00:00:00
please uh join me um welcoming venod
00:00:02
kosla to Stage
00:00:04
venod
00:00:07
oh how are you how are you doing legend
00:00:12
legend let your winners
00:00:17
[Music]
00:00:19
ride and in we open sources to the fans
00:00:23
and they've just gone
00:00:25
[Music]
00:00:28
crazy
00:00:31
wow thanks Legend how are you Legend
00:00:34
venod good to see you just by way of
00:00:36
background I think we all know venod um
00:00:39
venod is a legend and a
00:00:42
friend uh his Indian army officer dad
00:00:45
wanted his kid to join the Army and
00:00:46
venod decided instead to go into
00:00:48
business after reading about Intel and
00:00:51
getting inspired by co-founder Andy
00:00:53
Grove an immigrant who got um funding
00:00:55
for his startup in Silicon
00:00:57
Valley venod later got a masters in
00:01:00
biomedical engineering from Carnegie
00:01:01
melon on a full scholarship Stanford GSB
00:01:04
rejected him twice he finally got
00:01:06
accepted and received his NBA 1980 and
00:01:08
he worked at several startups some of
00:01:10
which failed and then one of them worked
00:01:13
out pretty good Sun Micro Systems
00:01:15
started in
00:01:16
1982 with his Stanford classmates Scott
00:01:19
McNeely Andy beckim and um Bill
00:01:22
Joy uh venod um raised $300,000 in seed
00:01:26
capital from Venture firm Kleiner
00:01:27
Perkins Cofield and buyers and within 5
00:01:30
years son made a billion dollars in
00:01:32
annual revenue he became a GP ecliner uh
00:01:35
where he helped create an NextGen which
00:01:37
he sold to AMD for 28% of its market cap
00:01:39
the first successful Intel
00:01:41
microprocessor clone company and his
00:01:43
largest return to date maybe we'll talk
00:01:46
about this today was Juniper Networks
00:01:47
where the firm's $3 million investment
00:01:49
in the 90s earned $7
00:01:53
billion not bad in 2004 he launched
00:01:55
coasta
00:01:57
Ventures 2006 I pitched him and he
00:01:59
rejected me me as an investor in my
00:02:02
company I came back in 2011 and he led
00:02:04
my series B and I had three competing
00:02:06
offers and I picked V so um ven's been a
00:02:10
great partner by the way I will also
00:02:12
give you guys I had offers from um I
00:02:15
guess I'll say it Founders fund and
00:02:16
andron Horowitz for my series B venod
00:02:19
gave me the lowest valuation and I still
00:02:21
took his offer entry price matter did he
00:02:24
add more
00:02:25
value vod's whole point was VCS
00:02:28
typically add negative value we're going
00:02:30
to put a senator on your board that we
00:02:32
know that knows this and that that firm
00:02:34
and the other board members they brought
00:02:36
and the partners they brought to the
00:02:37
table for me were incredible and he's
00:02:39
been a great um mentor and adviser I'll
00:02:42
also say I always gave people advice on
00:02:44
Veno because I get a lot of reference
00:02:45
calls like should I take money from
00:02:46
venod I heard he's pain in the ass
00:02:48
sometimes and I'm like let me tell you
00:02:50
something about venod I've never told
00:02:51
you this in public but I said the thing
00:02:53
about venod is you go into a meeting
00:02:56
with venod and he's like the Oracle in
00:02:58
Neo The Matrix you go in the kitchen and
00:03:01
he's like you know she says no thyself
00:03:03
she pulls out the chocolate chip cookies
00:03:05
she said he says am I the one she says
00:03:07
you're not the one kicks him out he has
00:03:10
to ultimately decide what the right
00:03:12
decision is and venod will cast for you
00:03:14
a vision of all the things that you can
00:03:16
become and and push you and try and make
00:03:19
you extend your vision for your business
00:03:21
and what is possible and through that
00:03:23
thought through that conversation
00:03:25
through the visioneering that he can
00:03:26
help you do as an entrepreneur sometimes
00:03:28
you need to listen and sometimes you
00:03:29
need to ignore the [ __ ] advice he
00:03:31
gives you but I will say that um the
00:03:34
role he plays is is really important and
00:03:36
I've seen it work across some of the
00:03:38
most important companies in Silicon
00:03:40
Valley in the world today so venot
00:03:41
thanks so much for being can I just
00:03:42
interrupt with one comment yes you know
00:03:45
advice to all the entrepreneurs most
00:03:47
we's want to be your
00:03:49
friend and so they are hypocritically
00:03:52
polite instead of brutally honest in a
00:03:55
way that can actually help you make a
00:03:57
better decision in in that's sort of
00:04:00
religion for me I prefer brutal honesty
00:04:03
to hypocritical politeness and when
00:04:05
people don't do that they are actually
00:04:07
harming the entrepreneur and having the
00:04:10
entrepreneur like them more that's not
00:04:12
my game and is that what you mean when
00:04:14
you say V most VCS add negative value
00:04:17
well they take a very shortterm
00:04:19
approach uh just going through a
00:04:23
situation companies being recapped they
00:04:26
asked us to participate I
00:04:28
said you are hanging on to revenue that
00:04:31
is bad Revenue if you get rid of it and
00:04:35
restart the company with positive
00:04:38
Revenue in the right technology approach
00:04:42
will partipate and even lead but they
00:04:45
hung on to revenue since the beginning
00:04:47
of this year by the way three years ago
00:04:49
I asked them to take a technology
00:04:51
approach as opposed to a people approach
00:04:53
to the business they did not because
00:04:56
Revenue came fast the technology took
00:04:58
time to develop
00:05:00
and they didn't invest in it now they're
00:05:02
going back and I just got an email today
00:05:05
saying they are happy to take our
00:05:08
approach of really investing in the
00:05:10
technology lower burn rate not try and
00:05:12
show higher Revenue so they can be
00:05:15
acquired instead build a company yeah
00:05:17
that's fundamentally the kind of thing I
00:05:19
run into constantly so I I want to start
00:05:23
our conversation with AI um I'm going to
00:05:25
kick this off you invested in open AI
00:05:28
you've been in inting in AI technology
00:05:30
as people are framing it today for some
00:05:32
time can you share with us a little bit
00:05:35
about how you became an investor in open
00:05:37
AI in the context of the AI trajectory
00:05:40
that you've been seeing for some time
00:05:42
and also maybe give us a little color
00:05:44
because we haven't had this question
00:05:44
answered on our show about how open AI
00:05:47
went from nonprofit to for-profit and
00:05:48
what the future is for open AI well let
00:05:51
me tell you a more General story first
00:05:53
starting with the
00:05:55
internet in 1982 when we started Sun we
00:05:58
bet on on the
00:06:00
Internet only company to bet on the
00:06:03
internet 1996 since you brought up
00:06:06
Juniper the internet was growing but not
00:06:10
yet clear the the Senior Management of
00:06:14
every major TCO in the United States
00:06:16
told me they would never adapt tcpip on
00:06:20
for the internet and I said fine will
00:06:23
build Juniper to build tcpip Cisco told
00:06:27
me they would never do tcpip above what
00:06:30
is probably your home home service today
00:06:33
oc2
00:06:35
never um and so we built
00:06:37
tpip why because we've seen the internet
00:06:41
on an exponential and the flat part of
00:06:44
the exponential looked very
00:06:46
uninteresting and nobody believed in it
00:06:49
we invest in Juniper it was a
00:06:52
2500x return for us at Kleiner one of
00:06:56
the largest returns in Venture at least
00:06:58
when uh Wall Street journ J did a piece
00:07:00
on the snap IPO of largest Venture
00:07:03
returns ever but we believed in
00:07:06
something and did that how it relates to
00:07:09
open AI probably year 2000 I gave a
00:07:15
interview with the New York Times it's
00:07:18
my son just sent it to me so I remember
00:07:20
it I said AI will be powerful enough we
00:07:24
don't know
00:07:25
when but it'll we'll have to redefine
00:07:30
what it means to be
00:07:32
human and I understand Steven Wolfram
00:07:35
had some interesting comments in that
00:07:37
area so happy to talk about 10 years ago
00:07:41
I wrote two blogs the first one was
00:07:43
called do we need doctors with a premise
00:07:46
that AI would replace the expertise of
00:07:49
doctors and let them do the things
00:07:52
humans do really well so 20% doctor 80%
00:07:56
AI is the thing I defined um and I wrote
00:08:00
another blog called do we need teachers
00:08:03
because the only way to scale education
00:08:06
at an equal level for every kid was with
00:08:08
AI that was 20 January of 2012 so about
00:08:13
12 years ago so four or five five years
00:08:16
ago late 2018 is when I started talking
00:08:19
to Sam the opportunity came up to do
00:08:22
something with open a eye I said this is
00:08:26
a clear B I couldn't predict when things
00:08:28
will Emer
00:08:30
how much capability will happen how fast
00:08:32
but you didn't know need to know that
00:08:35
whether it was in 3 years 5 years or 10
00:08:38
years I was pretty convinced it would be
00:08:41
pretty phenomenal and there'd be
00:08:44
practical applications along the way
00:08:46
where it would have a large impact long
00:08:48
before we got to AGI so we placed the
00:08:51
back and we placed a long-term back you
00:08:53
know this audience is going to hear from
00:08:56
Bob mguard you know 5 years ago I played
00:08:59
was a bet on Fusion for the same reason
00:09:03
if we get it to work these are humongous
00:09:06
returns big bats but reasonably high
00:09:09
probability of success initially but
00:09:11
very high payback and I was so convinced
00:09:14
in open AI I placed twice uh our initial
00:09:19
investment was twice the largest
00:09:22
previous investment I'd made in 40 years
00:09:25
in Venture Capital 2x the largest bet
00:09:28
I'd ever placed initially in a company
00:09:31
how did that go from nonprofit to
00:09:32
for-profit and raise Venture money well
00:09:36
you know those are
00:09:39
details we don't understand but I would
00:09:41
say the foll figure those out there are
00:09:44
plenty of
00:09:46
nonprofits in
00:09:48
Europe that have for-profit arms right
00:09:52
right it's a very good model to sustain
00:09:56
and generate profit for a nonprofit to
00:09:58
scale like I think the European model
00:10:01
some of these examples are really really
00:10:04
great Ikea Bosch they have very large
00:10:07
nonprofit ownership and I think that's a
00:10:09
very good model to scale I mean I just
00:10:12
mentioned do we need doctors my son's
00:10:14
building a primary care doctor AI my
00:10:18
wife's doing a primary care or tutor
00:10:21
AI as a nonprofit it really doesn't
00:10:25
matter how you do it the question is how
00:10:27
can you gather the resour resources to
00:10:29
scale the social good you're tring to do
00:10:32
whether with a doctor or with a tutor
00:10:35
for every kid on the planet there's a
00:10:37
lot of discussion about um how jobs
00:10:41
might be displaced because of AI you've
00:10:43
seen I guess three or four of these
00:10:46
amazing um efficiency uh Cycles what's
00:10:50
your take on that because this one does
00:10:53
feel qualitatively different uh and it
00:10:55
does seem to be moving at a faster pace
00:10:57
so the speed at which AI is moving Ming
00:11:00
uh and just the the Leaps and Bounds of
00:11:02
how it's making people augmented you're
00:11:04
saying 80% AI 20% doctor um what do you
00:11:07
think this does to jobs and do you have
00:11:09
those concerns of displacement that a
00:11:11
lot of people seem to be concerned about
00:11:13
I I think it's a genuine concern okay um
00:11:17
the path to getting to that point so let
00:11:21
me start with that 2025 years from now I
00:11:24
think the need to work will
00:11:27
disappear people will work because they
00:11:29
want to work not because they need to
00:11:31
work and nobody wants a job doing the
00:11:35
same thing for 30 years on an assembly
00:11:37
line at GM that there's a set of jobs
00:11:40
nobody wants to
00:11:42
do that we shouldn't really call jobs
00:11:45
they're
00:11:47
absolute treasury yeah so hopefully
00:11:52
those disappear I love my job so 25
00:11:54
years from now Health permitting I'll
00:11:57
still be doing the same thing like
00:11:59
I'm 68 so uh you know and Warren
00:12:02
Buffett's gone a lot longer than I have
00:12:06
I'm very passionate about what I do I
00:12:08
enjoy it so much I do it for free and
00:12:10
I'll keep doing it and other people
00:12:13
will there will be new jobs um what you
00:12:17
call jobs or what people get paid for so
00:12:20
is X Games popent a job I I don't know I
00:12:24
don't call it a job but it is a passion
00:12:26
that can
00:12:27
pay um so I do think the nature of jobs
00:12:31
would will disappear now societies will
00:12:34
have the the transition from today to
00:12:38
that vision of not needing to work if
00:12:40
you don't want
00:12:42
to will be very messy very political
00:12:47
very uh I would say controversial and I
00:12:51
think some
00:12:53
countries will move much faster than
00:12:57
others I I suspect let's take something
00:13:01
I'm very hawkish on China they will take
00:13:04
a tanaman square tactic to enforcing
00:13:08
what they think is right for their
00:13:09
society over a 25e period we will have a
00:13:13
political process which I prefer to the
00:13:15
Chinese style you like the debate over
00:13:19
the tanks well debate always improves
00:13:21
things but it also does slow down and
00:13:24
certain societies may
00:13:26
choose not to adapt these
00:13:29
uh and I think the societies that do
00:13:31
that will fall behind and I think that's
00:13:33
a
00:13:35
massive uh disadvantage for our
00:13:38
political system we may choose not to do
00:13:40
it we've done it before but we've also
00:13:44
seen massive transition before
00:13:46
agriculture was 50% of us employment in
00:13:49
1900 by 1970 it was 4% so most jobs in
00:13:55
that area the biggest area of employment
00:13:57
got displaced we did manage the
00:14:00
transition well cost went down and
00:14:01
productivity went up yeah so I do think
00:14:06
we will have to rethink so the one thing
00:14:08
I would say
00:14:10
is capitalism is by permission of
00:14:13
democracy and it can be revoked that
00:14:16
permission and I
00:14:18
worry
00:14:21
that the this messy transition to the
00:14:24
vision I'm very excited about may cause
00:14:27
us to to change systems I do think the
00:14:31
nature of capitalism will be will have
00:14:33
to change I'm a huge fan of capital
00:14:35
capitalism I'm a capitalist but I do
00:14:38
think the set of things you optimize for
00:14:41
when economic efficiency isn't the only
00:14:45
criteria will have to be increased so
00:14:49
disparity is one that will be increased
00:14:51
I first wrote a piece of 2014 in fores
00:14:56
it was a long piece that say AI will
00:14:59
cause great abundance great GDP growth
00:15:02
great productivity growth all the things
00:15:05
economists measure and increasing income
00:15:07
disparity that was 2014 and I think I
00:15:11
ended that to close this answer off with
00:15:14
we will have to consider Ubi in some
00:15:18
form and I do think that will be a
00:15:21
solution but meaning for human beings
00:15:24
will be the largest problem we face
00:15:26
which was my followup which is if you
00:15:27
give people Ubi do you have
00:15:30
concerns of you know the idle mind and
00:15:33
what happens we've seen in other states
00:15:36
when 25% of young men maybe don't have
00:15:39
jobs their minds wander and bad things
00:15:41
happen and so what do you think the
00:15:44
societal issues will be if we do offer
00:15:48
Ubi um I think we will have issues but
00:15:51
we will eliminate other issues you know
00:15:55
there's pros and cons to everything it's
00:15:58
weird very hard to predict how complex
00:16:00
systems
00:16:02
behave and it's very hard for me to sit
00:16:04
here and say I have a clear vision of
00:16:06
how that happens but I for one am
00:16:10
addicted to learning I know other people
00:16:13
addicted to producing music uh all of
00:16:16
those are reasonable things for people
00:16:19
to find meaning in or personal
00:16:22
meaning yeah there's others like I said
00:16:25
just want to perfect their participation
00:16:27
in in a certain sport or I love X Games
00:16:31
because wasn't even remotely considered
00:16:34
a sport not that long ago uh so things
00:16:38
passions will emerge and I think if
00:16:40
people have very early in their life the
00:16:43
ability to pursue passions I hope that
00:16:45
adds the meaning human beings will need
00:16:48
and in some sense that's what I was
00:16:50
talking about in 2000 when I said we'll
00:16:53
have to redefine what it means to be
00:16:56
human your assembly line job won't
00:17:00
Define you for your whole
00:17:02
life that'd be a terrible thing yeah in
00:17:05
the VK in the AI landscape just going
00:17:08
back to that for a second um you have a
00:17:10
lot of these big companies who seem to
00:17:12
have been caught flat-footed so they're
00:17:14
acting pretty aggressively I think even
00:17:15
yesterday there was a a leak of
00:17:17
something where meta trying to really
00:17:19
superpower llama 2 and then create some
00:17:22
high performance Cloud that they
00:17:23
effectively want to give away um we
00:17:25
talked about this a little bit on the
00:17:27
Pod but the idea seem seems like the big
00:17:29
companies want to Scorch the Earth on on
00:17:31
the foundational model set then you have
00:17:33
some other big companies that are
00:17:34
working on Silicon so can you just walk
00:17:37
us through in your mind how you've
00:17:38
organized where the where the value is
00:17:40
getting created where there'll be
00:17:42
product value but no economic value etc
00:17:44
etc yeah look uh going back to openai
00:17:48
part of my interest in funding openai
00:17:51
was when we looked at it when I looked
00:17:54
at it there was two major centers of
00:17:56
excellence in AI there was Google gole
00:17:59
they had Deep Mind and Google brain but
00:18:01
Google and then there was an effort at
00:18:04
Biden and it was very worried the
00:18:07
Chinese might win that uh I'm hawkish on
00:18:11
China I have no trust there um and I
00:18:15
loved Graham Allison's book uh destined
00:18:17
for war I'm a complete believer in that
00:18:20
thesis
00:18:22
um but I thought it'd be valuable to
00:18:26
create a third Center and that was part
00:18:28
of the motivation in funding open eye
00:18:32
I'm really glad meta is doing it and
00:18:35
others are trying to do it and you know
00:18:37
there's efforts in every country every
00:18:40
country wants their National AI I think
00:18:42
more competition especially in the set
00:18:45
of countries subscribing to Western
00:18:48
values is good and I think that will
00:18:52
sort of but I would say the following
00:18:55
almost all the
00:18:56
efforts I have seen are very limited
00:19:00
they're all trying to scale
00:19:02
llms and you know I spend my time
00:19:06
looking at what besides llm will
00:19:09
play an important role I haven't seen
00:19:12
one effort among the majors that isn't
00:19:16
following the same model yeah more
00:19:18
Nvidia gpus scale the parameters more
00:19:22
data um instead of saying are there
00:19:26
other methods so so I'll mention you
00:19:29
know we made an investment in a symbolic
00:19:31
logic idea would that be dramatically
00:19:34
additive in in you know just like open a
00:19:37
ey didn't look that interesting for 5
00:19:39
years ago 2018 late 18 is when we really
00:19:44
uh made our decision to invest it got
00:19:47
took a while to get
00:19:48
organized but I'm saying what else would
00:19:52
look like that in 2028 yeah you know
00:19:56
would it be symbolic logic All In in
00:19:58
anybody doing that probalistic
00:20:01
programming is it completely otherwis
00:20:05
like multi-agent systems is there are
00:20:09
going to be other axes in which we will
00:20:12
see progress said differently do you
00:20:13
need to make sure there are other axes
00:20:15
so that we we're not looking back a few
00:20:17
years from now and there's just massive
00:20:19
Cuda lock in and nothing can happen
00:20:21
without one gatee keeper is that well
00:20:23
the great thing about Venture and
00:20:25
Venture capist is they like to play long
00:20:27
bats yeah so I'll place all those long
00:20:30
bats on alternative ways to learn yeah
00:20:33
and not replace but add to llms I think
00:20:36
is the more likely scenario though who
00:20:38
knows yeah um but even among just if
00:20:44
imagine llms do it
00:20:46
all it's good that there's multiple
00:20:49
plers and so more big companies pop up
00:20:52
great uh I also do think governments
00:20:55
will have their own efforts China of
00:20:58
course has one but plenty of other
00:21:01
Rumblings of governments trying to put
00:21:03
together an effort in my country tell us
00:21:07
the and that's all good tell us
00:21:09
upleveling away from just AI but broadly
00:21:11
speaking Venture give us the State of
00:21:14
the Union of venture capital in
00:21:15
September
00:21:17
23 so let me give you a perspective I've
00:21:20
been doing Venture for some 40
00:21:24
years in that time I have not seen
00:21:29
and people get surprised at this one
00:21:31
example of a large
00:21:33
Innovation that came from a large
00:21:35
company or
00:21:37
institution so the only thing you can do
00:21:41
not
00:21:43
one the the the only two examples I
00:21:47
could find was in the early 70s which
00:21:49
was if you go back 50 years when I was
00:21:52
Bank of America put debit and credit on
00:21:56
a plastic card like that wasn't even a
00:21:58
major Innovation but I'll give them that
00:22:01
but I couldn't think of examples like
00:22:03
that apple with the iPhone maybe um that
00:22:07
was Steve Jobs I think it was a Founder
00:22:09
Le I should have said founder Le
00:22:11
Innovation okay got it is a more
00:22:13
accurate description I like that you
00:22:16
know did General Motors do driver L cars
00:22:19
or
00:22:20
weo did
00:22:22
uh would somebody at Avis do Uber would
00:22:28
somebody at Hilton do Airbnb right would
00:22:33
would somebody and Elon speaking next
00:22:35
but somebody at blacked Martin do SpaceX
00:22:39
or rocket lab we investors in rocket lab
00:22:42
so I like it a little more
00:22:45
um they do have a superior strategy but
00:22:48
SpaceX has done a great job uh but
00:22:54
think think of every large social change
00:22:59
is there a large company that play you
00:23:02
know the only thing somebody came up
00:23:04
with was debt structur debt structuring
00:23:07
that the bankers did uh and I
00:23:10
said they didn't take any risk right
00:23:13
right no ination they took a risk with
00:23:15
somebody else's money other people's
00:23:17
money easy to play with whether you're
00:23:19
uh in finance or in government uh other
00:23:22
people's money is easy uh but my idea is
00:23:26
innovation comes from group like this
00:23:29
and I'm completely convinced I lost the
00:23:32
original question but this is where
00:23:34
Innovation will come
00:23:35
from on that yeah so any large area
00:23:39
you'll be hearing from Bob mumar you
00:23:41
know four five years ago invested in
00:23:43
Fusion now we may fail I think we are
00:23:47
much more likely to succeed sitting here
00:23:49
today than fail but it's possible can
00:23:52
you talk about that just while you're on
00:23:54
Fusion that there we've got both Bob and
00:23:57
David from here helon tomorrow morning
00:23:59
yeah and there's broadly call it six or
00:24:02
some number of General architectural
00:24:04
approaches to solving the yeah there's
00:24:06
six approaches a dozen companies I think
00:24:09
it's well now there's 70 but yeah it's I
00:24:11
mean it's just ballooning but how do you
00:24:13
think about the role that a commonwealth
00:24:16
or even an open AI I guess how big of a
00:24:19
role does Capital play in getting there
00:24:22
because what made what makes
00:24:23
Commonwealth different is how much
00:24:25
Capital they've raised Billion Dollar
00:24:26
Plus at this point open AI similar the
00:24:31
um the decision as an investor how much
00:24:34
of it is predicated on this team or this
00:24:36
architecture let me tell you the
00:24:38
Commonwealth story before Bob has a
00:24:40
chance to tell you is he here by the way
00:24:41
to
00:24:44
myad I'm sure he's here somewhere
00:24:46
because I'm meeting him in half an hour
00:24:48
okay
00:24:50
backstage U but here's the thing they
00:24:54
may have raised a lot of money it took
00:24:57
double double digigit Millions to handle
00:25:00
the single biggest risk in building the
00:25:03
architecture right which was can you
00:25:05
build a 20 Tesla magnet yeah the magnet
00:25:08
right and Bob and I had this discussion
00:25:11
I
00:25:12
said if Fusion doesn't work but you can
00:25:15
build a 20 Tesla magnet we got lots of
00:25:18
interesting applications in nuclear
00:25:20
medicine yeah you know a fusion reactor
00:25:23
is a particle accelerator that technical
00:25:25
breakthrough we can pivot somewhere else
00:25:27
yeah nuclear Med MRI machines lots of
00:25:29
areas so let's go build with double
00:25:32
digit Millions which is not a huge
00:25:34
amount in the context of the upside of
00:25:38
maybe a bigger upside than the 2500x
00:25:41
Juniper had if you solve the fusion
00:25:43
problem there's still a 10x upside if
00:25:46
you just solve the MRI machine problem
00:25:48
so you're thinking of probabilities well
00:25:51
in today we not but but when we stored
00:25:54
it that was clearly to look at off RPS
00:25:57
right like where else could we go if we
00:26:00
only had double digit Millions we didn't
00:26:02
we built the magnet if we didn't build
00:26:04
the magnet we failed that's rent of
00:26:06
capital so how do you know that that bet
00:26:08
that investment at this point is still a
00:26:11
good investment or you keep putting more
00:26:12
capital in if there's now six competing
00:26:15
architectures any one of which could
00:26:16
have their own breakthrough that could
00:26:17
yield better economics so we've
00:26:20
continued to invest in it mostly because
00:26:23
I think Bob's just a phenomenal
00:26:25
CEO and if any body can make this happen
00:26:29
and the world needs it to happen Bob
00:26:32
can but I'm also optimistic and as you
00:26:35
know we invested in Sam too he invested
00:26:37
in
00:26:38
helon I think there's more than one
00:26:40
approach that could be successful in
00:26:42
fact i' guess 15 years from now we'll be
00:26:45
looking at one more than one company
00:26:47
that's successful Sak you had a question
00:26:49
I wanted to make sure you got to ask it
00:26:50
well let's let V finish that that point
00:26:53
I thought it was finished yeah anything
00:26:54
else so my view is large problems like
00:26:59
that yeah you'd think G would solve not
00:27:02
a chance not a chance right seens not a
00:27:05
chance um you know you take any of these
00:27:09
one of my favorite you take large
00:27:11
problems one of my biggest beefs is the
00:27:14
traffic in cities I actually think we're
00:27:17
very much on track to replace most cars
00:27:20
in most cities in 25 years how by
00:27:24
building a different kind of public
00:27:26
transit system anybody want to invest in
00:27:28
it I'm really excited about it um you
00:27:32
know we went the way of autonomous
00:27:36
cars for consumers or taxis aut
00:27:41
autonomous cars in public transit will
00:27:46
increase through the only thing fixed in
00:27:49
cities is Lane wdth you have a road
00:27:51
certain size you don't destroy buildings
00:27:53
on both sides so if you can increase
00:27:57
passenger per hour through it and by the
00:27:59
way bicycles don't we love bicycles but
00:28:02
they don't scooters
00:28:05
don't public transit autonomously driven
00:28:08
in small parts will make public transit
00:28:12
on demand so unscheduled on demand any
00:28:15
time of the day or night uh and point to
00:28:19
point you don't stop for anybody else to
00:28:21
get up it's a beautiful way to do change
00:28:24
cities and I think it'll happen there's
00:28:27
no question about it so like uber pool
00:28:29
but with a larger Mini Bus type format
00:28:32
no no no you don't want to go to mini
00:28:33
bus because then you have to stop for
00:28:35
other people to get out so just if it's
00:28:38
two or four person Vehicles it's faster
00:28:41
than your private car it doesn't stop at
00:28:44
traffic lights it's on demand you don't
00:28:46
have to park it I could go on and on uh
00:28:50
I don't want to spend too much time on
00:28:51
one particular idea but all I'm saying
00:28:54
is whether it's Fusion or public transit
00:28:57
and we're doing 3D printed buildings uh
00:29:01
another one of my favorites is a
00:29:03
new and you were asking about this a
00:29:06
music model to produce music that is not
00:29:09
trained on any YouTube or public music
00:29:12
yeah so completely IP free of any
00:29:16
constraints like I'm really excited
00:29:18
about little outfit in Australia and
00:29:22
they've just um produced a released a
00:29:25
product that would be the mid Journey
00:29:27
for
00:29:28
music really exciting what's your
00:29:31
favorite genre of music what do you like
00:29:33
to I I'm not a huge music fan to be
00:29:35
honest I Adit it but you know I I love a
00:29:39
great challenge when this entrepreneur
00:29:41
came to me and said hey in some period
00:29:44
of time say 10 years and this was four
00:29:47
or five years ago and long before AI was
00:29:50
a big thing he said I want to produce a
00:29:53
music a song a top 10 music song
00:29:57
Untouched by a human I said I love a
00:30:00
challenge like that let's go it
00:30:02
literally took me half an hour to say I
00:30:04
in no diligence didn't need all that
00:30:09
David yeah I think where tth was going
00:30:11
to go a minute ago about the state of
00:30:13
the union and VC is that we talk a lot
00:30:16
on the pot about the fact that we just
00:30:17
been through an asset bubble that popped
00:30:19
uh last year yeah and it you know really
00:30:22
inflated during coid with all the money
00:30:24
printing that went on and it may have
00:30:25
been inflating before that with these
00:30:27
Zer policies going back I don't know 15
00:30:30
years um I guess my question to you is
00:30:33
how much fakeness do you think that
00:30:35
created in the ecosystem um you know or
00:30:39
is there any way to even quantify it I
00:30:41
mean I think we see now that the size of
00:30:43
venture funds is Contracting that new
00:30:46
funds are going to be smaller probably
00:30:48
there's a lot of VCS who shouldn't even
00:30:49
be in business I think you'll be happy
00:30:52
about that um you know we've estimated
00:30:54
on the Pod there's 1400 unicorns right
00:30:56
now we've tried to sort ort of get all
00:30:57
are half of them fake um I guess my
00:31:01
question to you is how much did this
00:31:03
asset bubble sort of inflate everything
00:31:05
in VC and where do you think it stands
00:31:08
right now you know it inflated things a
00:31:11
lot in certain areas and it deflated
00:31:14
things a lot in certain areas but let me
00:31:17
give you two analogies to understand
00:31:19
this you know investors only and there's
00:31:21
a fair number of investors I hear here
00:31:23
only care about two emotions fear and
00:31:25
greed and they bounce between the two
00:31:29
and there's nothing in the middle and so
00:31:31
the key to being a good investor is
00:31:34
staying focused on the long term in the
00:31:36
middle so anytime you get a hype cycle
00:31:40
you'll get inflation and then you will
00:31:43
get deflation but let me ask you the
00:31:45
following question 1998 there was a do
00:31:49
com bubble or
00:31:51
bust tell me and there was clearly a
00:31:54
change in Wall Street prices and value
00:31:57
valuations and all that there was
00:31:59
absolutely no bust in the amount of
00:32:02
internet
00:32:03
traffic so if you look at internet
00:32:06
traffic you can't tell when the dot bust
00:32:09
happened so I'm saying the reality of
00:32:12
using the internet didn't change just
00:32:14
because people had inflated values and
00:32:16
then deflated them that's just a fear
00:32:19
reaction and a grade reaction same thing
00:32:21
happened with the original bubble in the
00:32:23
1830s in England if you got the right to
00:32:26
build the railroad between two cities
00:32:29
you could offer scripts on the market
00:32:31
big bubble then a big collapse but more
00:32:35
railroad was built in the 10 years after
00:32:38
the bubble collapsed in England in the
00:32:41
1830s than before my point is the
00:32:44
reality of the underlying business if
00:32:46
you're adding value does not change that
00:32:49
much it can accelerate a little bit or
00:32:51
slow down because less resources
00:32:53
available but fundamentals is a good way
00:32:57
way to invest sort of the Warren Buffet
00:32:59
method not the shortterm hey he's
00:33:03
valuing it twice so let me pay up and
00:33:06
then you'll pay twice as much as that or
00:33:08
soft M will uh later uh
00:33:13
um I think I think it's a question of
00:33:16
you can't time the market you as long as
00:33:19
you know that this is a an investable
00:33:21
technology an investable Trend timing
00:33:24
what price you're coming in at I mean
00:33:26
this is an old ad in in Market you know
00:33:28
it doesn't really matter what price
00:33:30
you're coming at it you know in
00:33:31
inflationary times you pay a little bit
00:33:34
more but you get less dilution later
00:33:38
because other people are investing and
00:33:40
then when it gets uh goes into a
00:33:42
deflationary cycle you just have to
00:33:44
learn to respond quickly and good
00:33:47
investors will help guide entrepreneurs
00:33:50
correctly through all those phases we've
00:33:52
seen a lot of that in climate investing
00:33:54
yeah for sure you know very low
00:33:57
valuations look when we invested in
00:33:59
impossible Foods it was a $3 million
00:34:02
valuation pre money crazy idea crazy
00:34:05
idea nobody thought plant proteins
00:34:08
wasn't a word um another one of my
00:34:12
favorite projects is replace soy as a
00:34:14
protein Source on this planet and I
00:34:16
think we are well on our way to do that
00:34:18
right I'm very excited but at that level
00:34:22
it doesn't matter whether valuation in
00:34:25
the end goes up or down by a factor of
00:34:27
two it doesn't really matter if you're
00:34:30
adding fundamental value and long-term
00:34:32
things but I just want I know that your
00:34:34
long-term view is critical to making
00:34:37
Silicon Valley work I think that we all
00:34:39
kind of appreciate your leadership your
00:34:42
support your mentorship and um and the
00:34:44
fact that you're willing to make those
00:34:47
big bets in very low probability
00:34:49
outcomes but if they work the return is
00:34:51
so much more um than that that that
00:34:54
multiple um that ratio that you're
00:34:56
getting and it matters so much I I just
00:34:58
want to add it's just inspiring how much
00:35:00
you share with all of us who are behind
00:35:02
you and how much we've all learned from
00:35:04
you and you've I didn't mention it but
00:35:06
you've take a lot of time for jamath
00:35:07
you've taken time for me in my career
00:35:09
and I I just think it's like a great
00:35:11
gift we end I can end with a quick story
00:35:13
I when I when IED we really do have to
00:35:15
jump I emailed I I mailed paper mailed
00:35:19
been trying 40 or 50 people to try to
00:35:21
get a job in VC in 2000 the only single
00:35:24
person sent me an handwritten rejection
00:35:27
letter VK
00:35:29
hey give it up thank you
00:35:32
everybody [ __ ]
00:35:34
yeah another stand wow thank you thank
00:35:38
you sir apprciate great
00:35:41
[Music]
00:35:49
great let your winners
00:35:54
[Music]
00:35:55
ride
00:35:57
and instead we open sources to the fans
00:36:00
and they've just gone crazy with
00:36:01
[Music]
00:36:09
it
00:36:12
Bes my dog taking
00:36:18
driveway
00:36:20
man we should all just get a room and
00:36:22
just have one big huge orgy cuz they're
00:36:24
all this useless it's like this like
00:36:25
sexual tension we just need to release
00:36:28
[Music]
00:36:33
some we need to get
00:36:36
[Music]
00:36:47
mer
