This Week in Startups: OpenAI taps Jony Ive for “iPhone for AI," FTC/Amazon + Roam Around CEO Shie Gabbai | E1820

This is crazy. It's a crazy concept. I'm not saying she's crazy, but I'm saying it's a very

bold, audacious way of looking at the world. I would say it's also a little bit narcissistic

that you, Lena Kahn, are going to be able to predict future harm. I know I'm super humble,

having been in the industry for 25 years and invested in 400 companies. I can tell you

that she cannot predict the future. Nobody can. This Week in Startups is brought to you by

SuperGut is the only nutrition brand clinically proven to improve digestion, balance blood sugar,

sustain energy, and manage weight. Save 25% on their delicious shakes, bars, and prebiotic mix

at supergut.com with code TWIST. Catalog. Stop wasting time and money with expensive

design firms and unreliable freelancers. Get fast, three-day turnarounds for a flat monthly fee

with catalog. Get $1,200 off right now at trycatalog.com slash TWIST and Codecademy. Build

the future you want to see with Codecademy. Codecademy Pro helps you learn everything you'll

need to shape what comes next in the tech space. Try it free for 14 days. Visit codecademy.com

slash TWIST. All right, everybody. Some quick news here at the top of the show before I do an

interview with Rome AI, one of our great incubator investments. Open AI made a lot of news this week.

First, there's news that Open AI is an advanced talk with Johnny I of Apple fame and Masayoshi

Sun from SoftBank to raise a billion dollars to build the iPhone of AI. I don't know if this

would be a form factor of an iPhone if I'm being honest. I would find it very surprising if Johnny

I've wanted to come back and just do an iPhone. I think this will be some very interesting wearable,

and this is the interpretation by the Financial Times, by the way, who broke the story.

If you don't know, Johnny I've, when he left, Apple started a firm called Love From. It's

like a design firm. He's done a couple of projects. He's obviously a genius and was

Steve Jobs' partner in crime. They've had some brainstorming sessions. Sam Altman and Johnny

I have, according to the FT, it's very early stages and Masa would lead a billion dollar

seed round according to the reporting. I think multiple outlets have talked about this.

It's unclear if this would be a new company or it would be an Open AI project or where this would

sit. It sounds to me like this would be part of the Sam Altman collection of companies,

and I think they would probably make it its own company and then just give it access to

Open AI's stack, but you never know. They could do anything now because in our second story,

Open AI is discussing a secondary share sale that would value the company at 80 or 90 billion

dollars, according to the Wall Street Journal. That's 3x the valuation from earlier this year,

which was 30 billion, which was stunning. It seems like Open AI wants to sell a couple of

hundred million dollars in shares, and that they're on track to do a billion in revenue in 2023.

A billion in revenue is pretty extraordinary because the product costs $20 a month. That's

240 a year. If you had a million people using the product, that would be 240 million. If they had

4 million people paying for accounts, that would be a billion dollars-ish. Let's just break that

down for a second. There's 100 million people signed up, I think, for Open AI's product,

ChatGPT4. If 4% converted, that would be a very large number, but it is possible

that they could be generating something in that range. Of course, there's the API. Maybe

half their billion dollars in revenue, if that's in fact true. It would be shocking, but it wouldn't

be unexpected that they were trending towards that because the product is so helpful. So many

startups are using it. Keep in mind, though, that Open AI is 49% owned by Microsoft,

and then Sam says he hasn't owned any shares personally, but he controls the nonprofit.

Anyway, this would be a big gain for Microsoft once again. Microsoft has made some big gains

from their investments over time. Sounds really interesting. The multimodal mode,

which launched from Open AI that we talked about with Sunny earlier in the week,

would make it very believable that they have something here, and that something might be

the ability to use glasses, audio. I could see this being anything from sunglasses to

something like the personal communicator on Star Trek that you put on your vest, like a clip-on

to your jacket where you press it and you talk to it, and then AirPods. So the possibilities here

are endless. It could be something on your arm. So imagine if you had something on your arm that

was a display, something in your ears, and then a pair of glasses. You have some really interesting

wearable because if it's multimodal, well, you might want to hear some answers. You might want

to see some answers in your glasses because you're doing an activity, and you might want to tap

on your wrist here if you were wearing like a bracelet. So if you look online at wearables,

about 10, 15 years ago, there was a huge wearables moment in startups, but the technology wasn't

ready because AI didn't exist. And really, you need a keyboard, right? It's hard enough to use

your iPhone typing on that keyboard. The keyboard's gotten better. It does guess your words a little

bit better. But generally speaking, people are not writing long emails on their phones.

It's just starting to be the case that maybe doing voice to text is starting to work. I know I use

voice to text. If I'm driving, I can send a text message or I will have Siri read me text messages

when I'm driving or walking. So all of this put together is, I think now is the time. And so this

sounds like a great idea. This is a very interesting thing. Is it a coincidence that all of this

OpenAI news dropped in the same week? You got the iPhone for AI with MASA and Johnny Ive. That's

really interesting. OpenAI launching multimodal and OpenAI discussing a secondary share. All of that

together, to me, sounds like a massive MASA push. I'm not saying MASA is the leak here, but

he likes to make bets on number one. Uber, WeWork, DoorDash, all of that. And in OpenAI,

Johnny Ive, Colab is like, yeah, that's just like Peak Silicon Valley. And MASA being the backer of

it is insane. I mean, we're talking about like this is building a super team. It sounds like a

super band, right? Like David Lee Roth with Slash, with Robert Plante. Or Chris Bosch and Dwayne

Wade and LeBron taking their talents to Miami. This is an absolute super team. MASA loves a big

swing. He would love to put a billion dollars into this project and have it have a 1% chance or a

5% chance. And you know, hey, listen, MASA just had that big liquidity event with Arms IPO. That's

got to take a little edge off. And maybe MASA is getting frisky again. And I'm here for it. I like

it. I like an exciting dynamic space. To me, this is huge, huge. You know, it's like bringing the

X-Men and the Avengers and the Fantastic Four back together. It's going to be great and great for

technology, great for consumers and great for this podcast where we talk about all these things. So

really interesting moment in time. Well, we'll keep track on all this. And of course, remember

Mondays are this week in AI with Sundeep Madras. So everybody's talking about how

you know, all in is a great episode and they look forward to it and it's event viewing. Everybody's

now saying Mondays on this week in startups are event viewing. So life is more fun with MASA in the

mix. Life is a lot more fun when MASA is in the mix. That is 100% true. You have heard me talk

about my weight loss journey and the role that SuperGut has played in it. SuperGut is just an

amazing product. They were a huge hit at the All-In Summit. I love their bars. I love their

shakes. In fact, here it is. I was just drinking this shake. My wife makes a beautiful shake for

us every morning. I love the chocolate brownie bar and they have unflavored free biotic mix

that you can add to anything. My wife puts it in my coffee. You know, she's trying to keep me

healthy. She wants to keep me around for a little while. Can't blame her. I'm a fun time. There's

a reason why SuperGut is so good for weight loss. Okay. The SuperGut products, you may not know

this. They mimic the effects of ozempic by boosting your GLP1 hormone. And this will help quell

your hunger and it's going to boost your metabolism. That's the combo you're looking for. Products

taste great. Tons of great flavors, strawberry, chocolate, all the stuff you love. But here's

the thing. This is all proven scientifically. The Sultan of Science himself, David Freberg,

is the co-founder of this company. They conducted a placebo-controlled clinical trial with

Stanford. You may have heard of that university last year and that's been published in the medical

journal Diabetes, Obesity and Metabolism. The results, amazing. Participants lost weight just

like me. They lowered their blood sugar and they improved their metabolic health. And also,

I helped with digestion. So it tastes great. It's going to help you lose weight. If you want to

improve your gut health, you want to drop a few pounds or you just want to feel sustained energy

throughout the day. Do what I did. Go to supergut.com and use the promo code TWIST to get 25% off.

All right. The FTC and 17 states in other news have brought an antitrust lawsuit against Amazon.

We knew this was coming for contacts. Lena Khan has had it in for Amazon for a long time. She

don't like Amazon. However, she's been losing a ton. Lena Khan, of course, is the chair of the FTC

and she was put in place by the Biden administration, which hates big tech, as we know,

and Khan is, of course, the youngest FTC chair in the history. She's only 32 years old or she was

only 32 years old when she was elected to the position, I think about two years ago. So she's

probably 34 now. So she's been a longtime critic of Amazon. She hates it. When she was at Yale

school, Khan wrote a very notorious paper, Amazon's antitrust paradox. And basically,

she wants to change the typical benchmark for antitrust. Does this harm the consumer?

And she argued that Amazon should not get a pass on its monopolistic practices because it

delights consumers. Anybody who uses Amazon is delighted. So Lena Khan got a lot of attention

for saying, no, that shouldn't be the test. Consumer harm should not be the test. She

wants to change the law and she wants the law to be, and her view is we should protect consumers

in the future as if she is some clairvoyant and she can predict the future and she can

ensure that there is not future harm, right? This is crazy. It's a crazy concept. I'm not

saying she's crazy. What I'm saying is it's a very bold, audacious way of looking at the world.

I would say it's also a little bit narcissistic that you, Lena Khan, are going to be able to

predict future harm. I know I'm super humble having been in the industry for 25 years and invested

in 400 companies. I can tell you that she cannot predict the future. Nobody can. The second anybody

thinks they can figure out how they're going to define future harm and they think they're a precog

from minority report if you've seen the movie. It's just the height of arrogance. I think we

want to protect consumers, but the idea that you're going to be able to predict that Amazon's

going to be Google or Google's going to be Microsoft or Microsoft is going to be Apple,

none of us know. Now, I like the idea of going after tactics. To be clear, I think that this

lawsuit has got a lot of tactical stuff in it. Even though Lena Khan's philosophy has been

future harm, and I'll just quote her here, we cannot cognize the potential harms to competition

posed by Amazon's dominance if we measure competition primarily through price and output,

she wrote. She thinks instead regulators should look at how companies like Amazon use low prices

to undercut rivals on their way to controlling all this critical infrastructure. Okay, great.

I think that's interesting, but again, it comes with this arrogance that you'll be able to predict

the future, which you will not. Of course, we can make arguments all of us all day long about who's

going to win AR. Is it going to be Google's glasses? Is it going to be Oculus? Is Google going to

release a new version of Google Glass? Is it going to be Microsoft? Who knows? The idea that Amazon

won cloud computing and it wasn't Microsoft or Oracle that won cloud computing or Google is crazy.

Nobody knows who's going to win. Lena Khan's a bit of a loser in terms of her track record here,

and I think she's a loser by design. She said she wanted to be much more bold and lose a bunch of

times in order to win big. She wants to take bigger, bigger swings. She wants to attack tech

companies. She wants to stop them. That's the agenda here. You can debate it all you want,

but she said that. She wants to take bigger swings and get bigger results. She wants more shots on

goal. Just a quick recap, the FTC sued Metta and lost over their acquisition of a VR app maker

within Unlimited. The FTC's argument, Metta having two beat Saber-like products would give them an

unfair advantage in the VR space. The FTC lost. They've got an antitrust lawsuit of Instagram and

WhatsApp acquisitions that occurred decade ago. I mean, people kind of expect the FTC will lose

this. You never know. FTC versus Illumina, they lost, but it's technically ongoing. It's a biotech

company that they blocked the acquisition of. Then the FTC versus Microsoft, they lost this.

Technically, it's ongoing, but the FTC lost this case famously against Microsoft to stop the $69

billion Activision Blizzard. They lost the case against Microsoft to stop the $69

billion Activision Blizzard acquisition. The FTC is appealing all this stuff.

Maybe if you are a fan of breaking up big tech and stopping their ascension,

I guess that it's noble that she's trying. I've said that we should really be looking at the

tactics. Let's take a look at the Amazon suit here and what they're alleging. This is what

they're alleging in the suit, that they're using anti-discounting measures that punishes merchants

for offering lower prices elsewhere. If Amazon sees that, I don't know, some provider of cables

like Anchor, if they see them offering lower price goods elsewhere, they can bury the sellers

far down in search results, which is basically the same as killing them. They're on the platform,

but they don't get seen. They want to stop discounts off Amazon so that Amazon has the

lowest price. By the way, I think a lot of distributors negotiate for this like

Walmart and Target. Then they say they're using the market control to keep prices artificially

high and lock sellers into its platform, which then harms rivals. The way they

claim this is happening is that they are controlling the competitor's prices,

effectively raising them for consumers by controlling the buy box on their site. If a

product is offered for less on another site, Amazon reviews the buy box button, etc. Then

coercing sellers on Amazon by forcing them to use their fulfillment service. I've heard this is

actually true. If these things are true and proven, these could be very easy sanctions.

Amazon will do exactly what Microsoft did, which is cut them off at the pass, basically

change how search works, change how fulfillment works, just make little tweaks. If you remember

Activision, there was a question, would Microsoft allow Activision to be on other platforms and

not have exclusives for Xbox, their platform. They basically said they're going to do that.

So the most charitable interpretation of Lina Khan's success is that by her rattling sabers and

filing these lawsuits, that she's getting tech to behave better. That might actually be a pretty

good legacy. She might lose the cases, so she loses these battles, but maybe she wins the war

because people start on a tactical level, allowing a fairer marketplace. She did do that with the

Activision case. She got Satya Nadella to guarantee for 10 years that Call of Duty would be on other

platforms. In some ways, even though she's racking up the L's, she might be racking up a

lot of little W's. Other things that they're claiming is that Amazon degraded customer experience

by replacing organic search results with paid ads. Well, we all see that when we use the service,

but I don't understand that that's illegal. You could say the same thing about Google. You

do a flight search on Google, a product search on Google, it's all paid ads.

The Amazon, of course, disputes all this, quotes from Amazon's rebuttal. We fundamentally

disagree with the FTC's allegations, which are in many cases wrong or misleading, and with the

overarching and with their overreaching and misguided approach to antitrust, which would harm

consumers, hurt independent businesses, and up and longstanding and well-considered doctrines.

I think, put all this together, there's plenty of rebuttals, and I think this might be a turning

point for Lena Kahn because it's been underwhelming what she's done here, but she could be making

progress on scaring big tech. Just maybe the saber rattling will get people to behave better

because they don't want to be in her crosshairs. On that case, maybe that's her big strategy. I

don't know. She's been invited to come on the All In Pockets. She's certainly invited to come

here on this week in startups, but I don't know that she wants to sit in for a hard discussion

with hard questions. I think she'd probably like to do interviews only with anti-tech journalists,

which there are many for her to choose from. My prediction right now is that Lena Kahn loses

the Amazon antitrust case, and I think she's going to lose it. Badly, and I think Amazon

is going to fight, and Amazon will just tweak their model. I think this era of antitrust

is going to fail in terms of the lawsuits, but I think she'll win in terms of getting people

to behave better tactically. Mixbag for Lena Kahn. Next up on the program, roam around. Stick with us.

When I see a startup that's got bad design, it's hard for me to take it seriously in today's day

and age. Why? Because having great design, having world-class design, it's kind of table stakes

today with consumers, with employees, with partners, and of course with investors. But design

agencies, they're very, very expensive, and there's a lot of back and forth to figure out what you

want. It's not for startups. You can't spend a quarter million dollars, a half million dollars,

getting some design agency to do your look and feel. And it's hard for you to hire somebody

full-time because that's a full-time staff position. So let me tell you about an amazing

service. It's called Catalog, a full-service design studio tailor-made for startups. And Catalog does

all of your branding. They do UI and UX. They'll do landing pages, pitch decks. They'll do it all.

It's half the price of a typical design agency, but they offer a full team, including a dedicated

project manager and a design director. They've worked with hundreds of startups. So if you have

a problem, chances are they've already solved it. You can stop and start back at any time it's a

monthly subscription. And Catalog will never waste your time. They do three-day turnaround, okay?

That's more than enough. Stop wasting time. Stop wasting money with all these expensive design firms

and unreliable freelancers who disappear. What you want is fast three-day turnarounds for a

flat monthly fee with Catalog. And you're going to get $1,200 off right now at tricatalog.com

slash twist. That's $1,200 off at tricatalog.com slash twist.

All right, everybody. I am in New York City for the SCIF conference. If you don't know what the

SCIF conference is, it's like the number one travel conference in the world. You can go to skift.com

and I got to meet with all of the executives in that space. We had a really thoughtful discussion.

I spoke at it and a lot of the discussion was on how AI, language models, etc., are going to

impact the travel industry. My big observation was I think society is going to become more wealthy

because people are going to work less three, four-day work weeks because AI is going to do a lot of

the heavy lifting, which means there'll be more time to travel. That being said, the travel industry

is going to change dramatically because the model that we currently use of using Google and

making documents and bookmarks and maybe using a bunch of different Google Flights, Kayak, Expedia,

etc., to look at grids and sort through our choices is going to change forever. How is it

going to change? I believe it will change because of startups like Roamaround. Roamaround attended

our launch accelerator in April of this year, 2023, and they were part of launch accelerator's 28th

class. They have a very simple idea, which is itineraries for travel can be done by AI

really well. The product proves that out with me today is Shai Gubai, the CEO of Roamaround.

Welcome to the program, Shai. Thanks for having me. How appropriate during Skip Week.

Yeah. Did you go to the conference this week or are you know of Skip, I guess?

Yeah, I know of Skip pretty well. I mean, they're obviously mammoths in the industry.

They've also interviewed us back in March. I've been in close contact with Justin and obviously

I've been following Rafat and his journey for months and years. Rafat worked for me back in

the day at Silicon Valley in the 90s and I'm a small investor in Skip, which is just a great

data source if you're in the travel industry. So I thought I'd have you on and then

just maybe show your product and talk about the progress and then

you know, there are some big issues here about who's going to win

travel and how will travel change and then why should you even exist if there's these

incredible language models everywhere and just the journey of doing a startup in the AI space.

So maybe we just start with a quick demo. If you're watching this, great. If you're not watching it,

we'll do sports casting and explain what happens and we'll describe what's happening on the screen.

You can always use Spotify, which has our videos. If you are listening to this week in

startups, there's a button or you can turn the screen and see the video or you can go to YouTube

and just type in this week in startups and you'll find our channel. Hit the subscribe button,

hit the bell, you get an alert every day or two when we either go live or we have a new show posted.

So go ahead, Chey. There's your app and it says roamaround.io is the URL and I just typed in

two days in Barcelona. So you typed into this chat interface two days in Barcelona and then

describe what we say. So rather than spending hours or weeks playing a trip, you just type in two

days in Barcelona, we create a complete and actionable itinerary. And what I mean by actionable

is that you could actually book it right then and there. If it's not perfect, you can iterate it.

So you can say it's our anniversary and we'll recommend romantic beach walks, hot air balloons,

et cetera. From there, you could also say something like out of beach day and we'll

recommend day three in Costa Brava. And then once it's perfect, you can say translated to Spanish.

It'll translate it and then you can quickly share it via SMS, email, WhatsApp, whatever.

Great. And so these itineraries and the AI maybe talk a little bit about what is the difference

between using chat GPT or Bard to do this and using roamaround? What's the best pitch for why

people will use roamaround versus Google's Bard or just the native chat GPT interface?

Totally. So there's a few things that play here. Number one is if you were to go to,

let's say Bard or chat GPT and say, I'm going to Costa Brava, it would say while you're there,

you should definitely go for a bike ride. Well, now you have to actually start researching what's

a good place to rent bikes. Are they going to scam me? Are they highly rated? Are they in stock for

the day that I'm actually traveling? What we do is we do that for you where before we show you

something, we've already queried all our API providers and ensured that they're in stock

highly rated. And within one clicks, you no longer have to start researching. Within one click,

we take you to a provider that's highly rated and you can book it right then and there.

So that's number one. The second thing is when it comes to data cutoff and hallucinations. So

the way we actually work is when you say two days in Barcelona, what we first do is we actually

start querying all of our points of interest providers, trip advisor, get your guide, etc.

We get a wish list of the top 100 things to do in that city. And then and only then do we query

chat GPT and say, of these 100 things, create a compelling itinerary for me. And so what that

does is two things. Number one is it solves the hallucination issue. So a funny story is before

we implemented this little fix, if you were to say I'm going to Tucson, Arizona, it would say

while you're there, you should check out the Pearl Harbor Museum. Yeah, the reason it hallucinated

that is because the ship was called the USS Arizona, right? So because we limit it to the

things that we know are in that city and in stock, we solve the hallucination issue. And then the

other thing is is we solve the data cutoff issue. So we're not going to query restaurants that have

been closed for two years. Since September 2021, like chat GPT, we query, let's say open table,

we say what's in stock or what's available for reservation on these dates, and only then do

we include it in itineraries. So when you're doing verticalized AI, whether it's travel or it was

medicine or it's legal, there is a lot of nuance to what data you're going to query and then what

you're going to advise and what you're going to send back to the user. We saw this week, multimodal

come out from open AI, which basically means you could take a picture of something and then get

text back or vice versa. Obviously, we use text right now, but we don't get images back. And then

Bard, Google's product now includes Google flights, for example, and your Gmail, etc. So maybe you

could talk a little bit about what you think those two advancements will bring for roam around.

Totally. And then another one actually is the meta AI glasses, right? So combining meta AI glasses

plus the multimodal, right? You can already see a walking tour taking place here, right? Taking

shape, right? So you're walking through... Explain how that might work. We're in Barcelona,

what happens? Yeah, so you're in Barcelona, you're wearing the glasses, you see Sagrada Familia

and you're like, okay, this is an incredible building. What's going on here? You take the images,

you feed that to ChatGPT, you feed that to roam around, you say, you are this all-knowing

travel advisor. What am I looking at here and what's the context here? And so it'll tell you

it's this famous architect in Spain. It's been in construction for 100 years and for another

200 years from now. It could tell you all of these things. And then going to Google Bard,

I think the most interesting thing is Gmail and Calendar. So one thing right now we're in the

discovery phase, right? You have to prompt roam around to say, hey, I'm going to Barcelona. It's

more interesting that it could say something along the lines of, hey, I know you and your wife

have been meaning to go to Barcelona. I see that you have an open week in July. I have all of these

places that you've already bookmarked on Instagram. This is an itinerary. Do you want me to book for

you? And you could even say, hey, this is a great weekend to go because I see your calendars open and

travel costs are down. And it's before the peak season. So the costs are going to be lower.

And this hotel's got this great thing. So it's almost like having not just a travel agent,

but a travel agent plus your personal admin, if you had an executive assistant,

acting in concert to optimize. And then it could even learn over time what your preferences are.

Hey, you like to walk because it has information on how many steps you did, or you told it you

like to walk, or it knows you like to do active things. So it says, hey, you can go kitesurfing

in this area on the coast of Spain. So this has got unlimited potential and it feels quite disruptive

to me. Totally. And we're already laying the groundwork for that. So in the example that I gave you,

it says it's our anniversary, right? So now we know when your anniversary is. So next year,

we can say, hey, we saw that you booked this on your anniversary. Do you want to maybe check out

this on your next year's anniversary? Or if you were to iterate the itinerary and say,

recommend vegetarian restaurants, okay, well, now we know that you're a vegetarian, right?

And so like what you're saying is, is correct, is this all knowing travel agent that's a personal

admin for you. And it sounds kind of like far fetch, but think about, you know, 50 years ago,

who had chauffeurs, right? It was the ultra wealthy, the powerful businessman, right? And now, everyone

with the click of a button have a personal driver, right? So that was the promise of Uber, yeah.

Exactly. And you don't have to tell the driver where you're going. You don't have to reach into

your pocket. It's a personal chauffeur. And so the idea is just as Uber brought chauffeurs to the masses,

we're hoping to put a travel agent in everyone's pocket. And it's even more than a travel agent

because this is, like I said before, they understand some things about you that maybe even a travel

agent wouldn't be able to query. And you could even, I think, when I was speaking at this travel

conference, talking about how I think experiences are such a big part of travel today and that,

you know, explaining how you want to feel after your vacation. Do you want to feel inspired?

Do you want to feel like you learn something? Do you want to feel exhilarated? You know, for me,

I want to learn some new, I learned how to play pickleball and I learned how to efoil on my last

vacation. That was a lot of fun for me. You know, I got to experience deep powder skiing in Japan

last year. All of these things are peak experiences. So I'm curious if you've given thought to

hey, I have this amount of money. I've got $5,000. I've got 10 days. I want to feel

renewed, refreshed, you know, equanimity and exhilaration from some activities.

Now explain to me what I should do. Have you thought about that working backwards from an

emotional state and with other criteria? Yeah, I mean, what you're describing is something along

the lines of like in our roadmap, it would closely align to like the personalities. So

we did it in a way that we don't want any friction before we show you the itinerary, right? Right

now, we're still proving ourselves. So a lot of these other, let's say, travel planners,

before you actually jump into the itinerary, they're asking you all of these questions. And what you

see is that people are bouncing, people are falling out of the funnel, right? Because you're

adding too much friction. So for us, it's like, we're two. And then we show, we prove ourselves

and say, this is a really great itinerary. And then from there, you can iterate as much as you

want. You can say, I'm traveling with the family, I want to go shopping, or I want to feel exhilarated,

I want to do extreme sports. But first, we want to prove ourselves and then earn that right to

actually delight you. And so the initial business model here is you get an affiliate fee, there's

all these affiliates in travel, travel is right next to e-commerce as the number one and two

affiliate opportunities. Every time somebody books a flight, you can make a couple of bucks,

books a hotel, you can make tens of dollars per night, books and activity, you can make single

digit dollars, I think. So that's the affiliate. Oh, tell me. So our bread and butter is actually

activities. Hotels, you do earn money on flights, you earn almost nothing,

nothing. But on activities, right now, we're making a percent on each activity. Our median

activity is 265 bucks. So, yeah. So, but we're actually moving towards a model that will give us

20%. But you would have to have a direct relationship because the people who provide

the affiliates, those activity marketplaces take 25%, take rate typically, you get eight

points of that 25 for bringing somebody. If you did it yourself and you went direct,

you had the kite boarding relationship directly, you can then ask them for 20%, which is 20%

less than the 25% that the majors take. Exactly. So, that's stage one. Stage one is the 8%,

stage one and a half is the 20%, right? Stage two is what we're actually doing right now is

we're shifting into the enterprise. And we can talk about that in a second, but stage three is

actually AI model as a model, right? So, or as a service. And so, the idea is right now we have

millions of interactions, millions of itineraries, millions of clicks, thousands of conversions.

And so, we're able to train our model and say, this is a good itinerary, this is a bad itinerary.

And so, we're going to be fine tuning our model such that in the future, travel startup isn't

going to be using, let's say GPT-4, they'll be using GPT-4 hyphen Romaround, right? And we'll be

selling them our API keys. Got it. And I'm curious, knowing what you know, and we'll get into the

B2B side of this in a minute, how are the different platforms that you can use shaping up? Obviously,

Facebook has their open source one. If you want hugging face, there's a ton and there's Falcon

from the UAE and there's Bard and there's going to be a lot of providers here. Is Chat GPT-4

just far and away the best one to work with? Or are you starting to see parity across the

different services that are available? And what's your take if we're sitting here two, three, four

years from now? Will these language models be generally commodified? And just like you could

store something on Google Cloud, you can easily store it on Rackspace or, you know, any other

cloud provider? Yeah, so just to be clear, we never intend to create our own model, right? We'll

be forking or fine tuning one of the larger players, but we don't want to be competing with

Sam Altman, right? That's not our game. In regards to what's the strongest model, it probably changed

since the start of this conversation, like three or four times. But when we started and like the

multiple times that we've tested, GPT-4 is the strongest model for our use case. Not just because

it's the one that is strongest in terms of output, but also in terms of like the infrastructure,

right? They were the first mover and therefore you start seeing all these plugins, right? So in

order to do what we discussed earlier of this like proactive booking AI agent, you need the

infrastructure, you need these APIs to be tying in together. That isn't to say that that meta and

Google aren't going to have that, of course they will. And they, they have the advantage of also

like the user profiles. And so once they actually start releasing that, we'll be really interested

because that's where we could really create this bespoke itineraries at scale that we're

really interested in. But in our first tests, GPT was by far the best.

Everybody knows technology is moving at a blistering pace, but Codecademy's goal is to help

you stay a step ahead. So whether you have some experience or you're brand new to coding, Codecademy

offers everything you need to build the tech career you want. Gain specialized skills in

artificial intelligence, full stack engineering, data science, cybersecurity, and more. Right now,

twist listeners can try Codecademy Pro free for 14 days. That's two weeks to check out

everything Codecademy has to offer, like more than 400 courses covering in-demand languages like

Python and JavaScript, and new courses in generative AI, large language models, LLMs, and

more. You can also get step-by-step guidance on career paths, if you're looking to switch jobs,

real world projects to build your portfolio, and individual prep to help you stand out in a

competitive job environment. Those interviews are getting tougher and tougher. And once you get the

job, you can work towards your next promotion with guided skill paths. So start your free trial

today. Visit Codecademy.com to try Codecademy Pro free for 14 days. That's C-O-D-E-C-A-D-E-M-Y.com

slash twist. Codecademy.com slash twist. I met with all these business-to-business companies and I

buy a show of hands in this room full of 500 people. I said, how many people use chat GPT

in the last week? And like two out of three hands raised. It was pretty amazing. Then I asked how

many people use Bards connection to Google Flights, which came out last week, and two people raised

their hands. So it is the case that I think the travel companies, you know, they're very busy

running their businesses. It's a very challenged environment. Maybe there's a recession coming.

Who knows? The consumers don't have money right now, or they're starting to run out. And the

revenge travel of the last three years was kind of over post COVID. They need help. So on a business

to business side, what's your strategy here? Because hotels and Airbnb providers and activity

providers, they don't have developers sitting around. They don't have startup energy. They may

not have the cycles to do AI in travel. So how do you work with them? And what's the business-to-

business plan here for white labeling? What roam around does? Yeah. So it's actually interesting

that you bring up Google because the travel industry as a whole has a love hate relationship

with Google, right? More hate than love? You'll have to ask, and I'm sure it depends,

and I'll try to stay politically correct. So I can actually like, I would love to share just

one quick slide here, right? Because I think it's pretty telling. So an interesting thing

happened, right? We were minding our own business and joining the consumer side of things, and hotels

and airlines actually started reaching out to us, DMOs as well. And we're like, why is this the case?

That was that last piece? You said DMOs? Yeah. So it's like destination marketing organization.

So it's like visit Miami, right? So it's like the city. The economic city travel group. It's

economic council, tourism, all those different folks. Yeah, come to Barcelona group. Yeah.

Exactly. So they're reaching out to us and we're trying to figure out why that's the case. And so

it turns out that it's because the OTAs have commoditized airlines to the point where airline

brands or hotel brands are just a footnote on the user's journey. So consider how you currently

interact with your airline. They say, okay, here's your boarding pass. Then you go enjoy your entire

vacation. And they say, okay, here's your baggage, right? And so going back to Google, it's pretty

telling that Google gives more real estate to emissions than to brand. Think about like the

dynamics at play here, right? So here is North Atlantic Airways. Their name has been truncated

and Google has highlighted in green 11% emissions. So I think that that shows you what the dynamics

are here, right? Like Google spends more energy, for lack of better term, virtue signaling than

they do for highlighting the brands, right? Yeah, they don't care about the brands. They care about

the click and the money they will make from the click, right? That is their mission.

And I guess arguably their mission is to satisfy the user. Of course, if the user gets a perfect

answer, then they don't click around, which is the great paradox of Google's business model.

If they give you an answer, then maybe you don't click on an ad. And so there's a natural tension

there. Right. And I think it's like, it's even more surprising since like, you know, you're an

eco-friendly person, right? Can you really contextualize what 431 kilograms? Yeah, like

literally going through the amount of CO2 an airline is using is the most silly thing for

consumers to contend with. Like they can have no impact on this. This is ridiculous, right? Like

literally no person in the world is looking at the amount of CO2 that their airplane takes

and then making the choice based on that. It's peak virtue signaling nonsense. I mean,

it's nice to have, don't get me wrong, but that would be the equivalent of telling me like my

hamburger CO2 emissions and then I'm not going to eat the hamburger at Shake Shack. Like there's

other ways for me to make a better contribution to society. And the point isn't that they're

showing this information that's useful to some people, I guess, but they're doing it in lieu

of the brands. And so that shows you maybe why there's this tension between airlines and

and OTAs such as Google, right? So essentially what like this, it comes as no surprise that

airline loyalty keeps degrading. This is actually from Skift maybe two or three days ago.

And so what we're doing to combat this is the same product that is the roam around product

that's been enjoyed by millions of users. We're actually white labeling it for all hotels and

as well airlines. So this is going to be a home run. We had an investment in this space that

didn't work out, but it was using humans to be an outsource concierge and obviously the humans

were a big part of the expense and part of the reason it didn't work out training and finding

humans. But this is an incredible one because if the hotel's in Brooklyn, you just need to know

that area if it's in Arizona. This is so brilliant. Great job. Exactly. And so it's like imagine you're

saying like the airline is not just telling you like your baggage is here. It's also telling you,

okay, your reservation at this restaurant is now. What's the dress code? This is the dress code.

Okay. And don't forget your tour is tomorrow. So you're interacting with the airlines brand

throughout your trip. It's no longer just a footnote. It's no longer just a pipe.

You're actually interacting and reinforcing the brand throughout the trip. And so this is what

it looks like. It's the exact same roam around but without our brand. Looks great. Yeah. Great

job. So for anybody who's running a hotel, if you want to do this, how can they email you and

get this for their brand or anybody who owns it? Yeah. So enterprise.roamaround.io. It's

24 seven concierge, your brand, your fonts, your logos, your voice, and most importantly,

no code, right? Give them your, give them your email. What's it with your email so that they can

email and you can close the day at roamaround.io. S-H-I-E at roamaround.io. Yeah. Listen,

I'm an investor in the company. I'm trying to get you a couple of customers here. I want to move

any friction from it. What a great idea. I think, you know, the, you know, the both of the businesses

are interesting. The B2B business is fascinating because you don't have to bring the customers.

They have the customers already and you can get the value to the customers and they're not going

to build it anyway. It gives them a competitive advantage and it would fall into the win, win,

win. The customer wins, the hotel wins and you win. So what a great idea. And it gives them an

advantage because direct booking is amazing for these hotels. I was, I had lunch with a very,

very, very, very, very famous hotel magnet and they were just talking about their relationship

with hotels.com, Expedia, et cetera, and what percentage they take versus direct and all of

these hotels are trying to get people to book direct because they don't have to give that 20%

to whatever they're forking over to these travel agencies and they're just trying to get everybody

in the habit of going direct. And this would increase the chances of people going direct

because you get this great value. Yeah. So if you're ever an end user and your hotel is helping

you plan your entire vacation, is it constantly being reinforced throughout your trip? The next

time you're planning a trip, you're most likely going to be going back to that hotel. That's

providing you so much value. That's the idea. Fantastic. Awesome. Well, listen, it's so great

to be in business with you. Congratulations. You did really well in the accelerator. Everybody

loved your idea. I love your idea. We've invested, I think, twice in the company and you've got some

other investments. So stay focused and really delight both of your customers, that enterprise

customer. You have to delight them and you got to ultimately just absolutely delight, save time,

save money and delight those end customers and you're well on your way to doing that.

The reason we loved investing in your company was we saw you have product velocity. You really

have been making roam around better and better every time we see the product. So congratulations.

Everybody check out roamaround.io or enterprise.roamaround.io depending on which bucket you fall

into. And we'll see you all next time on This Week in Startups. Bye-bye.

Machine-generated transcript that may contain inaccuracies.