00:00hi everyone welcome to the Asics & Z
00:02podcast today's episode features Chris
00:04Dixon now a general partner on a 6 & Z
00:07crypto interviewing allod Gil the author
00:10of a new book just out called the
00:12high-growth handbook scaling startups
00:15from 10 to 10,000 people Elan is an
00:18investor or advisor to numerous tech
00:20companies as well as co-founded color
00:22genomics he was also co-founder and CEO
00:24of a company that was acquired in the
00:26early days of Twitter where he also then
00:28became a VP and he also worked at Google
00:31where he started their mobile team and
00:32worked on Adsense the discussion that
00:35follows covers everything from executive
00:37hiring to product management to late
00:39stage financing and more the two alsa
00:41then discussed broader market and tech
00:43trends ranging from the continuation of
00:46mobile cloud to machine learning to
00:48crypto and finally to lunge evety both
00:51near-term and further out in the future
00:52but they begin with the challenges of
00:55scaling there's sort of famous books
00:56like Peter teals book which is from zero
00:58to one kind of the early stages of a
01:00startup and I think a lot of blogging
01:01and other kinds of stuff has been out
01:03there from like Paul Graham and Fred
01:05Wilson and things about fundraising and
01:07product ideation and things like this
01:09but there hasn't been much written about
01:10all the later stages I mean the airplane
01:12I think there's you know Peter chills
01:14book is zero to one and really the hope
01:15is to have a book that's talking about
01:17when you go from 10 to 10,000 you know
01:18something's working you have product
01:20market fit and then what are all the
01:22different things that break if you think
01:23about it the surface area of an
01:25early-stage company is actually really
01:26simple it's don't run out of money don't
01:29fight with your co-founder and find
01:31product market fit and that's that's all
01:32you have to do and you'll be successful
01:33to some explodes of course very hard to
01:35do it's incredibly hard but you know
01:39those are the three things and at the
01:41late stage there's all sorts of things
01:42that can break executive hiring buying
01:44companies for the first time whether you
01:46should go public how do you do late
01:48stage fund raises how do you build a
01:49product management or like there's all
01:51this complexity that starts to kick in
01:52and so this book is really meant to try
01:54and address those leader stages
01:56something's starting to work what do you
01:57do you know it's interesting because I
01:58use three based in New York and now
02:00based in Silicon Valley and there's a
02:02lot of you know the companies that get
02:03to product market fit they're not just
02:05in Silicon Valley right there's plenty
02:06of companies in New York and other
02:07cities other places what I've noticed is
02:10different about Silicon Valley is
02:12there's a whole bunch of people who have
02:13done like scaling the one to end phase
02:15and so I kind of describe it as like the
02:18second stage rocket you know that's sort
02:20of the first stage is finding product
02:21market fit and then can you get the
02:22second one layered on there and one of
02:25the advantages of Silicon Valley I think
02:26people talk about how there's more
02:28engineers and there's more investors but
02:31those are frankly everywhere but the big
02:32difference is VP marketing VP sales beep
02:35engineering and then all the product
02:37managers and just sort of all those
02:38layers in between somebody was telling
02:40me a European founder was telling me
02:41that they felt that often European
02:43technology companies will cap out at a
02:45certain stage and it's in part market
02:47dynamics around the EU market but in
02:49part it's just driven by the fact that
02:51you tap out in terms of where the
02:53executive team is capable to take things
02:54given their experiences so there's a lot
02:56of misconceptions are missing around
02:57scaling startups what do you think are
03:00the biggest changes in the company as
03:01the organization grows I mean the naive
03:04view is you're 50 poised you have a
03:05hundred people that's just twice as many
03:07and you've got to do some you know more
03:08management but of course what really
03:10happens is the communication patterns
03:12break down and a whole bunch of new
03:14things have to be developed there's
03:15three or four big chefs so you touched
03:17on a really key one which is the way
03:18that you communicate is fundamentally
03:20different particularly if you're high up
03:22in the company and part of the reason is
03:24you know if you're 10 people you can
03:26just sit in a room and talk to everybody
03:27if you're a hundred people there's two
03:29to three layers between the CEO and
03:31every other person but as you start
03:32hitting 500 a thousand people the layers
03:34multiply and the ability for information
03:37to move up and down the stack shifts and
03:38that also means that the modes of
03:40communication has to change right you
03:42can't go and have as many skip level
03:44meetings not everybody's gonna check to
03:45the all-hands anymore you have people in
03:47their national offices and so you really
03:49have to move the modes of communication
03:50and a pretty deep way the second thing
03:52is you really have to focus on building
03:54out a strong executive staff and not
03:57just that you're gonna have entirely new
03:58functions that you as the CEO of never
04:01dealt with and you have to figure out
04:02how to a higher CFO for the first time
04:04how do I think about customer support at
04:06scale and so you have to start
04:08understanding how are you gonna hire all
04:09these really talented people and
04:11convince them to come in and you know
04:13build out their respective orgs the
04:15third is that you start adding new
04:17processes and you have to have a certain
04:19set of lightweight processes for
04:20strategic planning for deciding how to
04:22budget that a thousand people it's more
04:24about how do you start allocating
04:26to the right areas and then lastly the
04:28surface area what you're doing expands
04:30so dramatically that you have to really
04:32learn how to delegate and manage at
04:34scale so if your internationalizing and
04:37you're dealing with different time zones
04:38in a much larger orb you have to deal
04:40with that if you're buying companies you
04:41now have to have people who can deal
04:43with the integrations of those companies
04:44if you're launching multiple product
04:46lines how do you think about deprecating
04:47the old ones how do you do cross sales
04:49bundles etc and so the complexity keeps
04:52sort of notching up across the board
04:53yeah one thing I've noticed so for
04:55example if you take your typical like
04:56former program or engineer founder and
04:59let's say they're doing you know a b2b
05:01kind of sales company and then they have
05:03to at some point hire sales leader and
05:06marketing and all these other things
05:07they have never done that before right
05:09and so they just have no idea what
05:10greatness looks like and so you know how
05:12do they do that number one and number
05:14two there's a tendency I think to apply
05:17some of the same filters that one might
05:19use in the first ten people when you're
05:20hiring product so for example I've heard
05:22multiple times engineer founders wanting
05:24salespeople who know how to program or
05:26something right which is probably not
05:28compatible with generally great
05:29enterprise sales skills and so what kind
05:32of advice do you give for people who are
05:33hiring in these roles for jobs I've
05:35never frankly even worked maybe with let
05:37alone hired for I think there's
05:38basically two things that are sort of
05:40good generic advice and obviously the
05:42only good generic startup advice is that
05:43there's no good generic startup advice
05:44right so take everything with a grain of
05:46salt but the first is don't reinvent
05:49everything I think if you're an
05:50innovative especially technical founder
05:52you want to rethink what's my org
05:54structure ultimately you know there's a
05:56reason that corporations have been
05:57around for hundreds of years and
05:58actually work and there's certain org
06:00structures that you can innovate on but
06:01ultimately you should find certain
06:03experienced people to do certain things
06:04where the function needs that experience
06:06so that's sort of a meta point I think
06:09in terms of finding an executive there's
06:11a few steps that you can take the first
06:12one is you should go and meet all the
06:14best people of that function it's not
06:16necessarily people to hire its people to
06:18learn from so what does a great CFO look
06:20like and then I would just ask them what
06:23would you look for in a person that
06:24would fit my company and what sort of
06:27experience would you look for what
06:28interview questions should I ask them
06:30how should I get to know this person how
06:31will I know if they're good in the
06:32function second once you get that
06:34feedback I'd write up a job rack and I'd
06:36circulate it to all the people who are
06:38going to be interviewing this individual
06:39because they don't have any context
06:41either and they have no idea what to
06:42look for and so I found for example when
06:45you're hiring a BD person or if somebody
06:47to run BD an early company will have a
06:49mix of people from engineering and
06:51product in other orgs and they may have
06:52no idea what to actually look for and
06:54they may think it's a salesperson versus
06:55a BD person versus something else and so
06:57just defining that tightly so everybody
06:59knows these are the questions we're
07:01gonna ask these are the characteristics
07:02we're looking for allows both clarity of
07:05interviewing but then it also allows for
07:06clarity of discussion of the candidate
07:09once you make that hire and then the
07:10last thing is you should figure out how
07:12to UM board them effectively and part of
07:14that will be working with them closely
07:16to figure out what they think their role
07:17should be and what the function should
07:18be what they think it should evolve as
07:19and then mixing in your own opinion and
07:22making sure that you don't feel
07:23overwhelmed by their expertise but
07:26you're willing to defer when they really
07:27understand something you don't one thing
07:29when the process I've observed is
07:31companies that don't think they need
07:33let's say a CFO and what we like to do
07:35is we do what we call calibration where
07:37we introduce them to three great CFOs
07:38who probably aren't looking for jobs and
07:41almost invariably the founder comes back
07:43and says oh my god I didn't know that
07:46someone could do those things and like
07:47all those problems that would solve for
07:49me and it's completely eye-opening right
07:50because if your background is
07:51engineering or whatever you've just
07:52never seen greatness and CFOs VP sales
07:55etc right yeah absolutely I think the
07:57services model that Anderson horses
07:58built is really valuable for founders as
08:00are scaling and the CEO role because to
08:02your point out the whys that we'll have
08:02no idea founders have a bunch of
08:04questions about hiring use x-wind should
08:05they hire them what should they look for
08:07there's lots of debate about how high
08:09the bar should be and you know what kind
08:11of athletes versus veterans or something
08:14where do you come down on those issues
08:15so there's a few different sections of
08:17the book that talk about both executive
08:18hiring as well as managing your
08:20executive team and how do you go about
08:21it and I think there's a few key tenants
08:23in terms of bringing on execs or
08:24alternatively promoting people from
08:26within I think Ben Horowitz may touch
08:28about the inside cameras in his book or
08:29in other articles you should always be
08:31thinking of the person is the right
08:32person for the next 12 to 18 months not
08:34the person for the next five years and
08:36so I think a common mistake is that
08:37people will end up hiring you know
08:39they'll have a 20% engineering team and
08:41they say hey in five years we'll be 500
08:43people so let's hire that five hundred
08:44percent engineering team person no they
08:46hire way too far ahead and that's a
08:47person who's out of a big company is
08:49used to managing a big organization I'll
08:51basically show up be incredibly boring
08:53and either try to over hire and hire a
08:54super senior staff or just turn out
08:56second people tolerate good versus great
08:59hires for too long and so they'll have a
09:01good VP marketing versus the person
09:04who's truly exceptional and a great fit
09:05for what they're doing and it's almost
09:07like being on a life raft and you're
09:08like I have five spots and who are those
09:11five spots it's not this this giant
09:12yacht you know cruising down the
09:14Mediterranean that you can have hundreds
09:16of people on and so really your
09:17executive team should be this life raft
09:18you should ask if this person's good but
09:20not great how do I find great sometimes
09:22by the way those people will be
09:23cofounders other people who were early
09:25you're loyal to them they were really
09:28helpful in beginning and they may still
09:29be great an organization but maybe
09:31they're not the best beeping right I
09:32don't know that's the tension I've seen
09:33sometimes I think in general dealing
09:35with sort of what are called the old
09:37timers in the book is a key question
09:39because they've proven their loyalty to
09:41the company they were essential in the
09:42early days they have all the cultural
09:44contacts they have all the trust of the
09:46founders which is really important and
09:48valuable currency in terms of getting
09:49things done and a subset of them are
09:51gonna act really badly when all said and
09:53done they're going to drag their feet
09:54they're gonna argue with new hires
09:57they're gonna effectively throw sort of
09:59a prolonged mini fit and part of that is
10:03just driven by the fact that they used
10:05to have enormous influence and that
10:06influence is diminishing part of that is
10:08it's a very chaotic environment and part
10:10of that is not understanding what's
10:12coming and what's coming is enormous
10:14growth potential for everybody who's
10:15involved with the company and so I
10:17remember when I joined Google it was
10:19about 1,500 to 2,000 people and over
10:21three and a half years I grew to 15,000
10:23people and the first time I went through
10:25hyper growth I was caught up in all the
10:27chaos if I had a peer who got promoted
10:30above me I'd worry of what about me and
10:31the second time I went through it it was
10:34a Twitter and they bought my startup
10:36when Twitter was about 90 people in over
10:38two and a half years I went to 1,500
10:40people and the second time through it I
10:42just thought you know what the company's
10:44gonna grow in all these different
10:45directions and that means that from a
10:47career perspective I'll be fine and I
10:48should just do good work and keep my
10:50head down and try and be helpful and so
10:51there was a product reorg that happened
10:54on the Twitter team that was a little
10:55bit messy and dick was asking different
10:58product managers what they wanted to do
10:59and I said well I just want to help the
11:00company like put me wherever you want I
11:03don't care I just wanna help the company
11:04and then two months later I got promoted
11:06to be a vice president there and I think
11:07it was because I just put everything
11:09else aside and said where can I help so
11:11I think that's a really important thing
11:12put aside the angst focus on the big
11:14picture and realize that every six
11:16months if the company is in hyper growth
11:18it's a different company if it goes from
11:2050 to 200 people the five hundred people
11:22over six to 12 months increments it's a
11:25completely different company there's new
11:26opportunities somebody's gonna go and
11:28need to open international offices
11:29somebody's gonna need to launch a new
11:30product somebody's gonna be needed for
11:32new functions and you can be that person
11:34if you do good work and keep your head
11:35down and the other thing you should
11:37expect as an old-timer is your role is
11:39gonna shrink and then it'll expand and
11:41you need to be okay with that shrinking
11:43period you need to be okay with you
11:45losing responsibility before you
11:47suddenly gain tons of responsibility
11:48because if you're the only designer you
11:50were designing everything but you're
11:52gonna start hiring other designers and
11:53you're gonna have to give things away
11:54but a year later they're gonna have to
11:57promote somebody to run a chunk of the
11:58design team and that may be you if you
11:59act the right way and so I think really
12:01keeping that in mind is crucial so let's
12:04talk a little about product management
12:04what are some of the key learnings that
12:07people should know about there there's
12:08very few very strong product
12:10organizations in Silicon Valley Google
12:13has a great product or Facebook has a
12:14great product order there's a handful of
12:15them and part of that I think is because
12:17often technical founders either create a
12:20suitor role for that and other functions
12:24are basically filling the gap or in some
12:26cases it's back to the earlier point
12:28around you never see excellence in the
12:29role so you don't really understand how
12:30to build it out and so often the
12:32evolution of a product organization will
12:34be the one of the founders will be
12:35really the product manager for the
12:36product and then it'll evolve into a
12:39state where there's gaps and they're
12:41trying to fill the gaps around product
12:42and so they'll pull in somebody from
12:44another team it could be marketing it
12:46could be sales it could be business
12:47operations it could be something else
12:48and those people may be very exceptional
12:50running process and running schedules
12:52but they may not have the product
12:53insights and the experience to really
12:56know okay this is how you think about
12:59this is how you think about a PR D or
13:01writing a spec for the product for the
13:02first time and here's a set of processes
13:05that help me engage all the stakeholders
13:06around the company to really sort of
13:08drive a product forward so eventually
13:11what happens is if those process you
13:13start to break down or that approach
13:15starts to break down they'll end up
13:16eventually hiring in somebody who has
13:18the product experience to build out an
13:20it's verdict it's professionalized
13:21Facebook and Google they obviously are
13:23hiring great people but I think it's
13:25also the way they empower those people
13:27and design the organization that makes
13:29some great product organizations yeah
13:30they ended up hiring Jonathan Rosenberg
13:33and externally to run product for Google
13:36and what he ended up doing is I think
13:39three key things number one is he really
13:41started mentoring a lot of the key
13:43product individual contributors who went
13:45on to actually run most of the product
13:47or at Google in the early days so Susan
13:50Wojcicki who now CEO YouTube Salar who
13:53similarly had run YouTube and before
13:54that the ads team Marissa Meyer who
13:57became CEO Yahoo see how these people
13:58who were sort of all-stars in terms of
14:00the things that they've accomplished and
14:01very early on they were basically
14:03individual contributors on the product
14:04org and they were really learning the
14:05ropes and some came from engineering
14:07backgrounds and some came from marketing
14:09or other backgrounds back to that mix of
14:11people sort of joining and so one is he
14:14mentored people actively to is he really
14:16focused on the process side of it what
14:18are the check points for launching a
14:19product how do we approach it and then
14:21lastly he actually built a training
14:22program called the associate product
14:24management product manager program where
14:26they basically recruited really really
14:28smart people out of in the Google case
14:30top universities and they basically said
14:32we're gonna train these highly technical
14:34really smart people to do product
14:35management and we're gonna select for
14:36people who in our interview process seem
14:39to exhibit very strong product instinct
14:40and so that group ended up including a
14:43variety of people who've really gone on
14:45to impact product organizations all over
14:47Silicon Valley so this will go to the
14:49question of what is a great product
14:50manager just first what is product
14:52management cuz I think it gives a few
14:53sometimes with project management on the
14:55one hand design on the other hand yeah
14:57it's a good point Ben Horowitz has a
14:58good blog around good product manager
15:00bad product manager from sort of the 90s
15:01the tries to encapsulate enterprise
15:03product management you know ultimately
15:05the role of a product manager is four
15:07fold number one is to really own the
15:10generation of the product requirements
15:11and what's the user need how all this
15:14actually get used how's it
15:15differentiated what are the key
15:17components that make this a good product
15:18so one is sort of owning that view of
15:21the product and roadmap and sort of
15:23independence for example an enterprise
15:24company typically they'll sit between
15:26sales and engineering and kind of
15:28mediate the - yeah there's typically a
15:30natural tension between product and
15:32and a sort of a second role that they
15:35often play particularly in enterprise
15:36companies is they're almost a buffer
15:38between engineering and the rest of the
15:39world but they're also a communication
15:41device both in and out in other words
15:43they'll help synthesize a lot of the
15:45different feedback from sales and
15:46directly from customers want this
15:48feature I want that and the salespeople
15:49are kind of incentive to say yes the
15:51engineers on the other hand want to keep
15:53the product focused and get their job
15:55done and so there's a tension right and
15:57the product manager sits in between and
15:59kind of prioritizes it figures out what
16:01really matters exactly and on the
16:03engineering side there's tension with
16:04product because engineers may think a
16:06feature stupid because they're not
16:07necessarily hearing the direct feedback
16:09from the customer and they may not
16:10believe oh the customer really needs
16:11this or they may want to go down the
16:14road of building something that may be
16:15more complicated than what's really
16:16needed to future-proof better for other
16:18reasons and if product pushes back on
16:20that there may be tension there as well
16:21and so the role of a product manager in
16:23part is also to navigate all that
16:25tension between the different
16:26organizations and end up with a view of
16:28what's really needed for the product
16:30despite all these various inputs great
16:32so let's talk about in your book you
16:33talk about late stage financings there's
16:35been a lot written and sort of blogged
16:36and other things about early stage and I
16:38think people at least in the
16:39subcommittee kind of understand Series A
16:40and B but there's a whole kind of new
16:43new I say because it in the old days ten
16:46years ago you'd go public or something
16:47and now there's this whole kind of other
16:49set of investors what are the things the
16:52main things that founders should know
16:54about that yeah I think there's a few
16:56key things that founders should consider
16:57when doing a very late stage ground and
16:59many of these rounds in some cases are
17:01tied to secondary in some cases are not
17:02so the first is there are some value
17:05added late stage investors and so just
17:08like in the early stages not all money
17:09is equivalent and the late stages there
17:11may be people who will continue to buy
17:13stock as you go public and therefore
17:15help support the price of the stock in
17:16the early days as things start using so
17:18called crossover investors the IT
17:20reprise fidelity and the like they may
17:22have interesting insights in terms of
17:24how public markets will think about them
17:26and therefore how they should be
17:27thinking about navigating their business
17:28early on or the sets of reporting or
17:29controls or other things that they want
17:30to build in where they may have real
17:32insights into markets or differential
17:34networks that you normally wouldn't get
17:35access to and so I do think that there's
17:37investors who can help in a more
17:39traditional sense they just get involved
17:40later in the life of a company so the
17:42first question is just whether or not
17:43this money will be helpful it doesn't
17:45have to be frankly the
17:46because in some cases in many cases you
17:48have escape velocity but optimally you
17:50have capital onboard that's always going
17:52to help the second thing is that you
17:55don't want to get too far ahead of
17:57yourself in valuation and I think often
17:59at the later stages there's a temptation
18:01to do that because the odd thing is the
18:04higher the valuation of the company and
18:06the better known the brand the more
18:08sources of capital exist and so what you
18:11see is on the late stage is some
18:12companies start building in very
18:14complicated structures because they want
18:16that very high valuation and a firm may
18:18be willing to give it but then they
18:19build in extra preference or
18:21participation for the investor or things
18:23that effectively are really a debt
18:24instrument masquerading is equity
18:26they're really locking in a return just
18:28being preferences multiple liquidation
18:30preferences ratchets all sorts of kind
18:33of quote structure yeah yeah exactly the
18:35last thing that I think is interesting
18:36about lay stage capital is the degree to
18:38which it allows companies to keep going
18:40whether they should or not and so I
18:42would posit that about half of the
18:44unicorns are not worth a billion dollars
18:46more in other words 50% of the companies
18:49that are valued in the private markets
18:50as being worth more than a billion
18:52shouldn't be worth more than a billion
18:53and that day of reckoning keeps getting
18:55pushed out because of so much capital in
18:58the markets and so much liquidity sort
19:00of pushing things ahead so there's been
19:02big shifts in terms of late stage
19:04financing and how you get that done
19:05there's also been big shifts in terms of
19:07M&A what do you think that what do you
19:09attribute that to I think it's maybe
19:11three things one is private market
19:13valuations continue to go up
19:14second I think that there is a
19:17generation of founders who haven't
19:19really made the transition mentally that
19:22often as a company scales you're moving
19:24from a product centric organization to
19:27one that's both product and distribution
19:28centric this is actually something that
19:30Marc Andreessen talks about a little bit
19:31in the book and in terms of a discussion
19:33that we have where you know if you look
19:34at the old model you look at Cisco or
19:36Microsoft or any of these companies what
19:38they would do is they'd have over than
19:40anything else they'd get distribution
19:41due to that and then what they do is
19:43they go and buy companies or build
19:45things and push it against that
19:46distribution channel and continue to
19:47sort of iterate that also happen with
19:49the next wave I mean Google did that
19:50really aggressively with everything from
19:53Google Maps to Android to Chrome
19:54Facebook did it and so the traditional
19:56model is you buy stuff and you
19:58distribute it to your channel and you
19:59realize you're also just
20:00Bhushan company and I think that a lot
20:02of folks today have less of that mindset
20:05of hey we're also a distribution company
20:06and therefore we should be buying things
20:08aggressively to distribute I think a
20:11third area is internal pushback but
20:13maybe it's stronger now around two
20:15things so the argument is often well you
20:17should wait or we shouldn't buy it
20:19because it'll take us two years to
20:20integrate it and often the answer may be
20:22don't integrate it you know it's okay
20:23maybe if it runs independently you have
20:24a series of API is talking to each other
20:26and then secondly there's often
20:28arguments around comp if we had 10
20:30million dollars instead of buying this
20:31company we could hire 50 engineers and
20:33therefore we should hire those engineers
20:35but it's often really hard to find 50
20:38incremental engineers and then allocate
20:39them to that thing that you buy often
20:41you allocate them elsewhere I think I
20:42agree with all that and I would add to
20:43it that some of these big tech companies
20:46have shifted their salaries so dramatic
20:48like they basically broke they
20:49traditionally incumbents would have sort
20:52of salary bands and that was one reason
20:54sort of to leave and then they would
20:55also then they could resource a laries
20:59at companies like google some of this
21:01came out and like the court documents of
21:02what they're paying like the way Moe
21:03engineers and it's like you know it's
21:05like the price of an annual price of a
21:07start-up acquisition or something so
21:08they've also just dramatically rethought
21:10their pay scales which has hurt the kind
21:12of aqua hire market I think there's
21:13probably different analysis of different
21:14stages like I co hire and then sort of
21:16tech tuck-ins would be more the
21:17distribution I think the takeaway though
21:19for founders I think it was probably
21:20never a good idea to build a company to
21:22sell but it's even more so not a good
21:25idea today you need to first hit product
21:27market fit and then you need to figure
21:28out how to scale and you just have to
21:30deal with those issues right absolutely
21:31but I also wonder how much of it is
21:33cyclical or sort of secular change right
21:35is it like a fundamental change in the
21:36market I think you're raising two really
21:37good points one is what's the why now of
21:39why to buy a bunch of stuff and often
21:41those mirror technology waves and so
21:43since we just went through cloud mobile
21:45and social as three separate waves
21:46you actually saw companies subsequently
21:48buying talent for those areas I remember
21:50when Twitter bought my startup back in
21:522009 shortly thereafter there was a big
21:55wave of buying mobile companies and
21:56every company was trying to buy mobile
21:57teams cuz iOS was new yeah and then it
22:00was AI and then it was you know whatever
22:02so maybe the only area where people are
22:04very aggressively buying right now is
22:06machine learning and so maybe there's a
22:08lack of why now I think part of it too
22:10is people aren't really hiring corporate
22:12development people early in the life
22:14these breakout companies of this last
22:16batch or last generation and I think
22:18that may also be a reason like there's
22:19nobody who's just thinking about it
22:21non-stop and often you see these
22:23transitions or if you hire a great deaf
22:25person selling the company will go and
22:27buy a bunch of companies and it's a
22:28little bit chicken-or-egg but I actually
22:29think having somebody who's thinking
22:31about it strategically and can aggregate
22:33all the input from the different teams
22:35in the company and take it to the
22:37executive team and say hey here's the
22:38five things we should buy and why
22:39actually makes a big difference and so
22:41coinbase is a good example where over
22:42the last you know 12 months I think
22:44they've made five or six acquisitions
22:46and in part that may be driven by them
22:47hiring Emily Choi to come in and you
22:50know really Drive that function in a
22:51smart way so one thing I like to think
22:53about a lot is sort of where are we in
22:55the history of tech we're still in the
22:56middle I guess of the sort of mobile
22:57revolution but obviously sort of mobile
22:59phones and all that and cloud computing
23:01and SAS social what's the big thing over
23:05the next 10 years clearly those trends
23:06are gonna continue but will we layer on
23:09more trends we do seem recently to be
23:11headed for this kind of consolidation
23:12around the big five tech companies will
23:15that continue will there be new things
23:18that kind of come out obviously you and
23:19I both have an interest in crypto as an
23:21example as kind of this new thing that's
23:22kind of potentially quote-unquote
23:24disruptive yeah it definitely feels to
23:26me like there's four areas that are
23:28really interesting right now the first
23:30to your point is a continuation of the
23:31trends of the last decade around cloud
23:33and mobile and then sort of social being
23:35a primary consumer platform at this
23:37point the second area really is the
23:40crypto world and at least for me a lot
23:42of the really interesting areas in that
23:43are around forms of money store value
23:46privacy tokens etc as well as sort of
23:49the smart contracting world
23:50securitization of everything and then
23:52lastly the NFT or the non fungible
23:54tokens so persistent digital goods that
23:56you can mix and match in different ways
23:58across media games etc so crypto is
24:02super interesting third area is the sort
24:05of systems and semiconductor layer of
24:07machine learning so if you look at
24:09Nvidia and NVIDIA GPUs are really the
24:12primary basis for a lot of sort of
24:15performant machine learning models and
24:16Google obviously launched CPUs a couple
24:20years ago tensor processing units a
24:21specialised ASIC that's really meant to
24:23help accelerate certain types of machine
24:25learning models as well from a hardware
24:27and so I think that's a really
24:28interesting pretty wide open area and I
24:31think that's pretty transformative and
24:32it's a way to effectively index machine
24:34learning because I think it's very hard
24:35to invest in quote-unquote a machine
24:37learning company and then lastly I'm
24:39personally very interested in sort of
24:40longevity and anti-aging technologies
24:42and by that I mean true biopharma which
24:45is translating existing science that's
24:47existed in labs for 10 20 years and is
24:50trying to turn it into drugs that can
24:51extend lifespan pretty dramatically
24:53so the continuation of mobile and cloud
24:55we've gotten through obviously kind of
24:57the first wave of that what would wave
24:59to will be the most sighting things
25:00about wave - I think there's two or
25:02three components of wave to you that are
25:03really interesting number one is on the
25:06south side I still think there's a dozen
25:07billion dollar-plus companies to be
25:09created where you're just taking
25:10something that everybody's building
25:13internally over and over again or is
25:14turning it into an API to factor out a
25:16thing that people are building over and
25:18over and make it an API exactly and I
25:20think if you took apart a fortune 500
25:22company and asked what are all the
25:23different things that they have to do
25:25repetitively you could come up with a
25:27dozen of these so I think that's super
25:28interesting I do think there's a lot of
25:31services that are sort of mobile first
25:33in really interesting ways I think
25:35there's again reinvention of a lot of
25:36different areas you and I could talk
25:37about crypto for a full podcast or more
25:39so I won't go too much into that but I
25:41guess one big question is where do you
25:43think we are in the evolution of it like
25:44what are the next kind of key milestones
25:46that you'd want to see I think of it as
25:48there's two things are sort of the
25:49prices and then there's the real
25:50innovation and the real innovation has
25:51been strong but I think there's a lot of
25:53things that need to happen yeah it's a
25:55great point you know I think the primary
25:57use case that is clearly working as some
25:59former store of value slash speculative
26:02asset and there's lots of things that
26:03are claimed to be working that aren't so
26:05for example nobody's really building
26:06scalable apps on a theorem right now and
26:08I think the theorem community is clear
26:10that they need to be able to scale that
26:11platform and so one key thing is
26:13infrastructure and scalability and so a
26:15lot of people are obviously focused on
26:16that right now you know ultimately I
26:18guess the way that I would think about
26:19it at a macro level that just abstract
26:21out a level is it reminds me a lot of
26:23the late 90s where there's a few
26:26fundamental core things that people are
26:27clearly doing on the internet and then
26:29there's tons of speculative stuff that's
26:31gonna go to zero and so in the early
26:33days of the Internet you needed PayPal
26:35to be able to buy things you needed to
26:37be able to start doing commerce Amazon
26:38and eBay and some of the early sites you
26:41and then later Google for media and
26:43things like that but most comm companies
26:45didn't work out the interesting thing is
26:47that you basically had a speculative
26:48bubble couple to real value creation
26:50I mean Amazon the most controversial
26:52company that era of course is worth far
26:54more than its peak back then so it took
26:56a long time to get back there but yeah
26:57and then there was new companies like
26:59Google that ended up going public much
27:00later Facebook didn't even exist and
27:02then ended up becoming a major you know
27:04to your point about earlier it said you
27:05had to have both apps and infrastructure
27:08in that case the infrastructure was
27:09things like broadband and mobile phones
27:11and things like that but you know you
27:12couldn't have had Netflix and YouTube
27:14and a lot of the rich experiences the
27:17people expect today without a lot of
27:19infrastructure built out right I mean
27:20Netflix was literally mailing DVDs right
27:22and so it's shifted pretty radically to
27:24becoming a major content producer and I
27:26think in crypto we're gonna see the same
27:28thing where a most scripted projects
27:30from today are gonna go to zero there's
27:31a few that are gonna be incredibly
27:32valuable and/or TR and that's gonna grow
27:35but I think most are gonna go to zero
27:36but second I think as you have this
27:38technology built out you're gonna see
27:40ideas from today that failed ten years
27:42later work there's this game I think was
27:44called bomb it was a board game that was
27:47popular popular but somewhat popular
27:49among certain people in like the early
27:502000s and they had cards that were all
27:53of the they're called bad idea cards and
27:55they're all the bad ideas of the 90s and
27:57it was literally like if you go through
27:59the cards it's literally every top
28:01there's like ride-sharing yeah selling
28:03pet food online internet money you know
28:05group deals like mobile gaming it's
28:08literally all the quote unquote bad
28:09ideas actually turned out to be really
28:11good ideas they were just way ahead of
28:12their time and not built right yeah I
28:13think we're gonna see the same thing in
28:15crypto where there's all these ideas
28:16that are early and there's no
28:17infrastructure and in 15 years are gonna
28:19be amazing and then the third category
28:20is machine learning is a silicon machine
28:23like that's interesting cuz historically
28:25you have some big new trend and what
28:27they startups do you build
28:28infrastructure software etcetera now all
28:30of this stuff is either open source or
28:32like the algorithms are published openly
28:34as papers a lot of the code implementing
28:36those algorithms is very high quality
28:37kind of open source stuff that's in turn
28:40monetized through the cloud service
28:41providers right so it's kind of hard you
28:44know now that's a start-up you can go
28:46and you can take this stuff and you can
28:47apply it so you can build machine
28:49learning power at accounting software
28:50which is a very interesting area we've
28:51got a bio fund that's doing a lot of AI
28:54by Oh investing but you're right that
28:56there's no way if you want to just sort
28:57of bet on as an investor or as an
28:59entrepreneur bet on machine learning
29:00it's very hard to yeah and I think every
29:02major technology wave also had a major
29:04silicon company created in it and right
29:07now to some extent that's Nvidia for not
29:09only graphics processing but also for
29:11machine learning purposes but if you
29:13look at what a machine learning ASIC
29:14really needs it's really fast i/o and
29:16then it's a lot of matrix multiplication
29:17and if you look at the surface of a GPU
29:20that's actually dedicated to those
29:21functions is quite small relative to the
29:23overall chip and so there is room to
29:25create things that are ten hundred times
29:26better or faster or more performant from
29:29a power perspective by creating custom
29:31Asics per ml and so I think that's
29:33the other thing with ml writers is the
29:34training side which usually happens in
29:36data center and then the deployment side
29:37which will happen at some edge device
29:39and they have you know size and power
29:42constraints and other kinds of things
29:43that require specialized hardware that's
29:46a great point inference side of machine
29:47learning will have its own specialized
29:49chips potentially and if you look at the
29:51companies in the market they often
29:52segment themselves by whether they're
29:54involved with the training to your point
29:55or the inference and if you get into a
29:57world where you really have broad-based
29:59self-driving cars or robotics or other
30:01things then these chips become really
30:02important and then your fourth category
30:04longevity there's 20 years of evidence
30:06the aging is a biological process you
30:08can perturb and there's three or four
30:09main lines of evidence number one
30:11there's caloric restriction so if you
30:13clerkly restrict you extend lifespan
30:14animals organisms to is there's certain
30:17genes that if you knock out in model
30:18organisms they'll live to three times
30:20longer they'll be healthy adults the
30:21whole time and then will crush out so
30:23for example in C elegans
30:24there's a pathway called the DAF pathway
30:26that is involved in regulating lifespan
30:28and actually worked on that directly for
30:30my PhD third there's what's known as
30:32para biases which is exchange young
30:33blood and old animals and vice versa and
30:35it's clear that you can increase the
30:38rejuvenative capacity of older animals
30:41by giving them young blood and actually
30:42it looks like those factors in old blood
30:44that are screwing up functions if you
30:45give old blood to a younger animal it'll
30:47actually be less functional in terms of
30:49regenerating a damaged liver or muscle
30:51or the like and then lastly there's at
30:54least two fda-approved drugs that in
30:56multiple organisms will extend lifespan
30:57between five and thirty percent one of
30:59them is rapamycin which in mice for
31:01example extend I spent 10 to 30 percent
31:02and it also extends lifespan and flies
31:04in worms and other organisms so it's
31:06evolutionarily conserved
31:07the other one is metformin which is a
31:09tenth most prescribed drug in the world
31:10than today beauty's drug so it does look
31:13like there's clearly things that can
31:14modify the pace at which different types
31:17of organisms including mammals actually
31:19age then a lot of people think that's
31:20kind of science fiction or something
31:22they make jokes about the blood
31:23transfusions from young people or
31:25I think it's as near-term as any biotech
31:27product is in some sense because it
31:29takes time to develop there's a lot of
31:31factors that have been cloned out of
31:32blood that are actually really valuable
31:34drugs today - examples of that would be
31:36insulin which again is just a factor in
31:38blood which impacts the biological
31:39process that people now inject as a
31:41standalone and then there is Epogen
31:43which is amgen's big blockbuster drug
31:46which helps you generate red blood cells
31:48and similarly the way that was
31:49discovered was they've literally looked
31:51for what's the factor in blood that lets
31:53anemic housed B non anemic is something
31:56that have red blood cells and so these
31:58aging molecules are the same thing it's
31:59not some magical thing about blood is
32:02just there's proteins floating around
32:03and maybe some of them do good stuff the
32:04hard part has been that because you're
32:07looking at things like aging and
32:08longevity it takes a really long time to
32:10measure it and so the pace at which
32:12people have been innovating on these
32:13things is slow but also biopharma in
32:16general tends to be focused on oncology
32:17or cancer and a few other areas and
32:19they're not really thinking about how
32:21can i push disease back in general you
32:23know with a drug so I do think there's a
32:25big market gap which creates a really
32:27interesting opportunity what do you
32:28think about the societal implications of
32:30if we do develop longevity technology
32:32the societal implications are massive
32:34and I think there's all sorts of ethical
32:37conundrums because if you look at the
32:38kuhnian paradigm of scientific change it
32:41basically says you need the old guard to
32:42die in order for scientific progress to
32:44happen and so imagine if no professor at
32:47any university ever retired or imagine
32:49if any of these characters who you know
32:52have had a very negative impact on their
32:54country right no we can flip it around
32:55of course and find all the positives
32:57yeah but it does lock into idle
33:00structures and it does lock in power
33:03structures and so from that perspective
33:05one can argue that it could really hurt
33:07self renewal in a global system and
33:10therefore is it you know quote unquote
33:12bad the other sort of really interesting
33:14societal dynamic is around how you think
33:18about your own life so say that you had
33:21to viewers were you basically
33:2230-something for a hundred years you'd
33:24think about marriage and having kids
33:26differently when would you do that would
33:27you do it early would you do it late
33:28would you do it multiple times would
33:29they'd stay in grad school - they're
33:31like inertia people who do that so the
33:34people who stay in grad school - their
33:3540s where they stay till they're 140 do
33:38you try to make money early in life so
33:39you could spend it through the rest of
33:40your life or do you wait until late and
33:42gather experiences do you start thinking
33:44about having multiple careers over the
33:46course of your life and you also think
33:47about social mobility so you know if
33:50you're disadvantaged early does that
33:52extend for 200 years and what does that
33:54mean and so I just think there's these
33:55really deep societal implications of the
33:58ability to extend my spend fascinate
34:00well great well thanks a lot for joining
34:01us e laud ye lads the author of the new
34:03book high growth handbook alright thanks
34:04so much thanks for my shot me