Latest / AI for Founders with Ryan Estes / Can this AI replace Amazon for smart shoppers?

Transcript

Can this AI replace Amazon for smart shoppers? · AI for Founders with Ryan Estes

0:00
Welcome back to AI for Founders. I'm your host Ryan Estes and
0:03
today's episode is packed with insight, energy and some serious
0:07
fresh takes on the future of e-commerce.
0:09
Not only that, but we talk classic Buffalo Bills.
0:12
I love it. I'm joined today by brilliant
0:14
founder of Tanita, the next Gen. AI shopping platform that feels
0:19
more like a conversation with your stylish, all knowing best
0:22
friend than a typical online store.
0:24
Tanita isn't just a search bar. It's shopping assistant that
0:27
guides you through every decision.
0:28
It teaches you what to look for and curates products in real
0:31
time, whether you're hunting for the perfect black sweatpants or
0:35
Patagonia alternatives for fly fishing.
0:38
With a background at Google and a deep obsession with fixing the
0:41
broken UX of online shopping, Seva's building something bold,
0:45
elegant, and deceptively simple. We talk AI, user behavior,
0:49
monetization strategy, raising 5 million, and even fake Versace.
0:54
So buckle up, this one's a gem. But first, some business.
0:57
Let's face it, in today's world, attention is currency.
1:00
If you're a founder, there's no better way to get attention than
1:03
by sharing your story on top podcasts and your niche.
1:06
That's where Kick Caster comes in.
1:08
Kick Caster is the industry leader in podcast guest booking.
1:11
We've helped thousands of founders land conversations on
1:14
the world's most influential podcast.
1:16
Whether you're looking to drive traffic, raise funding, or build
1:20
undeniable authority, we got you at Kick Caster.
1:23
It's not just about interviews, it's about relationships and
1:26
distribution and building a content flywheel that fuels your
1:30
entire brand. So stop hoping people discover
1:33
your product. Go tell them the story.
1:35
Visit kickcaster.com to apply today and let's get you booked.
1:40
Kick Caster podcast guesting done right, That's Kick Caster
1:43
by MO Burst. OK, everyone loves to talk about
1:47
vision, the pitch, the funding round.
1:49
But you know what gets 0 love? The actual product build, that
1:53
magical, terrifying phase where nothing works, the docs are out
1:56
of date, and you've asked your dev team how's it going?
1:59
For the 47th time today on Slack, That's where Code Story
2:03
comes in. It's a podcast for startup
2:05
founders who know that building The thing is the thing.
2:08
Hosted by Noah Labheart, Code Story goes deep with tech
2:11
leaders who've been through it, shipping real products,
2:14
surviving messy launches, and figuring out what actually works
2:18
and when. You're on the edge of chaos.
2:20
If you're building now or ever plan to, you'll want these
2:23
stories in your ear balls. Search code story wherever you
2:27
get your podcasts or visit codestory.co.
2:30
Because founders talk vision, but builders builders tell code
2:34
stories. AI live tonight Shining stars
2:40
Computer life AII take the flight in the glow of neon
2:48
bright Siva. Thank you so much for joining me
2:51
for the interview. Thanks for thanks so much for
2:53
having me, Ryan. Wonderful to be here.
2:56
Wonderful to meet you as well. OK, let's just start right into
2:59
the product and we'll expand from there.
3:01
So briefly describe what your product does to me like I'm five
3:06
years old. OK, Tonita is a topping search
3:11
website, so you can go to tonito.co and search.
3:15
At the moment we sell clothing and apparel and accessories, so
3:19
I don't want to assume that you're a 5 year old.
3:21
But you can go search for, you know, men's polo charts, or you
3:25
can search for sunglasses or you can search for a nice dress for
3:28
a cocktail party or anything that you want.
3:31
And the AI will guide you through how to pick one of those
3:35
things, how to what are the decisions?
3:37
What are the things you need to think about when buying a pair
3:40
of sunglasses or a sweater or a cocktail dress?
3:44
It'll help you make informed, good choices.
3:47
It'll show you products. It's not just going to talk to
3:48
you all the time, it's going to show you products.
3:51
It's a nice, interactive way to learn about what you're shopping
3:54
for and make good, informed decisions.
3:58
Amazing. This is so cool.
4:00
And I put in I'm at tonita.com or.co right now and and I'm
4:05
putting in what I am in the market for shopping for which is
4:09
always black sweatpants. OK.
4:13
So I entered in there, it's got preferred gender, preferred fit,
4:16
specific material it's going through and kind of just taking
4:20
kind of my preferences and then just link to to different
4:24
vendors on the backside. It links to different vendors.
4:26
You can click on one of the products, you can see the
4:28
details, or you can actually go to the store and buy one of the
4:32
products as well. You could do both.
4:34
You could ask Tonita to educate you.
4:37
You know, if you didn't know anything about black sweats,
4:39
sweatpants, which probably doesn't apply to you, but if you
4:41
didn't, you can always ask. Tell me, how do I pick a good
4:45
pair of black swim pants? You can, you can take that and
4:48
this is brilliant. And yeah, this is kind of like a
4:52
interesting area of AI I'm really interested in, which is
4:56
basically it's it's redefining the UIUX of absolutely
4:59
everything. One common pain point that I
5:02
have and probably people share this is everything has its own
5:05
app and then you have to go figure out how every app works.
5:09
Or especially if we're talking about e-commerce, like, OK,
5:12
where's my shopping cart? How do I filter in every single
5:15
different one? Whereas this has given me
5:17
exactly what I want. It's like, you know, maybe I
5:20
have a preference of brand, I can enter it, but really what
5:22
I'm looking for black sweatpants give me everything all at once
5:26
and then allow me to continue through one specific interface.
5:30
Exactly. And the interface couldn't be
5:32
simpler. And that's the most important
5:34
point, right? You could search with keyword,
5:36
you could actually talk to it. You could say things like I'm in
5:40
the market for a pair of sunglasses, what should I look
5:43
for, right? Some of us don't even know what
5:45
to look for. You know, many people wear
5:47
sunglasses regularly, routinely, they know what to look for.
5:50
I'm not one of those people. I have a D key on pad from 20
5:53
years ago that I used for driving occasionally.
5:55
And recently I broke. So I went to buy sunglasses on
5:59
Tonita. And because I wasn't sure, I
6:01
said educate me. And Tonita told me here are the
6:04
things to look for. You know, you can look for the
6:06
shape of the frame. That's one of the important
6:08
considerations. You can look at the color, the
6:11
lens type, you can look at other features.
6:14
Does it have UV protection, for example?
6:17
These are some things that I didn't even think to ask.
6:19
And that's where I think shopping site Victoria makes a
6:23
lot of sense because often times it's really overwhelming when
6:28
you're shopping for something that you don't know too well
6:30
about. And there is just so many
6:32
choices. And how do you make a mental map
6:35
of what all the options are? How do you navigate them?
6:39
How do you browse through them? How do you ask questions?
6:41
In the middle, you can say, why do I need UV protection?
6:44
And then, you know, the reader will tell you, the AI will tell
6:47
you, you know, you might be going on a hike, long hike on a
6:51
sunny day. And you know, if you're in
6:53
California, for example, there are some UV rays that that hurt
6:57
you. So it's like talking to a
7:00
superhuman concierge who knows all the answers to all the
7:06
questions, like any good it would.
7:08
And it's backed by a really large catalog of products.
7:12
So you're actually not just talking in that strike, you're
7:15
actually looking at products as well.
7:18
And that's huge too, because a lot of times if it's not in like
7:24
if it's not a Costco and it's not on Amazon, if I strike out
7:27
there, I don't know where to go shopping.
7:29
So this is given some good options.
7:32
Now the, the, the genre of AI assisted e-commerce is
7:36
absolutely wide open and there's so many places you can go.
7:40
What I love is this simple sophistication of what your tool
7:43
does. Now, just off the top of my
7:46
head, I can think of AI being predictive of what I'm going to
7:49
need. It can say like, OK, we're we
7:50
got to start thinking about the fall.
7:52
Here's some scarves based on what you search for, what you
7:56
purchased in the past. Here's some things you might
7:57
like. There's kind of the predictive
7:59
nature of AI you could use. There's also the generative fact
8:03
of like I could put in a picture and it could show me where I
8:06
could buy those elements or like just weave that into like social
8:10
media and things like that. So as you're thinking about kind
8:13
of the product road map for Tanita, what kind of additions
8:17
are you planning on bringing to the platform?
8:19
All of the above, all of the above.
8:21
Maybe you should come be a product manager with us.
8:23
I think there's a lot of good ideas there.
8:25
And, and you know, we've been thinking about this long and
8:27
hard and every one of those is actually super, super critical,
8:31
right? The proactive recommendations of
8:34
products based on things that you viewed, things that you said
8:38
in the past, things that you said you liked.
8:40
For example, you might have said that, hey, I'd like to shop for
8:43
sustainable clothing and that's something that don't need, I can
8:46
remember and show you over a long period of time, the right
8:50
kind of products that match that description.
8:52
Or you can shop for certain kind of shirts, let's say, and
8:58
similar ones or related ones come on a sale.
9:01
And that's a good time to let you know that you know, these
9:03
are these are now available. So and you'll be the right
9:07
person to to show that to. So the predictive power is
9:11
absolutely critical. Anticipatory things are super
9:14
important and simply simply allowing the user to to express
9:18
themselves freely I think I said is a major part as well.
9:23
Agreed. Very cool.
9:25
What's the business model? How are you getting paid?
9:27
Not at the moment. So we are, we will be launching
9:31
paid recommendations, for example, the kind we just talked
9:34
about because these are highly, super targeted to your taste
9:40
based on what you search for, what you browsed in Toledo.
9:43
So paid, paid Promort insurance is the primary way of
9:47
monetization there we're working on.
9:51
Why not just take a Commission from the retailers?
9:53
That seems like a pretty way easy way to monitor.
9:56
It's a reasonable way to do as well.
9:57
And there are some standard vehicles like affiliate linking
10:02
that would work as well. And we want this to be somewhat
10:04
clear to users when we are working directly on behalf of
10:09
the user based on what they've just said and when we are
10:12
showing something that we think based on AI that they might
10:16
like. And this is this is a line that
10:18
we're trying, trying to be very clear about, you know, show that
10:22
this is a recommendation. It's a paid recommendation based
10:25
on what you've been browsing versus here's a mismatch,
10:29
mismatch of results. You know, some of these might be
10:32
paid links, some of these might be good results for what you've
10:36
searched. And we don't like to do that
10:39
mixing without clarity for the user, but that's a that's a
10:44
reasonable model as well. And you know, we might do some
10:46
of that down the line as well. You bet.
10:50
I just put in heavyweight black T-shirt because I weigh those,
10:53
wear those seven times a day. And some great recommendations
10:57
too, because it's hard to find a heavy duty heavyweight black tea
11:01
that isn't too boxy, but there's a great one here from
11:04
abbreviated Apparel. Amazing.
11:06
So so right now you're kind of pre Rev. So a couple of
11:10
questions. You know, what's a benchmark
11:13
that you want to begin? Are you looking for numbers of
11:16
users? Is that a important KPI for you?
11:19
And then secondly, how are you financing Tonita now pre Rev?
11:24
Yeah, users, the number of users, the quality of their
11:27
engagement, I think these are the two biggest things that
11:30
we're focusing on. And that, that is what's going
11:32
to be our focus for the next quarter or two at least because
11:35
that's where we set the foundation.
11:38
And as an, a, a shopping engine, you know, as you, as you know,
11:43
the use of a, A in search is still in it's infancy, right.
11:46
But it's, it is just a long way to go.
11:48
So we're still figuring out what resonates with users, what
11:52
people really like, how people like to shop.
11:54
And it differs so much from user to user.
11:58
And even for the same user, it changes a lot depending on how
12:01
much they know about something versus how little they know
12:03
about something. So we're figuring out what are
12:05
the right ways to help users understand and browse and shop
12:10
online. And because of this, our first
12:13
goal really is to grow the user base and get good usage of the
12:18
product and, and really understand how users use the
12:21
product and you know, produce something that delights users.
12:25
And once we set the foundation, we go for monetization and and
12:30
other other type metrics. 100% how have you guys financed so?
12:35
Far we are. We are a venture funded startup.
12:39
We we are funded by 3B, CS and a bunch of angels as well.
12:43
Amazing. How much have you guys raised so
12:45
far? We have raised over $5,000,000
12:48
and we've been, we've been building busily and we've
12:51
probably on going to be raising soon later this year.
12:56
Fantastic. That's that's exciting.
12:58
How big is the team? Team is very small, still under
13:00
10 engineers, 1010 employees and all of us are engineers to
13:03
actively building. We yeah, we do have some design
13:07
collaborations and others as well to get the product up and
13:10
moving. Very cool Siva, what brought you
13:13
to this project? What what is your history as far
13:17
as like developing developing tools and and products?
13:20
Yeah, I've never built a product before.
13:23
I have been at Google for 15 years where I worked in the
13:27
research organization, primarily working on algorithms, machine
13:31
learning and artificial intelligence.
13:33
So mostly working on the technology end of it, not so
13:36
much on the product end of it. So this is a new experience for
13:39
me, but what really brought me to this is the confluence of
13:44
maybe two things, right? One is the growth in AI,
13:48
specifically language and vision AI, right?
13:51
Because we've seen in the last 15 years, in fact, well before,
13:56
even well before that GPD came on the scene, we, we've known
13:59
that language understanding is getting better and better, but
14:02
vision is computer vision is getting better and better.
14:04
If you upload a picture of a sweater to to Donita, for
14:08
example, we can tell you exactly what kind of sweatshirt is, what
14:10
kind of pattern it is, etcetera, etcetera.
14:12
But the technology pillars have been getting much better.
14:16
That's one of the pillars. The other pillars, I say shopper
14:21
myself. You know, I buy things all the
14:23
time and it's been no end of frustration for me because to
14:28
buy a $20 thing, I often open 20 tabs and I'm clicking through
14:32
each one of them. And as I do each one of these
14:34
things, no, I totally wait. As I'm doing each one of these
14:37
things, I'm learning some phrase or some concept that I didn't
14:40
know before. And I open one more tab because
14:42
I need to Google what that is and learn about it and then go
14:45
back to the tab and say, you know, this is what it is.
14:48
You know, I'll give you an example.
14:50
I was looking for a kitchen knife set once a few years ago,
14:54
just before Foundingtonita. And very soon, as I said, I had
14:58
some 20-30 tabs and I'm looking at multiple sites, you know, all
15:02
over the place, different interfaces, different ways of
15:05
searching in each one of them. And somewhere I encountered a
15:08
phrase that I didn't know before.
15:10
There's something called a santoku knife.
15:12
And I didn't know that thing existed.
15:14
So now we need to know what it is.
15:15
And once we knew what it is, I needed one of those things, but
15:18
I needed the ability to say, I need a kitchen knife set that
15:21
includes A santoku knife and also has a tomato knife because
15:25
we use a lot of tomatoes. So, and there's just no way to
15:29
say these things any, any website, including Google or
15:32
Amazon or what have you. And this is that was the other
15:35
motivation. So on the one hand of a
15:36
technologist working in search and a A for many decades now, I
15:41
could, I could see where the trajectory of AI and search and
15:45
shopping was going. And as a user, it is a very
15:48
personal frustration for me that I'm not able to shop the way I
15:50
think about things. I wanted to bring them together
15:53
and that's how Toyota came to be.
15:56
Amazing. I love it especially and it's
15:58
such a bold and brave move if you're not used to like doing
16:02
products to dive right into it. How did you kind of secure
16:06
meetings for raising that round and how did you reach out to
16:09
your network to kind of facilitate and close the close
16:12
the? Round, yeah, yeah.
16:13
Reached out quite quite broadly to a lot of friends across the
16:17
industry saying, hey, you know, you know me, I'm starting this
16:20
thing and connect me to the right set of folks, both from
16:26
immediate circles and friends of friends and, and a lot of them
16:29
and one of them connected me with South Park Commons, where
16:34
they spent about 1/4 as a as on their on their program.
16:38
It's a, it's a, it's, it's what they call a -, 1 to 0
16:43
experience, you know, founders who know that they want to start
16:45
something, but they don't even know exactly what they want to
16:47
start. And that was a great experience
16:49
for me at South Park Commons. And that again, brought me some
16:53
other connections as well. And South Park Commons also
16:56
ended up being our lead investor for the state round.
17:00
Even your your expertise, when you were pitching, did anybody
17:05
say like, hey Seva, you don't want to start a brand like come
17:08
work for me, come work on my projects.
17:11
Happens all the time happens, you know, before I started after
17:16
I started. And to this day, I keep getting
17:17
those calls very frequently. And you know, some of them are
17:21
very tempting, of course, but but you know, I, I really think
17:25
we have, we have something unique here.
17:27
And I really want to see this fine.
17:31
Amazing. OK, I just threw Tonita A
17:33
curveball here. OK.
17:35
I said I want a Patagonia alternative to fly fishing
17:39
waders. These are kind of clothes.
17:42
It came back with a price range for me.
17:45
I said under 200 bucks and it's got the men's Kennebec stockings
17:49
by LL Bean, 189 bucks. Does it pass your test?
17:54
I passed my test. That's amazing.
17:55
I'm gonna keep you at it though. I'm gonna.
17:57
I'm gonna try and break Tenita with some of my requests here.
18:00
Please do. Please do.
18:01
We love it. We love it when when users do
18:03
this. That's how we learn.
18:04
That's how we learn what's working and what doesn't work
18:06
right. And it's cool.
18:09
Now talk to me about go to market and where you found your
18:12
first users. Where are you?
18:13
Where are you looking to to grow an audience?
18:16
Yeah, we. Started to create Tony as a
18:19
broad product, but we're slowly kind of converging on this idea
18:22
that we want to find an appeal to users somewhere in the 20 to
18:26
40 age range who have, who value their time and who have, you
18:33
know, good shopping needs across the spectrum.
18:36
We are not specifically going at one gender or the other.
18:41
You know, it's, it's a, it's a neutral side works for, works
18:44
for everyone because it's clothing an apparel.
18:48
The spectrum is broad. So we're not going super
18:50
targeted that, you know, you know, women 30 to 34 or anything
18:54
particular like that. But a fair amount of our our
18:58
outreach through say social media promotions and others paid
19:02
paid media and others are all you know in that in that range
19:06
20 to 40 users with some interest in shopping again
19:12
depending on the exact knobs that Radius advertising consumer
19:16
was to use. You bet.
19:20
Now, one thing I like about the the product too is it's, it's,
19:23
it's very open now. Is there a temptation
19:26
considering that like shopping is 90% women to like skin it in
19:30
a way that's very feminine or skewing towards certain trends?
19:35
How, how do you approach kind of like your entry into the market
19:38
and kind of like your brand positioning?
19:40
Right, right. That's, that's such a great
19:41
question, right. On the one hand, if you think
19:44
about shopping e-commerce as a whole, it's a humongous market,
19:48
right? It's a massive, massive market.
19:50
On the other hand, to go to market, we need to be super
19:53
focused on delivering a great experience for, you know, 1
19:57
segment of it. We've taken a position,
20:01
obviously, we've taken a position not to cover all of
20:03
e-commerce. So sadly, we're not even doing
20:07
anything beyond retail. And even within retail, we are
20:10
sticking to clothing and apparel and accessories.
20:12
Recently we've added cosmetics to the mix as well.
20:16
And slowly we're adding some HomeGoods coming in as well.
20:19
So we're we're starting on one corner of the retail e-commerce
20:24
spectrum around clothing and apparel.
20:26
And within that we didn't want to, you know, segment too
20:30
strongly so that it creates a very skewed brand perception.
20:35
We try to try to avoid that very carefully because we really
20:39
would like to need that to be the ultimate shopping
20:41
destination for everyone, and we chose to go category restricts
20:46
rather than audience and consumer restrictions.
20:51
Amazing. And that's got to be a tough
20:55
decision, right? Because yeah, it is.
20:58
Talk about that a little bit more.
20:59
How did you come to that conclusion?
21:02
Really just didn't want to pigeonhole ourselves into very
21:07
narrowly saying that we, we had lots of lots of ideas there.
21:10
For example, we, we thought about saying how about let's
21:13
just go scour the web for all sustainable clothing, which
21:18
would be reasonable, right? But it's such a small inventory
21:22
and you know it's going to be niche and it, it, it wouldn't
21:26
even come close to passing what Larry Page would call the
21:29
toothbrush test, right? And it's not something that you
21:32
use everyday. So we need to strike a balance
21:35
where we can make a reasonably large number of users think
21:42
about Donita on a reasonably frequent time scale.
21:45
So ideally we would want a very large number of users thinking
21:49
about Donita every day or something.
21:50
But even as far as commerce needs go, the moment we skewed
21:56
too narrowly into something like sustainable clothing or, you
22:01
know, urban women clothing or something very, very niche like
22:05
that, there is a danger that there's a danger that, you know,
22:10
it's going to be hard for us to learn what works and what
22:13
doesn't work. And the reason, the main reason
22:15
for that is because we're also simultaneously trying to figure
22:19
out how AA works for topping, right?
22:22
And if we Stew the user base too narrowly, we'll only learn from
22:26
that one one set of users, right?
22:28
And how that particular demographic uses Tunida.
22:32
And obviously there's a wide range in how people of different
22:35
ages and education levels and income and geographic locations
22:40
and so on use. Not everyone uses in the same
22:43
way. Younger people tend to be
22:45
texting all the time. They're much more comfortable
22:46
texting. So we did some surveys and
22:48
market analysis and we found that users under the age of 30,
22:53
for example, said they would use something like Donita at a much
22:58
higher rate than any any any other use group.
23:03
Again, that's not surprising, but they are the, you know, this
23:05
generation, you know, dens, Ethan millennials live on
23:09
texting platforms of one kind or the other 24/7.
23:12
So for them, they are, they are sort of AI native in a way that
23:17
other demographics probably are not.
23:19
So we wanted to skew in that direction, but not super finely
23:23
segmented any further. Fantastic.
23:27
I love it. Now I detect a slight accent,
23:31
which I believe I'm pretty good with accents.
23:33
I believe you're from Brooklyn, is that right?
23:36
No, I'm not from Brooklyn. You got me there.
23:40
Yeah. Tell me your story.
23:43
How did you find yourself in Silicon Valley?
23:45
Yeah. I I grew up in India, went to
23:47
college there, and I went to not to Brooklyn, but to Buffalo, NY
23:53
from Graduate School. Go Bills.
23:56
Go Bills. Still a Bills fan.
23:58
I was dead for three of the four Super Bowl losses, if you know
24:02
about them. I definitely do, man, I love
24:06
that team too. Thurman, Thomas.
24:09
Bruce Smith. You and I can geek over bills of
24:13
the early 90s and talk about John Elway and the Broncos as
24:16
well. And yeah, we can go on.
24:18
Yeah, absolutely. This episode is brought to you
24:23
by Moeburst, the AI first growth engine trusted by iconic
24:27
startups like Calm, Robin Hood, Reddit, Uber and Google.
24:32
Moburst is a digital, full service, mobile first marketing
24:35
agency that helps companies to scale and become category
24:38
leaders. Whether you're finding product
24:40
market fit or ready to accelerate growth, Moburst
24:44
delivers creative, data-driven solutions that help with scale
24:47
and efficiency and impact. At Moburst, we're not just
24:51
guessing what works, we're predicting it.
24:53
As an AI first agency, Moburst uses cutting edge tools to
24:57
analyze user behavior, adapt creative in real time, and cover
25:02
hidden growth opportunities. The result?
25:04
Marketing that's faster, smarter, and built to convert
25:08
from early stage startups to global brands.
25:10
We've helped hundreds of companies overcome complex
25:13
challenges, unlock new audiences, and optimize every
25:17
metric that matters. At MO Burst, we believe that
25:20
creative thinking, advanced technology, and data are the
25:23
core drivers of success at scale.
25:25
If you're serious about scaling your business with precision and
25:28
purpose, it's time to talk to MO Burst.
25:31
Visit moburst.com and Let's build something category
25:34
defining together. Absolutely.
25:39
That was the time I was doing my Graduate School.
25:42
So I did my Graduate School, then I worked in computer
25:45
science research. I was a professor for a few
25:47
years. Then I worked at IBM Research
25:50
for a number of years, doing research in the deeper ends of
25:55
mathematics and computer science.
25:57
And search started to happen. Internet and search started to
26:00
happen. And initially I was studying a
26:02
whole bunch of mathematical questions related to search.
26:05
How do you rank search results in better ways than than were
26:09
known at the time? And that got me interested more
26:13
and more into that that part of the Internet experience.
26:18
Around 2005, I went to Google to work there as a retail
26:22
scientist. I worked at Google for a number
26:24
of years since then in 2021 is when I left Google to to found
26:31
Donita. That was that was the moment
26:33
when I had this very clear idea that a was happening.
26:36
But you know, all along the journey was fantastic, going
26:40
from being a researcher in a very obscure field of research
26:45
that, you know, probably no more than people in the world care
26:49
about to, to working on something super mainstream like
26:53
shopping using natural language that is something potentially
26:56
everybody would care about. So that that transition has been
27:00
really beautiful and really rewarding.
27:04
What an incredible career. You know, in the, the AI hype
27:07
cycle right now is just at peak levels.
27:10
You know, everyone's really excited or really scared.
27:13
One of the two having gone through kind of the rise of web
27:17
one, point O, the Dotbom, web 2, point O cloud, and now we see
27:23
the rise of, of AI. How do you put enthusiasm from
27:28
what you're seeing in tech broadly in comparison to to
27:32
other kind of bubbles in tech? That's it's a fantastic
27:36
question, also a loaded question.
27:38
So obviously there have been multiple cycles of, you know,
27:44
quote, UN quote hype, but under underneath all the hype and
27:48
drama, we've also had some truly beautiful developments, right?
27:52
The Internet happened. The Internet is such an amazing
27:54
thing that, you know, until then, communication between
27:58
humans and humans and other organizations like business and
28:02
others was just so broken. And now it's just so much, so
28:05
much better. And that was a big gain for
28:08
connecting all of us together and and making new experience if
28:12
possible. And then the next one came, I
28:15
think, and obviously that led to the.com boom and bust and, you
28:19
know, people went overboard. I think things settled, but the
28:23
next revolution came with, you know, the smartphone and that
28:28
took computing to US 24/7, right?
28:31
Always on. We're constantly checking
28:32
something or the other. We're connected, hyper connected
28:35
with other friends and family and also work, which is, which
28:40
has got, you know, both pluses and minuses and and that took us
28:45
down a path. And that was real, Real in every
28:50
imaginable way, right? Because it affects the vast
28:53
segment of the world's population because so many, I
28:58
would bet that a significant fraction of the world's
29:00
population carries the cell phone may not be 50%, could be
29:03
20%, but that's still, you know, billions of people that would
29:08
never say true revolution, not just not just the hype.
29:11
Then there were subtler ones like cloud computing.
29:14
Cloud computing, you know, most average humans on the planet
29:18
wouldn't know or wouldn't care what cloud computing is.
29:21
And yet internally within the feed, that was such a big
29:25
enabler. You know, people like me, for
29:27
example, who have no expertise in, you know, building or
29:31
spitting up a small data center or whatever it is, couldn't have
29:36
conceived of doing a startup even 10 years or so ago, right?
29:41
Because all of that was a big bottleneck.
29:43
And cloud computing totally enabled, enabled all that.
29:46
You know, we can just pick our favorite cloud provider, get
29:50
down to coding and build and ship a product really quickly.
29:53
So that was this kind of quiet revolution happening as well.
29:58
And then there is AI, which I think is probably the biggest of
30:03
them all. But again, it didn't happen
30:05
overnight. It was an overnight success 30
30:09
years in the making, right? It's it's been coming for a long
30:13
time. And also it's a very different
30:15
one qualitatively. By that I mean this, right?
30:19
If you look at the, you know, 30 year history of the Internet
30:22
until say chat DPK, right, things were evolutionary.
30:27
You know, we had a sense for what next year is going to bring
30:29
up and the year after that and the year after that.
30:33
The one exception was the birth of the Internet and maybe the
30:35
iPhone. iPhone was such a beautiful device that it it was
30:38
a significant quantum jump and of course other others caught up
30:42
and things things were smoothened again.
30:45
But a A was truly revolutionary in the sense that we're now
30:51
interacting with our devices in ways that we couldn't even think
30:56
about five years before a happened.
30:58
And we are asking questions and it goes both ways too, right?
31:01
One is our computer systems are able to understand us a lot
31:05
better than, you know, any system could do before Google
31:08
was getting better and better, but still nowhere near, you
31:10
know, this, this level of fluency.
31:13
And in the other direction, our computers are able to tell us
31:15
things in much more organized manner.
31:19
It's not like just giving you a link to go see something, but
31:22
actually condensing a whole amount of knowledge and
31:24
transmitting to you in a very concise and more powerful way.
31:29
So this, this one is, you know, it's probably the biggest of
31:33
them all in terms of the revolutionary breakthroughs.
31:37
You know, obviously there's a lot of hype in the media and so
31:39
on, but I think it has the potential to live up to all that
31:43
and all that and more actually, because there are things that
31:47
we're going to see happen in the next decade that that are going
31:51
to be truly transformative. Writable technology is
31:57
ubiquitous now, but take me back when you were a kid, like there
32:02
was no way you could have known when Scott Norwood was blowing
32:06
and breaking the hearts of all the Buffalo Bills fans that that
32:11
computers and technology would be where it is today.
32:13
How did you get introduced and what was your your early
32:17
attraction to technology and computing?
32:20
Yeah, I think there was a very precise moment.
32:23
I think after my 10th grade in high school, there was a summer
32:28
program. Somebody said, hey, they say
32:31
this is a small town in India. And, you know, somebody said,
32:33
hey, there's this computer class that somebody's teaching.
32:36
And, you know, we go there to look at what it is.
32:40
It's a really tiny little room with one small section of it
32:44
closed out and air conditioned because he couldn't even air
32:47
conditioned the whole place. It's the tiny little sub room
32:50
that's air conditioned. And they're they're like 3
32:53
benches for about 10 students to sit.
32:55
And you know, I was bored and I said, OK, let me try this thing
32:58
and signed up. And it was a class and basic
33:02
that entire summer. All they did was learn basic
33:05
programming and we would get to use this computer maybe about
33:09
one hour every week. So which means that all of the
33:12
stuff has to happen on on paper. It was also the time, you know,
33:15
because it was high school, I started reading, reading much,
33:19
much more widely and a reasonable bit of fiction,
33:23
including science fiction. And some of them obviously have,
33:26
you know, beautiful themes of futuristic.
33:30
Then there are I think if you go dig up, I'll find some of those
33:33
where they're talking about AI like systems as early as in the
33:37
80s, right. And so I think that that the two
33:41
things came together, this idea of a utopian future where
33:46
amazing things happen. And that's on the one hand, you
33:49
know, things that are very hard to even wrap your head around.
33:54
On the other hand, there is this one hour of basic programming on
33:57
the Spectrum computer, which was super concrete, as concrete as
34:00
you can imagine. It was just fun, right, Doing
34:04
them together. And that I think led me to
34:06
computer science and then Graduate School wide right by
34:11
Scott Norwood. As you pointed out, you know,
34:13
but the phrase that's etched in every build span January.
34:17
But yeah, that was the that was the introduction to technology.
34:22
Cool. You sure got it right.
34:24
OK. When you were a kid, what was
34:25
your favorite sci-fi franchise? Was it Dune?
34:28
Was it Star Trek? Was it Star Wars?
34:31
What? What were you?
34:32
I would. Say Star Trek was the closest.
34:35
I do have this. I do have this, I think in
34:40
sci-fi. I think that's a distinction.
34:41
I think true sci-fi fans will maybe kill me if I say this
34:44
incorrectly, but there's this thing of a high and low kind of
34:48
thing. The high ones being so bizarre
34:50
that they have nothing to do with how the human society or
34:54
this planet are functions. Whereas the low ones are a
34:57
little bit more like sort of like our worlds, but with super
35:00
powers. I, I, I'm drawn to the ladder.
35:04
So I like Star Trek, I like Harry Potter, for example, you
35:07
know, different kind of fiction. But again, it's, it's kind of
35:10
grounded in our values and systems, but also there is some
35:13
magic happening, right? That's the kind of stuff that I
35:16
love. So even though Harry Potter's
35:18
probably not a science fiction franchise, that's probably the
35:21
one that I love the most because to me, to me, and again, that
35:25
goes back to my love for AI, give me a magic, right?
35:29
And magic is something that that I love to see.
35:35
So true, the the best AI products just work like magic
35:38
and they spread like wildfire because of it.
35:40
Exactly. Yeah, Yeah, OK.
35:43
OK, Are you ready for the lightning?
35:45
Round. Let's do it.
35:47
Let's do it now before I do. I just tested Tunita again this
35:52
time I want to see the guardrails.
35:54
So if people what people might not know about me is that I love
35:59
fake Versace. I got I got like 6 pieces of
36:05
fake Versace and I I rock it much to my kids chagrin.
36:08
But I put it in there. I said OK I want Versace
36:12
knockoffs find me that. And it said you know what I
36:14
can't find you image or fakes or any kind of copies.
36:18
But what I can do is find Versace alternatives in this
36:21
price range. Fantastic.
36:22
Fantastic. That's music to my ear.
36:24
That's music to my ears. And that's we take a lot of
36:27
effort to get this fate. And again, I'll not say that
36:30
we're getting it right 100% the other day.
36:31
But for the most part, I think we want to do the reasonable
36:35
thing right. And when a user is looking for
36:38
Versace knockoffs, they don't mean it in the sense of, you
36:41
mean fake bootleg products. I wonder what they want.
36:43
Something that has the right look and feel, the right style
36:46
maybe. And yeah.
36:49
Exactly. Yeah.
36:50
It doesn't have to say Versace. I just want like black and gold
36:54
with chains on it and cool plants, you know, maybe like a
36:57
tiger. You should get there.
37:02
The lightning round let's Let's assume that today Tanita was
37:06
acquired for a billion dollars. What do you do tomorrow?
37:13
Make us worth the billion dollars.
37:15
Somebody's buying us for a billion dollars.
37:17
We really want to make them succeed in ways that they think,
37:21
you know, we really got it cheaply.
37:25
To me acquisition the best model is YouTube, right?
37:28
People say the $1,000,000 for a billion point $6 for YouTube.
37:31
Now the credit say it's a phenomenon.
37:32
That's that's the kind of acquisition we'd like to be and
37:34
that's what we'll work to work. I love it.
37:38
Now also, before we continue with the lightning round, I want
37:41
to tell the listeners to go over to YouTube because we are
37:44
actually going to get insight from Seva exactly which mistakes
37:50
to avoid and which mistakes to perhaps are necessary for
37:53
founders to have a career similar to Seva's.
37:57
Now back to the questions. If you could build $1,000,000 a
38:02
company that does $1,000,000 in 12 months right now, what
38:05
company would that be? $1,000,000 in revenue.
38:09
Yeah, I think we'll take Tonita there very soon.
38:13
That's that's a really there. I love it.
38:18
If you were in college today, do you feel like Josh Allen should
38:23
get a second MVP this year? Or even better, what would you
38:27
study in college? Let's let's get this clear.
38:30
George Allen should almost surely get the second MVP and
38:33
hopefully Super Bowl ring as well.
38:35
That's what I'm rooting for. But.
38:37
I'm I'm rooting for it, too. It's his year, man.
38:40
The Chiefs must go down, period. They had a fun team to watch as
38:44
well, but I'm rooting for the Bills all the way.
38:47
What an imagine again, I think 30 years of computer science and
38:52
I want to do computer science all over again.
38:54
I'd probably do physics a little bit more seriously because I
38:57
think quantum computing is something that I've been
38:59
interested in quite a bit over the last few years.
39:01
But computer science has so many beautiful questions, you know,
39:05
touching on just not just mathematics and technology and
39:08
electrical engineering and all that, but also in terms of
39:10
philosophy and how it connects to humanities and, and all of
39:13
that, the field of computer science, the discipline of
39:16
computer science, if you all so beautifully, I would do it all
39:19
over again. As a computer scientist of 30
39:24
years, when you think about quantum computing, what are use
39:28
cases that you're really excited about?
39:30
I think scientific use cases like drug discovery potentially,
39:34
I think it's something that quantum computers might come,
39:37
might contribute to. The lines are not super clear
39:40
today that things we know how to use quantum computers for super
39:44
well are things like breaking today's cryptography.
39:48
But that's not the part that it fakes me.
39:51
Quantum computers, I think have not been understood in in many,
39:54
many ways. And I think things like
39:56
searching large search spaces and figuring out things like
40:00
it's a new drug that can solve cancer or whatever.
40:03
I think that's the kind of thing that I believe quantum computers
40:06
could become useful for, yes. You know, is there ACEO or
40:12
another founder that you follow closely or inspired by?
40:17
Yeah, quite, quite a few of them.
40:18
I think I'll say the PEO of perplexity.
40:22
You know, it's a very, very young person, but the energy
40:25
with which it's founded, the company and it's been driving it
40:27
is obviously very, very impressive.
40:29
His name is Arvind Srinivas and that company came out of nowhere
40:33
and there's no one of the hottest startups.
40:37
Nice. Was there any particular
40:39
technique he deployed? Like how did how did he reach
40:42
that kind of traction? I think he's been thinking about
40:45
a A working in AA. And when that GPT happened and
40:49
GPT 3 happened, I think he woke up to the fact that, hey,
40:52
there's a whole new way to summarize web pages.
40:55
Now let's put a search engine on top of that and not just give
40:59
links to web pages, but actually read and summarize them.
41:02
And that was such a simple, clean idea.
41:04
And I think because of its unique and clean focus, they've
41:08
been doing a lot better than one would expect somebody to be
41:11
doing with competition from Google and GPD as well.
41:16
Yeah, Von Miller was recently released to free up cap space,
41:21
good or bad. OK, I think he's he's served
41:25
this role admirably. Still leaves the defensive hole,
41:30
but but I think I'm OK with that.
41:32
I'll take it. If you could live in any
41:38
historical era besides this one or the future, which would it
41:41
be? Oh, let's say it's this one or
41:45
the future. I would say maybe I'm, maybe I'm
41:51
being a bit too myopic here. Maybe I would say the 70s.
41:55
That was the kid in the 70s. I would probably, I would like
41:59
to have been adult in the 70s because I would have like, I
42:01
would have seen computer science grow from, you know, Ground
42:05
Zero. You're at least the fifth person
42:09
that is said the 70s. What?
42:10
What makes you nostalgic about the 70s?
42:13
What do you miss? Because as a, as a computer
42:15
scientist, every classic thing we learn in computer science
42:18
happened in the 70s, right? And that would have been a, you
42:23
know, and it was, it was still a bit of a niche discipline.
42:27
It was not like a, a where hype cycle was big.
42:29
It was still a very focused field of study.
42:32
And I think I would have really enjoyed being part of that.
42:36
Amazing wrapping things up here. Where can people connect with
42:40
you online? LinkedIn.
42:43
I'm on Twitter as well. What are what are good places?
42:47
Great. It's been so fun to get to know
42:50
you a little bit. Siva I, I encourage everybody to
42:53
go check out taneda.co. It's, it's beautiful.
42:56
It's a very elegant and simple solution, but boy does it get it
42:59
right. Pretty incredible.
43:02
Congratulations. Thank you so much for handing
43:04
it. Wonderful.
43:04
Talking to you about, yeah, yeah, computer science,
43:07
technology, but also about football and Buffalo Bills.
43:11
Hey, go, Bills. Go Bills.
43:15
OK, now let's record the little YouTube segment.
43:19
So I'll tee it up for you. Be like, hey, here's some
43:21
mistakes to that to avoid as a founder.
43:25
Is that is that OK? Yeah, that's that's, that's
43:27
good. That's great.
43:28
Do you want us to do a quick preview of that before we record
43:31
that or go right in? Well, it's recording.
43:35
I can always chop it up. So don't worry about it.
43:38
It's always. It's just a rough.
43:39
Take OK, OK, sounds good. Sounds good.
43:43
Hey. What's up everybody?
43:44
I'm here with Siva from tonita.co, and he's going to
43:47
give us a little bit of tips of the trade of mistakes to avoid
43:52
as a founder. Siva, take it away, all right?
43:55
So we founded Tonita a few years ago and we've been building
43:59
Tonita out over the past couple of years and we are have a beta
44:02
version in and now we're going to go fully live with the new
44:05
design very soon. Ryan asked this interesting
44:08
question. What are some of the mistakes
44:11
that you made that you tell other founders not to make?
44:14
I think that's excellent question.
44:18
So I think the first thing I'll do is think through what you
44:22
want before you raise money. I think building something,
44:27
prototyping something, sometimes it may be easier to raise money
44:30
because to get some idea off the ground you need some money.
44:33
Maybe there is an Angel round or a seed round or something.
44:35
But before you go in and raise the serious capital, get
44:40
something, something working. Second mistake I would say I
44:45
made are things that I didn't do as well was went down a few
44:49
rabbit holes in the early days that I think with in retrospect,
44:53
I should not have done some of that was investing a lot of
44:56
effort into designed before the idea was clear and that that
45:02
just turned out to be energy and money spent on something that
45:07
you know, just just didn't pay off in the way it could have
45:11
been paid off. So I think we had a clearer idea
45:14
of what it is that we are building.
45:17
We probably should not have gone down the fine details of every
45:21
pixel on the screen. So I think that was that was 11
45:24
mistake that I made that I probably wouldn't make.
45:28
Make your first hires super carefully when in doubt, And I
45:33
think it's a little it's a little hard to say this.
45:35
When in doubt, do not hire the person that you're considering
45:38
because you know, I do have suddenly some engineers that
45:42
turned out really great, even though I had some reservations
45:45
about them. But sometimes look for the great
45:48
fit and see. And this person, I think the
45:52
most important question to ask and I didn't ask at least in one
45:54
or two cases, was can this person handle the uncertainty
45:58
and the ambiguity that often comes with working at a startup?
46:02
And that's a question that I didn't ask.
46:03
And I think that probably was well worth asking.
46:06
The last one I would say is when the right time comes, double
46:12
down on things like design. You know, we launched Anita and
46:14
Beta with with a reasonably minimal design, which is OK, but
46:19
for a consumer product, it's important that design really
46:22
shines. And I think that we didn't put
46:24
enough time and energy into it early on this time.
46:26
So I did both mistakes there. One was putting too much into it
46:30
too soon and not putting enough into it when the time was right.
46:34
I think these are some things that are worth worth paying
46:37
attention to. Seva really appreciate words of
46:41
wills wisdom. Awesome.
46:44
Hey founders, let me ask you this.
46:46
Are you building with AI or getting left behind by it?
46:51
If you're serious about scaling smarter, faster and bolder, you
46:54
need to check out AI for founders.co.
46:58
It's your go to resource for cutting through the noise and
47:01
actually using AI to grow your startup.
47:04
Each week, AI for Founders drops a sharp, no fluff newsletter,
47:09
dives deep with expert guests on the podcast, and hosts monthly
47:13
workshops where you can get your hands dirty with the real tools
47:17
and strategies. Whether you're pre seed or
47:20
series B and beyond, launching something new or just staying
47:24
sharp, AI for Founders gives you the edge you need to win in 2025
47:28
and beyond. Hit AI for founders.co to
47:32
subscribe to the newsletter, tune into the podcast, and grab
47:35
your spot in the next workshop and it's all free.
47:39
Build smarter, move faster, think bolder with AI for
47:43
Founders. Now back to the show.