00:00well thank you for coming this morning
00:02we are trying something a little bit
00:06different this startup school year we
00:09are not just having our weekly to
00:14lectures but we are having some
00:16conversations with notable people and I
00:20couldn't be happier to have one of the
00:22most notable startup people here he'll
00:27get embarrassed when I say anything
00:29complimentary so I'll try to avoid it
00:31Paul Graham my friend and the the
00:34founder of Y Combinator back in 2005 and
00:39we're just going to talk and hopefully
00:42it won't be boring Paul yes Jeff so you
00:49and I got to know each other
00:50strangely at a it was kind of a start-up
00:53at the time Yahoo back in the late 90s
00:5798 yeah guys know he wrote Yahoo Mail
01:03Y Combinator has the people who wrote
01:05Yahoo Mail and Gmail that is weird right
01:08Paul Buchheit who who actually has more
01:11of a claim on actually writing Gmail
01:16than I do have a claim on writing Yahoo
01:18I wrote I wrote some important parts of
01:20Yahoo Mail but Paul was kind of by
01:24himself doing G almost and we had we
01:27were a little more intentional with
01:29yahoo mail and had had a team creating
01:32something called rocket mail but it is a
01:33little weird that both of us ended up at
01:38YC years later uh but Paul was at Yahoo
01:43I was at Yahoo because we created
01:45something called rocket mail and we were
01:47purchased in 97 but a Paul's company was
01:49bought in 98 and so how can you you had
01:55this weird idea for a product for
01:58something you thought people would would
02:01want in in the mid-90s which had to do
02:05with like SAS before SAS was a thing how
02:08did you come up with this idea for for
02:10via web the company that Yahoo bought
02:13what we made was something that is now
02:15called a web app I believe at the time I
02:20was the first one and that's why the
02:22company was called via web because it
02:24via the web and the way I remember very
02:29vividly we thought back then everybody
02:32thought that writing software was
02:33identical with writing software that ran
02:36on the client and we're still for those
02:39of you who remember the mid-90s the
02:41client meant Windows right and like we
02:44knew how to write software for UNIX and
02:46we liked UNIX we did not know how to
02:49write software for our windows but we
02:51knew enough to know that we did not want
02:54to learn and so we were really really
02:59highly motivated to figure out how to
03:03write software without having to write
03:05software to run on Windows and so we
03:08thought to ourselves well maybe well you
03:10know the way I first thought of it
03:11actually was that you would have this
03:13you would have it was an online store
03:15builder right and I thought that you
03:19could you could send it updates by email
03:21that was I thought you can send it
03:24updates by email right um and it would
03:27change your website and then I remember
03:30thinking well if you Coster us for
03:32shopping carts or whatever if you can
03:35use SMTP why not HTTP to update your
03:41website whoa you could cut the software
03:44could run on the server and you could
03:46control it by clicking on links in the
03:49browser would that even work right but I
03:52remember I was like sleeping on a
03:53mattress on the floor in Roberts spare
03:56bedroom and I remember I sat up in bed
03:58when I thought of this like like the
04:00letter L right it was actually like one
04:03of those scenes in a movie I'm like oh
04:04my god we could avoid having to write
04:06software on Windows at all
04:09we never at that point we were never
04:11thinking about the benefits of web apps
04:13which are great as it turned out we were
04:17just thinking we could like actually do
04:19this without ever having to learn
04:21pointing windows you did yeah at ER that
04:23like you could deploy then without
04:25having to ship software did none of that
04:27avoiding windows we could actually write
04:30thank-you gates never have to learn
04:32windows because like the Netscape people
04:35had already gone through the pain of
04:37making Netscape work on Windows
04:39hey just like that was like a hole onto
04:42Windows you know and people could get
04:44through the hole and click on links in
04:46the browser and control the software and
04:47so we immediately we weren't even sure
04:49this would work because it was so weird
04:52but we immediately sat down and tried to
04:54write this really really clunky website
04:58builder that worked by clicking on links
05:00it was terrible but it did actually work
05:02it could make a website and I'm like
05:04holy this actually works that's
05:07weird that like an innovative idea like
05:10that just sort of comes out of no I mean
05:13it scene doesn't seem innovative now
05:14because all websites work that way but
05:16the first one like that's that's going
05:20from 0 to 1 how do you like it's very
05:22common happen right very common for
05:24ideas in retrospect they see him obvious
05:27and you and Matt you sort of historians
05:30sort of straighten out all the kinks in
05:32the development and it just seems like
05:34oh they had this you know it's like this
05:36idea I always hate it when people
05:38represent startup ideas as light bulbs
05:40you know because as well as being the
05:42most cliched metaphor imaginable it's
05:45also false you don't just like have this
05:47idea and then realize it right no it's
05:51more like you have this sort of inkling
05:52that something you could do something
05:54that somebody else hasn't done before
05:55and it's probably a bad idea
05:58but you're too lazy to learn Windows and
06:00so you want to do it that way right um
06:02and so or like who would or like Zuck
06:06starting Facebook you know just like
06:08let's see what happens right the way
06:11implausible if this would ever be a
06:12startup which is good for you all
06:16hopefully some of your ideas seem really
06:17implausible right I love it when ideas
06:21seem implausible in the right way over
06:23the years I mean I already had just from
06:26being a hacker as sort of a good nose
06:28for this sort of thing but it has become
06:30refined over all these years of dealing
06:33with startups and when when people have
06:35I was talking to some people yesterday
06:38and they were talking about two possible
06:41and one of them seemed so it seemed like
06:45naughty not like it was it was it was
06:48sort of maltreating anyone it was just
06:51sort of taking advantage of something
06:53that seems like it shouldn't even be
06:55possible right like like making this
06:58software run on the server and just like
07:00never making software that seemed to the
07:02users to run on their computer but
07:04actually was just talking to them
07:05through the browser that kind of
07:06naughtiness and then I said like pick
07:08that one because it seems it seems so
07:10outrageous it seems so hilarious it
07:13seems like there's all these ways we
07:16have talking about what's a good idea
07:17and when you try to match up ideas with
07:20them it's always sort of a complicated
07:22you know it's like is this idea naughty
07:24or is it is it is it is it like not
07:27quite intuitive or not so javea and it's
07:30really always hard to to take any
07:33particular idea and say is that the one
07:35was potential well you can't tell in
07:37fact that's another thing um it's not
07:41just that the outcomes of startups are
07:45hard to predict I think to some degree
07:47they're actually indeterminate there's a
07:49huge amount of luck involved you know
07:51and so I think even if even if Y
07:55Combinator were as good at picking
07:58startup so you could possibly get I mean
08:01in fact you have like 150 startups per
08:03batch of which maybe 5 will become giant
08:05or something like that
08:06hopefully high that would be good yeah
08:08well depends on your definition of giant
08:10right it's a power-law distribution
08:12we're like if for some definitions of
08:14giant one of it one is taking any
08:16batches will think if five or any
08:18definition of giant yeah we'll be happy
08:20so talking about I like oh yeah the
08:26indeterminacy of yes even if you were
08:29able to pick perfectly that doesn't mean
08:31you could get the batches down to only
08:33five startups you probably have to pick
08:35twenty I think even if you were perfect
08:38so it just it's just it's hot it's hard
08:41like it's hard to know there's this
08:43there's a strange mix of extraordinary
08:45hard work an idea that actually has
08:49something to it but is not obvious it's
08:52hard for founders to know yeah you don't
08:54you don't know yourselves I mean I mean
08:58you don't know comma yourselves not that
09:00you don't know yourself so that it's
09:01also true so and that's why it's good to
09:06be a hacker because if you're some
09:08business person you know and you think I
09:11am going to work on business
09:13opportunities right well I don't know if
09:15this is a business opportunity so I'm
09:17not gonna work on it um whereas someone
09:19who's a hacker who's just doing things
09:20for the fun of it or because they don't
09:22like Windows yeah laziness to this is
09:29actually in some cases really helpful at
09:31least some kinds of some kinds of later
09:34I'm a little scared about this meme
09:36escaping into the wild right
09:38can't you just see well yeah you just
09:40see it's it's truly they said laziness
09:42is good he said from the sofa as he ate
09:44another doughnut you know it would be
09:48better if people thought laziness wasn't
09:50good because the glaze enos the kind of
09:52the people to you that's good they do
09:53and they would do anyway even if they
09:55thought it's weird yeah well using
09:58laziness not in the normal sense of
09:59laziness right because if you're late
10:01there's there's no such thing as a lazy
10:03founder that it builds a big company
10:05there's but but a certain sort of it's
10:09sort of a reticence to do something for
10:11some reason I'm calling it lazy
10:12weightings not really it seemed like
10:14sort of slaps it makes you it you know
10:17if you have a natural hatred for
10:19gratuitous slap slaps like then what
10:23that that produces elegant solutions so
10:26so in 1995 ish I was like trying every
10:30single idea I could possibly try to
10:32start some internet business and I was
10:34doing it all by myself which was awful
10:37and and what I didn't know how to do was
10:40to find people to do it with and I kind
10:43of ran into a bunch of people and I had
10:45no clue whether they would be good
10:46people to start companies with and I was
10:51thinking about your experience you said
10:52you were like sleeping on Roberts well
10:54you started you started via web with two
10:56fascinating people Robert Maxwell and
11:00trevor Blackwell who later robert morris
11:02robert morrison and trevor blackwell
11:05Morrison Trevor Blackwell and that's a
11:08that was a weird mix of names anyway um
11:11how did you pick them as co-founders how
11:15did that happen and and when did you
11:16decide they would be good co-founders
11:18were they good co-founders um alright so
11:23they're extremely worked they were
11:25extremely truculent co-founders who were
11:27very good at programming but I remember
11:30Robert was like not into the whole
11:32startup thing the entire way like
11:35whenever he was on the board and
11:36whenever we got and whenever you get an
11:38acquisition offer the board has to at
11:40least consider it right and so we would
11:42get these crap lowball acquisition
11:44offers we'll give you like two million
11:46dollars in our stock right was like yes
11:49no no Robert would say like we would
11:51vote on it and Robert would always say
11:53well you know I have to be honest here
11:54like this would at least mean we could
11:56stop working on this thing and so yes I
12:00would take this deal right I made a deal
12:04with Robert that if he ever made a
12:05million dollars out of out of via Web he
12:08would get an earring Robert it's not the
12:10kind of guy who would enjoy having an
12:12earring let me tell you um and so as
12:16soon as the deal closed me and Trevor
12:18frog-marched him to the place in Harvard
12:21Square where people get their things
12:22pierced never heard this before he got
12:25him an earring there's photographs on
12:27the internet if Robert Morris with an
12:28earring that he didn't keep it when his
12:31but not for long we should have
12:33specified that I never thought I never
12:36thought to specify you know I thought
12:37you'd actually get the million dollars
12:38no no I really actually did but I
12:40thought he would wear it for a bit
12:42um when his fiancee saw it she fell on
12:45the floor why did you choose why did I
12:48try killing Trayvon ders okay so I chose
12:52Robert because he was my co-conspirator
12:54in everything and I don't mean I was
12:57always I was always the lead and he was
12:59just the co-conspirator when he did
13:00things I would be his co-conspirator to
13:02you so like the internet worm if 19 you
13:05guys know Robert Morris's is pretty
13:10famous all on his own
13:11yeah he invented buffer overflow and I
13:16remember when he told me about it and I
13:17said wow what a cool idea she totally
13:19do that he was the first famous hacker
13:21who really got in trouble I think yeah
13:23really in trouble he was the way it's
13:24true he was the first person to be
13:26prosecuted under the Computer Fraud and
13:29Abuse Act of 1986 which so if you're
13:33looking for a co-founder look at people
13:35who have been prosecuted because it
13:36yeah user he is actually a convicted
13:38felon okay this is another mean we don't
13:43necessarily want to escape out there
13:44that convicted felons are what you want
13:46to look for as co-founders they're
13:48convicted for that maybe it was funny
13:50because the FBI like like law
13:52enforcement kind of does things by the
13:54book right like imagination is a big
13:58thing in in like Sherlock Holmes stories
14:01but not in actual everyday law
14:03enforcement and so they have these the
14:06FBI agent told me like they have these
14:08motive they look for sex drugs money
14:13and so like Robert did it out of
14:15curiosity and that's just think it's
14:19just not on the list hey they really had
14:22a hard time figuring out what the hell
14:23was going on I think even within the
14:25government there were lots of people who
14:27did never understood what was going on
14:29so back to Robert and Trevor whoo-hoo oh
14:33so why did I pick Robert Trevor yeah
14:35Robert Robert Robert I picked because I
14:40would I would do everything with him
14:41right we had all these schemes and so
14:44you knew you could work with him I mean
14:46I could work with him when I knew it was
14:47really good he was a really good
14:48programmer he could program as fast as
14:50he could type of course this was in C
14:53which is a lot of very verbose language
14:56but but he could program as fast as he
14:59could type he was amazing he would like
15:01if he didn't like what the operating
15:02system would do it he would edit the
15:04source and recompile it and then it
15:07would do what he wanted
15:08so truculence aside he was exceptional
15:10like when he showed up at Harvard
15:12undergrads were only allowed to have
15:14accounts on like the official undergrad
15:17computing system off in the Science
15:18Center and there were all these real
15:19computers in the cs department and let
15:22aiken lab and so they wouldn't give him
15:25an account there and so he just walked
15:29you know switched it to single user with
15:32him as the super user gave
15:33cellphone account and then and switched
15:35it back hey it was curious yeah well he
15:39also wanted to counts on the right
15:41machines he got kicked this whole thing
15:45will be about Robert Robert such an
15:46interesting character I shouldn't talk
15:48about it may be more interesting than
15:50anything else we can talk about he got
15:52kicked out of Harvard as an undergrad
15:54for reconnecting Harvard to the Internet
15:58um Harvard Harvard had been one of the
16:02early internet nodes that's so funny to
16:05like like get interval for disconnecting
16:08a school from the internet connecting it
16:10can happen for connecting a school where
16:15well we're still trying to figure out
16:20oh I should tell you though about how we
16:22tell us about the disconnecting really
16:25fast no no it's okay so we have all day
16:27we don't do it no we don't okay how long
16:30have we have we have enough time we have
16:34the same watch more or less Robert okay
16:43it was one of the very early internet
16:46nodes and by the time Robert arrived in
16:49as an undergrad the connection had died
16:52from bit decay that's like no one
16:54noticed that's how unimportant the
16:57internet was back then or rather ARPANET
16:59as it was then called verily an Internet
17:02right so RT M spent one whole semester
17:04working on getting Harvard reconnected
17:06RT M is Roberts yes yes Robert and he
17:09like didn't do any work for any of his
17:10classes and he got such bad grades he
17:12got kicked out for a year so we actually
17:15used that later as a recruiting
17:16technique because we knew there because
17:18there might be other people who got
17:19kicked out for spending all their time
17:20on some project and so we actually went
17:23around Harvard and posted up this poster
17:25saying did you get kicked out working on
17:28some project of your own we'd like to
17:29hire you I think there's probably so we
17:32got a really grid programmer that way
17:34um there's probably something made tons
17:36of money off of options cuz we got
17:37acquired journey here right and you like
17:40enjoyed life out of school so much I
17:42don't think he went back until he was 25
17:48so um all right so why did why Trevor
17:52why Trevor well so you picked Robert
17:54Robert was what was the guy you did
17:56everything with and he was great okay so
17:58after we'd been doing before the wave
18:00when you picked Robert did you know he
18:02was gonna be a pain in the ass in so
18:04many ways as you like or was that a
18:06surprise to you well I knew he was a
18:08pain in the ass I didn't know the
18:10company was gonna be so hard and it
18:12would bring out so much penis so you
18:14didn't realize how hard tell you the way
18:17we got Trevor was very indicative yeah
18:19so a month in we'd been working on this
18:21company for a month and Robert rebels he
18:25says we've been working on this company
18:27for a whole month and it's still not
18:28done and I thought hmm maybe we need
18:39more programmers so so I said all right
18:43Robert who's the smartest person you
18:45know in grad school and he said Trevor
18:48right and I said Trevor right and
18:51Trevor's one of those people who seems
18:52of a lot smarter than you seems so but
18:57it's Trevor is actually super smart and
18:59so we went and recruited Trevor and he
19:02said oh okay that's that's a really good
19:05Trevor but almost not a Canadian Trevor
19:08has two modes oh okay and oh I'm sorry
19:11which is what he does when he does
19:14something that breaks everything which
19:15was very common very common and so we
19:20got Trevor I recruited Trevor to work on
19:22the thing he does nothing for two weeks
19:25at least as far as I can tell he just
19:26disappears and then two weeks later he
19:29says come into my office and he's
19:30rewritten all our goddamn software in
19:32small talk typical yes um so I said all
19:39right Trevor we're not using that but do
19:43um and so Trevor like Trevor was like
19:46this super productive hacking monster
19:50but you didn't know that in advance you
19:54just trusted Robert when Robert told you
19:57he was really smart well if Robert said
19:58he was the smartest he's deaf
19:59the smartest but there's kind of a deep
20:01lesson in that and how you how you pick
20:04whether whether it's employees or
20:05co-founders it's you find someone you
20:07trust and then find someone they trust
20:10yeah although here's an interesting
20:12point that works better for some
20:15qualities than others like intelligence
20:17the intelligent people can judge other
20:19intelligent people but trustworthy
20:22people cannot judge other trustworthy
20:24people in fact trustworthy people are
20:26often fooled by untrustworthy people so
20:28like Kate quarto is like our architect
20:30who did this is like super trustworthy
20:32but she's always fooled by easily sneaky
20:35schemers you know who take advantage of
20:37her so I guess that means you take
20:39recommendations from people but then you
20:42have to add your own no no no you can
20:44totally dress but if Robert says
20:45someone's smart and I disagree I think I
20:48must be wrong right yeah but you need to
20:51just works for some qualities better
20:53than others let's hold right you want
20:54them to be trustworthy when they join
20:55you as a co-founder so you have to add
20:57that for the trustworthiness yes this
20:59whole thing of like networking doesn't
21:01work it doesn't you have to have someone
21:03you can judge whether people are
21:04trustworthy which I'm I am not even good
21:07at I am not good at the person who does
21:09that for Y Combinator is Jessica the
21:11social radar the secret power behind
21:14everything so you did hire people at via
21:17web - yes what did you learn about
21:19hiring in your like your first startup
21:22like so you learned that they they if
21:25they were getting lousy grades your
21:28start working on a project that seemed
21:29good which is which is kind of
21:31counterintuitive a lot of things we seem
21:33to do it.why see end up being
21:35counterintuitive startups in general are
21:37very very counterintuitive and so that's
21:41actually why C exists if it was obvious
21:44what to do and starting a start-up we
21:46wouldn't why C couldn't really add much
21:48by teaching people how to do it right
21:50and so I often say I often in fact
21:54practically every day if I do office
21:56hours I'm reminded of this that like
21:58what Y Combinator does is tell founders
22:00things that they ignore right so we tell
22:04people don't hire to you fast and then
22:07they go off and hire to you fast and
22:08then they come back later and say oh I
22:10wish we'd listened but
22:12all the things we tell people we tell
22:15people the counterintuitive stuff not
22:16the obvious stuff and then what
22:18counterintuitive means is it sounds
22:21wrong and so they go with their gut and
22:23do the wrong thing instead and then
22:26hopefully catch the mistake in time why
22:30is it that startups are so
22:31counterintuitive you know that would be
22:34if that would be very much writing an
22:36essay about that would be worth writing
22:37an essay about why startups are so
22:39counterintuitive I don't know I don't
22:41know I don't know it's I could come up
22:45with some theories but I have a feeling
22:47the answer is so interestingly
22:48complicated that would be unlikely to
22:50come and it's really just we were just
22:52talking about one of the things that
22:53that you you did write an essay about
22:55which is somewhat counterintuitive when
22:57you start a startup and you're like okay
22:59now I'm gonna grow and then you write a
23:01whole bunch of stuff but you've already
23:03software that helps you grow and and
23:05you're at a whole si says don't do that
23:07do things that don't scale in the
23:10beginning yeah do things that don't
23:12scale because growth is maybe you can
23:16start by saying what it means to do
23:17things that don't scale because I'm
23:18afraid the best thing you did not
23:20already I'm sure you've all read all as
23:22well as I say but this is a this is
23:24actually very important because most of
23:26you all have early startups and when you
23:29do an early startup it's become our
23:31mantra YC and I've seen it again and
23:34again you should not try to do too much
23:36in the beginning you should do things
23:37that don't scale what doing things that
23:39don't scale means specifically is doing
23:42things in a sort of handmade artisanal
23:46painstaking way that you feel like yeah
23:50it would be great if you could do things
23:52that way forever but you and you're in
23:54the back of your mind you think to
23:55yourself well there's no way we can keep
23:57doing this and become giant right and
23:58what do things that don't scale means is
24:01do those things early on anyway because
24:05you know if you don't do them you'll
24:07never be big and you've got nothing to
24:09lose and also you learn a lot from it
24:12you learn a lot and so I discovered a
24:15lot of this things that we teach
24:18startups at Y Combinator or you teach
24:21startups the Y Combinator I'm retired I
24:23don't know if you know that
24:25but a lot of things you guys teach
24:26startups is are things that I hit myself
24:31and didn't realize that they were
24:33actually common startup lessons so one
24:36of them is doing things very manually
24:39for your early customers like it's so
24:41important to get early customers that if
24:43you have to do a ton of manual stuff
24:44that's ok you'll learn a lot from it did
24:47you guys do manual stuff oh yeah well
24:49totally and we thought oh we're doing it
24:52wrong this is so lame and in fact it was
24:55exactly the right thing
24:56we made an online store builder you
24:59could build a store on the internet and
25:00sell stuff and we would go to would be
25:05would be customers and say would you
25:08like to use our easy online store
25:10builder and they would say no and we
25:15would say but you want an online store
25:16right and they would say yeah and and
25:19they would say well what if we use we
25:21would say what if we used our software
25:23for you to make an online store and then
25:25you could have it would that be good and
25:27they'd say all right and so we'd like we
25:29can't get anyone to use our software but
25:31at least they're willing to let us use
25:32it for them and it seemed it seemed so
25:37lame I learned so much about like
25:41direct marketing as it turns out it's
25:43called we were members of the DMA the
25:45Direct Marketing Association that's what
25:47you know in every business card carrying
25:49member the DMA were we were in every
25:53business they have a name for the
25:54business that's not the name Outsiders
25:56call it like fast food no no no they
25:58don't call it that they call it fast
26:00casual within the industry right and in
26:03in the catalog business that used to
26:05call the direct marketing industry right
26:07um so that's what the dma meant back in
26:10those days the catalog business we
26:12signed up for every catalog you know his
26:15catalogs you get in the mail we would
26:16just like write a wide right and ask for
26:18more well you had this like bookshelf
26:20full of every catalog ever yeah it seems
26:24like almost like if you were building a
26:26new search engine and what you ought to
26:27do is have a have your box out there and
26:29then just take all the searches manually
26:32and like get results and throw them out
26:34there and see what see what happens I'm
26:35trying to think what we did for mail
26:36that was manual in the beginning we we
26:39sort of a manual ad server like we
26:42didn't really have ad serving technology
26:43we just throw at like I wrote it so it
26:46wasn't really good and and we would just
26:50throw them in there that there's a deep
26:52lesson there for four for four people
26:54though like what happens if you think
26:57you know the solution and you build lots
26:59of stuff for it you're going to be wrong
27:01because you just don't know and having
27:03to use our software myself made it much
27:06better because I would be using our
27:07software to make someone's website right
27:10and I was also the guy who wrote the
27:12site builder and so I would be using and
27:14think to myself this is inconvenient
27:16it's like going in the middle of
27:19building their site I would change the
27:20software and I mean ship as in like MV
27:26and like that's like that's the great
27:28thing about server based software ship
27:30is the UNIX MV command so well know CP I
27:36would want to keep a copy but but I
27:43would like change the software in the
27:45middle of using it and then go back to
27:47working on their website and boy did
27:49that make that that made the software
27:50much better that I had to use it yeah so
27:57um so picking co-founders you want to
28:04pick the best people who you can somehow
28:05believe are trustworthy
28:07it's probably helpful for a lot of
28:08people I think a lot of people taking
28:09the class or sole founders now yeah why
28:12is it that that sole founders that being
28:16a sole founder is so difficult Wow um
28:19there's so many different reasons like I
28:24think the hardest part is more L there's
28:27no one there's like if you have multiple
28:29people they keep one another going when
28:31things are going badly and there's no
28:34that's what's hard for me I remember 95
28:36I'd be like I know this internet thing
28:38is gonna be real but damn oh so you were
28:41sole founder yeah yeah
28:44that's why his hair is greater than mine
28:47the yeah like being a sole founders hard
28:51like it's hard because there's a million
28:52reasons why what you're doing isn't
28:55going to work and mostly you're right so
28:59it's easy to convince yourself to used
29:02used to say this and I always remember
29:04to used to say look you have two choices
29:06as a founder do you remember what you
29:09would say after that well there's so
29:10many different cases this versions of
29:12this what will he would say you could
29:13either quit or get rich mmm yeah die or
29:16get rich like the company's a lot of
29:20versions well that's what happens with
29:22startups mostly the company either dies
29:24or it's something succeed you get rich
29:27and you know getting rich can mean you
29:29know you have a nice little business
29:31where you're the boss and you have 20 or
29:3250 employees or it could mean your
29:34Airbnb right they're both good outcomes
29:38not as good for your investors
29:40necessarily but for you it really good
29:42and and doing things that don't scale
29:44that's hot you know and internalizing
29:46what that really means I was thinking
29:47about you know one of the it's the sort
29:49of things that don't scale we talked
29:51about at YC all the time we say you're
29:53supposed to for success you talk to
29:56users and write code you you talked and
29:59if you think about it talking to users
30:01especially as sort of the founder
30:03doesn't really scale very well but
30:05that's what it forces you to do if you
30:07have to do everything it doesn't scale
30:09it forces you to learn right in the
30:12beginning you only have ten customers
30:14you want to grow to 20% you want to go
30:1710% next week 10% weeks an ambitious
30:19goal you only got to get one more
30:21customer you can go out and do that very
30:22manually right and then next week you
30:26have eleven customers you have to get
30:271.1 customers right so this is basically
30:30one right you just keep going out there
30:32and doing things manually and as long as
30:34your growth rate is good it doesn't
30:36matter how small the number is because a
30:39grow a constant growth rate means
30:41exponential growth and that means the
30:44base number will soon take care of
30:46it's kind of thinking about what makes
30:50companies good what makes companies fail
30:53and I was thinking about maybe how we
30:55could talk a little bit about some of
30:56those those aspects of
31:00of how people can kind of figure out how
31:02they're doing here we have thousands of
31:04companies out in start-up school and
31:06trying to figure out what to do next and
31:09whether they're doing the right thing
31:10and whether they're failing and whether
31:11they're on the right path
31:13what makes companies fail most of the
31:16time is poor execution by the founders
31:18right um we we often talk to startups
31:23who are worried about competitors and
31:25one good event one advantage of why is
31:28he having funded so many companies you
31:30have a really large data set and like
31:32how many companies have been killed by
31:34competitors one more than one out of
31:38nineteen hundred right um so I tell
31:43startups that basically you have the
31:46same protection against competitors
31:48that's that light aircraft have against
31:51crashing into other light aircraft when
31:53flying through clouds
31:56you know what the protection is space is
31:59large hey this little Cessnas don't have
32:03any radar in them you know you can't see
32:06in clouds it's this it's it this falls
32:10under the rubric of counterintuitive
32:12advice because one of the things that
32:13happens all the time is people come up
32:16and say oh my god there's another new
32:18company what should I you know that's in
32:20exactly my space what should I do
32:22right and it's sort of like if you're
32:24running the hundred meters right and
32:26suddenly there appears another lane with
32:28another runner in it what should you do
32:30you run as fast as you can just like you
32:33presumably were or hate if they're
32:35better than you they'll win and if not
32:37then then you'll win and that's a good
32:42reminder could everyone please silence
32:43your phones if you haven't already yeah
32:49what else like like one of the one of
32:53the things we do a lot is try to judge
32:55Wright companies in it founders Nick
32:59who's gonna succeed and and and you guys
33:04are doing the same thing except you're
33:06trying to pick yourselves hopefully what
33:10what what is it that
33:12then make startup succeed that makes
33:15startup succeed well you have to start
33:17with a good founding team ideally not
33:21just one person that's my quote you can
33:23do it with one person well I couldn't I
33:25couldn't III I ended up I ended up
33:28having to join other people like I
33:30actually think oh that's majority that
33:32was not rock you know I was no way like
33:36I couldn't do it I'm not built that way
33:38I think most people aren't built that
33:40way that's that's my questions about
33:42Robert and Trevor that you know it's
33:44just so difficult I would not try and
33:47start a start-up just with just myself
33:49as the founder hmm I don't think I could
33:52do it yeah it's lonely enough anyway
33:54think you can but I couldn't so um you
33:59need a good solid founding team who know
34:01one another well and who can work
34:02together just like stuff pause they're
34:04like knowing one another well why does
34:06that even matter because what happens in
34:13will shake your relationship if there's
34:17any sort of flakiness or uncertainty or
34:21disloyalty or whatever like lingering in
34:23there it will emerge under the stresses
34:25of a start-up yeah it's like a marriage
34:28only harder you spend more time with
34:32that purpose other differences there's
34:33there's potentially yeah okay so you get
34:42a good founding team you get someone who
34:46you can stand stress with so there's two
34:51categories of stuff there's two
34:53categories of answers we could give here
34:55like there's the answer you give to
34:56investors about how to pick startups and
34:59there's the and then there's like only a
35:01subset of that is useful to founders
35:03right so like if like if you you know
35:06you could tell investors you want to
35:07pick smart people right but founders are
35:09as smart as they are if they party what
35:11can we tell them get smarter right um
35:13and actually it turns out it's not that
35:14important to be smart um it's much much
35:17more important to be determined if um if
35:20you imagine if you imagine this
35:22hypothetical person who's like a hundred
35:26hundred out of 100 for smart and a
35:27hundred out of 100 for determination
35:29right and then you start taking away
35:31determination you know it doesn't take
35:34very long before you have this sort of
35:35ineffectual but brilliant person um
35:38whereas if you take someone who's like
35:40super super determined and you start
35:42taking away smartness bit by bit I mean
35:45eventually you get to some guy who owns
35:48a lot of taxi medallions but he's still
35:49rich right or like a trash hauling
35:56business or something like that but like
35:58you can take away a lot of smart yeah
36:02you take a lot of smart and and add and
36:05take away a bunch of ethics at the same
36:06time and then you get that guy right he
36:15yeah I've been thinking about what like
36:18zoos thought that would Trump would
36:20actually teach an important lesson yeah
36:21for startups it's all about
36:23determination I was think about these
36:25axes we we think about and those are two
36:28like of the intelligence axis and the
36:30determination axis and and obviously if
36:33you were really intelligent really
36:35determined that's awesome there's by the
36:37way only one person has to be super
36:39determined right right so like you know
36:43you can have one person who really wants
36:44to start the one the company and look in
36:46you and Robin exactly who after a month
36:49is like what we're still working but
36:51that depends on this tight bond where
36:53you can drag that person along with you
36:55otherwise they'll go away but I was
36:56thinking there's actually another axis
36:58which is you have to have the ability
37:00somewhere on your team to to to project
37:08to think of but because your idea is
37:09gonna suck and then it'll suck less and
37:12less and so you have to have the ability
37:13to think what where should we like to
37:15navigate like you have to have some sort
37:17of creativity that allows you to to to
37:22like make the right choices and you know
37:25actually you might not have to be that
37:26creative if you care enough about users
37:29you can just follow what what MIT what
37:31will make users happy the way the way a
37:34scientist follows the truth and
37:36eventually without much thinking on your
37:40the need to grow will give you this
37:42product idea that's actually the product
37:45that's the result of evolution so maybe
37:48you're either sort of Steve Jobs and can
37:51intuit what customers need or you just
37:54have to have the skill of actually
37:55talking to customers and understanding
37:57what they're saying which isn't always
37:58obvious you know Steve Jobs his trick
38:00was he satisfied users to accept he was
38:04the user it's like I don't want to have
38:06any phone jack anymore so no more phone
38:09jacks for anybody right maybe he was
38:11just lucky that he his his desires were
38:14generic enough that it was closer
38:15everybody yeah yeah yeah he was living
38:17in the future yeah like he didn't really
38:19like the internet except for reading the
38:21New York Times so he never really got
38:22that right hmm interesting
38:27you work didn't you work you work at
38:29Apple at one point after you got bought
38:31did for this it was months what was it
38:33like what was Steve like could you tell
38:34from where you were in Apple what Steve
38:36was like yeah I got to talk to him a few
38:39times he um he's different tell us tell
38:45us how are their useful lessons you can
38:48learn from him or Zia one off I wouldn't
38:54presume I only met with from a handful
38:57of times and I wouldn't presume to know
38:58too much about them yeah you had a lot
39:00of indirect evidence and stories and
39:02stuff like that yeah like do you ever
39:04see founders at YC you think to yourself
39:06oh this is like Steve Jobs yeah I don't
39:09think we'd I don't think we'd fun to
39:11Steve Jobs at YC well that's not good
39:14yeah and the reason is especially when
39:17Steve Jobs was starting off he was an
39:19 oh I think actually you would
39:21find him because Steve Jobs would make
39:23you find him maybe right maybe we try
39:25like we try to avoid people who who seem
39:29like they're and I think like
39:30and and and like we like you know you
39:33wrote a whole essay on on being nice
39:35yeah the Ronco principle right there's a
39:38couple couples yeah be nice yeah and
39:41called be good be good right and and I
39:43don't think he would have passed the
39:45test especially when he was younger
39:47which is I worry about that because it's
39:50true that it's usually good
39:52to be nice and friendly and benevolent
39:58maybe if YC had caught him early we
40:02could have helped maybe maybe that would
40:04be buddy maybe that's part of what made
40:06him successful in the beginning I don't
40:08know I don't know it worked for Donald
40:10Trump Donald or did it that's a
40:14different kind of that example right
40:16yeah we should know what like in
40:20business determination matters more but
40:22like business is not identical with
40:23startups I want to leave these these
40:27guys some time to ask questions but I
40:29want like what are they gonna get wrong
40:33what what mistakes will these people
40:35make that maybe maybe hearing them might
40:39help it probably won't because you won't
40:43listen but what you will get wrong is
40:47that you will not pay enough attention
40:49to users you will have you will make up
40:51some idea in your own head that you will
40:54call your vision and then you will spend
40:58a lot of time thinking about your vision
41:00in a cafe by yourself and right build
41:05some elaborate thing without going and
41:07talking to users because that's doing
41:09sales which is a pain in the ass and
41:10they might say no and and you'd be way
41:14better off finding someone anyone who
41:17has a problem that they will pay you to
41:20fix and fixing it and then seeing if you
41:23can find more people like that best case
41:25is if you yourself have the problem
41:27right but like you will not ship fast
41:30enough because you don't want to face
41:31the you you're embarrassed to ship
41:34something unfinished and you don't want
41:37to face the the the likely feedback it's
41:41humbling and humiliating when they tell
41:43you how shitty what you built actually
41:45is right but that's the only way you get
41:49it better yeah so you will shrink from
41:52contact with the real world or contact
41:54with your your users that's the mistake
41:56you will make yeah there's a Paul Graham
42:00ISM in there which is all around
42:01launching its launching before you're
42:06tell people the story of its Reid
42:08Hoffman actually who harped on that one
42:11early not me he there's I think he said
42:14if you're not embarrassed what with what
42:16you launched rather than you launch too
42:18late I remember when I was I was joining
42:23Y Combinator and at the same time this
42:25is not necessarily good idea of
42:26launching imagine k12 this this EdTech
42:30version of YC and and I came in to talk
42:33to Paul and he's like so how's it going
42:35with imagine k12 because he helped out a
42:38lot and I said well it was probably
42:40called something else well no no no
42:42still it was by then you'd already he
42:45actually gave us the name of Magic eight
42:46twelve and and so it was called a magic
42:48a twelve but it's like when you get a
42:49launch and I was like well we're going
42:52on our PR plan and we're finishing the
42:55software so this was like you know we're
42:57we're gonna have a summer batch and this
42:59was probably in February or something
43:01and and I said so we have like you know
43:03probably in like six weeks he said one
43:05he just launched when he just launched
43:07now and I was like damn he just did it
43:12to me and we did we launched we launched
43:14like within a week we launched and so
43:17that is a mistake you'll make you'll
43:18wait too long you'll you'll you'll be
43:22we launched YC probably in less than a
43:26week of having the idea of course
43:30launching it was merely saying was
43:32merely posting a website and and we
43:36didn't have we didn't have any software
43:37in the beginning people we just had this
43:39ascii form that people would fill out
43:42and reply to us by email the first
43:45couple batches the applications were by
43:48email we would print them out and pass
43:52them around amongst the partners and
43:55like we would all right grades on the
43:57top like we were grading papers and then
44:00we would sort this heap of papers and
44:02that was the application press talking
44:03about doing things that don't scale try
44:06to do that with thousands of
44:08applications now it would be harder so
44:09we have lots of software that does that
44:11because eventually you scale you know
44:13eventually you you build software when
44:16you learn what software you ought to be
44:20it'd be fun to spend a lot of time more
44:22time talking about how why she got got
44:25started but but I think rather than that
44:28let's some how do you have left we can
44:31probably go to around 12 and and or all
44:34videos that get too long or like no one
44:37wants to watch anyway so so we've
44:40already been going for about 40 minutes
44:42or so so let's open it up to any
44:43questions you guys might have yeah are
45:14you saying that if you get if you
45:15recruit your co-founder your friends as
45:18co-founders you might lose them as
45:19friends I don't think that's a big
45:22danger I don't think you have so the
45:24question was should I have my friends be
45:27Cole founders and then maybe risk losing
45:29them as friends or should I just take
45:30someone off the street where I they're
45:32not I don't have the risk of losing them
45:34as friends but I they might break my
45:36company is that was that your question
45:39do I get that right I already answered
45:41it dude come on I just wanted to make
45:43sure they heard the question too just
45:44because I repeated the question how do
46:00you deal with different commitment
46:01levels of co-founders and if you do have
46:04to fire a co-founder what should you do
46:06okay so the way you deal with different
46:08commitment levels is you ask yourself
46:11would I rather have 30% of this person
46:16or a hundred percent of some other
46:17person and in my case I would rather
46:19have like 10% of Robert Morris's brain
46:21than a hundred percent of almost anybody
46:23else's so is an easy decision
46:26how do you could fire co-founders in Y
46:29Combinator the answer is you
46:31you ask carolyn levy y combinator as
46:37there's so much experience with dealing
46:39with crap like firing co-founders and
46:41like you just go and talk to them and
46:43they say here you do this and this and
46:46and it's probably online too so you
46:48could google it really is it YC
46:50published how to follow we've we have
46:53published we've actually if you look at
46:56some of the talks Carolyn has given and
46:58Christie I'm pretty sure they've covered
46:59that particular issue one way you deal
47:01with it by the way is making sure that
47:03someone has a little bit more stock
47:07perhaps like it's usually better to have
47:09somewhat equal shares not not but not
47:13completely equal and if one person has
47:1651% the other person has 49% then they
47:19can you know deadlocked yes so when you
47:23talk about launch is launching to like
47:25private bailout a couple of people good
47:27enough what do you have you just launch
47:28so what does launch really mean it means
47:31you have actually okay so the question
47:32is is launching in a private beta to a
47:36few people good enough for you to have
47:37to launch the general public I will take
47:39launching in just a private beta to a
47:42couple people because that's closer to
47:44the that's closer to full launch than
47:46like just sitting thinking of your idea
47:54wouldn't we come out of the currently
47:57raising in the project maybe a little
48:03bit rate so basically you asking about I
48:06SEOs yeah what do you think about I CL
48:08is Paul I have I haven't I know nothing
48:12about this world um I've heard there's a
48:15huge amount of crypto money floating
48:18around um so like if there's money
48:22floating around it's often a good idea
48:23for startups to try and get it beyond
48:28that I don't know but but I will say
48:30like one of the things we've seen a lot
48:31at Y Combinator is it's not always right
48:33to raise too much money and even if you
48:36don't it like the IPO thing by
48:38definition of - it's always wrong yeah
48:40the the ways a lot of money and the
48:44thing about there's a couple reasons for
48:45that one you dilute more than you
48:47necessarily need to but the other reason
48:49is that money just sitting there has
48:53sort of a gravitational it makes you
48:56suspend it and it makes you do stupid
48:58things so it's like these companies that
49:00have raised a billion dollars on icos
49:02and I don't know if they'll ever have a
49:03real company where you get like you get
49:06a bunch of LPS and raise a billion
49:08dollars and then you just hire so many
49:09more people than you actually need
49:11right I know that's the difference when
49:17I see oh I get it but just having a
49:19billion dollars when you only barely
49:21have an idea is not necessarily you know
49:24magic leap raise 2.3 billion dollars and
49:26they seem to have almost nothing by the
49:28way that is the trick for getting your
49:30question answered I learned that a long
49:32time ago you raise your hand as your fin
49:35as people are finishing the preceding
49:37question and and I see that guy's hand
49:39over there he knows the trick go ahead
49:48my best tips for a great user interview
49:51um well you want to figure out but not
49:56just what they think is wrong but what's
49:57actually wrong like what's missing in
49:59their life right so just start talking
50:02to them like say what would you like to
50:03do what would you like to be able to do
50:06that you can't and they might just you
50:08know they might they'll they'll tend to
50:10give you an answer that's like a subset
50:12right um like you know if you were
50:15asking someone about email they'd like
50:17to say I like you know I it's really
50:23important to me to be able to mark
50:24emails as unread right and like what
50:27does it mean unread that whole unread
50:29thing is a sign that the whole inbox is
50:32being misused it's what's really going
50:33on is it's a to-do list you know and
50:36that's why you have to mark items as
50:37unread so they stay on your to-do list
50:39um and so you could so you start asking
50:43them with what's wrong and then you try
50:45and figure out what they really mean
50:46what they're really getting at which
50:48they themselves might not know you know
50:50and then you start asking them so what
50:53if you could do such-and-such what if
50:54you could do such-and-such right so
50:56start by asking them questions about
50:58their life and then get hypothetical on
51:00them okay we should ask someone on this
51:03side you've taken over so you just keep
51:07calling them you're doing great
51:08but you you pick you'll go all right
51:10right there in the way corner how do you
51:26justify the financial risk of launching
51:29a shitty product okay
51:34so the risk of launching early is not as
51:37great as the risk of launching late it's
51:39not like there's only risk on one side
51:41right there's a risk of launching late
51:43too so you got to have a rule of thumb
51:45about when to launch and my rule of
51:47thumb the way I like I have this phrase
51:53for it that didn't become as sticky as
51:54Minimum Viable Product because it
51:56describes what the Minimum Viable
52:00but it's a quantum of utility right um
52:04launch as soon as you have a quantum of
52:07utility which means as soon as you're
52:10there's one person in the world who is
52:13glad that you launched because now they
52:14can do something that they couldn't do
52:16when there's when you if you don't if
52:18you have something that if you launched
52:19it no one would be happy then you're not
52:23ready to launch right as soon as there's
52:25at least somebody out there who would
52:27say oh here's this thing I can use
52:29because I can do X and before that I
52:31couldn't do it then you should launch as
52:33soon as you have a quantum of utility it
52:35sounds like there's some wolves utility
52:36function like if it makes too many
52:38people happy and like one person happy
52:40in like a hundred really unhappy maybe
52:42it's not ready but if you you know Pablo
52:45Kaito he says if you have ten people who
52:47really love your product that's a good
52:49place to be yeah all you need is if you
52:51that's why if there's ten people who are
52:54totally totally launch unlike and nobody
52:57else cares that is perfectly fine that's
53:00great okay well the danger of building
53:25okay so the question is how do you
53:26decide between building what people need
53:28and building what people want right well
53:32the problem is what you really want to
53:34be doing is build something people will
53:36choose to use or buy more becoming your
53:39user or something like that and so
53:41here's a counter example to the whole
53:44need versus one thing what people need
53:46is healthy food what people buy is
53:49unhealthy food so if you're a business
53:53which one do you want to do okay I mean
53:56yeah ideally you want to make healthy
54:00food that also tastes good right but
54:02really um you can't you can't push this
54:06you can't push this too far I mean I
54:09myself would not start some company
54:10making junk food or something like that
54:12but if you are making food
54:14don't like or something where need and
54:17want are very different be real careful
54:19about picking need you know or you'll go
54:22out and you know your goal will be to
54:24cleanse the world of sin or something
54:25like that it's it's easy to delude
54:28yourself as a founder to think you're
54:30like you yeah the other danger is what
54:33they need is what you think they what
54:35you what you're is implicitly talking
54:36about is what you think they need
54:38cleanse the world of your definition of
54:40sin right yes yeah normal as a founder
55:02to be to see multiple different
55:03solutions yeah sure absolutely you just
55:06have to pick one pick the one that you
55:09think could get could get usage quickest
55:16what's the founder DNA founder DNA
55:21bothers you mean what makes a what makes
55:27yes it choose to be one what makes
55:35someone a good founder a combination of
55:38positive and negative things I think
55:40positive things are things like
55:43determination and you know willingness
55:46to try new things but then there may
55:50also be negative things like if if you
55:54are I think if you have worked for a
55:58large company for twenty years you might
56:01not be a founder unless you were forced
56:03to for visa reasons because if you were
56:06the kind of person that would that would
56:07make a good founder you wouldn't be able
56:08to stand working for a large company for
56:11twenty years right and so actually we've
56:17noticed certain companies where the the
56:20alumni of these companies tend to make
56:21bad founders and it's because no one who
56:24would make a good founder could have
56:27for these companies for very long okay
56:32maybe we should pick one on this side
56:33okay what do you what are you price
56:50during lunch well you can probably guess
56:54if you know your business you know
56:56roughly what you should charge so you
56:58should probably just go with whatever
56:59you'd guess if you want to be more
57:01scientific about it you could try
57:02talking to it talking to some potential
57:05users most of the time you have some
57:07sort of tame users who are actual users
57:10but are also friends of yours or
57:11something like that or family and you
57:14tell me honestly like what would you
57:15actually pay for this right um you can
57:18always change your prices later though
57:20if you want to lower your prices no
57:22one's gonna complain and if you want a
57:24grant if you want to raise your prices
57:25you just grandfather your existing users
57:28which if you have exponential growth
57:30will always be a tiny subset of your
57:32total users and then no one will
57:33complain about that either so don't
57:35sweat about it too much
57:36one way I think about that is it just
57:38seems like in the beginning what you
57:41need more than anything is customers and
57:43you want them to pay so you will need to
57:45pick a price that gets you customers so
57:48you'll learns because you learned so
57:50much from those customers and you're
57:53don't lose don't leave you customer
57:55price yeah cuz in the beginning you just
57:57want growth because the customers teach
58:00you the first customers don't just give
58:03you money they teach you all right two
58:05more questions you choose you you pick
58:08to you okay right up front here angel
58:11investing so my name is Karen and I'm
58:13with chat goose which is a chat manager
58:16so three tips for dealing with and
58:20finding angel investors if you don't
58:22have you know the original goals etc
58:24well we never talked about fundraising
58:26at all well maybe I'll have to do this
58:27whole thing again that's good fun
58:30fundraising as it is it as a as a Nakia
58:33is the thing to occupy your thoughts as
58:35a founder is overrated it's better not
58:37to be spending all your time thinking
58:38about fundraising we come to it much
58:41we don't need to all right so the
58:43question was how to find angel investors
58:46well that's the part that's the thing
58:49I'm particularly ill-suited to answering
58:51that question because Y Combinator is
58:52like this machine for funneling angel
58:55investors into all the startups that it
58:57accepts so like how do you find angel
58:59investors if you don't it's like asking
59:01a pilot how to you know walk along and
59:03track on the ground I know the least
59:07about that um here's the thing angel
59:11investors are looking for you
59:12so get out there go to I don't know well
59:20there's there's lots of startup events
59:22the angels are ancient like they're not
59:24necessarily good angels but angels are
59:26out there looking and the look the best
59:29way to actually meet angels is to get
59:32one to invest in you and then to
59:34introduce you to people actually I know
59:35the answer to this I know the answer
59:38find find people who work at startups
59:41and ask them to introduce you to their
59:44investors that's the way to do it that's
59:47good that's a good answer
59:48okay last question Paul you get to you
59:50get to choose maybe from this right
59:51there is why is he thinking of going
01:00:05after high school students to fund well
01:00:08I don't know because I don't think on
01:00:10behalf of YC anymore but you better not
01:00:12be because that would be an evil thing
01:00:14to do there at wise high school students
01:00:17could there are plenty of high school
01:00:18students who could start successful
01:00:21startups but they shouldn't just because
01:00:25you could start a successful startup
01:00:26doesn't mean you should because if you
01:00:28start a successful startup like that
01:00:31means like the you know the Footloose
01:00:33and francy fancy-free days of your life
01:00:35are over hey that like you're going for
01:00:39that company but we have funded kids in
01:00:42high school but only because they're
01:00:44already going like like it's that we're
01:00:47not encouraging that I agree you think
01:00:50to go to high school encourage kids to
01:00:51like do this incredibly hard thing
01:00:55fun and like like being kids that's
01:00:57that's awful I actually think that it's
01:00:59MoMA Chur optimization to write what you
01:01:02should be doing when you're in high
01:01:03school and even college is you should be
01:01:05figuring out what the options are
01:01:07we're not picking one option and running
01:01:08with it right and you know it just turns
01:01:13out that usually people who are in
01:01:16school and then stop to start companies
01:01:21it's like you know there are dropouts
01:01:24from Harvard we know the classic
01:01:26examples but if you have like if you
01:01:28drop out and just go do it and it's
01:01:30you're on a path that's one thing but
01:01:31most people like they're dabbling and
01:01:33they have a fallback and that's not a
01:01:35good way to be a startup founder you
01:01:37have to be a hundred percent committed
01:01:39and so we'll fund people in college
01:01:41maybe even high school if they're
01:01:42already clearly a 100% committed
01:01:44otherwise no whatever Trent I think that
01:01:47trend you're talking about is that is
01:01:49that you graduate from high school and
01:01:50why bother getting a college degree
01:01:52because you're already a fantastic
01:01:53hacker so I'm going to hire you and and
01:01:56if that person wants to enter the
01:01:57professional workforce well that that's
01:01:59great I don't know I was I don't think I
01:02:02would it would have been good for me it
01:02:03wouldn't have been good for me either
01:02:04but you know like go off and start just
01:02:06some startup I just think between 18 and
01:02:0822 you grow a lot as a human being who's
01:02:12human and humane and it's not clear what
01:02:15what society will turn into if we if all
01:02:18those folks enter the workforce it's
01:02:20it's good to like mess around and do a
01:02:22whole bunch of different things in your
01:02:24early 20s if you could have work yeah
01:02:25whether it's whether this messing around
01:02:27doing a whole bunch of things takes the
01:02:29form of college or something else just
01:02:31like don't be careful because like
01:02:34starting a startup is like catching a
01:02:36dragon by the tail if it works it like
01:02:40be careful at what point in your life
01:02:42you do that Paul thank you very much for
01:02:46coming in thank you deaf
