YT YouTube2Text

YouTube2Text → Transcripts → Mark Pincus: Why No App Has Gone Viral in Years

Mark Pincus: Why No App Has Gone Viral in Years — Transcript

by James Altucher · 13,267 words · 1,879 segments · language en · Watch on YouTube

Open in transcriber Summarize with AI Download SRT/VTT

Full transcript

0:00You know, everyone's so jaded that
0:03you're trying to find like an OMFG
0:05moment, like trying to find a vein and
0:07you just put crazier and crazier [ __ ]
0:10until [music]
0:12somebody starts clicking on it and then
0:15it's like a perfect trade win. One of my
0:17mantras that's still [music]
0:19repeated in the hallways at Zinga is all
0:21new fails. And that might sound like a
0:24beatdown. If you start with that mantra,
0:26you won't be disappointed [music]
0:28because there's lots and lots of
0:30statistical proof that all new fails.
0:32Like the app store last year where a
0:34100% of new apps failed, right? If you
0:37can perfectly copy a proven successful
0:41app and no one thinks it's a copy. That
0:44is the [ __ ] magic trick.
0:47This isn't your average business podcast
0:49[music] and he's not your average host.
0:52This is the James Alterer show.
0:56>> [music]
0:59>> Great book by the way.
1:00>> Thank you.
1:00>> I want to get right to what the two
1:02pages that I find the most fascinating
1:04and they're facing each other. One says
1:06build failure machines. The other says
1:08don't even start with an idea. And I
1:10love these two concepts a lot. Like I
1:14love the concept of building a failure
1:15machine.
1:16>> Yeah, me too. It's easier said than
1:20done, you know. I think I I would have
1:23thought by now with AI that everyone
1:26would do that and I can't say that I've
1:30even seen one good example yet. So I
1:33think that's still a hill to take.
1:35>> Why don't you give us give us a good
1:37example? I think our best innovations
1:39come out of dire need and you know
1:43ruthless need to survive necessity and
1:47in the very beginning of Zingga you know
1:50we were living day to day and week to
1:52week and literally you know if something
1:56changed on the Facebook platform or
1:58MySpace we would have like been wiped
2:00out of existence and some were I mean
2:02Christmas vacation killed the number one
2:06game. I think it had been the number one
2:08game on Facebook. It was called Jetpack
2:11Joy Ride. And people didn't feel the joy
2:14when they came back from Christmas. It
2:15was just gone. Um so we were aware of
2:18that. But so it meant that every day
2:23really mattered and we we couldn't be
2:25wrong. And so we just would test things
2:30with I wrote these text links that we
2:32would put at the top of the poker game
2:34and we would offer you a thousand chips
2:37for clicking on them. you know, it was
2:38very direct and cheesy and we just, you
2:43know, looked at the clicks and we just
2:46kept testing until we got a lot of
2:49clicks on something and we didn't build
2:52anything until it got, you know, huge. I
2:56mean, we were looking for on the order
2:59of 25% of players were clicking on
3:03something, you know, before we were
3:04going to build it. So, someone once told
3:06me an idea. You sell these online
3:08courses, but you don't know what online
3:10courses to create first. So, you say,
3:12"Hey, for $25, here's like an online
3:15Spanish course, and if people don't
3:17click on it, you don't make the online
3:18Spanish course."
3:19>> Yeah, 100%.
3:20>> So, is it similar to that idea?
3:22>> Yeah, it was 404 page not found. You
3:25know, that was what you got. We didn't
3:26even have time to put up like a landing
3:29page, which arguably we should have, you
3:32know, registered your interest or
3:33something. It was just you just thought
3:35like the page didn't load, but it
3:38actually did do what it was supposed to.
3:39And
3:40>> what were example things you were
3:41testing out with those? Well, like what
3:42text links were you putting up there?
3:44>> Well, it turned out, and this is a not
3:47so humble brag, that I wrote the the
3:50most successful copy on the text. So, I
3:53don't know. I have four sisters and you
3:57know three daughters and I just seemed
4:00to channel like my inner teenage girl
4:02who was seemed to be a lot of you know
4:05the most engagement on Facebook at that
4:07time and MySpace and I would misspell
4:11things. I would have bad grammar. I
4:14would just kind of really quickly type
4:18something in like it was a text message.
4:21I mean, now I think people see that and
4:23it's cheesy, but yeah, anything. But
4:25just it would it would be like somebody
4:28telling you about a new game, a new word
4:31game, or like your friend was. And I was
4:33always I like to say I'm always like
4:35trying to find a vein, you know?
4:36Everyone's so jaded that you're trying
4:39to find like an OMFG moment, like trying
4:42to find a vein and you just put crazier
4:45and crazier [ __ ] until
4:48somebody starts clicking on it and then
4:51it's like a perfect trade win. Like you
4:53can set your sales by it because once
4:5522.8% click, 22.8% always click on it.
5:00Like you just it doesn't change. That's
5:02a beautiful thing about consumer. It
5:04just with enough numbers, you know, the
5:07only thing that changed was the LA
5:08election results, but normally consumer
5:10numbers stayed consistent.
5:12>> Well, well, what was an example text
5:14though? Like what were you specifically
5:16like?
5:16>> I don't know if I can remember my texts
5:18from 2007, but um but I'll say that I
5:22also sent these really successful
5:24emails. That's when emails worked. And
5:28one thing I did for poker on April
5:30Fool's Day, I sent a really successful
5:34email saying something like it was
5:36cheesy, but it was like, "You just you
5:40just won a billion poker chips." or I I
5:44I actually I don't remember, but it I I
5:48wrote like I wrote these I would just
5:51try to come up with just weird
5:56crazy nonsequittor things that you just
5:58part of part of what I've understood
6:01with consumers is that we have to like
6:05it's really hard to like shock and
6:07surprise people because we're just so
6:10used to every marketing message and it's
6:14like that's why misspellings work.
6:17Things that made it feel like there was
6:19a human there. It just it had to be
6:21unexpected. It had to be like something
6:25just you that doesn't seem right. Like
6:29it's you're we're looking at patterns
6:31and our brains the more we see a pattern
6:34that we expect the more we ignore it.
6:37And so even if it doesn't make sense,
6:41like like a billboard on 101 to the
6:43airport, if you put up a marketing
6:47slogan that just doesn't quite make
6:49sense, it'll roll around in people's
6:51minds. Like it just,
6:54>> do you know what I mean? It used to be a
6:55good slogan, you remember? And now it's
6:57one that like kind of is just weird.
7:02Like why would they why would they do
7:04that? But but my point was that the
7:07failure machine point was
7:10put up get to a really consistent way
7:13that you can just effectively
7:17in the the fastest least effort way you
7:20know be testing you know 100 ideas a day
7:25or more and our mantra was you know test
7:27more ideas in a week than the industry
7:29tests in a year and that wasn't hard
7:31because I mean the only thing better
7:34than competing with the game industry in
7:362007 was probably like competing with
7:38NASA or you know the defense industry or
7:41the government
7:41>> and yeah because I feel like with Zingga
7:43you were able to basically test lots of
7:47new features on all of your games all
7:48the time. How many games did you have
7:50going at any single moment at your peak
7:52in on Facebook? It's a good question
7:54because we at our peak we probably had
7:5714 global studios and each studio
8:01probably had one major franchise and
8:05then a couple other games and then we
8:08also had studio I in India which is
8:11where we sent all of our games to die.
8:13Any game that got below $100,000 a day
8:16in revenues, we shipped off to India.
8:18And they were amazingly innovative. And
8:21they actually became a center of
8:22innovation because there's a great thing
8:24that happens when nobody cares. You've
8:27get to just try [ __ ] and nobody no one's
8:30going to stop you. And so the the
8:33unexpected things we got were we'd ship
8:35these games off to India to die and all
8:37a sudden they'd start growing. India is
8:39where we found out that subscriptions
8:41worked. you know, they they actually
8:45were became a real center for innovation
8:47and and it was just because, you know,
8:50you could [ __ ] around once the game was,
8:52you know, in our book, you know,
8:54inconsequential.
8:56That's when people, you know, can just
8:59do whatever they want with it. When it's
9:00it's very hard to innovate on like your
9:03your Tiffany brand, you know, that that
9:05the team gets afraid of [ __ ] it up.
9:09>> Yeah. And I I feel like in the gaming
9:12industry, you can do that kind of like
9:13testing, particularly online gaming,
9:14because you can make changes on the fly
9:16and see human user behavior. What other
9:19companies do you see kind of failing
9:21fast? Maybe not hundreds of ideas a day,
9:23but you know, lots of ideas a year.
9:26>> And I'll I'll challenge what you just
9:27said. It used to be that in the gaming
9:29industry, you could test tons of ideas.
9:31We've actually gone backwards because
9:34now we've gone from web gaming which to
9:37me was the real renaissance in game
9:40innovation, game development to now
9:43mobile game development on Unity is a
9:46lot like the old game industry because
9:49it's compiled code. It has to be, you
9:53know, uploaded to the app store,
9:55downloaded. You can AB test text and
10:00little bits of content, but features,
10:03you know, the feature testing in
10:04development has has really slowed down
10:06and moved backwards. And the mentality,
10:09it's not even worth having a weekly
10:10roadmap meeting anymore because it's too
10:13hard for a team to change a game in a
10:16week at this point. So, you have a
10:17monthly roadmap meeting. It's moved from
10:19like a weekly cadence to monthly. So,
10:21that's that's sad to me. But, who else
10:24innovates? I mean we we do see Facebook
10:27who or sorry Meta I still call them
10:29Facebook they hired so many Zinga
10:32people. They hired our whole data
10:34science team, all of our PMs and Meta
10:37has become very a center of excellence I
10:42think for for testing and and data
10:45centric but they they seem to focus much
10:48more though on monetization
10:51things that drive engagement and
10:53monetization and less consumerf facing
10:55features. I've seen Dolingo do some cool
10:59stuff testing bringing out and testing
11:02you know, game mechanics, but I I I
11:06don't know. Who have you seen? I I have
11:07trouble pointing to anybody.
11:09>> Yeah, I feel like AI will and and you
11:14know, I'm questioning your earlier point
11:16that AI, you haven't seen AI do this,
11:17but I feel like AI because you can
11:19program apps in a few hours, like
11:22significant apps, not like toy apps. You
11:24can really program interesting things in
11:27just a few hours using these AI coding
11:29tools. This should be a place where
11:31people fail very quickly. You can launch
11:32an app a day that could that formulate
11:35that 30 years ago would have been a full
11:36business. You can you can launch, test,
11:39and squash an app a day potentially
11:42using AI coding tools. But you're saying
11:44you haven't seen people do that, but
11:45it's probably just the beginning. Well,
11:47it is just the beginning. And when you
11:50say you can launch, test, and fail an
11:53app a day, not in the app store, right?
11:55I mean, you could you could jerryrig it
11:58to be on test flight and not exactly
12:01launch it, but have it turned on. So,
12:03that that's possible. You can do it as
12:07as a web app, but there's a question of,
12:11you know, where are you going to get
12:12people to try it? And that's why a lot
12:15of people have have gone back to Discord
12:18so you can kind of test things in
12:21Discord communities and I've done some
12:23of that. But we we really need well you
12:28need like a distribution sandbox. I mean
12:30I wish that we saw ChatGpt or you know
12:34Gemini. I wish we saw one of the big
12:38kind of LLM portals create an app
12:41sandbox. And I think you would once we
12:44see that we'll see an explosion in
12:47testing and innovation again.
12:49>> But I feel like you know the nature of
12:51virality is that it's e exponential. So
12:54you post a a link on Twitter and okay
12:57two users show it to two friends each
13:00and and on and on and just in a few
13:02iterations that you have a million users
13:04and or am I smoking crack here? Like
13:06>> you're smoking crack.
13:08>> Not the first time
13:09>> or Okay. Well, it's some good crack. Or
13:12I guess tell me what I'm missing because
13:15maybe I'm delusional. How when's name
13:19the last app that you can think of that
13:22blew up virally or or that became the
13:25last time an app became a top 10 or even
13:28top 25
13:30app in the app store? Well, not the app
13:33store, but like chat GPT as an example
13:35where I think wherever they had posted
13:37that link, it would have gone viral
13:39pretty quickly.
13:40>> Okay, but but let's think about that
13:42one. Okay, all you need is $10 billion
13:47and the top AI engineers in the world
13:51and I don't know, eight years of
13:53development and then you could become an
13:55overnight viral success and you can post
13:57your link anywhere and it'll work. I'm
14:00with you on that. I would love to be on
14:02that train, but I that's that's a tough
14:07example to point to all of us to follow.
14:09So, I don't want to pour water on this,
14:12but I I think at the moment consumer
14:15feels uninvestable
14:17uh and it's dead in the water. I mean,
14:20why would anyone do consumer now when
14:23they can clearly build, you know,
14:26enterprise or at worst proumer and, you
14:30know, with a get a smaller base of users
14:32that are willing to pay a lot of money
14:34and, you know, they're already
14:37interested in looking. It's it's pretty
14:40hard to get to the the mass market. I
14:44mean, they're they're not downloading
14:45new apps. They're not looking on Twitter
14:48for new apps to try. So, I'm not saying
14:52it's not possible, but I I'll tell you
14:55this, my rule of thumb for when a
14:57platform was investable at a consumer
15:00level, I used to call it 9010. And I
15:02think I forgot to put this in my book. I
15:04need a an addendum to the book because
15:06there's I keep realizing that
15:09>> Yeah, it's [clears throat] already out
15:11of date. Don't don't get it. Don't read
15:13it. That's my marketing message. So,
15:15I'll say don't buy my book. It's it's
15:17already out of date.
15:18>> That'll be like what?
15:20>> My rule of thumb, I'd called it the 9010
15:22rule.
15:23>> And it was if if a platform enabled an
15:28app to get to 10 million DAUs in 90
15:32days, it was investable. And so I'd look
15:35at platforms and say whether it was
15:37social networks or eventually mobile and
15:39say at the point that we can see at
15:41least one app that could get to 10
15:45million DAUs in 90 days that's
15:47investable. Now it sounds like I'm on
15:50another planet. I mean there's no
15:52platform today that enables that. It's
15:55true because you see you know first off
15:58platforms like Twitter and Facebook
16:00might limit virality. you know, they
16:02don't want necessarily things to go
16:03viral or the app store is, as you point
16:06out, it's just too crowded and discovery
16:07is really hard in the app store.
16:09>> Think about our mentality. I mean, how
16:11many new apps did you download today?
16:14How many did you download yesterday?
16:16>> Zero.
16:17>> Yeah. How many did you download last
16:18week?
16:19>> No, it's a good point. I have all the
16:20apps I want.
16:21>> Right. You're not in a discovery mode.
16:23So, that's we forget that it's like
16:27pushing a boulder up a hill, right? And
16:30when we all the first couple years we
16:32had our phones, we're like, "Oh, what
16:34does that? That's a new clock. Cool."
16:36Right? You're like, you're in the bar
16:38and your friend is showing you like some
16:41dumbass thing and you're interested in
16:44it. Now, it'd be like a joke if they're
16:47like, "Oh, did you see this new app?"
16:48You're like, "What?" You know, you're
16:51nobody's nobody's waking up looking. And
16:54in fact, I don't know if I have these
16:56stats exactly right. We can fact I wish
16:58we had real-time factchecking like where
17:00is AI? Where's this app? Um but last
17:04year there were like 40,000 apps
17:06launched in the app store and zero
17:09became top 10 and maybe one or two out
17:13of 40,000 made it to top 25 and then
17:15they didn't hold the position. So it's
17:18and it's this is a problem for Apple.
17:20They just don't get it. Okay, an app
17:23platform needs
17:25good. They need two things and this is
17:27Roblox just missed their earnings and I
17:30think this is why you need two things to
17:32be a really vibrant app platform. One is
17:35you need long-term franchises like think
17:38about like HBO how important the
17:40Sopranos was, right? For think about I I
17:43like to think of the analogy of
17:45streaming networks like Netflix. If
17:47Netflix had no franchises that we cared
17:50about, they don't have a guarantee we're
17:53coming back to to Netflix. Right. Right.
17:54like you, it's Sopranos. It's the next
17:57season of Game of Thrones that makes you
18:00come back to HBO. And when they stop
18:03having that, their players, their
18:05viewers are at risk. And they need spicy
18:09new. Like, they need both. But and
18:12Roblox, you know, was crushing it last
18:15year. And they had these apps that got
18:17up to enormous peak concurrent users.
18:21And Roblox saw their peak concurrence go
18:23from I think something like five
18:27million. They were on par with Steam to
18:30something like 25 million. And it was
18:32all because of one game called Grow a
18:34Garden. So they had a franchise. It was
18:36like Farmville. And then that thing fell
18:38over. And then there was another
18:40franchise but that was smaller. And
18:42another one was smaller. And now they're
18:44kind of in trouble because they don't
18:46have any franchises and they just have
18:48these kind of meme games that people
18:51play for a couple days and they're not
18:55totally in trouble because they're also
18:56kind of a social network. But but my
18:59point is Apple and I wouldn't say
19:01Apple's in trouble but I think they're
19:03more vulnerable because of this.
19:05>> This is very interesting. So it's it
19:06seems like the problem is like you were
19:08saying with distribution like there's
19:10everything is niched down. So, you know,
19:13everybody's interested in their there
19:15there's no one place where here's the
19:18movie theater. Everybody in town goes to
19:20this and
19:22>> I call that the cocktail party, right?
19:24So, the I this is also in my book, which
19:27you shouldn't buy
19:29>> or you should steal it. Steal from Abby
19:31Hoffman's steal this book.
19:33>> Steal this book. I I don't care, but my
19:36publisher might. So, this is a long-term
19:40instinct vein. in the cocktail party.
19:42And I think it started with Napster,
19:44which was the first time that we all
19:47just self aggregated on the web. The
19:50first time that we all just without
19:52Barry Diller, you know, some paid
19:56professional database in the middle, we
19:57all just connected to each other. It was
19:59like this pirate rogue thing, you know,
20:01it was kind of like Burning Man for the
20:04internet. And and it was cool. And you
20:07had a sense of there's something else
20:09going on. I'm connected to 4 and a half
20:11million other computers right now with
20:14nobody in the middle. And to me that was
20:16the beginning of this whole social media
20:18thing and this network of people not
20:20pages and it didn't really congre there
20:25was no center to it but it was this new
20:27version of a bulletin board just
20:29connecting all of us around music files
20:31and that in my mind led to you know
20:34friendster LinkedIn Facebook all that
20:37but those were cocktail parties that we
20:39came together and then you know was the
20:42mobile phone a cocktail party not
20:44exactly
20:45We've kind of we've fragmented and we
20:48don't really have a cocktail party and
20:50now it's like where are we hanging out
20:51now? We're hanging out on GPT and
20:55Claude, but that's a single player game,
20:58right? Yeah.
20:58>> They haven't figured out social. They're
21:01even kind of stopping social.
21:03>> And I don't know why, but I guess
21:06because they don't need it. Like right
21:08now their game is more compute and then
21:12more, you know, go after the coding
21:15opportunity and that's pretty clear and
21:17obvious. And if they don't do that and
21:19win that, it's like a a series of heats
21:21and if they don't win that sprint, they
21:23don't get to be in the next one. So even
21:25if they get that they're going to need
21:27to worry about consumer engagement, it
21:30just won't matter. They need to win this
21:33computing scaling and revenue war before
21:36they get to even think about holding our
21:38hearts and minds.
21:41Look, most of the customer service bots
21:44out there are 15 years old. They're
21:47horrible. You're constantly trying to
21:49like reach a real person. I can tell you
21:51from personal experience that Decagon is
21:54not like that. And look, if your team is
21:55dealing with more support tickets, more
21:58channels, and higher expectations from
21:59customers, you've probably thought about
22:01using AI for customer service, but most
22:04of the options out there are either too
22:06limited or take way too long to set up
22:09or are just old school. They don't work
22:11that good. And that's why I'm excited to
22:13tell you about Decagon. Just check it
22:15out. If you have a company and you need
22:16customer service, just check this out.
22:18Decagacon helps companies create
22:20personalized concier style customer
22:23experiences with AI agents across chat,
22:26email, voice, and SMS. They're available
22:29247. They feel natural to talk to and
22:32can resolve customer requests on their
22:34own so businesses can keep up with
22:36requests without losing their personal
22:38touch. Decagon plugs directly into your
22:40existing systems so agents can get live
22:43quickly. And because workflows can be
22:45updated using natural language, teams
22:48can make changes themselves without long
22:50engineering cycles. This is key. If you
22:52notice something in the interaction
22:53that's wrong, you just tell it, "No, no,
22:56no, say it this way." Because it's using
22:58the same technology as all these AI
23:00coding tools. It's like you're coding
23:01the customer service in real time. That
23:04means less time spent managing
23:07workflows, more time focused on
23:09customers, and you'll see the ROI in
23:12weeks. Also, Decagon gives your team
23:14full visibility into why agents make
23:16decisions and what's happening across
23:18every conversation, so it's easier to
23:21spot trends and improve the experience
23:23over time. Decagon helps power millions
23:25of conversations every day for brands
23:28you know and love, including Avis,
23:30Affirm, Fanatics, and Aura. Ready to
23:34transform your customer service? Go to
23:36decagon.ai/jamesde
23:40ag.ai/ AI/James to get a personalized
23:43demo and see what DecaGon can do for
23:46your team. Just get the personalized
23:47demo. Just see check Decagon out at
23:50decagon.ai/james.
23:52That's decagon.ai/james.
23:58It's a good point because there's there
23:59is of course if you if you let your
24:01imagination run wild, there's a lot of
24:02opportunities for social media on these
24:05AI platforms, but they don't need to
24:07play that game. And it should be
24:08mentioned you were there. You're like
24:10the godfather of social media. Like
24:13Tribe, I would argue Tribe was the
24:15second, you know, social media platform.
24:18Geio Cities maybe the first one funded
24:21by your buddy Fred Wilson and sold to
24:23Yahoo for billions and then you started
24:25Tribe right before like kind of MySpace
24:30and then Facebook of course blew up.
24:31What do you think
24:33>> made Tribe not a success? It was sort of
24:35a success at first like everybody had
24:36heard of Tribe. Like I had heard of it,
24:38but I wasn't using it. Like why didn't
24:40>> That's the problem. That's the problem.
24:43>> Well, why didn't I I didn't even really
24:44use MySpace, but I did use Facebook very
24:46aggressively right when I was able to.
24:49>> It's tribe and tribe is a major
24:54it should be a case study in what not to
24:57do. And it and it's a perfect case study
24:59in this these theories I have of trust
25:02your winning instincts but not your the
25:05idea that your ego is attached to and do
25:08proven better new like a scientist with
25:10a white lab coat because tribe
25:13should have been huge. I mean, it's hard
25:16for people to imagine today. You had to
25:18will yourself into failing at that
25:20point. Like, there wasn't just, you
25:23know, Friendster, Facebook, MySpace.
25:26There was Tagged, Bibo. Every social
25:29network kind of worked. People were just
25:32into it. You remember you got an email
25:34saying, "Mark wants to be your friend on
25:37X on on tagged or whatever." And you're
25:40like, "Oh, cool. What's that?" And you'd
25:42click on it. You know, these had 80%
25:44open rates, 50% click-through rates. You
25:47know, it had numbers that people can't
25:50imagine today. They wouldn't they
25:52wouldn't believe it, right? So, with
25:55Tribe,
25:56I had these three winning instincts and
25:59one losing idea. And I just like
26:01heroically, stoically stuck with it. And
26:04I I was ignoring my metrics. You know,
26:08it was a sinking speedboat. It was super
26:10viral. It grew way faster than LinkedIn.
26:13Like it was the tortoise in the hair.
26:15Like Reed had LinkedIn and LinkedIn just
26:17kept slowly slowly ticking up and had no
26:22engagement. And Tribe was this rowdy
26:24party. It was Burning Man. It was
26:28getting like way more installs a day,
26:30but it had no retention. It boiled down
26:33to a very small group people who loved
26:35it. and I got trust wrong where you know
26:40Friendster and Facebook and LinkedIn all
26:42got trust right. I didn't get that
26:45people were just putting themselves out
26:47on the public web and especially
26:50mainstream people, especially women were
26:53not comfortable with stranger danger
26:56with people they didn't know contacting
26:59them. And I made Tribe this just open,
27:03rowdy platform. And the reason I did it
27:05was cuz I was focused on Craigslist and
27:07the listings business. And I needed
27:09people to be able to contact each other
27:11to
27:13buy each other's couches and find
27:14roommates and jobs or I believed I did.
27:17And I was so focused on this listings
27:20business and Craigslist. And I thought,
27:23okay, I'm going to acquire my audience
27:24through social networking and virality,
27:26but then I'm going to move them into
27:29this jobs and listings business. And I
27:31ignored what my users were into, which
27:33was the tribes. So the idea of the urban
27:36tribes and networking through your loose
27:38ties through groups that you're loosely
27:41a part of that was gold and I completely
27:45ignored it and there was huge engagement
27:48and heat. I barely gave them any
27:51features and it just took off. And so
27:55many people have stories about meeting
27:56their wife through that or this amazing
27:59trip to Brazil through their mission bay
28:01group. And and I did I was not on the
28:05Paul Graham train of get a 100 happy
28:08users and then keep building what they
28:10want. I was I had all these happy users
28:12and I was like ignoring them and they
28:14were also extroverts posting a lot of
28:16dickpicks and you know it was
28:20>> extroverts do.
28:21>> Yes. It was there's there's no place on
28:24the well now there is I guess now
28:25there's like only fans and now there's a
28:28good business around it but you know
28:30this there's this these communities that
28:32nobody wants and they ended up on Reddit
28:36and other places and they ended up
28:39figuring out how to turn that into a
28:40business. But community always
28:42overdelivers
28:44when you enable it on the internet
28:46because people are lonely and if they
28:49can find new friends on the internet,
28:51they will and then they'll hang on to
28:53them. And that was true in all Zinga
28:55games and that was true with Tribe. But
28:58it didn't align with perfectly with my
29:00business model and it didn't align with
29:02the retention of mainstream people. So,
29:06you know, I it so many of my theories
29:10that I built into my book and my
29:12approach to to product management,
29:14product thinking came from the long
29:17painful failure of Tribe. It just it was
29:20my learning grounds. It just was
29:22amazing. It's interesting about the
29:24trust like Facebook had of course theedu
29:26so you knew hey these people are going
29:29to college just like you are and oh
29:31these people are also at Harvard just
29:33like you are. So you might run into
29:35them. So there's less like violation of
29:39kind of politeness when there's trust.
29:42>> Well, it's your tribe, right? You know,
29:45they they're at your same school.
29:46They're beyond some minimum boundary
29:50that they are they have a real identity.
29:54They they're part of this broader
29:57community. And that turns out to make a
30:00big difference. You know, years later, I
30:03discovered Rya and I with along with
30:06Reed, we put up all the money for Rya.
30:08And Rya
30:11got trust right also in a way that the
30:13rest of the dating apps missed, which
30:15was this human curation level that that
30:18a committee just like solo house had to
30:21vet you. You had to have a real
30:22Instagram account and humans checked it
30:25and accepted you or didn't. And so it
30:30made I like to say that dating apps were
30:32a 1 out of 10 experience and Rya was a
30:34three out of 10 and it just sucked three
30:36times less and they just took the
30:39stranger danger or that creepiness out
30:42of it to the point that and the other
30:44thing I want to circle back to because
30:46this is so relevant today as we think
30:48about how is consumer going to be
30:50reinvented? How is social going to be
30:52reinvented? The beginning of all this
30:54when Reed and I really spent all this
30:57time in 2002 brainstorming what he
31:00called web 2.0 was lead generation.
31:03Okay, this may sound counterintuitive.
31:07It's not the way you think about
31:08Instagram or a cocktail party, but it's
31:12the utility. You get entertainment for
31:14sure and that's the reason you come
31:16there. But the value you get out of a
31:19good cocktail party is often lead
31:21generation. It's this social serendipity
31:23that leads to
31:25a recommendation on a trip you end up
31:28going on or a job lead or a date or you
31:33know you don't know what you're going to
31:34get but you're around a bunch of
31:37interesting people some you know some
31:39you don't but it's really well curated
31:42and it leads to it's much higher signal
31:45for lead generation and that's what Reed
31:48nailed with LinkedIn and when you think
31:50about how will LinkedIn be reinvented?
31:53How will Instagram be reinvented? It's
31:56got to change this paradigm on,
32:00you know, signal to noise and and the
32:02return on your time. And in this AI
32:05agentic world, it has a huge opportunity
32:08too because it's almost like reinventing
32:11Napster meets LinkedIn because our
32:13agents could be going to cocktail
32:16parties all the time, networking
32:20and and kind of navigating this membrane
32:24of trust and access and information, you
32:28know, in order to to bring us back
32:30leads. That's interesting that that it
32:33boils down to to leads. I mean, you can
32:36similarly though right now you you've
32:39mentioned like the app marketplace is a
32:40place where things aren't going viral.
32:42Like it's really fascinating statistic
32:43that 40,000 apps launched last year and
32:45none made it to the top 10. What about
32:47if everybody went to their own AI
32:49curated curated app marketplace? like
32:52not necessarily just mobile apps but web
32:55apps, mobile apps, iPhone apps and so
32:58on. But it's just this is my AI curated
33:01app store.
33:02>> Well, first you got to have a desire for
33:04apps, right? I mean,
33:05>> yeah, that's true. Right. So, it's funny
33:08cuz Facebook Facebook's like the app you
33:11didn't know you wanted. Like I didn't
33:13realize, oh, deep down I wanted to stay
33:16in touch with my friends from second
33:18grade and see how their kids did in
33:19their soccer game last night when I
33:21hadn't spoken to them in 30 years. But
33:23the reality is you had these
33:24archaeological layers of your life that
33:27Facebook allowed you to drill down on
33:28which doesn't really exist anymore. Like
33:30nobody really uses Facebook anymore. No,
33:32I mean I I don't even I get on Instagram
33:35once a month and most people I talk to
33:39who don't use Instagram say it with a
33:40level of pride and it's like there's
33:43this weird thing that you can track you
33:46can you can look at the NPS scores. Do
33:49you know about net promoter score?
33:51>> No.
33:52>> Okay. It's it was a big deal for a while
33:55in consumer but the net promoter score
33:58is a very simple measure of the brand
34:01power like what's the likelihood that
34:03someone would recommend your service to
34:07someone else and eBay at its height and
34:10Amazon have hit like 80 so 80% of people
34:15feel positive and would recommend it.
34:18You can have a positive or a negative.
34:20And negative NPS means people actively
34:23tell people not to use it. Right? That's
34:25the cable company. That's the Democratic
34:28party. That's maybe the Republican
34:30party. That's probably any politician.
34:32That's the IRS. Right? There's a weird
34:34thing that you'll see, which is, and we
34:38saw this with with Farmville, and then
34:40we saw it again with Facebook and
34:41Instagram that Farmville had about a
34:44positive 40 NPS when people were playing
34:47it. But then after they stopped, it was
34:50like they'd given up cigarettes and we
34:52were like a negative 35 NPS.
34:54>> And and we saw the same thing. This I
34:57promise you the same is true with
34:59Instagram and you probably seen it with
35:01your friends, right? When somebody has
35:04it's like they've quit smoking. They
35:05become like a zealot, right? If someone
35:08has got is no longer using Instagram,
35:11they will say it to you with a level of
35:13pride. I don't use Instagram anymore,
35:15right? I'm I'm off of that drug, right?
35:17I don't waste my time on that. They
35:20don't say it like they miss it. They're
35:21like, "Ah, I wish, you know, I just
35:23don't have time for Instagram. My family
35:25got in the way. Life got in the way. I'm
35:27really trying to get back to Instagram."
35:30You know, nobody says that.
35:32>> You know, it's funny though because
35:33lately Instagram has replaced TV for me.
35:36Like, I go on Instagram every day for
35:38entertainment.
35:39>> But why? I mean, what is your content?
35:41Like, what? Take us through yesterday.
35:43Like, what was your Instagram? It goes
35:46back and forth. Like I see a lot of like
35:47history stuff or unusual facts and you
35:50get to dive down into them in the
35:51Instagram reels. And then there's
35:53Clavicular who's like this 20-year-old
35:56kid who talks about looks waxing and and
35:58it's gone viral. So So it goes back and
36:01forth.
36:01>> I'm with you cuz that's Twitter. Sorry,
36:04I I haven't gotten over calling it X,
36:05but that's X for me. I'm addicted to it
36:08and it's my social network. It's my
36:12news. I get stock ideas. I I learn about
36:17history. I have a whole community there
36:20of people I only know by their screen
36:22names. Yeah, I get that. I do. But then
36:26again though, most of the content now, I
36:28wouldn't I shouldn't say most because I
36:29don't really know. But on Twitter and
36:31Instagram, I feel like it's a lot of AI
36:33generated content. like give me 14
36:36unusual facts about, you know, the Civil
36:39War and now I have a whole Instagram
36:41reel, you know, generate images, source
36:44all the facts.
36:44>> I don't know about you, but the reals I
36:47find insidious because I get pulled into
36:50these reels and they start looping one
36:53into another and it just
36:56>> gets me and I don't like it and it's
36:59like that's bad, that's evil.
37:01>> Would you call that viral though? like
37:02some of those are getting like a million
37:04views.
37:05>> It's funny to say what's what's viral.
37:07It's it's being surfaced to you by an
37:10algorithm, right? Just like Tik Tok. And
37:13so it is it's a kind of virality, but
37:15it's a it's an algorithmic manufactured
37:18virality that we kind of don't know or
37:22understand. I guess you could reverse
37:24engineer it. Um it is it is a virality.
37:27It's like a machine-driven
37:30virality or it's a human and
37:32machine-driven virality. It's different
37:34than people wanting to actively share
37:38with each other. It's out there and and
37:40I'm sure there are I mean there we know
37:43there's people blowing up as influencers
37:47and podcasters and videos and stuff like
37:49that. Maybe, you know, we're thinking
37:52backwards when we say apps. you know,
37:54maybe there's a new kind of service
37:57that's not invented yet that will spread
37:59through that kind of mechanism that will
38:02but it's got to be something we engage
38:03with. And I keep thinking it's something
38:07that helps us flip to being generative
38:10and not consumptive because I don't know
38:12the value. Personally, I see little
38:14value in consumption and just getting
38:17you to watch something and consume it.
38:20To me, I want to give you something that
38:23is going to help you generate something
38:26you think is valuable.
38:28>> Maybe we're just missing times when
38:31there were fewer and now there's a lot.
38:34So, so for instance, in 20 years ago,
38:38there were just fewer people on the
38:39internet. Like when you started Zingga,
38:41there was basically that was the first
38:43time basically we hit a billion users,
38:45which doesn't sound like a few, but now
38:47there's like five billion users of the
38:49internet. Yeah.
38:50>> When I started this podcast, for
38:51instance, there was just a couple
38:53hundred podcasts and now there's 5
38:56million podcasts and
38:58>> so it's harder. Okay, there's 40,000
39:01apps going into the store last year.
39:02Maybe in 19 in in 2009 or 20 whenever
39:05they launched the app store, there was a
39:07thousand apps. I don't know how what the
39:09first
39:09>> not a thousand. I mean, honestly, if we
39:12look back, there was probably around
39:1640,000. But the difference is that the
39:19the there was many more people in even
39:22though there were less users, they were
39:25installing a lot more. They were
39:27downloading and installing more. And
39:29then development
39:32it exploded. It probably got up to
39:35500,000 apps and now it's it's crashed
39:39way down because it's, you know, it just
39:43doesn't it's not a very good
39:45distribution mechanism for anyone
39:46anymore.
39:47>> Yeah. So, what does what does the
39:48entrepreneur do? Like again, AI coding
39:51seems like a place to to create useful
39:54apps easily. And if something's super
39:57useful, so and we should talk about your
39:59your proven better new approach which is
40:03which is a great model for kind of you
40:05know a great filter for saying what app
40:08should I build? Oh, is it proven? Is it
40:10an improvement industry? Is it is it
40:11better than what existed before? And is
40:12there something new that you could kind
40:14of test? Uh, it seems like a great
40:17filter for for finding AI apps that you
40:19could build that might be not viral but
40:23useful enough that people will pay for
40:24it or or some there will be some
40:27business model around it. Well, in fact,
40:29we made a a GPT and a skill file
40:35out of first that chapter and then the
40:38whole book that does an amazing job on
40:42proven. It's it really accelerates you.
40:45I mean, it's like you shouldn't do
40:47anything without starting with that.
40:49It's just you should just we'll post it
40:52somewhere, GitHub or something.
40:54>> Well, maybe describe the the proven
40:56better new because it's really
40:57fascinating and I also like how you
40:59apply it to entertainment. It really
41:01works well when developing a TV show for
41:03instance. You can see it in every TV
41:05show out there. You know, oh, this genre
41:08is proven like a western
41:09>> or the mechanics. I mean, and it's not
41:11just in games when it's great to start
41:14learning this concept around games
41:15because it's easier to get your head
41:17around it, but it actually applies to
41:18everything even to enterprise, you know,
41:22products and even analog products. The
41:25idea is this. It's that there's well
41:29first of all you have some instinct and
41:32it's really important for you to
41:34identify and isolate what is the
41:37instinct that I'm feeling in my gut and
41:39write about it. get in touch with that
41:41instinct and then realize separate that
41:44from what you're instantiating that the
41:47way that you're productizing that
41:49instinct and just separate it because
41:52that is going to massively increase your
41:56odds of success or reduce your odds of
41:58failing for the wrong reasons. and then
42:01go out and say, "Okay, what are the
42:03closest products that have done some or
42:06most of what I'm excited about that I
42:09can see on the same platform for the
42:12same users and I'm going to like really
42:15be the get my PhD in that product. I'm
42:18going to get down to the pixel level. I
42:20mean, I'm going to I'm not going to let
42:22any second or click or pixel of that
42:25experience go unnoticed. I'm I'm just
42:28going to have such a fine microscope on
42:30it and I'm going to kind of rebuild that
42:34and think of the mechanics the what are
42:37everything from the first time user
42:39experience. I'm going to go screen by
42:41screen click by click into everything
42:44every feature whether you think of it as
42:47a feature or not. If there's an address
42:51book upload, like anything that makes up
42:54that experience, you want to capture and
42:57look at that and try to understand why
42:59is that proven? Why do people engage the
43:02most with that product in in that
43:06market? If it's very successful and it
43:08speaks to you, you can consider that
43:11proven, then you say, "What could I make
43:15obviously better about that product?"
43:18And it's not what you think is better
43:19because that's new. What 10 out of 10 of
43:23the existing users without question
43:25would say, "Fuck yes." And that's it's
43:28now free. There's no download. You know,
43:31it's half the price. It's something
43:34noticeably better. Something people care
43:36about in the product and it's noticeably
43:38better. And it could be very very small
43:42fine-tuning that if you think of our
43:44game Words with Friends, you could say
43:46that was just Scrabble. People said,
43:48"Oh, that's just a a copy of Scrabble."
43:52But it must have been more than a copy
43:53because it trounced Scrabble. Like we
43:56got up to 14 million daily active users
43:58and Scrabble got up to like 2 million
44:01daily active users and that was even
44:03years later. So there must have been
44:05something else that we that we got
44:08right.
44:09that [clears throat] was better. And so
44:11sometimes it can be the polish, you
44:13know, and in Farmville,
44:15we had better artwork around the crops
44:18and better math around the crops. So it
44:20can be something very small. We had less
44:22clicks. New is your novel idea. That's
44:26your innovation zone. That's that is
44:29what maybe you're most excited about.
44:31It's also what's most likely to fail.
44:33And one of my mantras that's still
44:36repeated in the hallways at Zingga is
44:38all new fails. And that might sound like
44:40a beatdown. If you start with that
44:43mantra, you won't be disappointed
44:45because
44:46I mean there's lots and lots of
44:48statistical proof that all new fails
44:50like the app store last year where 100%
44:53of new apps failed, right? I mean
44:55statistically you can see that all new
44:58fails. It doesn't mean no new products
45:02break through and it doesn't mean you
45:03shouldn't go for things. But it does
45:05mean that you have to be humble and
45:07curious and assume that this one new
45:11idea of yours probably will fail and
45:15isolate that. So part of this is don't
45:18fail for the wrong reasons. If you
45:21changed everything in the first time
45:22user experience and that wasn't better.
45:26So 10 out of 10 people don't thank you
45:28for changing that and it wasn't
45:31obviously your new innovation zone, then
45:34you're a shitty product maker. You know,
45:37junior product makers make that mistake.
45:39They just they want to be what they
45:42think of as a Picasso, but they're not a
45:43real Picasso because the real Picasso
45:45actually was a first he was a master at
45:48tracing and copying other people's art
45:50before he ever went and did his own art.
45:52And I like to say, you know, if we
45:55really respect our users and if we if
45:58we're really really ambitious, then
46:01we're going to kill our egos and we're
46:03going to say, I'm I'm going to define
46:06success not by respect from my peers,
46:09you know, not by what my ego wants to be
46:11fed, but I'm going to define innovation
46:15in the eyes of, you know, a nurse in
46:18Indiana. And does does she just vibe
46:21more with my product than some other
46:23product? And she can't even say why. And
46:25because my innovations were so little
46:28and unnoticeable. And that was words
46:30with friends. It was just such better
46:32polish and it was smooth and it was
46:36instantly social. And that was Zingga
46:39Poker. We completely copied the leading
46:43poker games. There was not a single
46:46thing you could detect that was
46:48different, but there was no download.
46:50So, it's like the they all were on
46:52Facebook before us. There was 10 poker
46:54games. They all were downloads. We
46:56weren't. Okay? So, they were losing 50
46:59to 80% of their users at the first click
47:02because you lose 80% of your users when
47:05you say download, 50% when you ask them
47:07to click. So, they already were behind
47:10me. And then you know so so it was like
47:14their game proven better because there
47:17was no download and the new idea
47:20happened to work on my first try and
47:22that's the way it goes you know when
47:24you're finally ready to do proven better
47:25new you don't have to it turns out your
47:28first idea works and that's not a
47:31promise but that's a weird thing that
47:34happened to me and my new idea was real
47:36people from you know Facebook from your
47:39work and 25% of the time people were
47:42joining their friends at the table. So,
47:44and that happened to work on the first
47:46try, of course, but but it's I don't
47:49know why people wouldn't, you know, use
47:51this framework and create a GPT or a
47:54skill file. At least it can do proven
47:56for you and benchmarking in seconds. It
48:00can go and do that at an A level, like
48:02as as good as the best Zinga PM. It's
48:06amazing at proven. That's what AI is
48:08great at. It's a B at better or worse
48:12and it was an F at new. So that's good
48:16news for us because it means that the AI
48:19isn't going to just replace us anytime
48:21soon.
48:23>> You know, Jay, I was thinking the other
48:25day, I remember back in the late 90s and
48:28early 2000s, I was trading, managing
48:30money, building businesses. There were
48:33spreadsheets everywhere, post-it notes
48:35on the wall. I was using a ton of
48:37different software programs that just
48:38didn't talk to each other. And every
48:40time someone asked a simple question
48:41like, "Hey, how much cash do we actually
48:43have right now?" It took three people
48:45and more to figure it out. Fast forward
48:47to today and I still see founders living
48:51that same nightmare. Except now, of
48:53course, the world moves at light speed.
48:55They say that every single day your
48:57business is late to AI. You fall two
48:59days behind. The competition isn't
49:01waiting. But here's the good news.
49:02There's actually a way to catch up and
49:04then pull way ahead. It's called
49:06Netswuite. Next, look, you probably
49:08already know Netswuite. It's an AI
49:10powered business management tool.
49:12Connects all of your data in one place.
49:14I'm talking financials, inventory,
49:16commerce, HR, CRM, whatever. Over 43,000
49:20companies use it. But Netswuite next,
49:23this is their next huge leap. AI isn't
49:26just bolted on. It's built into
49:28everything you do. It automatically
49:29surfaces custom insights throughout your
49:30day. Stuff you didn't even know you
49:32needed to see. like AI agents are
49:34working right alongside you solving
49:36problems, knocking out all the routine
49:38work that usually eats up a lot of time.
49:42And the best part is you can ask
49:43questions about your business. You can
49:45ask it like, "Oh, what was our growth
49:48last month? Do we need to hire new
49:49employees? Which sales context do I need
49:51to follow up on?" And on and on. It has
49:53all this data. It's not some generic
49:55one-sizefitit all thing. Netswuite gets
49:57customized for pretty much every
49:59industry you can think of. So whether
50:01you're running an e-commerce brand, a
50:02manufacturing business, it doesn't even
50:04matter. Your company's doing millions,
50:05hundreds of millions, it doesn't matter.
50:07Your business finally can use AI. You
50:09don't have to like figure out how to
50:10code it yourself. Just go to
50:12netswuite.ai/james
50:14right now. netswuite.ai/james.
50:17Build for every industry. Ready for
50:19every boardroom. netsweet.ai/james.
50:23Do it. You'll thank me later.
50:25I love this concept of proven equals
50:29copy the past. Like you mentioned in the
50:31book, how Picasso would literally trace,
50:34you know, the old masters in order to
50:36learn their style and maybe even, you
50:38know, emulate or copy or or steal. And
50:41this is such a hidden advantage that
50:43people psychologically
50:45don't normally subscribe to. They're a
50:47lot like and you can give another great
50:49example in the book of Slack versus I
50:51guess it was hip hip chat. Uh
50:53>> yeah,
50:54>> you know, where where Slack basically
50:55was the same thing. They just had, like
50:58you mentioned with these other products,
50:59they had smoother polish. They it was
51:00made by a game company. So, and Hip Chat
51:03was like an enterprise boring enterprise
51:05company. So, Slacker just had a more of
51:07a game-like feel in there. It had what
51:09about 8,000 installations the first day.
51:12>> Yeah. And Exactly. And this is actually
51:16an Easter egg for people, which is fun.
51:19If you can go and make a proumer
51:23enterprise app and you actually put in
51:26little doses of fun or you know game
51:30mechanics in it, you'll be shocked. I
51:33mean how how dumb it is. I mean I was
51:36talking to some of the team at OpenAI
51:39and they said and this always happens.
51:42Every one of these big hyperscalers has
51:44had this experience. It's not what they
51:46want. It's not what they build for but
51:48it's what they get. They said, "Oh my
51:50god, someone put a dumb, you know, pet
51:53game, a little AI pet game out on GPT
51:58and it's gone viral and everyone inside
52:00OpenAI is playing it and it's so weird
52:04and we don't really know why." That was
52:07the same thing at Facebook. They were
52:08shocked when games took off. They It's
52:11not what they built for or intended. And
52:15that's in a consumer space. You know,
52:17Slack proved that people want fun in the
52:22enterprise space. You know, we did a a
52:24joint credit card with American Express
52:26that I wanted to call the fun card. They
52:29called it some terrible name. Uh I think
52:32it was called the Discover Card, but
52:34every time it People [ __ ] love
52:37gambling, right? We love gambling. We
52:39have boring lives, right? It is [ __ ]
52:42boring to buy [ __ ] with your credit
52:43card. What if you could win something
52:47every time you brought your credit card
52:49out? Would you pick that over not? I
52:52think yes. And so did 20,000 other kids
52:56who signed up for the Discover card,
52:57right? Every time they bought something
52:59with it, they won something. And once in
53:02a while, it was something big. They they
53:04won something in our games. That turns
53:07out it worked. You can use these game
53:10also these game engagement mechanics in
53:12an enterprise app. You could reward
53:14people for coming back once an hour. I
53:17mean, every day. You could create
53:20leaderboards. I mean, all this [ __ ]
53:23works. It's proven. It's And it's Slack
53:26proved it in the enterprise. Nobody
53:28bothered copying them. And I'll just say
53:30this for your us for your users, for
53:32your listeners, the users of this
53:34podcast. It's it's a moral arbitrage in
53:37the Peter Tillian sense of a moral
53:39arbitrage. He loves a good moral
53:41arbitrage. This one is just sitting
53:43there for all of us. If you were willing
53:45to kill your ego, okay, if you're
53:47willing, and by the way, if you do this
53:49really well, the way Stuart did with
53:51Slack, your users, no one even accuses
53:54you of copying, right? That's the real
53:57master class of this. If you can
54:00perfectly copy a proven successful app
54:03and no one thinks it's a copy, that is
54:06the [ __ ] magic trick. like you just
54:09you're like look over here folks like
54:12look at my namaste when you first enter
54:15and my cuddly look and feel. Don't look
54:17over here where it's just 100% hip chat
54:21right I mean no one really accused them
54:23of copying hip chat except me and I and
54:26it was respect when I see that kind of
54:28level that is a masterclass in that is a
54:32brilliant product maker.
54:34>> Yeah. No, I I agree. And and what I like
54:37in this is you're sort of telling people
54:39you have permission to go back to the
54:42multi-billion dollar industries that are
54:45out there and just do it again and have
54:48fun with it. So,
54:50>> this one of, you know, I this is another
54:54thing I think I forgot to put in the
54:55book, but one of my life philosophies.
54:57>> You already made like part two, not even
54:59an addendum. You need a
55:00>> sequel. Yeah. Yeah. I'll just post it
55:02somewhere on a a medium or something
55:06[clears throat] or a Substack. Sorry, I
55:08don't want to sound so outdated. I'm an
55:09investor in Substack. I mean Substack.
55:12One of my one of my other my co-authors
55:14started calling these Marcisms is the
55:16more things change, the more they stay
55:17the same. And so even in an environment
55:20like this where it feels like nothing is
55:22yet proven. So you could argue in
55:24consumer AI very very little has been
55:28proven on if you call it a platform you
55:31know this technology platform with this
55:33audience then you go look at what was
55:35proven before because it will probably
55:38be reinvented the very first viral app
55:41on the social networks and pretty much
55:44every time one of these consumer
55:46platforms blows up are the I don't know
55:48if you remember these truth boxes or
55:51honesty box or crush apps. This is where
55:54they're they start with high schoolers.
55:56It's where you get a message. You know,
55:59someone you know has answered these
56:01questions about you. You know, find out
56:03what they think of you. Find out what
56:05three friends all said about you
56:07anonymously. And you click and it says
56:10you need to answer questions about three
56:12friends before we'll show it to you. And
56:14you put them in. And then they text
56:16them. And these are hyperviral and they
56:19always work. I mean it's what gas and
56:21tbh you know to be honest and Nikita did
56:25these on mobile and now I know a couple
56:28of startups that are doing these with an
56:30AI twist where you have the AI in the
56:33middle that's being the intermediary and
56:36is a little more smart and
56:38conversational and you're going to see
56:40those probably work you know again with
56:43an AI twist and that's that is a variant
56:46of of proven better new
56:48>> it also reminds me somehow Do you
56:50remember Genie which was also around
56:522007208
56:54it was build your family tree and by the
56:57way if they're alive put their email
56:58address here we'll send them to build
57:00their family tree and it went to 100
57:02million users in a week and I think
57:06Genie didn't quite work out but they
57:07were using internally Yammer to discuss
57:10the company that they invented Yammer as
57:13a tool to just discuss their own company
57:15and that sold to Microsoft for billions
57:17of dollars. I didn't know that. Um I
57:20know ancestry.com
57:22got pretty big and I think that's
57:25another great area. Yeah. That you could
57:28and I would encourage listeners to, you
57:32know, don't invent a new business. Like
57:35[ __ ] that. Go after big boring existing
57:39business like jobs, dating, you know,
57:42ancestry. find things that already have
57:45lots of users and money and see if you
57:47can add an AI twist to it or you know
57:50come up with a way to get viral you know
57:53through the kind of virality you were
57:54talking about because there's videos you
57:58know that the machine algorithm picks up
58:00but and then if you're going to copy
58:02them like copy them like a master and
58:05and don't just do one you know variant
58:07do 20 and don't build the 20 like first
58:12start with testing, you know, what
58:13clicks or which start with what videos
58:16get viral and then build your product
58:18from there. I mean, this is it's kind of
58:20thinking top of the funnel. Okay, I want
58:22to like find the key, then I'm going to
58:24go lower. And yes, I agree with what you
58:27said a while ago that AI can enable us
58:30to build prototypes or test things,
58:34create the videos, test things quickly.
58:36Unfortunately, it seems like people are
58:38using AI differently now. They're
58:40wanting to just build their prototype
58:42and they're excited about that one
58:43prototype and then they just keep going
58:45with it. So, I'm not seeing people I'm
58:49not seeing great examples of people
58:51testing and failing.
58:53>> Yeah, I guess that's right. Have you
58:54programmed any AI apps yourself using,
58:56you know, any of these coding tools?
58:58>> I I have and it's amazing how badly I
59:02did, how hard it was. Um, it's not ready
59:06for what I'd say the prime time of dumb
59:09[ __ ] like me. Like I tried Replet and I
59:14tried to make a website for my partner
59:16Hillary and Midjourney is magical. I
59:21mean, Midjourney has been magical since
59:23the first time I used it. The user
59:25interface through Discord is
59:29>> insanely
59:31hard and terrible. Yes. I I kind of hate
59:34Discord because of their journey.
59:36>> Yeah. But somehow I wouldn't change
59:38that. That's proven. Like if I was
59:39trying to compete with them, I'd make
59:41the same [ __ ] up, you know, weird ass
59:44thing because that's what they did and
59:45it's working. But but Midjourney
59:48literally makes me look more creative
59:50than I am. I mean, it's unbelievable the
59:53output of MidJourney. So, I made her a
59:55logo. She is she thinks she she the
1:00:00nature of the world spoke to her and
1:00:02told her her mission is to to be a
1:00:05garbage collector. Her mission I don't
1:00:07know if she feels okay about me talking
1:00:09about this but her mission is to find
1:00:11different ways to like collect garbage
1:00:14in the world and be like a vulture. And
1:00:16so I made her this really amazing logo
1:00:18on Midjourney with a vulture like flying
1:00:21over a city looking for garbage. And
1:00:24then I tried to make her a website with
1:00:26Replet, just a website. And I spent a
1:00:28week on it and I couldn't get Replet to
1:00:32move the logo and it kept covering the
1:00:34title of the website. I mean, literally
1:00:36a week. And yes, I am terrible. I suck.
1:00:40That's why I have a good mind for
1:00:41consumers cuz I represent the dumb
1:00:44[ __ ] you know? I represent I'm not
1:00:47willing to learn how to use Replet. I'm
1:00:49not willing to have someone teach me. I
1:00:53just don't care enough. And then I used
1:00:56Claude code and I tried to make an AI
1:00:58assistant and I spent two weeks on that
1:01:01and the Google connector broke every
1:01:04day. And so then I had to create another
1:01:08health check app that would check the
1:01:10health of that every day and then try to
1:01:14fix it. And then it would get stuck on
1:01:16these allows. And then I tried to have
1:01:18Claude coworker dispatch try to click
1:01:21allow for me and then it started to
1:01:24refuse to do that. So it was it was a
1:01:28good learning experience. I I did make a
1:01:31firsterson shooter game in like an hour
1:01:36that I played one time. Um I found my my
1:01:40altitude was Claude Co. What I find
1:01:43magical is dispatch skill files agents.
1:01:48You know, I've created I create agents,
1:01:52persistent agents, and I name them after
1:01:54my family. So, Carmen, my daughter, is
1:01:57my product manager.
1:01:58>> And so, what do your agents do? I'm I'm
1:02:00I'm actually having more success with
1:02:02the coding stuff on my own, but the
1:02:04agent stuff, I'm not sure. What do you
1:02:06use the agents for?
1:02:07>> I love the agents. They're I find that
1:02:10magical. So, I walk around and have long
1:02:14conversations with all of my agents and
1:02:16we have like team debates and I have
1:02:20Hillary is my head of design, Georgia is
1:02:24my my general manager, Wyatt is my head
1:02:27of engineering and most things start
1:02:30with Carmen because she's my PM and she
1:02:33does proven better new benchmarking, you
1:02:36know, she's somewhat cynical. She comes
1:02:39she comes back to me constantly
1:02:43pouring water on my ideas and telling me
1:02:46this has already been done and this a
1:02:48toss violation and she's she's kind of
1:02:52become the no department. I need to
1:02:54reprogram her a little bit but and then
1:02:57Wyatt is like the build it guy and so
1:03:01Wyatt and and Hillary is design. So I
1:03:04come up with an idea and then they
1:03:07debate like how to get it to be a
1:03:10prototype, you know, in 24 hours and
1:03:14it's throwaway like how do we make a
1:03:16prototype that is visual or just can do
1:03:20the function I'm talking about but in a
1:03:23way that you know only I can use it. we
1:03:25can't publish it to a website or it it
1:03:30doesn't go near any Google off or
1:03:32anything that's going to break. You
1:03:34know, I've learned my lessons, no APIs,
1:03:38but that has been really effective and
1:03:39and the most effective is just really
1:03:43walking around and brainstorming and
1:03:46helping me think through,
1:03:48you know, really accelerating the pace
1:03:50that I can think through ideas. Um,
1:03:53>> and what ideas are like are like passing
1:03:55the test? Like what are you working on?
1:03:57>> I just pulled the plug on a four-year
1:03:59project I talk about in the book. For 20
1:04:02years, I've been working on this vision
1:04:04of Earth. And again, my instinct is the
1:04:07metaverse. It's life at the speed of
1:04:08play. It's not just the book title. This
1:04:11idea that we're all going to live the
1:04:13way Elon lives. Like we're going to be
1:04:15able to tell an idea to the universe and
1:04:19magically it's going to bring back
1:04:22people,
1:04:23agents and people that can help us turn
1:04:26that into something real, test it. A
1:04:30testing machine will pop up for us. I
1:04:32mean, it's going to take all the
1:04:33drudgery out of this. And that's really
1:04:36the vibe that my book is about. And so I
1:04:39pull the plug though on I was building a
1:04:43a web-based browserbased
1:04:46game rendering engine called Stem Studio
1:04:49and we just open sourced it. So I think
1:04:53it's the most sophisticated 3GS
1:04:55rendering engine. There's no commercial
1:04:58opportunity for it anytime soon. And I
1:05:01said this the instincts I know are
1:05:04right. this idea is not going to hunt
1:05:06and I'm it's painful. It's a passion
1:05:09project. I'm pulling the plug. But it's
1:05:11built to do proven better new because
1:05:13basically you could build a flight
1:05:15simulator or you could take someone
1:05:17else's flight simulator. It's all stems
1:05:19and mod it and you could change it and
1:05:22use AI to you know make it multiplayer
1:05:24or say I want to take that and now you
1:05:27can land and have tank battles. So
1:05:29anyway, it's really cool ideas, but it
1:05:32wasn't going to work or I was going to
1:05:34have to go raise a lot of money. What
1:05:35I'm working on is I'm I'm really
1:05:39thinking a lot about what I call the
1:05:40social membrane. Like how are we going
1:05:43to network in this agentic AI world when
1:05:46our agents are our avatars, our agents
1:05:49are roaming the world while we're not
1:05:50there and and they're navigating trust
1:05:53with other people and they're they're
1:05:56finding us leads and and the first one
1:05:59of the first tests of that that I'm
1:06:01going to turn on in the next couple
1:06:02weeks. I've been building with a couple
1:06:05of hackers
1:06:06a a a network just for agents. So I
1:06:09said, what do agents care about? Our
1:06:11agents care about tokens and and
1:06:14effectiveness and token effectiveness.
1:06:16And so the first thing we're turning on,
1:06:18we're calling it slashwork and the first
1:06:21thing will be called slashchallenge.
1:06:23your agent will if you have a if you're
1:06:25a big token user, you'll be able to post
1:06:28projects you're doing or have your agent
1:06:30post projects to this agentto agent
1:06:33network and then other agents will be
1:06:35able to go and work on parts of that and
1:06:40they'll get evaluated and rated and
1:06:42there'll be challenges and there'll be
1:06:44leaderboards and then they'll get
1:06:45credits if they win. The idea is that
1:06:49agents will be out competing in
1:06:52different domains like QA or other
1:06:54things and if your agent wins and you'll
1:06:56develop better and better skill files
1:06:58around these areas or your agent will so
1:07:01it's going to start winning challenges
1:07:03and earning credits and then you'll be
1:07:06able to use credits to get swarms of
1:07:08agents you know a couple thousand agents
1:07:11to go work on something for you. So
1:07:14that's kind of the zone that I'm playing
1:07:16with. How could people find that when
1:07:18you unleash it?
1:07:19>> Uh, they can follow me on Twitter, you
1:07:21know, Mark Pink, M A R Kp P I N C. We'll
1:07:24post a link to this for people to try it
1:07:28probably,
1:07:30you know, by the end of next week or so.
1:07:34And we're thinking the first users will
1:07:36be like power users of like Claudebot
1:07:39and and Hermes, like kind of ML AI
1:07:43researchers.
1:07:45every ML research we've shown it to so
1:07:47far has really loved it. So it'll
1:07:50initially be like a gig marketplace for
1:07:53your AI, you know, kind of idle time. Uh
1:07:57but the idea is that eventually like
1:07:59agents love profiles and you'll have a
1:08:01profile and eventually they could start
1:08:04networking with each other, you know, so
1:08:06they're in little sub communities
1:08:09because they're interested in the same
1:08:10things you are or you know. So, I'm I'm
1:08:13just kind of [ __ ] around with it and
1:08:16exploring. Oh. Oh. Oh, another another
1:08:18thing I'll mention, sorry I forgot is by
1:08:20mistake I co-founded an enterprise AI
1:08:23company called Hivemind and is it came
1:08:26out of my social membrane research and I
1:08:29started researching it with this really
1:08:32talented ML researcher entrepreneur
1:08:35named Raton Cardas and he took it and
1:08:39ran with it and started building
1:08:42technology and a use case saying how do
1:08:44we apply this social membrane. How do we
1:08:47have a continuous learning system that
1:08:50deals with the human piece, which is the
1:08:52messiest piece, the input, the first and
1:08:55last mile of the AI is the human. And
1:08:58how do we not replace the human, but how
1:09:00do we start to continuously learn from
1:09:04the judgment of your best people? They
1:09:05they know all the edge cases. They know
1:09:07how to apply them in new ways. And so we
1:09:10built this, his team built this system
1:09:12and now he has it inside of like four
1:09:16large enterprises and it's delivering
1:09:19and he's building, you know, pretty big
1:09:22ARR and so it's I I can't say it's me
1:09:26because I'm a passive helper now because
1:09:30I didn't want to start an enterprise
1:09:31company.
1:09:32>> Well, but you but you we started off
1:09:34this podcast saying that that you feel
1:09:36that's where the opportunity is.
1:09:37>> It is. There's no question like, you
1:09:40know, that's where the puck is and it's
1:09:42probably where it's going to be in the
1:09:43next 18 months, two years. I'm
1:09:47passionate about consumer, so I'm going
1:09:50to keep tinkering on consumer. Do you
1:09:52look at something like Twitch and feel
1:09:54bad that you didn't do that? the idea of
1:09:57like not just playing games but watching
1:09:59other people play games,
1:10:00>> you know. I remember
1:10:03uh I remember you know Susan Wajiki who
1:10:08was a longtime friend she told me why
1:10:14aren't you focused on you know these
1:10:17streaming game streaming because you
1:10:20know they made up a shockingly big part
1:10:21of YouTube. It was like seven or eight%
1:10:24of all views and I just it was kind of
1:10:27hardcore gamers. It wasn't my zone of
1:10:29mass markets casual social gaming
1:10:32>> and I just kind of ignored although we
1:10:34were in the running to buy Twitch when
1:10:36Amazon bought them. We we had a
1:10:39competing offer. I can't remember if we
1:10:41offered 800 million or a billion. So we
1:10:44did in the end try to buy them but but
1:10:46we totally missed that opportunity. you
1:10:49you kind of focus on your zone and you
1:10:51know we missed Epic, Riot, you missed a
1:10:55lot.
1:10:55>> Well, it's interesting how Twitch now or
1:10:57YouTube you can watch like people
1:10:59playing poker and it does it's not
1:11:00necessarily these you know huge role
1:11:02playing games. It's you know just the
1:11:03classics that that you always you know
1:11:05were playing with.
1:11:06>> No, you're right. It's kind of somewhere
1:11:09in the zone of ASMR and it's just people
1:11:12are just kind of watching anything.
1:11:14>> Yeah. I just was never I was so focused
1:11:17on we our whole innovation zone was
1:11:20improving people's relationships and we
1:11:23just were into that like how do games
1:11:27help you move buckets in a in a
1:11:30friendship you know and the ultimate was
1:11:32getting to marriages you know out of
1:11:34games which we did get to and that
1:11:36opportunity was so big and it still is I
1:11:39mean as big as Twitch is I'll remind you
1:11:43you know they now oddly call all of
1:11:45casual gaming social gaming, which is
1:11:46weird because it's really not social.
1:11:48>> Yeah.
1:11:49>> Um, you know, they called it game video
1:11:51gaming and now when it is when it was
1:11:55now that it's not social, they call it
1:11:57social, but the gaming industry is $283
1:12:00billion and it's boring as [ __ ] and it's
1:12:04not social. So, it's that big and it's
1:12:07that bad. I mean, when we started Zingga
1:12:10in ' 07, the gaming industry was $23
1:12:13billion and it sucked. I mean, it was,
1:12:16you know, it was niche. It was people
1:12:18sitting on their mom's couch, you know,
1:12:20watching Tank Explosions. And now it's
1:12:23really big, but no one I know plays
1:12:26games, you know. I don't I play
1:12:27chess.com, you know, and Lee Chess, and
1:12:30that's it.
1:12:30>> I'm avid on chess.com and Lee Chess.
1:12:33>> You are?
1:12:34>> Yeah. Well, we'll have to I I don't know
1:12:36if I should play you, but do you know
1:12:38what your rating is
1:12:39>> on on Le Chess Rapid? I'm about 2250,
1:12:442,300. My highs are around 2400.
1:12:47>> Wow. I will tell you, and this is this
1:12:50is not even a humble brag. This is like
1:12:52a false brag. My rapid rating on leech
1:12:55chess I think is 2200 but it's only
1:12:58because I played like five games and my
1:13:04daughter was sitting next to me and kind
1:13:06of playing them with me and so we were
1:13:08like really telling each other moves and
1:13:10and I won them all and I stopped. I will
1:13:13tell you that my chess.com rating is
1:13:15like I think it's 850 and people are
1:13:19like shocked it's that low but it's it's
1:13:22like there's so many games that I'm
1:13:23putting the kids to bed and I'm sneaking
1:13:25and playing and then I have to like quit
1:13:27the game and you know but um I more Lee
1:13:31Chess is more my real ratings um because
1:13:34I'm I take it more seriously but I think
1:13:37I I think my real Leechess rating is
1:13:40probably like an 1800 00700.
1:13:43My daughter Georgia is obsessed with
1:13:45chess. I mean, she works with a chess
1:13:48tutor. She plays like three hours a day,
1:13:52especially in summer, like 5 hours a
1:13:53day. She goes to the park.
1:13:55>> She's a good player.
1:13:57>> Yeah, she's beating people who are like
1:13:582200s.
1:14:00>> What's her rating? What's her USCF
1:14:01rating?
1:14:02>> I don't know. She's only been in She
1:14:06went and played in one tournament in
1:14:08London and she did like shockingly well.
1:14:11She lost she lost a game that or she a
1:14:16guy convinced her to draw in a game that
1:14:19she would have won, but so so she has
1:14:21she's only 15. She needs to like get a
1:14:24little stronger emotionally, but she
1:14:26wants to spend the summer doing
1:14:28tournaments. She really wants to become
1:14:29a master or grandmaster.
1:14:32>> I'm uh in the US I'm a master.
1:14:35>> Yeah, she wants she's she's beating
1:14:38people with 2200s. So I I'll let's
1:14:41connect you, me, and her, but you'll
1:14:42have better games on Lee chess with her
1:14:45than me.
1:14:46>> Yeah. No, it's fun. I mean, I take
1:14:47lessons every week. I'm trying to get
1:14:49back to So, my peak in USCF was about
1:14:522250, and I'm trying to get back at an
1:14:55older age to where I was when I was
1:14:57younger, and it's impossible.
1:14:59>> Oh, really? Yeah. It's I I I don't I'm
1:15:04not I'm definitely not learning and
1:15:05picking up as fast as her, but I can
1:15:07still beat her. But it's
1:15:08>> it's more because like
1:15:10>> I'm her dad and I know her emotional
1:15:12weaknesses and I can just exploit those.
1:15:15>> That's brutal.
1:15:16>> But and I never let kids win.
1:15:18>> So
1:15:19>> you can't me either.
1:15:22>> Mark Mark Fis, author of Life at the
1:15:24Speed of Play. I thoroughly enjoyed the
1:15:27book
1:15:27>> and and I'll tell you I I had this like
1:15:31pit in my stomach the whole time writing
1:15:33like I just don't want it to be boring
1:15:35and there's so many business books that
1:15:38I find painful and I get annoyed at the
1:15:40author cuz I'm like
1:15:41>> yeah [clears throat] me too you're
1:15:42making obvious points you're doing this
1:15:44like playbook now it's I am ignoring
1:15:46some of the proven in it okay at my own
1:15:49risk but I just I didn't want like
1:15:52boring dry content and I I started
1:15:55learning learned by the end of it that
1:15:56whatever I had fun talking about and
1:15:59writing about I thought would be fun to
1:16:01read. And there is a trade-off in the
1:16:04book cuz some of the chapters like going
1:16:05deep on proven better new they're like
1:16:08they're more dry, they're more
1:16:10professorial but they're more valuable.
1:16:12And so there is always this trade-off
1:16:15between like the stories are more
1:16:18engaging and fun and you kind of learn.
1:16:22You know, my co-author, she wrote she
1:16:25she co-authored Ben Horowitz's book, The
1:16:28Hard Thing About Hard Things. And they
1:16:29just made half the book stories and half
1:16:32the book lessons. And in some ways,
1:16:34that's smarter and I was kind of more
1:16:36ambitious. So, I just didn't want it to
1:16:39be boring. I feel like you did that
1:16:41though. You told a lot of stories in
1:16:42this book and it's fascinating
1:16:44>> and it's kind of historical, too,
1:16:46because
1:16:47>> you really are going back to the 90s,
1:16:48the O's where all of this kind of first
1:16:50blossomed. I hate to be like, you know,
1:16:53the OG, but there's there are so many
1:16:56lessons, you know, to learn from from
1:16:59each era of this and there's so many
1:17:03proven ideas people have done before and
1:17:06just so much to
1:17:08Yeah. that that that
1:17:11when you when you kind of pull back the
1:17:13camera and and look at like, oh, people
1:17:16have done this before, you can
1:17:20you'll you'll be more successful and
1:17:23less just caught up in these little
1:17:26races, you know, like you don't have to
1:17:30build a coding agent, you know, just cuz
1:17:32that's
1:17:34massively working right now. or maybe
1:17:35you should, you know, but also there's
1:17:38so many other topics. We can do this
1:17:39again cuz I I think you're in Chicago.
1:17:42>> Yeah,
1:17:43>> that's cool. I'm I grew up in Chicago,
1:17:46so I have a lot of Chicago.
1:17:48>> So,
1:17:48>> Oh, yeah, that's right.
1:17:49>> How often are you here?
1:17:51>> I'm going to be there in the middle of
1:17:52July because I still have family there
1:17:54and we have a wedding.
1:17:55>> Let's do a podcast when you're in town.
1:17:57>> Oh, yeah. We could do it in person.
1:17:58That'd be fun.
1:17:59>> Yeah. And bring the chessboard. I'll
1:18:00bring the chessboard.
1:18:02>> Okay. and we'll bring Georgia so at
1:18:04least she can give you a serious game.
1:18:06>> Excellent. Excellent. Well, Mark, thanks
1:18:08once again. Life at the speed of play.
1:18:10Great book and I look forward to
1:18:12continuing the conversation.
1:18:14>> Yeah. Fun.

About this transcript

This page contains the full transcript of Mark Pincus: Why No App Has Gone Viral in Years by James Altucher, generated from the public captions YouTube serves with the video. The transcript has 13,267 words across 1,879 segments, with the original timestamps preserved so you can click any line to jump to that moment in the embedded player.

What you can do with it

Use the transcript to take notes, quote the speaker, build a study guide, generate a summary with ChatGPT or Claude via the YouTube Summary tool, or export it as a timed subtitle file with YouTube to SRT. You can also re-open it in the transcriber to translate the transcript into 100+ languages.

Free YouTube transcript tool

YouTube2Text is a free YouTube transcript generator — no signup, no daily limit. Paste any YouTube link and get the full transcript instantly, with timestamps, click-to-jump, translation to 100+ languages, AI prompts for ChatGPT, Claude, and Gemini, and exports to TXT, SRT, VTT, or Markdown.