Episode Transcript
Transcripts are displayed as originally observed. Some content, including advertisements may have changed.
Use Ctrl + F to search
0:00
That's. The story of are kind of
0:02
creative progress as humanity is like. people
0:05
build these tools. Which. Enables.
0:08
The next group of people to
0:10
focus on a higher level set
0:12
of problems. Just in the last three
0:14
weeks since launch, developers are built
0:16
over twenty thousand new activities on
0:18
the platform and that's generating four billion
0:20
minutes of use and interaction per
0:22
day is the scale is sort
0:24
of mind boggling. What's going on now?
0:27
I think something like ninety three
0:29
percent of Kinsey plays games were
0:31
back when we were kids. It was
0:33
super weird to be playing games.
0:35
We're now seeing some really significant
0:37
examples. Of companies being built entirely on discord.
0:41
Bill Gates. One sad that a
0:43
platform is when the economic value
0:45
of everybody that uses it exceeds
0:47
the value of the company that
0:49
creates it. That definition does that.
0:51
A pretty high bar, but the
0:53
few companies that surpass it, but
0:55
one company does come to mind
0:57
and that is disco. But officially
1:00
starting Twenty Fifteen can really be
1:02
traced back to Two Thousand and
1:04
Nine when Discord Co founder and
1:06
Ceo Jason Citron started building tools
1:08
and infrastructure for games. Fast
1:10
forward to today and Discord now
1:12
has over two hundred million monthly
1:14
active users. Some might even argue
1:16
that Metaverse is actually hear it
1:18
just doesn't quite look like the
1:20
sense. Now. In today's episode,
1:23
you'll get to hear from Jason
1:25
Alongside a sixteen see: general partner
1:27
Andre Mehta you actually sold his
1:29
company Ubiquity Sex to Discord and
1:31
Twenty Twenty One they're on set
1:33
up and ran Discords first dedicated
1:35
Felker platform including launching it's partnership
1:37
with Mint Journey all before joining
1:39
a sixteen he last year. You.
1:41
Can probably very quickly town that
1:44
Jason and Onge have this shared
1:46
history, especially because they got to
1:48
sit down together and are San
1:50
Francisco studio to discuss how discord
1:52
became such a thriving platform. but
1:55
what discord really do differently? Here
1:57
together they discussed community driven product
1:59
develop, So Jason himself went from
2:01
player to developer and their
2:04
focus on extensibility since the very
2:06
beginning. So with Discord's
2:08
recent release of embeddable apps, what
2:11
can we expect now that it's easier than
2:13
ever for a developer to build? If
2:15
I was to go back when I was in
2:17
college, which was almost 20 years ago now, like
2:19
none of that stuff existed. I mean, I built
2:21
games back then and it was like firing up
2:23
C++ and like reading the DirectX APIs and spending
2:25
a week trying to get a window to open
2:27
with a triangle on it. Prior
2:30
to this release, there was of course
2:32
already a flurry of new applications built
2:34
on the back of Discord like Midjourney
2:36
or Leonardo. So let's
2:39
find out what's next. As
2:45
a reminder, the content here is for informational
2:47
purposes only, should not be taken as legal,
2:49
business, tax, or investment advice, or
2:51
be used to evaluate any investment or security,
2:53
and is not directed at any investors or
2:55
potential investors in any A16Z fund. Please
2:58
note that A16Z and its affiliates may
3:00
also maintain investments in the company's discuss
3:02
in this podcast. For more details,
3:04
including a link to our investment, please
3:06
see a16z.com slash. I
3:15
am so excited for this
3:17
episode. Thank you for joining us,
3:19
Jason Citron, CEO of Discord, and dear
3:21
friend, former colleague, and probably
3:23
the person who I know who's tried
3:26
to start game studios the most number
3:28
of times and
3:30
ended up building several successful platforms
3:32
along the way. So
3:34
today we're going to talk about all kinds
3:36
of things focused on developers, infrastructure, the future
3:38
of the Discord platform. But before we get
3:40
to that, for folks who might not be
3:42
as familiar with the crazy story that led
3:44
to here, why don't we go back in
3:46
history. Let's start at the very beginning.
3:49
What was the vision for Discord when
3:51
you first started out? So way
3:53
back in 2012, I was sitting
3:55
around trying to think about what
3:58
could be an exciting business to be. to
4:01
build. And having spent most of my career,
4:03
in fact, all of my career and my
4:05
childhood steeped in video games and multiplayer games,
4:07
I had this hunch that multiplayer
4:10
gaming and gaming in general
4:13
was going to become much bigger than it already
4:15
was at that time. And back in 2012, gaming
4:17
was pretty big, but it was kind of at
4:19
the early innings of mobile and
4:21
still trying to figure out like where
4:23
was gaming going to go. And I
4:26
thought that there'd be an opportunity to
4:28
build a communications app for people who
4:30
play games that would span
4:32
all the platforms and all the devices as
4:35
gaming would become bigger and more cross platform. But
4:37
largely has played out that way. Today in 2024,
4:39
gaming is the largest form
4:43
of entertainment, bigger than music and
4:45
movies combined growing fast. And
4:47
people love to play games. It's gone mainstream. And I
4:49
think something like 93% of Gen Z plays games
4:53
where back when we were kids, it
4:55
was super weird to be playing games. So that
4:58
was kind of where it started. What
5:00
were the moments where you were playing games and you went from
5:02
being a player and a consumer of
5:05
games as a product to going,
5:07
you know what, the tools I'm using here could
5:10
be better. What was the moment where you
5:12
shifted from being player to developer? Well, I
5:15
fell in love with games when I was a little kid
5:17
because they were a way for me to
5:19
connect and spend quality time with people
5:21
in my life. And I remember sitting
5:24
with my dad, I must have been four
5:26
or five years old, this is like late 80s. And
5:28
he introduced me to this game called Where in the World is
5:31
Carmen San Diego on his old
5:33
like a Packard Bell computer. And I just
5:35
remember being so excited about coming home at
5:37
the end of the day, and being able
5:39
to sit with my dad and spend some quality time with him
5:41
exploring this world and trying to find this crazy lady. And
5:44
over the years growing up playing multiplayer games
5:47
on consoles and then on the internet, as
5:49
that became a thing in the late 90s.
5:52
And along the way, I met
5:54
someone who basically was like, you
5:57
know, I know how to make video games as a friend of mine, I was
5:59
like I was like, no, you don't. You
6:02
can't just make video games. And he was like, no, no, check it out. We
6:04
went to my computer and he fired up this thing
6:06
called QBasic and showed me how to draw a circle
6:08
on the screen. And I was like, oh my God,
6:11
that's amazing. I could make video games. And so that
6:13
was kind of when I became an
6:15
engineer and a programmer, and I learned how to code. And
6:17
then fast forward, I went to school, grew up,
6:20
and got into an opportunity where I was able to start a
6:22
company. And through the process of
6:24
building a game on the iPhone, I actually launched
6:26
a game the day the App Store opened in
6:29
2008. One of the first 50 titles on the
6:31
App Store. And that kind of took off like
6:33
crazy. As we now know, mobile has been the
6:35
biggest computing platform in the world. And through that
6:37
journey, realized that I had made a
6:39
fun game. It was called Aurora Faint. The
6:42
technology that we had built, which was kind
6:44
of like leaderboards, chat rooms, login, was
6:46
something that other developers really wanted. And at the
6:49
time, many game developers didn't know how to build
6:51
infrastructure. And I had learned how to do infrastructure
6:53
and also how to make games. So
6:56
we kind of spun out the backend tech and built this social
6:59
network for mobile gaming. This was probably 2009. It's
7:02
called OpenFaint. We opened up Aurora Faint. Lesson about
7:04
branding from that. No one knows how to spell
7:06
that thing. So that was
7:08
kind of the first moment when I started
7:10
building tools and infrastructure
7:12
for games. And that company did
7:15
pretty well. And then in 2012, after
7:18
I had kind of moved on from that, started
7:20
again building another game. But we began as a game
7:22
studio as well in 2012 called Hammer
7:24
and Chisel. And we started the
7:26
game because I thought that the path to building the
7:28
communications app would be to start with a multiplayer game.
7:32
I had this hunch that core long-form gaming was going to
7:34
come to mobile in a big way. And
7:36
so we started building a team-based competitive multiplayer
7:38
game on iPad at the time in 2012.
7:42
And one thing led to another. The game didn't
7:44
really work out. And so
7:46
it led us in late 2014, we started talking
7:49
about what if we just went to market directly
7:51
with the chat app for gaming. And
7:54
my co-founder Stan kind of had the insight for
7:56
what that concept could look like. And we started
7:58
building it in January 2015. and
8:00
then brought it to market in May 2015. And
8:03
that was kind of how it all
8:05
started. There's a theme emerging here,
8:08
right? Where at least twice now you've
8:10
approached building a product as an application developer. You
8:12
could think of a game as an app. And
8:15
you discovered along the way that there's a bunch of
8:17
really hard infrastructure that needs to be built first. Especially
8:19
when it comes to real time multiplayer gaming. The history
8:22
of computing is such that usually real
8:24
time gaming is one of the most demanding
8:26
infrastructure environments. And then you discovered that
8:29
there were a ton of infrastructure problems along the way. And then
8:31
you ended up actually building those tools for
8:33
other people to use. And have since built one
8:35
of the fastest growing,
8:37
biggest, real time communication platforms in
8:39
the world. Which is sort of insane to
8:41
think about the scale of Discord. But when
8:43
you play forward from that moment in 2015 when
8:46
Discord came out and today, how has
8:49
the community on Discord helped shape the
8:52
roadmap? The people actually using the infrastructure,
8:54
whether those are users or developers? Our
8:57
community, our user base has been part
8:59
of the conversation of what we're making
9:01
from the first day. When
9:03
we started talking about building Discord, of
9:05
course we played a lot of multiplayer games ourselves. So we had
9:07
a good sense for what the product should
9:09
be and how it should work. But as
9:12
anyone who is building a startup knows or
9:14
building products, we're building products in service of
9:16
other people. And so we immediately from the
9:18
beginning started talking to our friends and
9:21
their friends. When we showed it to them, even just
9:23
mock ups, what parts got them excited? Which parts seemed
9:25
confusing? And very quickly, once we
9:27
got a prototype off the ground, we
9:29
started giving it to our friends and having them
9:31
try it and seeing what they liked and what
9:33
they didn't like. And oh, crap, we had to
9:35
rebuild the voice tech three times. And we missed
9:38
an important set of features that we thought maybe
9:40
was not important, but turns out it was. And
9:42
so that was kind of part of the ethos
9:44
for how we built from the beginning. And
9:47
then over the years, we've continued to build products that
9:49
way in a sense that we always try to come
9:51
back to what are we hearing from
9:53
our customers, from our users, and the
9:55
different types of people who use our products. And then
9:57
how do we kind of muck that with what
9:59
are we? We excited to build for ourselves. And
10:02
then of course, what do we think will be great for us as a
10:04
business because we are a company. So along
10:07
the way, there have been many,
10:09
many, many moments when large shifts
10:11
have happened in our roadmap because
10:13
of customers. So I'll give you an example. Initially
10:15
when we built Discord, it was very focused on being
10:17
a voice and text chat app
10:20
for guilds, people who play games in groups
10:22
of like 15 people. And
10:24
actually the max group size on Discord, I
10:26
think was like 30 people, maybe 50, it
10:29
was pretty low. And we realized
10:31
pretty quickly from talking to people that they
10:33
wanted to use Discord as
10:35
almost like an IRC, like internet relay
10:37
chat kind of public chat room replacement.
10:40
And in that context, what we saw was people
10:42
were filling their servers up with 50 people and then they
10:44
were like, I can't add more people. What's going on? And I
10:46
was like, oh crap, we got to
10:48
make this work for folks. So we invested in raising
10:50
the cap and adding more infrastructure to support that. And
10:53
then developers started building moderation bots
10:56
and extending these Discord servers with
10:58
other capabilities that we never
11:01
even imagined that was made possible because we
11:03
had an open kind of API powering the
11:05
platform. So along the
11:07
way, many of these things happened. Like
11:10
these communities got big, Generative AI became
11:12
a big thing on Discord. The crypto
11:14
community was pretty big for a season
11:16
on Discord. But throughout all of
11:18
it, gaming and playing games with your friends and
11:20
hanging out with your friends was always
11:23
the main thing that people were doing, even
11:25
if they would go spend time in a
11:27
public community or futzing around with Generative AI
11:29
or something like that. Yeah, I think
11:31
one of the most underappreciated things about Discord
11:34
is that the product has found
11:36
a way to do two
11:38
things at once that almost
11:41
no other companies are able to
11:43
do at scale, which is have
11:46
a singular focus on a particular type
11:48
of user and their need. In this
11:50
case, what you said was
11:52
allowing friends to spend time together while
11:55
playing games, while also making
11:58
the platform and the product so exciting.
12:00
danceable, For. Other people to bring
12:02
their own use cases the platform and
12:04
I remember a couple years ago I
12:07
was talking to Stand and. He.
12:09
Brought up a screenshot of the first version of
12:11
the homepage you guys had put together and on
12:13
the front page on day one. You.
12:15
Had. A call out for integrations
12:17
and as the case, you hadn't open a
12:19
be I on day one as part of
12:21
the hero marketing and so clearly you are
12:23
thinking about making the platform extensible for all
12:25
other kinds of use cases. Ten years ago,
12:27
where did that come from and. Do
12:30
you doc? eluded? But the challenges of put
12:32
building a delightful product for users first while
12:34
also maintaining this extensibility for other kinds of
12:36
use cases that you may not have designed
12:38
for explicitly. Yeah. The extensibility
12:40
was built them from the beginning and it's
12:42
cool that you i'm back and looked at
12:44
I think if you go to the way
12:46
back machine of you can find it still
12:48
from like twenty fifteen. The idea was that
12:50
we knew that people. Are going to
12:52
want to? Build. Huston
12:55
integrations with different games as part
12:57
of thinking about like what the
12:59
group chat for gaming look like
13:01
so we're imagining. If you have
13:03
would say like and Eve online
13:05
corporation which they screw people playing
13:07
this outer space massively multiplayer game
13:09
or you're playing Final Fantasy Online
13:11
which is a fantasy adventure game.
13:13
These different games has. Data.
13:15
And things you might want to pull into your groups. Had
13:17
experience for the we knew we were not going to build.
13:20
All. Of these things for the hundreds or thousands
13:22
of games the people my care about. So that
13:24
was kind of one thing. We're like we've gotta
13:27
make it so other folks can integrate. Their.
13:29
Custom stuff from their games into
13:32
discord. And then related to
13:34
that was this insight that I have from
13:36
being observer in the gaming business for so
13:38
long. Which is that. In games
13:40
there's this concept of modeling where people
13:42
can mod games and with that basically
13:45
means as a developer will create a
13:47
game and then oftentimes ship with it.
13:49
The tools they use to make the
13:51
game I think it's software popular as
13:53
this in the early days with dunes
13:55
kind of person. I really remember getting
13:57
big in the call them wide files
13:59
his whole scene on my of the
14:01
download wide fi and they started packing
14:03
up and selling them and it added
14:05
a ton of a life's to the
14:08
game. And today now when we look
14:10
at the top titles that people play,
14:12
a lot of them actually began as
14:14
community driven mods like Counter Strike, Team
14:16
Fortress, League of Legends, Ceta Roleplaying Cost
14:18
The list goes on. Fortnight began with
14:20
version for of a mod that isn't
14:22
came from a game called Arma three
14:24
many years ago. So this idea that
14:26
give your community tools to create and
14:29
customize and extend. Your game or if
14:31
your software and are going to surprise you and
14:33
take it in places you never would have expected
14:35
was just telling the me seen how that's how
14:37
you make good software rice. So when we built
14:39
this court it's sort of seemed obvious to us
14:41
that we wanted to create any p I that
14:44
allow developers to extend discord to be more creative
14:46
with the to do things with it we never
14:48
would have expected. Some of the things that we
14:50
expected were like. Connecting. To the
14:52
eve online back and you can have
14:54
your own forum or pulling in world
14:56
boss spawns and as notifications but the
14:59
stuff we never expected was like. Gendered.
15:01
As a I, who could I guess
15:03
that might be. I'm in now. We're
15:06
seeing many years after you made those
15:08
investments in the craft of designing a
15:10
delightfully be I'd Fantastic tutorials are great
15:12
developer experience for people to mod discord
15:14
itself. We're now seeing some. Really?
15:17
Significant example: the companies being built and
15:19
rt underscored. The. Discord platform
15:21
has two hundred million monthly active users.
15:24
And. That's led to. Entirely.
15:26
New companies being built on top with the
15:28
open architecture describe people may be familiar with
15:30
my journey as one example. Yet.
15:33
And there's a few other ones to mid
15:35
journeys the most famous one. It's like a
15:37
canonical example of how this the habit like.
15:39
We have this open platform where we're allowing
15:41
people developers to customize and and expands. It's
15:43
I think the mid your any folks had
15:45
obviously working on them all for while they
15:47
try bring it to market as your the
15:49
ways and then they just tried the Discord
15:51
server and a discord bots and then this
15:53
was the example of exactly what I'm talking
15:55
about. I never would have predicted that you.
15:57
It has. Age and or the I in the
15:59
first place. Like what a crazy thing we've
16:01
created computer that can do the things and
16:03
then to that someone would figure out how
16:06
to take advantage of the magic of a
16:08
discord server. And. Our platform and build.
16:10
Such a cool experience And there's a handful
16:12
of these now. So yeah, it's been a
16:14
pretty cool thing to see. My Journeys.
16:16
One example of a gendered a model. It's a deck.
16:18
The image. Models. We've. Also
16:20
seen an explosion of Dex The Music Tools
16:23
as an example. How extensible discord as what
16:25
other kind of use cases are you seeing
16:27
emerge on the platform that you're excited about.
16:30
The. Tenor The I. Swear. Category
16:32
is exciting for us. the you know we
16:34
really come back to this idea of people
16:36
mostly on discord, spend their time hang out
16:38
with their friends and is kind of smaller
16:40
invite only spaces with less than sixteen people
16:42
per se. So in that context when I
16:45
think about generally I tools and using our
16:47
platform I really think about what are developers
16:49
creating that gives groups of friends more fun
16:51
things to do. and so one really cool
16:53
thing we see with the generally I stuff
16:55
is people take the bots and bring them
16:57
into their invite only servers and in the
17:00
in you them to create an explorer and
17:02
work on projects with their friends in a
17:04
more and a private setting as opposed to
17:06
being in the sort of the public chaos.
17:08
frankly have some of these large servers. but
17:10
other really cool experience as we see our
17:12
things like way to listen to music the
17:14
other soundcloud has a really cool bought that
17:17
you can use to play music when you're
17:19
and voice shots for example and we're working
17:21
with some other partners to try to bring
17:23
more music to the platform. There's a bunch
17:25
of games the people have made that are
17:27
pretty cool and many more coming. We just
17:29
actually launched. A couple weeks
17:32
ago a new kind of said a
17:34
capabilities for the platform that. Will. Allow
17:36
developers to go kind of beyond the
17:38
text box and build these. Rich kind
17:40
of visual interactive experience is powered by
17:42
Channel Five see Kabila web app and
17:44
essentially deploy it into the context of
17:46
discord. And so we're seeing lots of
17:49
really exciting stuff starting to get built.
17:51
That. I think we'll give people a lot really fun
17:53
things to do with their friends. Led Senate
17:55
Gop minutes on that. This is something you and
17:57
I spent years working on together. If we did,
18:00
The data and I picked fitness. Let's
18:02
take people looted behind the garden of
18:04
like what it takes actually go from
18:06
the moment where. We. Realize
18:08
that. Developers.
18:11
Wanted. To express their creativity beyond just
18:13
the command line like interface which is what
18:15
bots were initially designed around as a form
18:17
factor and going from there to expanding mint
18:19
are canvas for them to the whole screen.
18:21
really? with web apps. Yellow. Years
18:23
ago when the boss platform sorry to get
18:26
popular. There. Is actually a hacker project
18:28
at our company. so we do this thing every
18:30
are called Hockley where we basically stop or normal
18:32
work and everyone gets together. We have like a
18:34
little kind of creativity festival. cessna say describe it's
18:37
and weddings. Twenty eighteen was the year that I'm
18:39
thinking about less and like tents in our office.
18:41
Food like a whole cool cat was food and
18:43
stuff and one of the groups that year. We.
18:46
Had the idea of wouldn't it be cool
18:48
if we added in Html five canvas to
18:51
our ass platform in and people could like
18:53
may games and do other interesting stuff. and
18:55
says there was a team at Built this
18:57
and I was always like that's a really
19:00
cool idea, we've gotta explore that but. As.
19:02
Company Building goes. You have a long list
19:04
of ideas and you have to prioritize when
19:06
you get to them. So a couple years
19:08
later the time was right to look at
19:10
the idea and that's when we met you
19:12
actually working on your own start up at
19:14
the time. And. Work and a
19:17
building this as a standalone project and week
19:19
after some conversations I was like we should
19:21
just do together. So we acquired your company
19:23
and then we created a team and really
19:25
formalize this idea of like how do we
19:27
take our platform. From. Kind of
19:29
the text based era. In. Said
19:31
visual, rich interactive experience era.
19:34
And to start rather than just opening it
19:36
up we actually built a few games ourselves
19:38
to really test the platform, make sure that
19:41
it was designed well at work for player
19:43
that the interaction loops for good and that
19:45
to the couple years to kinda sorta I'll
19:47
out and now we're at the point where
19:49
okay right open up so it's and developer
19:51
previous it's been a journey. Yeah it's been
19:54
a fun journey. Yeah okay we're gonna girl
19:56
move a deeper they'll have. Okay I remember
19:58
one of the most exciting part. of
20:00
working at Discord was that
20:02
we had this incredible respect for
20:04
infrastructure. The company had
20:07
built its own WebRTC streaming service, had
20:09
built its own voice and video infra.
20:12
We built a bunch of serverless networking for developers
20:14
for the apps SDK. And
20:16
I remember there are these moments
20:18
in the product engineering cycle
20:20
when you're building infrastructure that you
20:23
inevitably have to ask yourself, well, what is this going
20:25
to be used for? And
20:27
I remember having debates with you about how to answer
20:29
that question and prioritize the most important
20:31
features. And so we came up with this ritual
20:34
of jam sessions, if you remember, where the team
20:37
that was working on tools and infra and SDKs for
20:39
the developers would come in and jam with you where
20:41
you would often role play the developer. So
20:43
today, fast forward, I'm going to ask
20:45
you to pretend we're in a jam session. And
20:48
we just launched the activities SDK, the embedded apps
20:50
SDK, and now you're a developer. And I'm going
20:52
to ask you now, what would you like to
20:54
go build with this entirely new set of capabilities
20:56
that you've been handed by Discord? So
20:58
I guess the way that I think about this is first, like,
21:01
what am I going to build? I try to think about what
21:03
do I as a consumer want? It's
21:05
almost like two steps even past the infrastructure.
21:07
It's like the developers are sort of thinking
21:09
like they're serving their customers. So they
21:12
have to take their hat off, right, and then
21:14
put on their consumer hat. And so for me,
21:16
it's starting to think about what kind of, you
21:19
know, games I might want to play and how they could
21:21
fit into my Discord servers in fun ways. And for me,
21:23
I really would love someone
21:25
to build some multiplayer titles
21:27
that don't require a lot of
21:30
time investment, but have really
21:32
cool moments of storytelling that you can
21:34
interact with your friends asynchronously around that
21:37
sort of weaves into the Discord experience
21:39
through text and with the visuals as
21:41
well. I have a small side project I'm
21:43
working on right now, which is a game kind of like this. It's like
21:45
a arcade shooter that
21:48
has kind of a leaderboard mechanic, but with a
21:50
modern kind of roguelike vibe on
21:52
it. Anyway, it's just like random stuff that I'm
21:54
making as a side both to sort of dog
21:56
food and test our platform. But part of this
21:59
is how I get myself in the mood. mind
22:01
space of what our developers want, what might their
22:03
customers want, so that I can give good feedback
22:05
to our team who is also doing this so
22:07
we can serve people effectively. One of
22:09
the things that I think you've always been pretty good
22:11
at doing is asking the question of what
22:14
is possible on Discord that isn't possible
22:17
elsewhere. And you're often
22:19
able to laser in and hone in
22:21
on the specific capabilities that are ultimately
22:23
technical primitives, but exposed
22:26
to users in a way that allows them to do something
22:28
with their friends they couldn't do before in a way that's
22:30
easier or faster or better or more convenient. And
22:32
that's how things like persistent stage channels happen, right? And
22:35
that's how forum channels happen, and that's
22:37
how the embedded app SDK happened. And
22:39
so as you're working on crafting your
22:41
side project right now, what in your
22:43
mind at the top two or three
22:45
things that the Discord-embedded platform
22:48
that you just launched really shines at that's
22:50
hard to do elsewhere? I
22:52
mean, the magic, I think, is the fact that you
22:54
have the social context of the space you're in, right?
22:57
So it's really, really easy for someone to
22:59
pick up a title, and then that game
23:01
can depend on the fact that there's a
23:03
group of people that are
23:06
connected to the person playing it, and you have
23:08
access to that data in a privacy-safe way. Unlike,
23:10
let's say, other games where maybe you log in
23:13
or you get a user to come into
23:15
your game, and then you have to get them to maybe build
23:17
a friends list or to invite their friends
23:19
to play. In the context of Discord, because of
23:21
the way the games work, they're
23:23
instant and they have the social
23:25
context, you could build a game
23:27
loop where one person plays, and
23:29
then immediately the result of that play
23:31
session is other people get exposed to
23:34
the game and they can play
23:36
it, and they don't have to go
23:38
do anything to be able to really set that up
23:40
because the friend brought it into the server. So
23:43
I think about cool mechanics like leaderboards
23:45
and challenges and these kind of things
23:47
that really, really depend on
23:49
having this group of people that have
23:51
access to this shared space
23:54
both synchronously and
23:56
then also on the go in bite-sized ways. So
23:58
it's a different kind of interaction model. So I
24:01
could open my phone, play a game, and
24:03
then when I get back to my computer,
24:05
I could pick it up, and then you
24:07
could play from your phone when you're somewhere else if you want
24:09
to come and compete with me when you have a few minutes
24:11
in your spare time. So this is
24:13
the kind of model that I'm playing around with with this title,
24:16
but part of what I'm so excited
24:18
about with these tools is to see the creativity that
24:20
other people bring. I know what would be interesting to
24:22
me, but much like when we
24:24
designed the platform for Discord
24:26
in the beginning, so many of the cool things
24:29
that happened with it, I would never have predicted.
24:31
So I'm most excited to see what
24:33
people do that I don't even think about.
24:35
I think it's worth taking a beat there to just
24:37
recap what you just said. Up until now, for a
24:39
developer to build in a cross-platform experience like that, it
24:41
takes years. And the end result,
24:44
I think, is what you're describing is that
24:46
a product or an app, whether it's
24:48
a game, like the one you're building,
24:50
or a non-game like Mid-Journey, can literally
24:52
go from zero to, I think Mid-Journey
24:54
is now at more than 20 million users in the server,
24:57
or about there. Yeah, I think their public
24:59
server is, yeah, it's about that size. It's pretty large.
25:01
Okay, there we go. In about,
25:04
in essentially less than a year. Yeah,
25:06
you should write the marketing briefer. Well,
25:09
I mean, you did, I guess you actually did. I
25:12
mean, you're spot on. I mean, I think that's
25:15
what's so magical about it is the trends that
25:17
entertainment and gaming in particular have been on for
25:19
the last decade are cross-platform. Like
25:21
people are more and more expecting games to
25:23
not be tethered to devices. It's
25:25
just a screen. Like you should be able to switch screens and play
25:27
the same experience. You want your content
25:29
and your progress to go across those screens. You want
25:32
your friends and your social graph to go across those
25:34
screens. You want the things you bought to go across
25:36
those screens. And these trends
25:38
are happening sort of in the world broadly in gaming
25:41
too, but what we're trying to
25:43
do is package that all together in a really easy
25:45
to pick up way, where not only
25:48
do the games have all those things, but we
25:50
also bring the social graph to the game. We
25:53
manage off for you because it's built into discord. It gets
25:55
deployed on every platform that we're on, which is most of
25:57
them. We have payments built in on the
25:59
platforms where the platform. forms don't and discovery and all this
26:01
stuff. So yeah, I'm just excited to
26:03
see what people can make. You know, that's the
26:06
story of our kind of creative progress as humanity
26:08
is like people build these tools, which
26:10
enables the next group of people to
26:13
focus on a higher level set
26:15
of problems. So we've extracted out
26:17
some of this infrastructure for folks so people can
26:19
spend more time on their gameplay
26:21
and the creativity of that then futzing around with
26:23
auth and building social graphs and
26:25
all this kind of stuff. One of
26:28
the most amazing things about building in an
26:30
openly extensible way is that you get the
26:32
combinatorial creativity, the explosion of your primitives, your
26:34
platform with other tools that may be coming
26:36
online or maturing at that moment in time
26:38
and sitting here in 2024, if you just
26:42
contrasted to the version of Jason in
26:44
college who was building his own games,
26:46
right? And you hand him all the
26:49
open infra of discourse developer platform and
26:51
the recent kind of explosion in these
26:53
creative generative models that can
26:55
allow you to turn text into images
26:58
and text into audio and allow code
27:00
generation and so on. When you look
27:02
at how the production pipeline of building
27:04
an entire experience like a game has
27:06
changed with generative models and you combine
27:08
that with the Lego blocks of discord,
27:11
what are the kinds of new game formats
27:13
or new types of interactive entertainment
27:15
that you think are possible today that just weren't possible
27:17
maybe when you were in college? I
27:19
mean, there's so much that has changed. It's
27:21
kind of amazing. Like you think about it and
27:24
today, between, like you
27:26
said, the generative AI tools, social
27:28
and distribution infrastructure like app stores and something
27:30
like discord, right? Plus the modern game engines,
27:33
we can't leave those out the kind of
27:35
game engine you can get off the shelf
27:37
today, whether it's Unity, Unreal, Godot,
27:39
phaser, all these things. They're just incredible. I
27:41
mean, if I was to go back when
27:43
I was in college, which was almost 20
27:45
years ago now, like none of
27:48
that stuff existed. I mean, I built games back
27:50
then and it was like firing up C++ and
27:52
like reading the DirectX APIs and spending a week
27:54
trying to get a window to open with a
27:56
triangle on it. Like, I think what ends
27:58
up happening is because people
28:00
can be so much more productive and
28:03
the markets are so much bigger. I
28:05
think we're going to start to see more
28:08
and more games that are focused
28:10
on more and more kind of niche topics
28:12
and niche mechanics. Because if you think
28:14
about it, big games get big
28:16
budgets, so they tend to be less risky.
28:19
Smaller games can be more
28:21
risky because the budget dynamics are different. You
28:23
could imagine a world where one developer could
28:26
build an entire game like Stardew
28:29
Valley, which was one developer, but I think it took him like
28:31
five or six years. The next Stardew
28:33
Valley might be built by one guy in a
28:35
year. And if
28:37
you imagine what that means is you may
28:39
get 10 Stardew Valleys and they
28:42
may all have different themes and topics. So we
28:44
may all just get more
28:46
entertainment that's customized to our particular sort
28:49
of proclivities because so many more people
28:51
are making games. So
28:54
the cost is down and then the markets are bigger,
28:56
so there's more people than ever who are looking for
28:58
this stuff. Yeah, this is also I think one of
29:00
the most underappreciated parts of Discord, which is it
29:03
has unlocked, paradoxically, niche at
29:05
scale. Through
29:07
the server context, there are now thousands
29:09
and thousands of niche communities on Discord
29:12
who have then found people who love
29:14
each of those niches globally. And
29:16
what always struck me when we were looking at
29:18
the activity in these servers and what kinds of
29:20
apps and bots they were using is the
29:23
extensibility of the platform allows those niches
29:26
to do things with the platform and build a bot or an
29:28
app that's custom designed for that
29:30
niche community's use case. I think there's
29:33
one you told me about a while ago, which
29:35
was the Harry Potter fan fiction server, right,
29:37
which had an app that that community
29:39
had built for the friends who are
29:41
huge Harry Potter fans to roleplay being
29:43
at Hogwarts. I can remember that one.
29:46
Yeah, yeah, yeah. So if I'm
29:48
asking you to channel your like inner will write,
29:50
for example, you know, one of the most successful
29:52
game genres of all time has been simulation, right?
29:55
The Sims, all the
29:57
tycoon games that allow people
29:59
to experience. this world
30:01
building desire where they're able to almost
30:04
use games as a tool for creativity. Where
30:07
the game is itself building and creating with
30:09
other people. Roblox of course has done a
30:11
phenomenal job at doing that in a pre-genitive
30:13
world. Minecraft. There's Minecraft,
30:15
Roblox, there's Sims. Given
30:18
how massive those genres were in a pre-generative world,
30:20
when you give the generative models to a
30:22
developer plus the insane distribution of Discord, 200 million
30:24
people on day one, do you
30:26
think we're going to see a new kind of genre there
30:28
of something that blends simulation, kind
30:31
of like the Sims with real life with your
30:33
real friends group? I mean it's
30:35
entirely possible. I think that these things are really hard
30:37
to protect. Right. And
30:39
as someone who makes more on the tool side of
30:41
things, what I think is definitely
30:44
going to happen is that because it's going
30:46
to be so much easier to create, we're
30:48
going to see more random stuff. So the
30:50
chances of something interesting happening I think are
30:52
going up. But the things that
30:54
sort of cause new kinds of games
30:56
genres to emerge are oftentimes changes in
30:58
distribution, business model, or
31:00
production capabilities. Right. And
31:03
so I think in this moment we're seeing some
31:05
of these things change, like Discord is offering a
31:07
different kind of distribution mechanism with
31:09
different context. Generative AI is
31:11
definitely changing the landscape for how people
31:14
produce games today. And so I
31:16
think we're probably going to see something interesting happen.
31:19
But yeah, I don't know. It's hard to predict. Yeah. This
31:22
is the most fun part about working on dev platforms,
31:24
right? You get surprised. Yeah. Okay. Just to
31:27
take a step back for context, the
31:29
Discord team finally launched in full
31:31
general availability after years of crafting
31:33
and honing the developer experience, the
31:35
embedded app SDK on March 18th. And
31:38
if I have my numbers right, just in the last
31:40
three weeks since launch, developers have
31:42
built over 20,000 new activities
31:45
on the platform and that's generating
31:47
4 billion minutes of user interaction per day.
31:49
And scale is sort of mind boggling. What's
31:51
going on? Why is this resonating so strongly
31:54
right now? And what are the top emergent
31:56
behaviors you were seeing in the first few
31:58
weeks? I think it's resonating because... because developers
32:00
intuitively understand a lot of this
32:02
that we've been talking about. They're
32:04
looking for distribution channels with
32:07
captive audiences on the other side with low
32:09
production costs, so they can explore their own
32:11
creativity and build products for themselves and their
32:14
friends and bring those to market. And the
32:16
games industry, I think, is in an interesting
32:18
situation right now in particular. So
32:21
this new channel, we thought people were gonna be
32:23
excited about it, and then at Game Developers Conference
32:25
a couple weeks ago when we announced it, I
32:28
was actually surprised at how much it resonated with
32:30
developers. We had a couple talks there, and the
32:32
lines were out the door. So as
32:35
far as what people are building, a lot of the
32:37
things that I know folks are working on have
32:39
not actually released yet since we opened that up, so
32:42
I don't wanna say anything that hasn't come to market
32:44
yet, but I think that
32:46
20,000 number sounds big, and it's exciting, but I
32:48
think in reality, what it signals to me is
32:50
a lot of excitement and curiosity, and we'll sort
32:52
of see how many of those come to market
32:54
and what that'll be. I suspect that there are
32:56
probably 20,000 people poking around,
32:59
I think there's thousands of them that
33:01
are actually really making something, and
33:03
so I expect we'll probably see hundreds of those things
33:05
come to market over the next six or eight months,
33:07
but without getting into specifics, some of them are
33:09
interesting new ways to stream games, and
33:13
some of them are new games, and some of
33:15
them are just interactive experiences, ways to enjoy different
33:17
types of entertainment together, and some of
33:19
them are silly things like comic book
33:21
related projects, and then there's a lot of stuff in
33:23
there that I actually haven't even seen, because there's so
33:25
much. I was talking to a
33:27
developer at GDC who is working on a Discord
33:30
embedded activity that is tinkering with some
33:32
of these new generative models that you were describing, and
33:34
one of the things he crystallized for me, which I
33:36
think kind of maps the experience we had in the
33:38
early days of mid-journey launching on the
33:40
platform, was this idea that game
33:43
development and sort of AI app development
33:46
have this very strong similarity in the
33:48
development process, which is you
33:50
pre-train a model, and then you
33:52
put it out with your community, And
33:55
then what the community does with the early days
33:57
of the model, and what outputs of the model
33:59
they prefer. Helps you then
34:01
sort of run a reinforcement learning loop.
34:03
To. Then improve the model as giving users what
34:06
they like. And. That's very similar
34:08
to the game sort of process right
34:10
where you put out a soft launch
34:12
title usually would live ups in South
34:14
Launch and then you start seeing which
34:17
parts of the multiplayer experience your community
34:19
likes and then you basically pipe that
34:21
into future live optimizers right? If you
34:23
had to describe why the discord. Platform.
34:26
As. Sound. So much
34:28
success with Gen Vi developers is a
34:30
fundamentally something similar about the game development
34:32
production process and a Ice. App
34:34
development that is so similar. That's.
34:37
Resulted in discord, basically hosting one of
34:39
the world's most successful consumer ai business
34:42
right now, which Comintern. Discussing Point:
34:44
I think both of those things make sense,
34:46
but actually think of using Mallow Bet there's
34:48
another interesting trend that this is kind of
34:51
part of which I think is this idea
34:53
of consumers wanting to be closer to the
34:55
people creating the things that they use in
34:57
their lives. And it's actually kind of a
35:00
broader I think dynamic of cocreation with consumers
35:02
as you're building products price. So I think
35:04
in the case of a I, there's quite
35:06
literally it like a direct feedback loop. Or
35:08
I think a lot of these models are
35:11
using sums up and thumbs down type reactions
35:13
on the outlets. To directly feedback and
35:15
improve the model. In the case of games it's
35:17
like a little one step removed were perhaps the
35:20
players are talking with the developers and they're using
35:22
the products and they're giving them feedback and I
35:24
know a lot of Debs used his court to
35:26
do early plate us and they hop on voice
35:29
sad and they show off build and they spend
35:31
time with their users but I also hear startups
35:33
doing it and other companies doing i two notches.
35:35
Games I think gaming in a I kind of
35:38
at the forefront of this but we see lots
35:40
of other tools. There's like a command line app
35:42
that I use as an engineer. Called
35:44
Work that has a Discord Server and they hang
35:46
out with their community in there and talk about
35:48
feature improvements and how to make their product better
35:51
and it just goes on on on. I think
35:53
the trend is actually that consumers. Want
35:55
to have. Say. And influence
35:57
over the products they build and it.
36:00
The car that as a product creator had
36:02
in a direct line of see back with
36:04
a tight feedback loop with your early adopters
36:06
really helps you say put your building and
36:08
make it better so it ends up being
36:10
i think this really powerful and a back
36:12
and forth right where you can improve your
36:14
product. People. Get excited about
36:16
it. You build evangelists and then. When.
36:18
You do go to market and sort a
36:20
long as you have this sort of built
36:22
in community energy that can help spread the
36:25
word around and sometimes the things will lead
36:27
into like Kickstarter is right, patriarch in other
36:29
stuff that is this whole sorta like community
36:31
driven product development saying. That. I think.
36:33
Discord. as part of or it may be helping
36:36
drive in some way but. It's another
36:38
one of those insisting, emergent. Things.
36:40
He coming back to the topic of like. He.
36:42
Build these you platforms and these tools
36:45
with something in mind. And
36:47
then other interesting things can happen with that in
36:49
a week. Again we really started are focused on.
36:52
Being. A place for people to come together and
36:54
play games with their friends. And that is still
36:56
the bulk of what people do today. But.
36:58
All these other interesting things happen right?
37:01
Company. Setting up servers to do code of
37:03
album with their consumers? Well, that's super cool.
37:06
There's. This company that I have the chance to work
37:08
with as an investor called Luma. It's
37:10
agenda to be. I modeled and before they even
37:12
had a website. Or a mobile app
37:14
the launched as a discord up and as
37:16
think what that. Resulted in was
37:19
in three or four days the at thirty
37:21
forty thousand people in the Comunismo a staff
37:23
of whom were from the games industry and
37:25
half who weren't. And
37:28
started. Using the model
37:30
in ways that didn't expect and allowed
37:32
them to realize that there was a
37:34
much broader set of uses for their
37:36
tool. And then that began a dialectic
37:38
that allowed them to then change the
37:40
focus or tweet that list of priorities
37:42
and they're frightened. Open road map than
37:44
ship that to that discord user base.
37:46
And when I saw that happening, I
37:48
realized there's this art in software development
37:50
of finding product, market said or non
37:52
retina. We used to talk about this
37:54
concept of remember of the highest expectation
37:56
user yeah, that a taxi And it
37:58
is remarkably. Hard. Find a check
38:00
sees the take time or the day
38:03
to get attention of users today. Early
38:05
on, but Discord is such a phenomenal duel at
38:07
aggregating these people in one place who care about
38:09
what you're building. That than the speed at which
38:11
you can interact with them is unbelievable. While.
38:14
The appeals process had three. Steps.
38:17
You idea with your community than your launch and
38:19
then you find a waiter. Actually Mana doesn't turn
38:21
entered sustainable business. You saw the first two parts
38:23
that journey what you think the last my looks
38:25
like on discordant. So. Our focus really
38:27
right now, starting with this indebted
38:30
ASS launch a few weeks ago,
38:32
is to bring to market the
38:34
full loop of Had A. As.
38:36
A developer billed as a sustainable growing business
38:38
on discourse buy and sell. Right now we have
38:40
all the parts and much of it isn't
38:42
developer preview and it can be rolling out over
38:45
the next few months. In that will begin with
38:47
how do you get your game listed in
38:49
our app directory and in our app launcher which
38:51
sees millions and millions and millions of people
38:53
everyday coming there a to find fun things
38:55
to do with their friends and ways to customize
38:57
their server and then once they have added
38:59
your app addy make my and so we
39:01
have in our own titles of were running payments
39:04
in. we have payments available to some apps
39:06
to days. You'll be able to directly monetize if
39:08
discourse on a phone that will run through the
39:10
phone payment systems on desktop. We have our own
39:12
stuff that we've built that you'll be able to
39:15
plug into seal get a sort of the expected
39:17
modernization has that you'd want out of the box
39:19
and then reengagement through there and then all of
39:21
back and dashboards in reporting is that that you'd
39:23
expect so you'll be able to build and launch
39:26
him on a ties into the whole loop as
39:28
a developer on discord and your apps for work.
39:30
And this is another one those things where when
39:32
we went in talks a lot of the people
39:34
who make apps and bought them to score. Today
39:36
many of them have been over the years
39:39
using off platform payment mechanism to try to
39:41
cobbled together whether it's to setting up there
39:43
are a website and implementing them to my
39:45
strife or trying to get people to go
39:47
to patron or whatever. but all the things
39:49
are super high friction for customers were the
39:51
gotta go off platform log and do a
39:53
whole bunch of nonsense and any of the
39:55
bill all the stuff A manages so what
39:57
we've done as we've created an easy one.
40:00
The solution inside a discord that works just
40:02
like you'd expect any off sort of work
40:04
were consumer can either purchase a one time
40:06
transaction in your app was make a subscription
40:08
seats ahead of decide how you want to
40:10
monetize your service for whatever makes sense for
40:12
you're. Looking. Back over the.
40:14
Than yours, that discord has been
40:16
around for this story of discord
40:19
has been consistently observing what the
40:21
biggest pain points. Of people
40:23
trying to communicate and do things. the love
40:25
with their friends and making just. Ten.
40:27
Times easier. And. This craft of
40:29
giving people a way to do what
40:31
they're already trying to do buy duct
40:33
taping are combining different tools all in
40:35
one place. While making for discord, doesn't get
40:38
become bloated, doesn't become slower, doesn't become.
40:40
More. Expensive has been this remarkable journey
40:43
of kind of. Ruthlessly
40:45
making. Where. People are trying to
40:47
do already easier and easier in one place. Yeah.
40:50
That's the journey of. I think most
40:52
great products and services are like how
40:54
do you make it so they were
40:56
some is trying to do is better,
40:58
faster and cheaper to sit right. Everyone
41:00
has the journey so we often talk
41:02
about removing objections are reducing friction in
41:04
the process for a person is trying
41:06
to accomplish something and whether that's a
41:08
and user who wants to open her
41:10
up and below quickly message their friends
41:12
or whether that's a developer who's looking
41:14
to build and deploy their app or
41:16
some creative project they're working on to
41:19
people. So. We just loves
41:21
that we get the opportunity to wake
41:23
up every day and help people spend
41:25
time with their friends and play games
41:27
in enjoy life mean that's what it's
41:29
about. I can't wait to
41:31
see where people both And maybe we'll check a
41:33
year from today and in so twenty thousand as
41:35
it's gonna be two hundred thousand. Accessibility of the
41:38
Mighty Rough Some pretty amazing stuff already on the
41:40
platform. Am so excited for first me as a
41:42
developer to get started on my psychotic this weekend.
41:44
Cost of Things have me us. Thanks for Common.
41:50
If you like this episode if you meet it
41:52
is or. How this goes out:
41:54
share with a friend or if. You're
41:56
feeling really ambitious? He can leave
41:58
us review free. Dot
42:01
Com/he is easy. You
42:04
know, candidly producing a podcast can sometimes feel
42:06
like they're just talking into. Avoid is if
42:08
you did like this episode. If you like
42:10
any of her episodes, Please let us
42:13
out. For Phoenix.
Podchaser is the ultimate destination for podcast data, search, and discovery. Learn More