negative. Whatever.

That's how you know a mid. And that's how you know a high performer. A high performer is like,

great, we got a nine. How do we get a 10? Great. We sold out for three months in advance. How do

we raise prices and get to sold out for four months in advance? Whatever it is. Just don't be,

don't accept reality. But you can celebrate. And I like to celebrate excellence. Take a minute,

pop the champagne quirks. Great job, everybody. Now, what lessons did we learn for next year?

Boom. Which is why I take negative feedback. I put it in a document called Lessons Learned,

not We're Idiots. So we used to have a document, Nick, when I did Lessons Learned after each event,

it used to be like the mistake tracker or problem tracker. Yeah. Like the anti-portfolio,

you know, people have done that. Yeah. And so I reframed it instead of it being mistakes,

the mistake tracker and things we screwed up. Tracker, which is how I used to say it,

I said lessons learned. Yeah. Framing matters. Right. So it gave people permission to say,

hey, I learned this lesson at this event. You know, people didn't know there was a VIP downstairs,

so they missed it, or the lines were too long at the VIP food. We should have had twice as

many stations and spread it out, right? This venue was too cramped. Like the first night's

venue was too cramped. Lesson learned. Lesson learned. Great. The other two ones were huge venues.

All right. You know, I've been on a health kick over the past year, and you know, I care about

data-driven solutions. And if you listen to this podcast, I bet you do too. So let me tell you about

Fitbot. This is a data-driven workout app that blends machine learning with exercise science.

Fitbot creates custom dynamic workouts programs based on your fitness goals, your experience,

and most interestingly to me, the available equipment. Let's say you got a bunch of kettlebells,

or let's say you're at some, you know, sparse gym at a hotel, or you're on vacation, you got nothing.

Well, Fitbot will maximize your fitness gains by varying the intensity and the volume between

your sessions and leverage the equipment you have or don't have, as the case may be.

You can customize the length of your workout, what muscles you want to target, and so much more.

So let's say you want to get a 30-minute workout in, and I want to do chest triceps and abs,

but I'm staying at an Airbnb. There's no equipment. Fitbot can create a perfectly optimized workout

for me based on these parameters, and it'll do it for you too. Check it out. It's amazing. The

design of this app is extraordinary. I was able to invest in it. That's how impressed I was with

it. Fitbot takes the guesswork out of fitness. Just open the app and start making progress.

You deserve it. Get 25% off your Fitbot subscription or try out the app for free

when you sign up now at fitbot.me-twist. That's F-I-T-B-O-D-dot-M-E-slash-T-W-I-S-T for 25% off.

All right. Multimodal, Sonny, Madra, the bestie, is something that I just got on my phone.

It's unbelievable. It works. I saw Robert Scoble take a picture of his spice cabinet and say,

what's in here? It made a pretty perfect list. I'm just like, you know what? There are people

who organize your cabinets and they cost $150 an hour and rich people favor this. Nonsense to come

and do their spice thing. You can just take a picture. I encourage everybody to do a chat,

GPT-4 shout-out, Sam Altman and Greg and everybody over there. Take a picture of a spice

drawer. It will tell you what's in there. Pretty great. Now you know what spices you have

in your cabinet, and then you can say, I have chicken. What can I do with these spices? Robert

Scoble did that tweet. It worked great. Fantastic. Let's keep going here. What's a multi,

what's going on with Multimodal? Multiple different input types, multiple different

inputs. Let me see this one for you guys, and then you guys can react to it.

Oh, you have one, Nick. Go ahead, pretty sure, Nick. Yeah, yeah.

Shout-out, pretty sure, Nick. Last week, box CEO and founder, Aaron Levy, tweeted this.

Shout-out, Aaron. Let's have him back on.

She put a screenshot of a landing page. And this was their own product, by the way, right?

Yeah. And he said, please suggest 10 ways to improve this interface for usability. And chat,

GPT gave it back, you know, a bunch of, you know, you could do breadcrumbs navigation,

a unified search bar, a file preview. So I interpreted that for us. And on this week in

startups, we just built- Oh, the new homepage?

Exactly. Basically, I just posted the screenshot in there, and I said,

this is a landing page of a website for a podcast on tech startups. Please suggest 10 ways to

improve this interface to maximize engagement from viewers and SEO.

Two things I always say. Yeah.

Exactly. And I'll just let you, can you read these?

Yes. It's a prominent call to action. Enhance the availability of CTAs such as subscribe

or listen. That was my number one thing that I was yelling at you guys about.

Provide short audio video snippets, teasers. That's a great idea.

Expand the episode descriptions with more keywords. Okay. That's pretty good.

Episode description. Interactive search. We have that. Social sharing.

Yeah. We don't have social sharing buttons. User reviews and ratings. That's a great idea.

Personalized recommendations. It's a really good idea.

Engage with community. Okay. It's a solid idea. Mobile responses. Sure. Yeah.

These are all great suggestions.

This was just for the landing page of the actual website. Then we went one step further.

And on the sales meeting today, we were like,

how can we make this new website that we're building better for partners?

So I took a screenshot of a specific episode page, right? So here's the episode page, right?

Last one you saw was the actual main homepage. This is an episode page from last week.

Yep. All right. And I said, my team produces a podcast about startups and technology. Every

episode gets its own webpage on our main website. Here's a screenshot from the most recent episodes

page. We have three advertisers per episode and would like to feature their logos and offers on

each episode page without having it be overbearing, but making it noticeable. Can you please suggest

some ideas on how we could integrate three advertising logos and offers given the current

layout of the page? Oh. So here are the ideas. Oh, sponsor ribbon, sidebar integration,

table of contents integration. Great idea. I love this one. Interstitial design.

As users scroll through the transcript, every few sections, you can integrate a horizontal bar

with one of the sponsors' logos and offers. That's what we need.

A gentle reminder without being too intrusive. That's a great suggestion.

I mean, it's literally if you hired somebody for $90 an hour, a UX designer, this is what they would

say. So it has, what's really the magic of chat GPT and large language models, I think, Nick, is

we spent the last 20 years of everybody trying to showcase their knowledge in order to get work,

every service person, whether you're a VC in service or a lawyer, accountant, UX designer,

you wrote blog posts. Here's how to make a perfect landing page. Here's how to make a podcast

landing page, whatever. Chat GPT has now put all that knowledge in. It knows how to spit it back

out to you in a very organized fashion, tailored to you. And all of those service people who try

to make content in order to get customers have now sealed their own fate. They have now killed

their business by sharing so much information that people who you would have hired you do not need

you. Congratulations, everybody. This is just pure job destruction. I saw a tweet this morning that

said, you know, like a little bit of expansion on that, that like Sam Altman OpenAI took all the

internet and basically made an API for it and basically charges you to consume it. Exactly.

Yeah, but it's like... They stole everybody's knowledge and now they're charging at $20 a

month for it. They've basically made all the knowledge... Access it one query at a time.

One query at a time. But there's a free version, Chat GPT. So in fact, they made it free for everybody.

You only have to pay for the enhanced features. Yeah, fair point. Fair point. I mean, I can't imagine

at this dollar level, $20 a month, that people are not going to own two or three of these,

even with the throttling. Like, it's so affordable to have this that I just told everybody to buy it

for $20. So we have $400 a month right now. I think maybe three people in my company are using it.

Literally, like Nick's using it. Presh is using it. I use it. I don't know if anybody else in the

company is using it. You've got to get on the enterprise so that you get them on enterprise

so you can see everyone's usage. Yeah, just like I see everybody's usage in Slack and then I keep

going, hey, we need to get these 10 people to use it. I don't use shame to use it. I use

professional development. So what I do is I look for the people. I just say, hey, listen,

have you done an email sequence? Have you used ChatGPT? Just be honest, yes or no. Did you use

it this week? People say, no. I'm like, okay, this person's going to do a Zoom call on Friday.

Please join it. And I just try to put a bunch of carrots out there and very few sticks.

You know what the intersection is, going back to that Slack idea for a second? If you created

like a AI native Slack, every time that you're typing something, it would just prompt you in

there to be like, hey, let me help you with this. And I'm like, yes. The co-pilot in Slack would take

the people who are the laggards and I would say half my company is laggards. No offense.

And they all listen to the pods. I don't mind talking to them, probably because I accept it.

But everybody who my company, who works for me, who's a laggard,

just like make your default screen, ChatGPT4, put on your home screen of your iPhone,

the ChatGPT4 app. Once it's on your home screen or it's in your dock,

you know, it's kind of like auto-loading Slack in the morning.

You can even do it as a widget on your lock screen if you want.

Yes, make it a widget. Yes. I think that's a really cool idea. Like, oh, here are things

in the company. Remember we talked about the negative sentiment? Hey, here are things that,

you know, Jackie, Matt and Ashley are noting or complaining about.

Here are some solutions for that. Here are some strategies, right?

One of the things I've been trying to get my team to do is, you know, you watch Y Combinator

and they are super chatty about their portfolio companies and they kind of do it as a group.

So they must be in a group chat where they're like, hey, I'm going to talk about this company.

And they just know to amplify and reply to each other to get that signal up. And Gary Tan is

amazing at it. And I'm telling my team, like, I'm the only person in our company talking about

our companies. Every day, I want a blog post starting today. This is my tirade this weekend.

I want every day, 250 days a year, somebody on the investment team to write a blog post about

our companies on the launch blog or the founding university blog. And it doesn't have to be a

long blog post. And you can do it in chat, GPT, you know, just, hey, here's the company.

Make it an X thread. And then make it, I'd said, and then make it an X Twitter thread.

And then put it on your LinkedIn. And then I had to tell everybody, please make your accounts

professional. We don't need your pronouns in your profile. You know, like if you were for me,

you don't need to put your pronouns in there. You don't need to put whatever else in there.

Just like, please focus your social media on being an investor if you're on the investment

team. It's not too much to ask. Like it's a pretty dope job. Your social media presence needs to be

serious. Private Instagram account, if you need that stuff, that's what you can do.

Yeah, that's I think X is for pro LinkedIn is professionalism. Be a professional. If you

can't be a professional, don't be on the investment team. Like you need to invest

in this, right? I don't know how you feel about pronouns

in your bio for business people. I mean, I don't tell people to take them out, but I will say

like if I see them in someone, I thought it looked silly. Yeah. And I don't think it looks

professional. That's a better way to say it. But when I do see someone that I'm working with,

like say like as a potential customer partner, it does it in my mind, I go, oh, you know, like,

yes, it just like it's, it would be like me being like in my profile, whatever, I'm, you know,

when I put that I was for Bloomberg, I changed my avatar to Bloomberg and I put my Bloomberg thing

and my Lord, I had advertisers, I had partners, LPs, founders, everybody in orbit, you're going to

get Trump reelected, you're splitting the boat from Biden, you remember how charged up it was.

It was a tough couple of months. I was like, you know what, I'm leaving it. And then I thought

about it afterwards. And I was like, you know what, the world does not need to hear me as a

business chime in on everything. Yeah. Full stop, right? Like coin based kind of approach.

So like we don't need to take a professional stand on everything.

Some things, sure, there'll be time for that. You don't need to take a professional stand

on everything. Chat, GBT, multimodal UI grades from both of you on it.

Let me see. If I were to hire somebody and pay them $100 an hour, I would have felt very good

if I paid $1,000 to a UX developer $100 an hour. 10 hours. If in 10 hours they give you

that in their report back, I would say, yeah, you know, I think two or three of these are

really actionable. Let's do it. Thank you. Here's $1,000. Now, I wouldn't pay $10,000 for that

information, but I would have paid $1,000. And I know somebody who's like from a big corporation

would have paid $25,000. But for me, I would pay $5 to $10 hours at $100 an hour, $75 an hour for

that. And I got it for free. It's an A plus. It's an A plus. That is one of the best things I've

ever seen chat GBT do. I'll be honest. What do you give me? Look at me with my grades today.

So people don't think I'm mature. I'm willing to give an A plus if I think the value is there.

That would save me from finding the consultant, two weeks of nonsense in Meshugana,

and then it saved me $1,000. If I could save a thou, even for J-Cal, saving a thou is meaningful

for me. That adds up. Yeah, I'm at like an A minus. What I want to see, you know, like going

back to the calculator demo, I want to see be able to implement it for you as well.

Oh, okay. I'm going to restate. I'm going to give an A. I'm not giving the plus. I'm saving the

plus for that. You're right. Because there is a, okay, and do it for me.

Okay. Yeah, good. We give them something to work on. You give it a minus. I give it A. That's the

first time we've had a flip. We've been, yeah. The first time we flip. All right. Give me one more.

One more demo. Okay. Before I do the demo, I actually want to give you a shout out on your

ads team because we were just talking about it. Okay. Shout out. So, you know, sometimes I go back

and watch the episodes and, you know, improvement or just see things like that. I kept seeing Vanta

and, you know, known about them for a long time. We actually tried to use them in my previous

company. They were pretty early on, but because I kept seeing them there, I actually used them

and we're in deep in a process with them and they're incredible. So, they're being our SOC2

compliance. Shout out to the Vanta team. Yes. Yeah. And it really, I know they're big advertisers.

So, it works. Yeah, it's working. What I tell people is if you're a product and add on this

week in startups is only $5,000. I've kept the price the same forever. It's almost $5,000 per

ad. So, if you do, you know, 10 ads is 50k. The minimum ad buy is five ads. So, to be engaged is

25k. The reason we do that and we have a minimum of five is people would do one ad. Let's say you're

on an episode and, you know, not a lot of people listen to that episode. And then there's another

episode that breaks out where you got to take a portfolio approach. Yeah, you don't get an ROI

but if you do five, a minimum of five, and one person buys your product and your product costs

$500 a month or more. That's where I tell people like it's kind of a no-brainer if it's a product

that business people use. Because if you spend 50k on ads and your product is 500 a month,

it's $6,000. And over the next three years, that one customer is going to be 18k. So, one customer

equals essentially 18 of 25 is pretty much break even. If you can attribute one, there's probably

two more, right? Because you can usually only attribute one in five or one in 10. Most people

don't use the URL. But like you, you just, you heard Vantavanta, Vantavanta, LinkedIn jobs,

LinkedIn jobs, LinkedIn jobs. You're like, okay, frick it, I'm going to test it. Well, I said,

let me go back to it because like, they're a little bit early in my first company. But now, and

I have to say they've done a great job. And so it was like really incredible. And you're going to spend

probably 25 or 50k with them. Easy, easy overtime. And so it's pretty easy to justify the ads. Now,

what I do tell people is if you're Audible and your product is only $12 for a book,

you know, I tell the Audible team like, you know, I don't know, maybe Tim Ferriss is a better

podcast, like some of that's less niche, you know, for a wider audience, maybe do Tim's podcast or

Joe Rogan's because or, you know, Bill Simmons might be better, although there's Spotify and they

do auto books now. But they sometimes they still do it because they're like, we just love you reading

the ads. It's like a branding thing for us. Okay, sure. Branding, I get it. You want to be

affiliated totally. But, you know, for if it's $25,000, you need to sell 2000 books to hit break

even or maybe you have to attribute 500 books. That's kind of a hard, that's a hard thing to kind

of do if you're a ROI based advertiser. So what you want to do, I like that math, that's, that's

helpful. Because I don't want people, my nightmares, people complaining they bought ads. So like,

we've had startups buy ads to try to get to me. I'm like, no, do not take their money. If it's a

seed state startup, don't take their money. What I do do is for our startups, three or four times,

we've had a little bit of extra inventory, like, especially over the, you know, 2022 when people

were cutting things, all of a sudden, we had some people asked to move their ads. So we let them,

we had a company go out of business, you know, that kind of stuff. And I just gave the ads for

half price to our founders on a first come first serve basis. And they like, they grabbed them.

And I was like, okay, but just so you know, like, don't do this if you know, like as a favor to me,

I'm just all, and these are companies that, you know, have 10 million in revenue. So it's

worth it for them. Awesome. For their mix. But anyway, so yeah, shout out to them.

I wanted to give a shout out to the team. Thank you. Last demo.

Vanta.com slash twist, you get $1,000 off your sock.

About a thousand bucks. That's the problem is most of my audience is so well healed.

You don't care about saving $500,000. I didn't even use the code. We should have used the code.

Of course. Vanta.com slash twist. Yeah.

All right. Okay. So back in Google, they've been doing amazing things. And what we're going to

talk about here is, so I just did an image search for Halloween table settings, right?

Great. Sure. And, you know, you typically get the same set of things here. But what you're going

to find is as we scroll a little window here, and this little window says, create something new.

See brand new images generated by AI. What? So if I hit this generally, yes. This is in images?

This is in Google search when you go to images. So you search for something.

This is new. This is what we do here. Take a look at this. See brand new images generated with AI.

This is sick. This is Dolly. Yes. And so you had said this a little bit earlier. And so you're,

you know, I couldn't transition right at that point. But I was, you're saying that all these

things you've been doing in the product, so they've done an incredible job here. So these are

on the left hand side for those listening. These are images that like Google has picked up through

its crawl process. But on the right hand side now, I have images that were generated by their,

you know, their image generator, which is, you know, likely palm or something like that. So,

very, very cool. Super handy. You know, it's limited in some ways. Like, you know, we were

kind of doing some demo this morning. And I think Nick wanted like a New York Yankees one. It

wouldn't do it. So it's obviously some, it's more aware of copyrights and other things like that.

But I've tried just a couple of little demos. Jacob, you can try one, but it's,

I thought it was really nicely done and how they've entered, you know,

inserted it right into their existing product, which is huge. So click it. Click it. I want to see

the results. This is huge. Remember, I said, right here on the right hand side, right here,

on the right, these are, these are all the generated ones on there. Oh, sorry, sorry, sorry, sorry.

I had my zoom screen. So look at this. A Halloween setting that features a green tablecloth,

orange, dinnerware and candles in the shape of jack-o-lanterns. This is unbelievable.

Yeah. And then you can click through these ones here.

Holy cow. Okay. Shout out to my guy. I just want to say shout out to my guy,

Sundar Pichai. I was just talking to him. He listens to the pod. I saw him at a social event.

He said to me, you're everywhere, J. Cal. I said, Sundar knows J. Cal. Okay. I'd met him before,

but I, you know, I always assume people don't know me or hate me. That's my own little neuroses.

I'm like, either this person hates me or they don't know me.

That's actually quite true as your friend on the outside. That's pretty true. That's a good,

you know, categorization. Do people not know me or they're like, F and J. Cal.

Yeah. Exactly. Well, I mean, the good news is anybody who knows me says to the person,

oh no, he's great. He's just, you know, he's got big opinions and they're like, really?

He is great. Okay. I'll give him a shot. But anyway, Sundar, I gave you very candid advice

in person. And now, and I told the teams, C minus D plus, the ratings were very low for me.

Okay. With good reason. You're putting out stuff that was incredibly low quality,

which I understand. You got to, you got to get stuff in Mark and get feedback. This,

this right here, this is, you ready for my score?

Let's hear it. It's not an A. It's not an A. It's an A plus. First perfect score.

This is A plus whoever on the Google team. Wow. Made this. This is A plus with a plus for extra

credit. Why? Not only are the results great, but conceptually, this is perfect. It's perfect

because you don't have to change the user's behavior. And you're gently communicating to them.

Whoever wrote the copy here creates something new. See brand new images generated with AI.

This is chef's kiss. You're not forcing people, forcing it down people's throat.

You're kind of just giving them that, Hey, there's a little option over here.

It's kind of like, Hey, there's a tasting menu. You don't have to do the tasting menu.

It's over here on the side. If you want it, you take it. Or you, or you could do Omokase,

trust me, but you don't have to do it. Wonderful. What a killer idea. This is what

Google Flight should be. And then you can go into an edit workflow. And if you want to

change it up, you can change the prompt. Exactly. And then regenerate it.

Wow. I mean, let me tell you something. If you're over at chat GPT-4 and you're open AI

and you're Bing, game on. Game on because Google has billions of users and it's game on.

This is how flights should work. When you're doing a Google Flight search or a Gmail search,

it should do a little thing saying, Hey, here's some other things you can do with Google Flights.

Ask Google Flights to find you a first class, find me a first class ticket to somewhere that's

has warm water during, you know, January 15th to January 30th. Find me the best business class seat

and affordable hotel where I can have the best chance at deep powder skiing February 15th to

30th. And it's like, Hey, Nisako, Hey, here, Hey, here. That's what we need, Sundar. And that's

where each of these product heads on Monday, the image team comes in Tuesday, the shopping team

comes in Wednesday, the Gmail team comes in Thursday, the YouTube team comes in and Friday,

you know, whatever team comes in enterprise, enterprise comes in. Yeah. And every day,

8am in Sundar's office, not 9am or 10am, 8am. And you got to be there at 745 to set up your demo.

8am Sundar walks in the room. Last week's progress was X. This week's progress is Y go.

And then if somebody comes three weeks in a row and it's not getting dramatically different,

you have to just do what Steve Jobs did. You ask the next Lieutenant, what should,

what should this look like? The person tells you what it should look like. You say, great,

make that. And then you look at the other person, you work for him. You work for her.

That's it. You're in charge now. Sundar, this is red alert. This is DEF CON. What's the highest

DEF CON? Yeah, it goes up. Yeah, it goes up. Yeah. Yeah. This is DEF CON 4. Like you, and this is

great work. So shout out your first A plus to the people at Google. This has been the C,

the C students have gotten an A plus. What do you give it? What do you give it?

On the same way, they're moving like A plus, like you said, integrated right in the product.

You don't even have to do anything separate. You can iterate from there. You don't have to,

you know, a lot of times when you're in Google search images, you're worried about copyrights

and all kinds of other things. Now you can get inspiration from the images around you.

I'm the same as you. I think right integrated right in.

I'm seeing this all over the place. You know why? A lot of the people who are doing the all-in

summit 150th episode, which is next weekend, I think, or this weekend, actually. This week.

This weekend. A lot of those folks are making their own artwork. And I can tell it's like

Dolly or ChatGPT4. So like the Stockholm one had a beautiful image of a bunch of dorks,

you know, hanging out, dressed like they're in the great Gatsby and, you know, in some famous

Stockholm city image show. All right. Defcon one is the highest level. It means nuclear war

is imminent. Her has already begun. Yes. This is Defcon one, Sundar. And anybody comes in and

they're treating it like Defcon five or four or three. It's Defcon one for Google. Take it serious,

folks. And if you can't take it serious, Sundar, you have my permission out the door.

Out the door. Nobody ever gets fired at Google. Now it's time to start firing people.

Is the JTRADing position in? You come two weeks in a row without progress? Done.

Increase the JTRADing position?

You know, listen, I might have to make a JTRAD on this. I might make a JTRAD. I may make a JTRAD

on Snapchat. I was talking to somebody on Snap. Snap's engagement is incredible.

And they keep growing. I'm going to take a look at Snap. This is one of my things I'm going to

do this week is look at Snap. You know what? I haven't traded on my JTRAD portfolio because

I'm very happy with the performance of everything except for Disney and Warner Brothers, which I

might double down on now that activist investor just bought a bunch. It might be time to double

down or I might be just keep buying into the crash. This could be bad consumer behavior.

So give me... We should actually do some JTRADs together here on the show based on AI.

Let's do that. Next week we'll do a JTRAD.

All right, everybody. This has been another amazing episode. If you need help with your AI

strategy and you're a large corporation with large data sets, you need to use definitive.

Definitive AI. I'm an investor. Sandeep is a genius as you all know. Go to definitive.io.

You go to definitive.io or your email sunny at definitive.io. It's absurdly expensive to work

with this firm, but they're the best. And if you need a partner, that's the partnership.

But it's only mid six figures or higher. So if you're looking to do something for 25K, no.

But if it's 500K engagement or higher and you're doing something legit,

he's got... What do you got? 40, 50 geniuses over there now?

No, we're not that big. We got a lot of AI now.

So... Okay. So 20, 30 geniuses over there. I can't say the name of your customers because

they don't want to talk about it, but I know... Yes.

Because if you would like to look at the Fortune 500, bing, bing, bing, bing, bing, a bunch of

things would light up. That's it. I'm going to make so much money. What did I put? I put 250

into this? You did. You did.

How many besties did you get? All of them.

Four for four. But four for four from the podcast, but even our chat group, you know...

We've got a couple there too. Yeah, full coverage. Yeah.

You got... I mean, people don't know this, but Sunny is the only company that I know of that has

four or four besties. Cafe X has three or four. Cafe X, Saks, Chamath and I all put money in

Cafe X. I'm surprised Freeberg didn't do it. Well, that's crazy.

Well, I didn't know Freeberg back then when we were doing the fund, but Freeberg did do Itza.

And I think actually I did talk to Freeberg about it and he thought it was competitive with Itza,

which I gave him the best piece of advice ever. Itza was all... It was like they were trying to

do automated food and it was like these incredible quinoa bowls. That's where the queen of quinoa

came from. He's come to the poker game and talking about quinoa. And I was like,

look, it's the queen of quinoa, which I got from the queen of California,

John Mayer song on the album Born and Raised. Because I happened to be listening to it at the

time he walked in. Oh, I said, oh, the queen of quinoa is here. I told him, listen,

I know you want the world to eat quinoa. Here's a way to get them to do it.

Flank steak, short rib, salmon with rare inside seared salmon, miso cod, and barbecue chicken.

Barbecue chicken, roast barbecue, it's sesame, roast sesame. I gave him how to do it.

Company didn't work out. But people I knew who were going to go there, they wanted to eat protein.

And then if you want to sneak the quinoa in, you put some short rib in there,

put a little seared tuna, you would have sold a lot more of that quinoa. You got to pull people

into vegetarianism. You can't force it. Anyway, lesson learned for everybody. All right,

everybody. See you next time. Bye bye.

Machine-generated transcript that may contain inaccuracies.