November 29, 2023 • 35 mins
This week we sit down with a highly respected product leader, founder, and CEO who is revolutionizing your work docs with Coda, the all-in-one collaborative workspace. We talk about:
🎙️ How YouTube, Coda, and Artsy each found product market fit and scaled
🎙️ The “Golden Rituals” that create culture and help teams cooperate
🎙️ Under the hood: how AI really works in the products you love
🎙️ When bureaucracy ruins innovation—and mechanical bull office parties
You’ll come away thinking differently about the routines that drive your work and the algorithms that shape your life.
Go Deeper:
📚 Shishir’s book-in-progress: Rituals of Great Teams
🌐 The Art of Framing Problems: Eigenquestions
💪 The Career Manifesto, Mike's full reading list, and other resources at mikesteib.com
Next Week’s Guest:
🚨 Michael Maslansky, founder and CEO of maslansky + partners, the world’s leading expert on language strategy and crisis communications. DM me your questions on Linkedin or call 213.419.0596
See omnystudio.com/listener for privacy information.
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Speaker 1 (00:04):
Welcome to Office Hours, where we sit down with the
chief executive shaping the world and answer your most pressing
questions about leadership, career, and life. Mike Steib and today
we are hanging out with my good friend Shashier Morotra.
Shasher is the founder and CEO of Coda, the venture
back startup that has created an amazing alternative to word
in Google Docs. He and I worked together previously at Google,
(00:27):
where Shashier ran product engineering and UX for YouTube. He
has previous experience at Microsoft and as an entrepreneur, sits
on the board of Spotify and as a thought leader
on how teams work and collaborate. He's also an all
around terrific guy and Shasher, I am thrilled to have
you on the show today.
Speaker 2 (00:47):
Welcome aboard.
Speaker 3 (00:50):
Thanks for having me, Mike. It's been so much fun
watching I think our journey started almost fifteen years ago
together and lots of great experiences in between.
Speaker 1 (00:59):
I'm glad you're here. As you know, our listener's sending questions.
There's stuff that they want to hear from you. And
first question is from Nina and Tucson, Arizona.
Speaker 3 (01:08):
She says you've both worked at some of the biggest
tech companies and now lead earlier stage companies.
Speaker 2 (01:14):
What have you found is similar?
Speaker 3 (01:15):
In what is different when the size of the company changes.
Speaker 2 (01:18):
And you share.
Speaker 1 (01:19):
Maybe just tell everybody quickly sort of your background. I
met you when you were coming out of Microsoft and
into Google, just so everybody has a good sense for
what you're all about.
Speaker 3 (01:28):
Yeah, I mean, I've definitely done the full gamut of
that big, huge company and the start from scratch in
just a room with a few people. So my career
I started out of college. I started a company called
some Trata, those in the data center automation space. I
then went to Microsoft. I spent a number of years
(01:50):
working on Windows, then Office and SQL server, and then
in a weird set of circumstances, I ended up at
Google working with my on television video. I ended up
running the YouTube group there from about twenty eight twenty fourteen,
really exciting time. And then I started Coda. And Coda
(02:12):
we started as a new company to reimagine the document.
We built an all on one document on the best
parts of document spreadsheets, presentations, and applications into a single surface,
so anyone can make a dot as powerful as an app.
Speaker 2 (02:27):
Tell everybody like, what is it about these companies that
are growing and changing the world and then something changes
and what have you seen in that part of your journey?
Speaker 3 (02:35):
So I joined Google in two thousand and eight, and
at the time I was I'd done to Microsoft. I
thought of Microsoft a little bit as my big company
education and my initial reaction to all the Google outreach was, yeah,
I'm not interested. I'm already at a big company. I'm
going to go do something small. I had all these
ideas in mind of what to go do, and as
(02:56):
guy Nam Jonathan Rosenberg was running product at Google at
the time, and he gave me analogy that the Hues gave,
which is that Google in two thousand and eight felt
like Microsoft in the in the early nineties. And I
think that was accurate that you know, at the time,
Google was really a one product company, just learning how
to take its culture and apply it across other markets
(03:17):
and divisions and categories and so on. At the time,
you know, it was just really searching ats, and you know,
we hadn't done you know, it just wought YouTube. We
had you know, started maps. As Chrome was about to
the launch Android, it was just still an early prototype
and all these things that now we take for granted
as being stills at Google weren't there.
Speaker 2 (03:37):
And now it wasn't a small company.
Speaker 1 (03:38):
This was post ipo, multi multi billion dollars of revenue,
tens of thousands of employees. But I remember to share
what I had come from a from a big, hundred
year old company, and when I came to Google, we
needed we needed fifty million dollars for this deal that
we wanted to do to jumpstart a business that had
no revenue. And so I was sort of imagining what
the bureaucratic process would be. And it was just going
(04:00):
to see Eric and Larry and Sergei on a Thursday,
and we got a yes in the meeting and I
was like, well, what's the follow up?
Speaker 2 (04:07):
And they were like, will you just spend it?
Speaker 3 (04:09):
Now?
Speaker 1 (04:10):
It's I couldn't believe the speed with which the company
was willing to work and the absence of bureaucracy and
middle management.
Speaker 2 (04:18):
It was really wild.
Speaker 3 (04:19):
Unfortunately, I don't think Google maintained that forever, but in
that period I think it was about as good as
it could get in terms. And I think Eric is
a wizard at managing chaos like that, and he was
just so good. I mean it said a similar story.
I came in and said, I need twenty people to
work on this and said, you know, do I go
to some a Google body, do I get it? I
(04:40):
do I need to get some you know, some official thing.
And I was told no, if you can convince twenty
people to work on it, then you have twenty people.
And it was like this interesting meocratic like there's just
sort of this assumption that if you screw up, those
other consequences and people will stop working on it and
so on. But there was you know, at the time,
(05:02):
Google's whole ethos was feed of innovation was all that mattered,
and Eric was excellent at you know, supporting people on it.
That was really wonderful.
Speaker 1 (05:11):
Everything came from a place of yes, but there's certainly consequences.
I remember we had a client event one time and
I ordered a mechanical bull. And when there's no process
in bureaucracy, there's no one to stop you from bringing
the mechanical bull into the office. So we've had clients
drinking and riding a mechanical bull, and in hindsight, like
a lot could have gone wrong.
Speaker 3 (05:32):
Yeah, yeah, I mean, it's definitely I don't know if
you could run a company the same way today as
as we did back then. But it's I mean, there's
a lot to be you know, the first era of
Google was this magic in a bottle search engine that
that that turned out to be, you know, really amazing.
But many companies never make it past that face, never
(05:54):
make it into multiple categories and multiple new products and
so on. And Google managed to make it through that casm,
which I think is remarkable.
Speaker 1 (06:01):
Right and was and you can also, you know, bureaucracy
emerges as a way to maintain you know, to avoid
risk because risk is too expensive. But we were a
company doing probably twenty million dollars a free cash flow
a day when you arrived, and there's a lot of
room to there's a lot of room to make mistakes
and be able to pay for those mistakes rather than
pay for them with added layers of bureaucracy. And yeah,
(06:24):
you know, and slowness and poor culture and all that stuff.
Speaker 3 (06:27):
Yeah, for sure, maybe I'll try to answer Nina's question.
I mean, I think when I think about and I
don't know if I would say there's large companies, there's
small companies. There's also small projects within large companies. I think,
you know, some of the stuff you and I worked
on together in the video space was you know, had
some advantages of being at the big company, like like
you mentioned, it was easy approvals, but we were really
(06:48):
you know, a typically.
Speaker 2 (06:50):
We're still starting out, that's right, Yeah.
Speaker 3 (06:51):
I mean you still had to define a market and
find a customer and and so on. And you know,
the best analogy I can give is, I had an
old boss at Microsoft give me this framework for thinking
about the stages that products go through. And you called
a joke, threat, obvious. Every product starts as a joke,
nobody believes in you, nobody thinks it's for real. Then
(07:13):
you become a threat, and everybody responds. They've got their
you know, they've got their response features or their response
go to market campaign or their response billboards, whatever it is.
And then at some point you become obvious. And when
you're obvious, everybody just kind of presumes you're going to
win and continue to win and they work around you.
Speaker 4 (07:30):
Right.
Speaker 3 (07:31):
And add I got to be at Google in particularly
YouTube through all three phases. When I showed up, it's
I don't think I can understate how much people thought
YouTube was Google's first big mistake. And we paid one
point six billion dollars for this company. We were losing
hundreds of millions of dollars a year. We were losing
(07:52):
you know, almost pennies per view.
Speaker 2 (07:54):
The only good content was content that they were not
allowed to have.
Speaker 3 (07:57):
There was we had a feeling somebody else there's.
Speaker 2 (08:01):
A huge Viacom lawsuit.
Speaker 3 (08:02):
Yeah, and so we we get this group together and
it was totally you know, you hear the analogy a
lot of pirates that navy. But you get this group
together that say, you know what, I think we can
do this. I think there's a I think there's a path.
And then at some point, I think about three or
four years in all that started working. We got profitable,
we figured out how to Google revenue, We got through
(08:24):
our agreements with the with the recording industry and with
the movie industry and so on, and all of a
sudden it became threat and everybody had their response to YouTube.
And then at some point it switched to being obvious
and and the and you know a lot of people
think of that as like really awesome, and it is.
I mean, this it's obviously the goal. It's like you're
(08:45):
building this company. You want to get to that obvious dage,
but there's downsides too. And when's the last time you
read a positive, positive press article about YouTube? I mean
the early days, we couldn't get anybody to write about
us at all, and then all of a sudden, now
all you can see is negative that negative things because
it's you know, these are sort of presumption that, of
course this thing is going to continue to be to
be big and useful and so on, So people will
spend all their time looking for the flat. So there's
(09:07):
lots of pros and cons of each. But I think
learning how to operate in all three phases, you know,
what do you do when the when the world thinks
your joke? How do you respond when they decide you're
a threat? And then what do you do to manage obvious?
And I think Google did a really good job of
taking that obvious stage and starting a bunch of jokes
and going and working its way back through that cycle.
Speaker 2 (09:28):
Well, this this actually probably is really well.
Speaker 1 (09:30):
In our second question, so Ian and Burlington Vermont asks us, Hello, can.
Speaker 3 (09:36):
You talk about your products and how you knew when
they had a product market fit?
Speaker 1 (09:40):
So Shashir here's where I admit to you and you
told me that you were going to do Google docs,
but they're also spreadsheets inside of it and other stuff.
Speaker 2 (09:46):
I was like, man, we.
Speaker 1 (09:48):
See, we see working on over there. So I originally
got the joke. I have come full cycle. Now I'm
a big fan of the product.
Speaker 2 (09:56):
But take take us through your your joke joke to
threat matrix.
Speaker 3 (10:01):
Yeah, yeah, I mean I think the Yeah, I mean
the early days of CODEA, I got that reaction from everybody.
They were building a new dock and most people would say,
why do we need a new doc? And I would
get all sorts of weird feature requests and people would say, oh, yeah,
that's really exciting. You know, word is missing this one
font that I really want, And I say, no, no, that's
not what we're working on. That's like, I don't think
(10:22):
there's a category there of like the font that's missing when.
Speaker 1 (10:25):
I try to download it as a PDF to print, Like,
I don't like that.
Speaker 2 (10:28):
It's two steps.
Speaker 3 (10:29):
Right, right exactly? You know, either get the why would
you bother? Or a list of because people can't picture
it right, It's like what what does what does it
mean to have a brand new product? In this category.
Speaker 2 (10:41):
And you know, everybody wanted a faster horse.
Speaker 3 (10:44):
Everybody want a faster horse, and the and and honestly,
like that sounds, you know, in retrospect sounds positive. Nobody
could picture it. We could that worked out well, but
you know the time, it just seems terrible. Right, people
look at your product and they and they just don't
know what they what they want. And you know, so
the early days of product market fit for us are
really tough. And Coda is this interesting product where we
(11:06):
often describe it as we have an incredibly low floored
and incredibly high ceiling. So if you download sign up
for Coda and you start using the app or use
the website, you'll it looks like a Google Doctor, the
blinking cursor, blank screen, just start typing and so on.
But it's got all the building blocks inside of it,
so you can assemble the things that you normally would
(11:27):
have had to use spreadsheets or presentation for, or things
that you needed to build full applications. And we see
people build you know, whole CRM systems and code our
inventory systems, and lots of tutors and small businesses use
it to build a clients. And there's all sorts of
instant things to get built, but it's this product that
looks like one thing and then operates in this totally
(11:48):
different way as it grows with you. And so early
on we would see people first off just not know
where to start and say I don't I don't understand
like I'm used to use to this other other tool set,
and then gradually we would get people through this journey.
I started our first customer. We decided there was another
side story we can come back to. So we decided
(12:09):
to start the company in stealth, which is not something
I recommend for every company, but it made sense for us.
So we weren't telling anybody what we were doing, and
so we were recruiting customers one by one. We recruit
this customer. It's an old friend of mine from Google
who had started a new company. There were five or
six people, and they said, hey, can you use CODA
to run your team? And so they start using it.
(12:30):
They ended up using their first use case with basically
a project management use case. They sort of planned out
their work in it, and we would watch our dashboard
and our dashboard they only six people in the company,
so our dashboard would go from zero to six, like
every day. It's like, how many of the six people
use the product? And I was like, that was it.
That was like the max Yx was six and one
day this goes from six to zero and we're like,
(12:51):
we're a very daily used product, Like it's hard to
use code and not use it every single day, and
so it grows to zero and like, ooh, that's not good.
And we wait another day it still says it's zero,
and I called my buddy, the CEO NOME, and I say, hey,
you know, I just wanted to check in how things
are going. And uh, you know, I noticed that you
guys haven't been using the product much the last couple
of days. Want to notice something happened or you guys
(13:13):
like doing an off site or something. And he says, yeah,
you know, I've been kind of avoiding calling you. And
he says I have good news and bad news and
I said, okay, well give it to me. Let's do
bad news first. So all the bad news is we
have a team meeting and the team said that if
I make them keep using CODA, they're all going to quit.
And I was like, well that's that's pretty bad news.
(13:36):
Feedback is a gift, like, that's pretty bad news. That
what's the good news And he says, well, the good
news is they've all fallen completely in love with your
mission and they totally get it. But they have this
list of things that you have to do before they
will consider reusing the product. And honestly, if you can
fix these in like two weeks, we'll start again. But
you know, up to you, this is good news and
(13:57):
bad news. And it was just like really like, you know,
first off, big kick in the stomach because you're like
your entire user base of apperates in one day and
then you've got this like deep conflict on you know,
do we do this set of things or go address
the waves on a roadmap or someone. But you know,
in my mind, for our product, product market fit felt
(14:19):
like we could tell, I mean, this is not a
product where you could force anybody to use it. There
was no like Lasadaisical users like you either used it,
loved it and evangelized it, or you gave up. But
the real moment for us when we started getting calls
that said, hey, I know you have me limited to
my team, can I expand it to this other team?
And there was in particular a moment with our first
(14:42):
lord customer was Uber, and it was a moment where
our main sponsor's guy named ukiyamasheet at ukis now they
had a product or Figma and carry coded there as well.
But he called up and he said, you know, I've
been assigned this project. I'm going to This was back
in when he eighteen. He had been assigned the project
to fix Uber's reputation with drivers. You may remember the
(15:05):
time period, A bunch of a bunch of issues with
drivers and and and uh. His job was to roll
out a feature a day to addressed everything drivers that
ever asked for. And he said, he said to me,
we've been using code for a bit, and I'm quite
convinced that if we don't have CODA we will fail.
And we do have CODA, then we'll succeed. And and
(15:26):
it was like this sign of like, oh we are
We've crossed into the pull, the feeling of people want
the product, that they're ready to evangelize it. And so
that that was that was our and that was probably our.
Speaker 2 (15:40):
That was your product market, that was your product market.
We're demanding it.
Speaker 3 (15:45):
Yeah.
Speaker 1 (15:45):
I went from threatening to quit to demanding the product
exactly so for for Ian and in my business, it
was much We got a bunch of Artsy is a
marketplace for art, and before I got to the company,
we've gotten tons of galleries around the world to upload
art to Artsy, and we've gotten tons of like millions
of people to come visit arts, see and look at
that art.
Speaker 2 (16:05):
We weren't selling it. So for us, product.
Speaker 1 (16:08):
Market fit really happened when someone was paying us at
a at unit economics that were scalable and would take
our business to profitability. And what we had to do
to get there was build globally a layer of transactional
capabilities so that you could not only find the art
and this was the hard part in the tech stack.
You could not only find the art, you could with
(16:29):
confidence click a button, buy the art and have it
show up in your home a week later. And then
all when the painting started showing up in people's homes
thanks to the products we had built, that's when we
really that's when we really had product market fit. And
it's just share as I'm sure you'd advise, it's it's
a little bit different for every company, but it's either
some moment when the users will go to the end
of the earth for it, right, It's not when somebody
says they like a thing. That's when you say, if
(16:51):
I took it away, how would you feel? Or it's
when someone opens up their wallet and says, you know,
this is so important to me that I'll I'll pay
you for it. M Devin in Rally, North Carolina says.
Speaker 4 (17:13):
I'm not a tech person and I would just love
to understand how all this stuff works, like the recommendations
on YouTube and in my music app, Like, can you
just take me through what's like under the hood of
these products?
Speaker 2 (17:24):
So just here you go.
Speaker 1 (17:25):
First, you've seen some really yeah, I mean you've been
on the end under the hood and some really awesome products.
Help somebody who doesn't code get an appreciation for what
is the thing doing when they're having a great user experience?
Speaker 3 (17:36):
Yeah, And it's interesting when I when I got to YouTube,
the uh you know, I meet people regularly and say
work on YouTube, and they would say, what do you mean?
Like you create content? Like what does that? What does
that even mean? Because I think people's perception of these
products is like what is behind the thing? You see
the recommendation, you see the video that's playing and so on,
(17:59):
it's like hard to tell what is what is all
these what are all the things that are happening there?
And so first off, there's there's a.
Speaker 1 (18:05):
Way which is buy by the way, non tech companies
are always like, well, we'll just hire some tech people. Yes,
we already have the market share, right, we already have
the brand. We'll just hire some tech people. You're like, okay,
good luck.
Speaker 4 (18:16):
Yeah.
Speaker 3 (18:16):
And I think it's hard, right, I mean I think
the feeling of like, it's a website, it plays video,
How hard could it be. It's like, well, there's thousands
of engineers sitting behind it building every element of what
happens there, and you know, really hard to like, we well,
I with the YouTube are are our benchmark? Said we
streamed about twenty percent of of the bits on the
Internet came to the YouTube servers. So it's like one
(18:38):
thing that can you can you stream us and a
video file to someone else across the world, Like maybe
can you do it and fill up twenty percent of
the Internet? Like now you've got a whole different scale
of things you look at. You know, YouTube was and
still is the number two search engine in the world
behind Google, and so you know, how do you answer
queries for people at scales? Like another great engineering challenge
(18:59):
AI and recommendation some one is A is a particularly
interesting topic the and I often tell people that there's
less magic than you might think. And you know, one
of my one of my favorite stories about this is
my my father is a computer scientist and I got
to college and I told them, hey, I'm going to
go to a internship at at on my TV had
(19:23):
this book called the I Lab, and he said, you
can't do anything at the I Lab. That's where all
the jokers hang out. And it's like, really, what does
that mean? And he said, well, you know AI, this
is back in the nineties. He says, AI is like
that's where all the stuff that doesn't work hangs out.
And he had a particular way of describe me, says,
just think about it everything. The moment it works, we
(19:43):
no longer think it's AI. And in those days, like
you know, I'll give like a simple example I would
give my kids is, you know, you're driving along, you
come to a traffic light. How does it know when
to turn right or cree? And you know, it's kind
of obviously because a censor underneath the underneath the pavement
and it says is there a car here or not?
(20:04):
And it decides on whether the turnout or green. And
because I can explain it, it doesn't seem like AI.
But if I just told you that I've built this
totally amazing system that automatically rops traffic through through an area, like,
it feels like AI. And so a lot of AI
is actually that it's actually things that when you explain
exactly what they do, they aren't really that complicated. So
as an example, you know, recommendations on YouTube. The predominant
(20:29):
signal for recommendation on YouTube is what's called co watch,
and it says that we make there's a big database
that says here's everybody and what they watched. And you
just look and say, people who watch this also watch this. Yes.
And you know, Amazon's pretty famous for a similar technique.
It's often called collaborative filter.
Speaker 1 (20:47):
In and slow realization that none of us is a
unique and beautiful snowflake, like you watch the you watch
these three videos and someone else did, you're both going
to watch the same fourth.
Speaker 3 (20:55):
You're going to watch the same fourth And it turns
out to be usually quite right. And you know, if
you go and you say you can now put on
a bunch of other signals of like people who are
like you in some other way may also watch something.
But it turns out like the core of a recommendations
engine like that is actually quite simple.
Speaker 1 (21:12):
Every episode now we've got an AI question, it's either
a is AI gonna help me make more money?
Speaker 2 (21:16):
Or is it going to put me out of business?
Speaker 1 (21:18):
So this one is just SORIYAH in Irvine, California wants
to know how is.
Speaker 3 (21:23):
AI impacting your companies and how is it impacting your products.
I'm going to come back to CODA in a second,
but I think it's probably worth just talking a little
bit about this wave of AI and why it's different
that I mentioned earlier. My dad's view was, you know,
anything that's labeled AI is the stuff that doesn't work.
Is the moment you understand how it works is no
longer called the I. All of that flipped and the
(21:45):
last ten years we've seen a wave of innovation that
stayed within AI, and so now we've got a much
much broader set of impact happening with the AI advances today.
The core one that everybod's excited about today is what
we've now referred to as generative AI, and it was
the core insight was started with a research project at
(22:09):
Google called Transformers, and again back to the idea that
some of these things are simpler than you might think.
So the Transformers paper came out of the group that
focuses on translation. And so if you think about translation
as a problem, so the interesting thing that is one
of the most valuable but also simplest problems to think
(22:31):
of in what we can do with natural language AI
is I'm going to take something that's written in English
and I want to turn into French. So what's the
best way to do that. You go read the Internet,
and you go find every web page that has been
translated between different languages, and you say, here it is
in English, here is in French, here is in Spanish,
heres in Japanese. On and you start to figure out
(22:53):
that what the powern's on. And the idea of Transformers
was they learned one very special trick was that they
could teach this machine to train itself by guessing the
next word in the sentence. So it would say, here's
the first ten words, guess the eleventh word, and so
for example, it would say these are the first ten
(23:13):
words in English. These are the first ten words in French.
Now guess the eleventh word in French, and it would
go quickly around the internet and say, ooh, I saw
those ten words over there. The eleventh word was this one.
And so I'm going to assume that the eleventh ward
in French is the same from that translated page and
dramatically oversimplifying what it has to do at a huge scale.
Back to the earlier point. But this idea of predict
(23:36):
the next word turns out to be you know this
I co watch signal is a really simple idea of
predicted next word is a really simple idea. Doing it
scale is really, really, really hard, and they figured out
the heart of the paper was figuring out how to
do it at scale. Now I think for most of
us this caused an eye to shift gears in one
(23:56):
very specific way. It went from a thing behind the
curtain to in front of the curtain and up till now.
AI was one of those things that like fancy engineers do.
It was one of these things that you couldn't you
could see the output of it, like you got better
YouTube recommendations and Spotify gave you better songs and so on,
but you didn't feel like you could interact with it.
(24:17):
And then chat GPT came out and all of a sudden,
we all felt like we could talk to this thing.
And so I like to say this way of AI
had actually two innovations. They had a fundamental AI innovation
with transformers and being able particular the next work, but
also has a UI innovation in that we all suddenly
felt like this is no longer product for developers as
(24:37):
a product for all.
Speaker 2 (24:38):
Right, Finally, AI had a good publicist, a I.
Speaker 3 (24:40):
Had a good publicist, right, and and the and that's
been you know, obviously you'd have to be living under
a rock to not not have seen that. The h
So that I think it's it's a much bigger change
than we've seen in the past with AI, and it
has caused every product in every company to rethink its approach.
So for CODA, you know, I think that there's a
(25:03):
few different layers of how we think about it. The
most basic layer is code of the documents. We have
to be able to do what we call be a
great writing assistant. So if the heart of this AI
was built around help you predict the next word, this
thing should be able to help you write, and certainly
the way that AI shows up in CODA first is
a writing assistant that helps you finish your sentences. You
can write your whole page for you. You can go and
(25:24):
check all your errors, go and check grammar and so on.
I kind of think of it as like, you know,
the next generation of spell Check, And I don't mean
that in a negative way. I mean that as a
it'd be crazy to have a writing surface these days
that doesn't doesn't have a writing assistant. And actually back
to the point, like we none of it would call
(25:45):
Spellcheck AI, but when spell Check came out, it was
absolutely AI. I mean, is the you know, the core
idea of like help me figure out how to present
myself better? Was what it felt like an AI promise.
They can kind of understand how it works, we can
call it AI. So that's the basic thing we do.
But there's two other things we do that I think
are maybe less intuitive. So the first is we realize
(26:08):
that AI in the workplace has an extra challenge, and
that is that I don't want something that just can
answer questions from the Internet. I need something that can
answer questions from my work and it turns out that
CODA is a particularly good place to do that. So
one of the features of CODA is that it's an
all one surface. We blend all these different surfaces together
(26:30):
in one place. We can also synchronize with every different
application in your company. It's about six hundred different we
call them packs, and six hundred different connectors that have been
built to connect CODA from everything from Slack and email
to Salesforce and Gira, and you can pull all that
into your documents. And so the first thing we did
was we said, we want to build an assistant that
works inside of your the core surface aer in all
(26:51):
day long, inside of your documents and actually understands your data.
And so this is that we spend a lot of
energy on integrating our integrations with our AI system. And
then the second thing we did is we said, it's
one thing to be able to ask questions. And it's
great to be able to come in this thing and
ask a question and say, you know what's not only
(27:13):
like you know what's the capital of France, but you
know who's the CEO of this company that I'm about
to talk to, and what's the most important meeting on
my schedule today or so on. So you can ask
all those types of questions. But the other half that
we want to be able to do is actually take action,
and so we put a lot of energy into making
an AI system that knows your actual work but also
(27:34):
is capable of performing real actions for you. What Open
Eye did is we moved from AI being for developers
AI being for users. We're trying really hard to make
AI product for makers, these people that live in the
middle of that world. So I don't just want to
chat with something and get answers. I got work to do.
I needed to know what I'm doing. I need to
do stuff for me. That's what we've been doing.
Speaker 2 (27:53):
So we've for our listeners who've heard previous podcasts, so
we spend a lot of time with Perkins Miller a
couple weeks.
Speaker 1 (28:00):
Ago on how the AI tools are also impacting productivity internally.
I've talked about how Copilot is helping engineers move faster,
how it's helping content creations teams make content a lot faster,
and I'll just I'll only reiterate here. The advice we've
gotten in the past is, as professionals, if you're building products,
I'm sure you're out there thinking a lot about how
(28:21):
AI feeds into your product. A lot of folks are
not thinking about how AI should be helping them do
a better job at their job. And I'm strongly encouraging
everyone get the add ons forh your browser. Spend a
lot of time in chat GBT, it won't be useful
to you the first ten times. You've got to start
to find the right way to prompt it and the
right way to put it to work. But you should
be able to shave ten hours off of your week
(28:42):
every week by.
Speaker 2 (28:43):
Using these tools properly, for sure. For sure, Oscar in
Saint Louis asks.
Speaker 4 (28:59):
At the stages when companies are moving fast and things
are very dynamic, how do you organize the team and
keep everyone aligned?
Speaker 1 (29:07):
And this just share is where you're big on the
rituals and behaviors in the culture of the organization itself.
Give everybody your overall thesis here and a couple of
concrete examples.
Speaker 3 (29:19):
Yeah, so maybe as context for everyone, I'm writing a book.
It's called Rachel's a Great Teams.
Speaker 2 (29:26):
And it's kind of a when's it coming.
Speaker 3 (29:28):
Out people, hopefully early next year?
Speaker 2 (29:32):
So do I sound like your publisher.
Speaker 3 (29:34):
When do we get into Yeah, exactly exactly. That said,
if you go to Richel's Great Teams dot com. I'm
writing it in the open so you can go. I
publish a chapter to time so you can go see
what they what they look like, and I'm about sixty
percent of it or so is available for people to read.
They actually kind of fund A couple of other people
are in their co editing with me, which is kind
of a nice way to to write a book. Anyway, back,
(29:58):
so the thesis of this book started with the conversation
I had with a friend of mine being Gordon and
people who don't know who Bing is. He is one
of the founders or the chief creative officer at Electronic
Arts and is now famous investor Zinga, Amazon, many great companies.
He and I sat on a board together and he
(30:18):
kept harassing the CEO with this question. He said, what
are your golden rituals? And the CEO kind of struggled
to ask a little bit, and at one point one
of us said, hey, Bing, what what's the golden ritual?
And he gave us this really clear rubric. He said,
great companies are the smallest of golden rituals. They have
three criteria. Number one they're named, number two, every employee
(30:41):
knows them by their first Friday, and number three they're templated.
And he immediately rattled off his examples. Amazon had six pagers,
Google has OKR, Salesforce's V two mom so and it
just turned out to be this really sticky idea for me,
this idea of ritual So this happened right before the
pandemic started. I started getting I started talking about this
(31:02):
on on a few podcasts and talked about it talking
to some customers, and people started sending me rituals and say, oh,
if you're really interested in this topic, you should hear
more about this. And it's become this like weird hobby
of mine. Like people have all sorts of normal hobby
life and normal hobbies too, but this has become my
like nights and weekends hobby is collecting rituals. And one
of the key things we did was we started doing
this dinner series. So every about three weeks, starting at
(31:26):
the beginning of the pandemic, we would get a group
of ten to fifteen people together and say just share
your rituals. And people from small companies, big companies, you know,
all sorts of different disciplines and functions and and and
so on. And at this point we've taken about a
thousand different rituals. We whirled them into about one hundred
of my favorite ones. And that's the that's what this
book is about. Rituals are great teams. And there's a
(31:48):
couple of interesting observations about this. You know, one the
first thing I would share is that, you know, one
of my favorite Stortes Darmesia, is the founder of HubSpot.
He comes to one whose dinners and he shared a
ritual called Flashtacks. It's kind of fun ritual. You can
go read more about it. But then you start getting
really excited about rituals. And he gave this analogy. Real
(32:11):
like he said, as a company, we build two products.
We build one for our customers, we build another one
for our employees. The one we build for our employees
we often call that culture. But if you ask somebody
to describe culture, they will often describe it with rituals.
They'll say, oh, we have this value, here's how it works.
And so from that perspective, rituals aren't just the operating
(32:32):
system of how you get stuff done, and how you
keep your team organized, keep everybody lying, and so on.
It's actually a two way mirror to your culture.
Speaker 1 (32:39):
I couldn't agree more with the thesis of the golden
rituals I've seen at work at my company too. I
have not read the first sixty percent of your book,
but I will for our listeners. I'll drop that link
in the show notes. You also referenced your Iigen questions,
which we're not going to be able to get into today's.
It's been so much good stuff we're going to drop
that in the show notes as well. There's been to
share a ton of amazing content today. I knew this
(33:02):
would happen getting together with you, and this has been
a blast.
Speaker 2 (33:06):
Thank you by both.
Speaker 3 (33:07):
Mike, Yeah, thanks for having me.
Speaker 2 (33:09):
This was terrific.
Speaker 1 (33:09):
Congrats on all the progress you've made with CODA. For
folks who haven't tried it, it's really phenomenal software. It's
not only makes it easier to create your content, you
actually get to better answers because of the tools that
you share in his team are building and dropping into
the show notes and check it out.
Speaker 2 (33:25):
I think you'll really enjoy it and you share. Thank you, buddy,
I'll see you again soon.
Speaker 3 (33:30):
Thank Mike.
Speaker 1 (33:37):
You know, Gagan, one of the things that I've observed
about just about all of our conversations is that the
folks who come in office hours have not only achieved
success through building great products or bringing great products or
services to market, They've also been extremely thoughtful about the
systems and processes and cultures that they create in their teams,
(34:01):
or in their companies, or even in their own lives.
And you'd be hard pressed to find someone who's as
good as Shasher is at creating these processes and as
you called it today, these golden rituals that helped to
organize a team, that helped to give everyone a shared
language and shared tools for solving problems.
Speaker 2 (34:24):
And if I leave you to.
Speaker 1 (34:24):
Think about something, what are those golden rituals in your organization?
And what are they for you in your life or
in your family. I know we have some that keep
my kids organized on the weekends. I know I personally
have some to make sure that my health and my
fitness get as much attention as my development.
Speaker 2 (34:46):
And my work and everything else.
Speaker 1 (34:48):
I'm going to link these in the show notes and
encourage you to check it out and do come back
and ask yourself, what should these rituals be in my
personal life and for my team.
Speaker 2 (34:58):
I think that you will find that you're.
Speaker 1 (35:00):
Able to unlock some productivity, some better alignment, some better
velocity in the things that you do. I want to
thank Youshier for coming on the pod. It's great to
see my friend again. And of course I want to
thank Jen, Cara, Jada, Meg Matt, the whole team at
Blue Duck Media for pulling this all together. I want
to thank Dylan and Sasha, Gay, Nathan and Christine at
(35:21):
iHeart and Ben and the team at William Morrison Devor
for all their support. Office Hours is a production of
Blue Duck Media and distributed by iHeart Radio.
Speaker 2 (35:30):
I will see you next week. Everybody, make sure you
stay on your grind.
Welcome to Office Hours, where we sit down with the
chief executive shaping the world and answer your most pressing
questions about leadership, career, and life. Mike Steib and today
we are hanging out with my good friend Shashier Morotra.
Shasher is the founder and CEO of Coda, the venture
back startup that has created an amazing alternative to word
in Google Docs. He and I worked together previously at Google,
(00:27):
where Shashier ran product engineering and UX for YouTube. He
has previous experience at Microsoft and as an entrepreneur, sits
on the board of Spotify and as a thought leader
on how teams work and collaborate. He's also an all
around terrific guy and Shasher, I am thrilled to have
you on the show today.
Speaker 2 (00:47):
Welcome aboard.
Speaker 3 (00:50):
Thanks for having me, Mike. It's been so much fun
watching I think our journey started almost fifteen years ago
together and lots of great experiences in between.
Speaker 1 (00:59):
I'm glad you're here. As you know, our listener's sending questions.
There's stuff that they want to hear from you. And
first question is from Nina and Tucson, Arizona.
Speaker 3 (01:08):
She says you've both worked at some of the biggest
tech companies and now lead earlier stage companies.
Speaker 2 (01:14):
What have you found is similar?
Speaker 3 (01:15):
In what is different when the size of the company changes.
Speaker 2 (01:18):
And you share.
Speaker 1 (01:19):
Maybe just tell everybody quickly sort of your background. I
met you when you were coming out of Microsoft and
into Google, just so everybody has a good sense for
what you're all about.
Speaker 3 (01:28):
Yeah, I mean, I've definitely done the full gamut of
that big, huge company and the start from scratch in
just a room with a few people. So my career
I started out of college. I started a company called
some Trata, those in the data center automation space. I
then went to Microsoft. I spent a number of years
(01:50):
working on Windows, then Office and SQL server, and then
in a weird set of circumstances, I ended up at
Google working with my on television video. I ended up
running the YouTube group there from about twenty eight twenty fourteen,
really exciting time. And then I started Coda. And Coda
(02:12):
we started as a new company to reimagine the document.
We built an all on one document on the best
parts of document spreadsheets, presentations, and applications into a single surface,
so anyone can make a dot as powerful as an app.
Speaker 2 (02:27):
Tell everybody like, what is it about these companies that
are growing and changing the world and then something changes
and what have you seen in that part of your journey?
Speaker 3 (02:35):
So I joined Google in two thousand and eight, and
at the time I was I'd done to Microsoft. I
thought of Microsoft a little bit as my big company
education and my initial reaction to all the Google outreach was, yeah,
I'm not interested. I'm already at a big company. I'm
going to go do something small. I had all these
ideas in mind of what to go do, and as
(02:56):
guy Nam Jonathan Rosenberg was running product at Google at
the time, and he gave me analogy that the Hues gave,
which is that Google in two thousand and eight felt
like Microsoft in the in the early nineties. And I
think that was accurate that you know, at the time,
Google was really a one product company, just learning how
to take its culture and apply it across other markets
(03:17):
and divisions and categories and so on. At the time,
you know, it was just really searching ats, and you know,
we hadn't done you know, it just wought YouTube. We
had you know, started maps. As Chrome was about to
the launch Android, it was just still an early prototype
and all these things that now we take for granted
as being stills at Google weren't there.
Speaker 2 (03:37):
And now it wasn't a small company.
Speaker 1 (03:38):
This was post ipo, multi multi billion dollars of revenue,
tens of thousands of employees. But I remember to share
what I had come from a from a big, hundred
year old company, and when I came to Google, we
needed we needed fifty million dollars for this deal that
we wanted to do to jumpstart a business that had
no revenue. And so I was sort of imagining what
the bureaucratic process would be. And it was just going
(04:00):
to see Eric and Larry and Sergei on a Thursday,
and we got a yes in the meeting and I
was like, well, what's the follow up?
Speaker 2 (04:07):
And they were like, will you just spend it?
Speaker 3 (04:09):
Now?
Speaker 1 (04:10):
It's I couldn't believe the speed with which the company
was willing to work and the absence of bureaucracy and
middle management.
Speaker 2 (04:18):
It was really wild.
Speaker 3 (04:19):
Unfortunately, I don't think Google maintained that forever, but in
that period I think it was about as good as
it could get in terms. And I think Eric is
a wizard at managing chaos like that, and he was
just so good. I mean it said a similar story.
I came in and said, I need twenty people to
work on this and said, you know, do I go
to some a Google body, do I get it? I
(04:40):
do I need to get some you know, some official thing.
And I was told no, if you can convince twenty
people to work on it, then you have twenty people.
And it was like this interesting meocratic like there's just
sort of this assumption that if you screw up, those
other consequences and people will stop working on it and
so on. But there was you know, at the time,
(05:02):
Google's whole ethos was feed of innovation was all that mattered,
and Eric was excellent at you know, supporting people on it.
That was really wonderful.
Speaker 1 (05:11):
Everything came from a place of yes, but there's certainly consequences.
I remember we had a client event one time and
I ordered a mechanical bull. And when there's no process
in bureaucracy, there's no one to stop you from bringing
the mechanical bull into the office. So we've had clients
drinking and riding a mechanical bull, and in hindsight, like
a lot could have gone wrong.
Speaker 3 (05:32):
Yeah, yeah, I mean, it's definitely I don't know if
you could run a company the same way today as
as we did back then. But it's I mean, there's
a lot to be you know, the first era of
Google was this magic in a bottle search engine that
that that turned out to be, you know, really amazing.
But many companies never make it past that face, never
(05:54):
make it into multiple categories and multiple new products and
so on. And Google managed to make it through that casm,
which I think is remarkable.
Speaker 1 (06:01):
Right and was and you can also, you know, bureaucracy
emerges as a way to maintain you know, to avoid
risk because risk is too expensive. But we were a
company doing probably twenty million dollars a free cash flow
a day when you arrived, and there's a lot of
room to there's a lot of room to make mistakes
and be able to pay for those mistakes rather than
pay for them with added layers of bureaucracy. And yeah,
(06:24):
you know, and slowness and poor culture and all that stuff.
Speaker 3 (06:27):
Yeah, for sure, maybe I'll try to answer Nina's question.
I mean, I think when I think about and I
don't know if I would say there's large companies, there's
small companies. There's also small projects within large companies. I think,
you know, some of the stuff you and I worked
on together in the video space was you know, had
some advantages of being at the big company, like like
you mentioned, it was easy approvals, but we were really
(06:48):
you know, a typically.
Speaker 2 (06:50):
We're still starting out, that's right, Yeah.
Speaker 3 (06:51):
I mean you still had to define a market and
find a customer and and so on. And you know,
the best analogy I can give is, I had an
old boss at Microsoft give me this framework for thinking
about the stages that products go through. And you called
a joke, threat, obvious. Every product starts as a joke,
nobody believes in you, nobody thinks it's for real. Then
(07:13):
you become a threat, and everybody responds. They've got their
you know, they've got their response features or their response
go to market campaign or their response billboards, whatever it is.
And then at some point you become obvious. And when
you're obvious, everybody just kind of presumes you're going to
win and continue to win and they work around you.
Speaker 4 (07:30):
Right.
Speaker 3 (07:31):
And add I got to be at Google in particularly
YouTube through all three phases. When I showed up, it's
I don't think I can understate how much people thought
YouTube was Google's first big mistake. And we paid one
point six billion dollars for this company. We were losing
hundreds of millions of dollars a year. We were losing
(07:52):
you know, almost pennies per view.
Speaker 2 (07:54):
The only good content was content that they were not
allowed to have.
Speaker 3 (07:57):
There was we had a feeling somebody else there's.
Speaker 2 (08:01):
A huge Viacom lawsuit.
Speaker 3 (08:02):
Yeah, and so we we get this group together and
it was totally you know, you hear the analogy a
lot of pirates that navy. But you get this group
together that say, you know what, I think we can
do this. I think there's a I think there's a path.
And then at some point, I think about three or
four years in all that started working. We got profitable,
we figured out how to Google revenue, We got through
(08:24):
our agreements with the with the recording industry and with
the movie industry and so on, and all of a
sudden it became threat and everybody had their response to YouTube.
And then at some point it switched to being obvious
and and the and you know a lot of people
think of that as like really awesome, and it is.
I mean, this it's obviously the goal. It's like you're
(08:45):
building this company. You want to get to that obvious dage,
but there's downsides too. And when's the last time you
read a positive, positive press article about YouTube? I mean
the early days, we couldn't get anybody to write about
us at all, and then all of a sudden, now
all you can see is negative that negative things because
it's you know, these are sort of presumption that, of
course this thing is going to continue to be to
be big and useful and so on, So people will
spend all their time looking for the flat. So there's
(09:07):
lots of pros and cons of each. But I think
learning how to operate in all three phases, you know,
what do you do when the when the world thinks
your joke? How do you respond when they decide you're
a threat? And then what do you do to manage obvious?
And I think Google did a really good job of
taking that obvious stage and starting a bunch of jokes
and going and working its way back through that cycle.
Speaker 2 (09:28):
Well, this this actually probably is really well.
Speaker 1 (09:30):
In our second question, so Ian and Burlington Vermont asks us, Hello, can.
Speaker 3 (09:36):
You talk about your products and how you knew when
they had a product market fit?
Speaker 1 (09:40):
So Shashir here's where I admit to you and you
told me that you were going to do Google docs,
but they're also spreadsheets inside of it and other stuff.
Speaker 2 (09:46):
I was like, man, we.
Speaker 1 (09:48):
See, we see working on over there. So I originally
got the joke. I have come full cycle. Now I'm
a big fan of the product.
Speaker 2 (09:56):
But take take us through your your joke joke to
threat matrix.
Speaker 3 (10:01):
Yeah, yeah, I mean I think the Yeah, I mean
the early days of CODEA, I got that reaction from everybody.
They were building a new dock and most people would say,
why do we need a new doc? And I would
get all sorts of weird feature requests and people would say, oh, yeah,
that's really exciting. You know, word is missing this one
font that I really want, And I say, no, no, that's
not what we're working on. That's like, I don't think
(10:22):
there's a category there of like the font that's missing when.
Speaker 1 (10:25):
I try to download it as a PDF to print, Like,
I don't like that.
Speaker 2 (10:28):
It's two steps.
Speaker 3 (10:29):
Right, right exactly? You know, either get the why would
you bother? Or a list of because people can't picture
it right, It's like what what does what does it
mean to have a brand new product? In this category.
Speaker 2 (10:41):
And you know, everybody wanted a faster horse.
Speaker 3 (10:44):
Everybody want a faster horse, and the and and honestly,
like that sounds, you know, in retrospect sounds positive. Nobody
could picture it. We could that worked out well, but
you know the time, it just seems terrible. Right, people
look at your product and they and they just don't
know what they what they want. And you know, so
the early days of product market fit for us are
really tough. And Coda is this interesting product where we
(11:06):
often describe it as we have an incredibly low floored
and incredibly high ceiling. So if you download sign up
for Coda and you start using the app or use
the website, you'll it looks like a Google Doctor, the
blinking cursor, blank screen, just start typing and so on.
But it's got all the building blocks inside of it,
so you can assemble the things that you normally would
(11:27):
have had to use spreadsheets or presentation for, or things
that you needed to build full applications. And we see
people build you know, whole CRM systems and code our
inventory systems, and lots of tutors and small businesses use
it to build a clients. And there's all sorts of
instant things to get built, but it's this product that
looks like one thing and then operates in this totally
(11:48):
different way as it grows with you. And so early
on we would see people first off just not know
where to start and say I don't I don't understand
like I'm used to use to this other other tool set,
and then gradually we would get people through this journey.
I started our first customer. We decided there was another
side story we can come back to. So we decided
(12:09):
to start the company in stealth, which is not something
I recommend for every company, but it made sense for us.
So we weren't telling anybody what we were doing, and
so we were recruiting customers one by one. We recruit
this customer. It's an old friend of mine from Google
who had started a new company. There were five or
six people, and they said, hey, can you use CODA
to run your team? And so they start using it.
(12:30):
They ended up using their first use case with basically
a project management use case. They sort of planned out
their work in it, and we would watch our dashboard
and our dashboard they only six people in the company,
so our dashboard would go from zero to six, like
every day. It's like, how many of the six people
use the product? And I was like, that was it.
That was like the max Yx was six and one
day this goes from six to zero and we're like,
(12:51):
we're a very daily used product, Like it's hard to
use code and not use it every single day, and
so it grows to zero and like, ooh, that's not good.
And we wait another day it still says it's zero,
and I called my buddy, the CEO NOME, and I say, hey,
you know, I just wanted to check in how things
are going. And uh, you know, I noticed that you
guys haven't been using the product much the last couple
of days. Want to notice something happened or you guys
(13:13):
like doing an off site or something. And he says, yeah,
you know, I've been kind of avoiding calling you. And
he says I have good news and bad news and
I said, okay, well give it to me. Let's do
bad news first. So all the bad news is we
have a team meeting and the team said that if
I make them keep using CODA, they're all going to quit.
And I was like, well that's that's pretty bad news.
(13:36):
Feedback is a gift, like, that's pretty bad news. That
what's the good news And he says, well, the good
news is they've all fallen completely in love with your
mission and they totally get it. But they have this
list of things that you have to do before they
will consider reusing the product. And honestly, if you can
fix these in like two weeks, we'll start again. But
you know, up to you, this is good news and
(13:57):
bad news. And it was just like really like, you know,
first off, big kick in the stomach because you're like
your entire user base of apperates in one day and
then you've got this like deep conflict on you know,
do we do this set of things or go address
the waves on a roadmap or someone. But you know,
in my mind, for our product, product market fit felt
(14:19):
like we could tell, I mean, this is not a
product where you could force anybody to use it. There
was no like Lasadaisical users like you either used it,
loved it and evangelized it, or you gave up. But
the real moment for us when we started getting calls
that said, hey, I know you have me limited to
my team, can I expand it to this other team?
And there was in particular a moment with our first
(14:42):
lord customer was Uber, and it was a moment where
our main sponsor's guy named ukiyamasheet at ukis now they
had a product or Figma and carry coded there as well.
But he called up and he said, you know, I've
been assigned this project. I'm going to This was back
in when he eighteen. He had been assigned the project
to fix Uber's reputation with drivers. You may remember the
(15:05):
time period, A bunch of a bunch of issues with
drivers and and and uh. His job was to roll
out a feature a day to addressed everything drivers that
ever asked for. And he said, he said to me,
we've been using code for a bit, and I'm quite
convinced that if we don't have CODA we will fail.
And we do have CODA, then we'll succeed. And and
(15:26):
it was like this sign of like, oh we are
We've crossed into the pull, the feeling of people want
the product, that they're ready to evangelize it. And so
that that was that was our and that was probably our.
Speaker 2 (15:40):
That was your product market, that was your product market.
We're demanding it.
Speaker 3 (15:45):
Yeah.
Speaker 1 (15:45):
I went from threatening to quit to demanding the product
exactly so for for Ian and in my business, it
was much We got a bunch of Artsy is a
marketplace for art, and before I got to the company,
we've gotten tons of galleries around the world to upload
art to Artsy, and we've gotten tons of like millions
of people to come visit arts, see and look at
that art.
Speaker 2 (16:05):
We weren't selling it. So for us, product.
Speaker 1 (16:08):
Market fit really happened when someone was paying us at
a at unit economics that were scalable and would take
our business to profitability. And what we had to do
to get there was build globally a layer of transactional
capabilities so that you could not only find the art
and this was the hard part in the tech stack.
You could not only find the art, you could with
(16:29):
confidence click a button, buy the art and have it
show up in your home a week later. And then
all when the painting started showing up in people's homes
thanks to the products we had built, that's when we
really that's when we really had product market fit. And
it's just share as I'm sure you'd advise, it's it's
a little bit different for every company, but it's either
some moment when the users will go to the end
of the earth for it, right, It's not when somebody
says they like a thing. That's when you say, if
(16:51):
I took it away, how would you feel? Or it's
when someone opens up their wallet and says, you know,
this is so important to me that I'll I'll pay
you for it. M Devin in Rally, North Carolina says.
Speaker 4 (17:13):
I'm not a tech person and I would just love
to understand how all this stuff works, like the recommendations
on YouTube and in my music app, Like, can you
just take me through what's like under the hood of
these products?
Speaker 2 (17:24):
So just here you go.
Speaker 1 (17:25):
First, you've seen some really yeah, I mean you've been
on the end under the hood and some really awesome products.
Help somebody who doesn't code get an appreciation for what
is the thing doing when they're having a great user experience?
Speaker 3 (17:36):
Yeah, And it's interesting when I when I got to YouTube,
the uh you know, I meet people regularly and say
work on YouTube, and they would say, what do you mean?
Like you create content? Like what does that? What does
that even mean? Because I think people's perception of these
products is like what is behind the thing? You see
the recommendation, you see the video that's playing and so on,
(17:59):
it's like hard to tell what is what is all
these what are all the things that are happening there?
And so first off, there's there's a.
Speaker 1 (18:05):
Way which is buy by the way, non tech companies
are always like, well, we'll just hire some tech people. Yes,
we already have the market share, right, we already have
the brand. We'll just hire some tech people. You're like, okay,
good luck.
Speaker 4 (18:16):
Yeah.
Speaker 3 (18:16):
And I think it's hard, right, I mean I think
the feeling of like, it's a website, it plays video,
How hard could it be. It's like, well, there's thousands
of engineers sitting behind it building every element of what
happens there, and you know, really hard to like, we well,
I with the YouTube are are our benchmark? Said we
streamed about twenty percent of of the bits on the
Internet came to the YouTube servers. So it's like one
(18:38):
thing that can you can you stream us and a
video file to someone else across the world, Like maybe
can you do it and fill up twenty percent of
the Internet? Like now you've got a whole different scale
of things you look at. You know, YouTube was and
still is the number two search engine in the world
behind Google, and so you know, how do you answer
queries for people at scales? Like another great engineering challenge
(18:59):
AI and recommendation some one is A is a particularly
interesting topic the and I often tell people that there's
less magic than you might think. And you know, one
of my one of my favorite stories about this is
my my father is a computer scientist and I got
to college and I told them, hey, I'm going to
go to a internship at at on my TV had
(19:23):
this book called the I Lab, and he said, you
can't do anything at the I Lab. That's where all
the jokers hang out. And it's like, really, what does
that mean? And he said, well, you know AI, this
is back in the nineties. He says, AI is like
that's where all the stuff that doesn't work hangs out.
And he had a particular way of describe me, says,
just think about it everything. The moment it works, we
(19:43):
no longer think it's AI. And in those days, like
you know, I'll give like a simple example I would
give my kids is, you know, you're driving along, you
come to a traffic light. How does it know when
to turn right or cree? And you know, it's kind
of obviously because a censor underneath the underneath the pavement
and it says is there a car here or not?
(20:04):
And it decides on whether the turnout or green. And
because I can explain it, it doesn't seem like AI.
But if I just told you that I've built this
totally amazing system that automatically rops traffic through through an area, like,
it feels like AI. And so a lot of AI
is actually that it's actually things that when you explain
exactly what they do, they aren't really that complicated. So
as an example, you know, recommendations on YouTube. The predominant
(20:29):
signal for recommendation on YouTube is what's called co watch,
and it says that we make there's a big database
that says here's everybody and what they watched. And you
just look and say, people who watch this also watch this. Yes.
And you know, Amazon's pretty famous for a similar technique.
It's often called collaborative filter.
Speaker 1 (20:47):
In and slow realization that none of us is a
unique and beautiful snowflake, like you watch the you watch
these three videos and someone else did, you're both going
to watch the same fourth.
Speaker 3 (20:55):
You're going to watch the same fourth And it turns
out to be usually quite right. And you know, if
you go and you say you can now put on
a bunch of other signals of like people who are
like you in some other way may also watch something.
But it turns out like the core of a recommendations
engine like that is actually quite simple.
Speaker 1 (21:12):
Every episode now we've got an AI question, it's either
a is AI gonna help me make more money?
Speaker 2 (21:16):
Or is it going to put me out of business?
Speaker 1 (21:18):
So this one is just SORIYAH in Irvine, California wants
to know how is.
Speaker 3 (21:23):
AI impacting your companies and how is it impacting your products.
I'm going to come back to CODA in a second,
but I think it's probably worth just talking a little
bit about this wave of AI and why it's different
that I mentioned earlier. My dad's view was, you know,
anything that's labeled AI is the stuff that doesn't work.
Is the moment you understand how it works is no
longer called the I. All of that flipped and the
(21:45):
last ten years we've seen a wave of innovation that
stayed within AI, and so now we've got a much
much broader set of impact happening with the AI advances today.
The core one that everybod's excited about today is what
we've now referred to as generative AI, and it was
the core insight was started with a research project at
(22:09):
Google called Transformers, and again back to the idea that
some of these things are simpler than you might think.
So the Transformers paper came out of the group that
focuses on translation. And so if you think about translation
as a problem, so the interesting thing that is one
of the most valuable but also simplest problems to think
(22:31):
of in what we can do with natural language AI
is I'm going to take something that's written in English
and I want to turn into French. So what's the
best way to do that. You go read the Internet,
and you go find every web page that has been
translated between different languages, and you say, here it is
in English, here is in French, here is in Spanish,
heres in Japanese. On and you start to figure out
(22:53):
that what the powern's on. And the idea of Transformers
was they learned one very special trick was that they
could teach this machine to train itself by guessing the
next word in the sentence. So it would say, here's
the first ten words, guess the eleventh word, and so
for example, it would say these are the first ten
(23:13):
words in English. These are the first ten words in French.
Now guess the eleventh word in French, and it would
go quickly around the internet and say, ooh, I saw
those ten words over there. The eleventh word was this one.
And so I'm going to assume that the eleventh ward
in French is the same from that translated page and
dramatically oversimplifying what it has to do at a huge scale.
Back to the earlier point. But this idea of predict
(23:36):
the next word turns out to be you know this
I co watch signal is a really simple idea of
predicted next word is a really simple idea. Doing it
scale is really, really, really hard, and they figured out
the heart of the paper was figuring out how to
do it at scale. Now I think for most of
us this caused an eye to shift gears in one
(23:56):
very specific way. It went from a thing behind the
curtain to in front of the curtain and up till now.
AI was one of those things that like fancy engineers do.
It was one of these things that you couldn't you
could see the output of it, like you got better
YouTube recommendations and Spotify gave you better songs and so on,
but you didn't feel like you could interact with it.
(24:17):
And then chat GPT came out and all of a sudden,
we all felt like we could talk to this thing.
And so I like to say this way of AI
had actually two innovations. They had a fundamental AI innovation
with transformers and being able particular the next work, but
also has a UI innovation in that we all suddenly
felt like this is no longer product for developers as
(24:37):
a product for all.
Speaker 2 (24:38):
Right, Finally, AI had a good publicist, a I.
Speaker 3 (24:40):
Had a good publicist, right, and and the and that's
been you know, obviously you'd have to be living under
a rock to not not have seen that. The h
So that I think it's it's a much bigger change
than we've seen in the past with AI, and it
has caused every product in every company to rethink its approach.
So for CODA, you know, I think that there's a
(25:03):
few different layers of how we think about it. The
most basic layer is code of the documents. We have
to be able to do what we call be a
great writing assistant. So if the heart of this AI
was built around help you predict the next word, this
thing should be able to help you write, and certainly
the way that AI shows up in CODA first is
a writing assistant that helps you finish your sentences. You
can write your whole page for you. You can go and
(25:24):
check all your errors, go and check grammar and so on.
I kind of think of it as like, you know,
the next generation of spell Check, And I don't mean
that in a negative way. I mean that as a
it'd be crazy to have a writing surface these days
that doesn't doesn't have a writing assistant. And actually back
to the point, like we none of it would call
(25:45):
Spellcheck AI, but when spell Check came out, it was
absolutely AI. I mean, is the you know, the core
idea of like help me figure out how to present
myself better? Was what it felt like an AI promise.
They can kind of understand how it works, we can
call it AI. So that's the basic thing we do.
But there's two other things we do that I think
are maybe less intuitive. So the first is we realize
(26:08):
that AI in the workplace has an extra challenge, and
that is that I don't want something that just can
answer questions from the Internet. I need something that can
answer questions from my work and it turns out that
CODA is a particularly good place to do that. So
one of the features of CODA is that it's an
all one surface. We blend all these different surfaces together
(26:30):
in one place. We can also synchronize with every different
application in your company. It's about six hundred different we
call them packs, and six hundred different connectors that have been
built to connect CODA from everything from Slack and email
to Salesforce and Gira, and you can pull all that
into your documents. And so the first thing we did
was we said, we want to build an assistant that
works inside of your the core surface aer in all
(26:51):
day long, inside of your documents and actually understands your data.
And so this is that we spend a lot of
energy on integrating our integrations with our AI system. And
then the second thing we did is we said, it's
one thing to be able to ask questions. And it's
great to be able to come in this thing and
ask a question and say, you know what's not only
(27:13):
like you know what's the capital of France, but you
know who's the CEO of this company that I'm about
to talk to, and what's the most important meeting on
my schedule today or so on. So you can ask
all those types of questions. But the other half that
we want to be able to do is actually take action,
and so we put a lot of energy into making
an AI system that knows your actual work but also
(27:34):
is capable of performing real actions for you. What Open
Eye did is we moved from AI being for developers
AI being for users. We're trying really hard to make
AI product for makers, these people that live in the
middle of that world. So I don't just want to
chat with something and get answers. I got work to do.
I needed to know what I'm doing. I need to
do stuff for me. That's what we've been doing.
Speaker 2 (27:53):
So we've for our listeners who've heard previous podcasts, so
we spend a lot of time with Perkins Miller a
couple weeks.
Speaker 1 (28:00):
Ago on how the AI tools are also impacting productivity internally.
I've talked about how Copilot is helping engineers move faster,
how it's helping content creations teams make content a lot faster,
and I'll just I'll only reiterate here. The advice we've
gotten in the past is, as professionals, if you're building products,
I'm sure you're out there thinking a lot about how
(28:21):
AI feeds into your product. A lot of folks are
not thinking about how AI should be helping them do
a better job at their job. And I'm strongly encouraging
everyone get the add ons forh your browser. Spend a
lot of time in chat GBT, it won't be useful
to you the first ten times. You've got to start
to find the right way to prompt it and the
right way to put it to work. But you should
be able to shave ten hours off of your week
(28:42):
every week by.
Speaker 2 (28:43):
Using these tools properly, for sure. For sure, Oscar in
Saint Louis asks.
Speaker 4 (28:59):
At the stages when companies are moving fast and things
are very dynamic, how do you organize the team and
keep everyone aligned?
Speaker 1 (29:07):
And this just share is where you're big on the
rituals and behaviors in the culture of the organization itself.
Give everybody your overall thesis here and a couple of
concrete examples.
Speaker 3 (29:19):
Yeah, so maybe as context for everyone, I'm writing a book.
It's called Rachel's a Great Teams.
Speaker 2 (29:26):
And it's kind of a when's it coming.
Speaker 3 (29:28):
Out people, hopefully early next year?
Speaker 2 (29:32):
So do I sound like your publisher.
Speaker 3 (29:34):
When do we get into Yeah, exactly exactly. That said,
if you go to Richel's Great Teams dot com. I'm
writing it in the open so you can go. I
publish a chapter to time so you can go see
what they what they look like, and I'm about sixty
percent of it or so is available for people to read.
They actually kind of fund A couple of other people
are in their co editing with me, which is kind
of a nice way to to write a book. Anyway, back,
(29:58):
so the thesis of this book started with the conversation
I had with a friend of mine being Gordon and
people who don't know who Bing is. He is one
of the founders or the chief creative officer at Electronic
Arts and is now famous investor Zinga, Amazon, many great companies.
He and I sat on a board together and he
(30:18):
kept harassing the CEO with this question. He said, what
are your golden rituals? And the CEO kind of struggled
to ask a little bit, and at one point one
of us said, hey, Bing, what what's the golden ritual?
And he gave us this really clear rubric. He said,
great companies are the smallest of golden rituals. They have
three criteria. Number one they're named, number two, every employee
(30:41):
knows them by their first Friday, and number three they're templated.
And he immediately rattled off his examples. Amazon had six pagers,
Google has OKR, Salesforce's V two mom so and it
just turned out to be this really sticky idea for me,
this idea of ritual So this happened right before the
pandemic started. I started getting I started talking about this
(31:02):
on on a few podcasts and talked about it talking
to some customers, and people started sending me rituals and say, oh,
if you're really interested in this topic, you should hear
more about this. And it's become this like weird hobby
of mine. Like people have all sorts of normal hobby
life and normal hobbies too, but this has become my
like nights and weekends hobby is collecting rituals. And one
of the key things we did was we started doing
this dinner series. So every about three weeks, starting at
(31:26):
the beginning of the pandemic, we would get a group
of ten to fifteen people together and say just share
your rituals. And people from small companies, big companies, you know,
all sorts of different disciplines and functions and and and
so on. And at this point we've taken about a
thousand different rituals. We whirled them into about one hundred
of my favorite ones. And that's the that's what this
book is about. Rituals are great teams. And there's a
(31:48):
couple of interesting observations about this. You know, one the
first thing I would share is that, you know, one
of my favorite Stortes Darmesia, is the founder of HubSpot.
He comes to one whose dinners and he shared a
ritual called Flashtacks. It's kind of fun ritual. You can
go read more about it. But then you start getting
really excited about rituals. And he gave this analogy. Real
(32:11):
like he said, as a company, we build two products.
We build one for our customers, we build another one
for our employees. The one we build for our employees
we often call that culture. But if you ask somebody
to describe culture, they will often describe it with rituals.
They'll say, oh, we have this value, here's how it works.
And so from that perspective, rituals aren't just the operating
(32:32):
system of how you get stuff done, and how you
keep your team organized, keep everybody lying, and so on.
It's actually a two way mirror to your culture.
Speaker 1 (32:39):
I couldn't agree more with the thesis of the golden
rituals I've seen at work at my company too. I
have not read the first sixty percent of your book,
but I will for our listeners. I'll drop that link
in the show notes. You also referenced your Iigen questions,
which we're not going to be able to get into today's.
It's been so much good stuff we're going to drop
that in the show notes as well. There's been to
share a ton of amazing content today. I knew this
(33:02):
would happen getting together with you, and this has been
a blast.
Speaker 2 (33:06):
Thank you by both.
Speaker 3 (33:07):
Mike, Yeah, thanks for having me.
Speaker 2 (33:09):
This was terrific.
Speaker 1 (33:09):
Congrats on all the progress you've made with CODA. For
folks who haven't tried it, it's really phenomenal software. It's
not only makes it easier to create your content, you
actually get to better answers because of the tools that
you share in his team are building and dropping into
the show notes and check it out.
Speaker 2 (33:25):
I think you'll really enjoy it and you share. Thank you, buddy,
I'll see you again soon.
Speaker 3 (33:30):
Thank Mike.
Speaker 1 (33:37):
You know, Gagan, one of the things that I've observed
about just about all of our conversations is that the
folks who come in office hours have not only achieved
success through building great products or bringing great products or
services to market, They've also been extremely thoughtful about the
systems and processes and cultures that they create in their teams,
(34:01):
or in their companies, or even in their own lives.
And you'd be hard pressed to find someone who's as
good as Shasher is at creating these processes and as
you called it today, these golden rituals that helped to
organize a team, that helped to give everyone a shared
language and shared tools for solving problems.
Speaker 2 (34:24):
And if I leave you to.
Speaker 1 (34:24):
Think about something, what are those golden rituals in your organization?
And what are they for you in your life or
in your family. I know we have some that keep
my kids organized on the weekends. I know I personally
have some to make sure that my health and my
fitness get as much attention as my development.
Speaker 2 (34:46):
And my work and everything else.
Speaker 1 (34:48):
I'm going to link these in the show notes and
encourage you to check it out and do come back
and ask yourself, what should these rituals be in my
personal life and for my team.
Speaker 2 (34:58):
I think that you will find that you're.
Speaker 1 (35:00):
Able to unlock some productivity, some better alignment, some better
velocity in the things that you do. I want to
thank Youshier for coming on the pod. It's great to
see my friend again. And of course I want to
thank Jen, Cara, Jada, Meg Matt, the whole team at
Blue Duck Media for pulling this all together. I want
to thank Dylan and Sasha, Gay, Nathan and Christine at
(35:21):
iHeart and Ben and the team at William Morrison Devor
for all their support. Office Hours is a production of
Blue Duck Media and distributed by iHeart Radio.
Speaker 2 (35:30):
I will see you next week. Everybody, make sure you
stay on your grind.
Host
Mike Steib
Popular Podcasts
On Purpose with Jay Shetty
I’m Jay Shetty host of On Purpose the worlds #1 Mental Health podcast and I’m so grateful you found us. I started this podcast 5 years ago to invite you into conversations and workshops that are designed to help make you happier, healthier and more healed. I believe that when you (yes you) feel seen, heard and understood you’re able to deal with relationship struggles, work challenges and life’s ups and downs with more ease and grace. I interview experts, celebrities, thought leaders and athletes so that we can grow our mindset, build better habits and uncover a side of them we’ve never seen before. New episodes every Monday and Friday. Your support means the world to me and I don’t take it for granted — click the follow button and leave a review to help us spread the love with On Purpose. I can’t wait for you to listen to your first or 500th episode!
Stuff You Should Know
If you've ever wanted to know about champagne, satanism, the Stonewall Uprising, chaos theory, LSD, El Nino, true crime and Rosa Parks, then look no further. Josh and Chuck have you covered.
Dateline NBC
Current and classic episodes, featuring compelling true-crime mysteries, powerful documentaries and in-depth investigations. Follow now to get the latest episodes of Dateline NBC completely free, or subscribe to Dateline Premium for ad-free listening and exclusive bonus content: DatelinePremium.com