00:00:00
making good progress honest on the on
00:00:03
the ship and the booster codename VFR
00:00:09
what does it stand for again well it's
00:00:13
like sort of a Rorschach
00:00:14
yes an acronym form it is very big and I
00:00:25
gave a presentation on this at the
00:00:27
International Astronautical Congress in
00:00:29
Australia last year design is evolving
00:00:32
rapidly we're actually building that
00:00:34
that should but that ship right now I
00:00:37
think right now that the biggest thing
00:00:39
that would be helpful is just barrel
00:00:41
support and encouragement I think once
00:00:43
we build it there will be well it will
00:00:47
have a a point of proof something that
00:00:50
other companies and countries can then
00:00:55
go and do then I think they will fill up
00:00:58
their game and they will build
00:01:00
interplanetary transport vehicles as
00:01:03
well there's a means of getting I'll go
00:01:06
and people turn from Mars as well as
00:01:08
turn from the moon other places in the
00:01:11
solar system that's really where that
00:01:14
there's a tremendous amount of
00:01:15
entrepreneurial entrepreneurial
00:01:17
resources that are needed and it's gonna
00:01:18
be harder and a lot harder in a place
00:01:21
like Mars or the moon I mean the moon of
00:01:24
Mars often thought of is like is this
00:01:26
some escape a escape hatch for rich
00:01:28
people but I it won't be that at all
00:01:30
it's and I think there's not many people
00:01:33
who actually want to go in the beginning
00:01:34
because all those things I said are true
00:01:36
but there'll be some who will for whom
00:01:39
the excitement of the frontier and
00:01:40
exploration exceeds the incentive danger
00:01:43
like marsh should really have great bars
00:01:47
the Mars bar right we are building the
00:01:54
first ship the first Mars or
00:01:59
interplanetary ship right now and I
00:02:02
think we'll be able to do short flights
00:02:05
short sort of up and down flights only
00:02:08
sometime in the first half of next year
00:02:09
this is a very big booster and ship the
00:02:14
lift or thrust of this would be about
00:02:16
twice that of a seven-five its it's
00:02:18
capable of doing 150 metric tons to to
00:02:22
orbit in and be fully reusable that we
00:02:25
can reduce the cost marginal cost per
00:02:27
flight dramatically by orders of
00:02:30
magnitude compared to where it is today
00:02:31
this question of reusability is so
00:02:33
fundamental to rocketry it is the it is
00:02:36
the fundamental fundamental breakthrough
00:02:37
that's needed that you can lease a 747
00:02:41
and do a return flight from Kevin
00:02:44
Bullock ah go from California to
00:02:46
Australia for half a million dollars
00:02:49
that sort of cost to Lisa 747 fully
00:02:52
round trip to Australia which is far to
00:02:55
buy a single-engine turboprop plane a
00:02:57
good one would know it would be about
00:03:00
one and a half million dollars and it's
00:03:02
tiny
00:03:02
Oh 747 so what that means is like a it
00:03:07
cost less to to use a giant plane with a
00:03:10
huge cargo for a long trip in it then
00:03:12
that costs way less than buying a small
00:03:15
plane or a short trip in the aircraft
00:03:18
world and the same actually is true of
00:03:20
rocketry ibf our flight will actually
00:03:22
cost less then then our Falcon one
00:03:25
flight bit back in the day um so that
00:03:28
was about a five or six million dollar
00:03:29
marginal cost per flight we're confident
00:03:32
that VFR won't be less than that is what
00:03:33
will enable the creation of a permanent
00:03:37
base on the moon any city on Mars and
00:03:41
that's the quality of like the Union
00:03:42
Pacific Railroad until you can get there
00:03:44
there's no way for all of the
00:03:46
entrepreneurial energy to do you can't
00:03:50
you can't do anything so if all the
00:03:52
flowers to bloom once you can get there
00:03:54
the opportunity is a mint and and then I
00:03:58
think it'll be it'll be amazing
00:04:00
the
00:04:00
all of this was to inspire you and make
00:04:03
you believe a game just as people
00:04:06
believed in the Apollo era that
00:04:07
anything's possible thank you at SpaceX
00:04:10
almost all my time is spent on
00:04:12
engineering and design it's probably 80
00:04:15
and 90% in order to make the right
00:04:17
decisions you have to understand
00:04:20
something if you don't understand
00:04:21
something you had a detailed level you
00:04:22
cannot make a decision in the US auto
00:04:24
industry the only two companies that
00:04:26
have gone bankrupt at least at some
00:04:29
point are Tesla and Ford every other
00:04:32
company but bankrupt or was failing and
00:04:35
got acquired there's only two companies
00:04:38
that haven't gone bankrupt and there's a
00:04:40
big graveyard of companies that did I
00:04:41
gave basically both SpaceX and Tesla
00:04:43
from the beginning a probability of less
00:04:46
than 10% of likely likely to succeed I
00:04:49
just kept wondering why we were not
00:04:51
making progress towards sending people
00:04:54
to Mars why we didn't have a base on the
00:04:58
moon just just wasn't happening year
00:05:00
after year yeah I was getting me down
00:05:02
and I look at the NASA website I was
00:05:05
like just where does it say when we go
00:05:08
to Mars
00:05:09
isn't that the genesis of SpaceX was not
00:05:11
to create a company but but really had
00:05:13
it how do we get NASA's budget to be
00:05:15
bigger I was initially to go and as I
00:05:17
cut more and more into what it would
00:05:19
take to do that
00:05:19
I learned that the fundamental issue is
00:05:22
actually the cost of access to space and
00:05:25
there's a gigantic difference between
00:05:27
the raw material cost of the rocket and
00:05:30
the finish cost the rocket so there must
00:05:31
be something wrong happening in going
00:05:35
from the constituent atoms to the final
00:05:38
shape and then unfortunately the the
00:05:40
Space Shuttle ended up being a counter
00:05:41
example of don't don't try to make
00:05:43
reusability work because spatial data
00:05:45
and enough credit costing more per
00:05:47
flight than an expendable vehicle of
00:05:48
equivalent capability there's no
00:05:50
question in my mind that if you could
00:05:51
reuse the Rockets it has to be true
00:05:55
reuse which means rapid and complete
00:05:57
reuse big orange tank which was also the
00:06:00
primary airframe was discarded every
00:06:02
time and the parts that were used were
00:06:04
incredibly difficult to refurbish if
00:06:06
it's not the only thing you're changing
00:06:07
between flights
00:06:08
farfel scheduled maintenance is the
00:06:11
and in the beginning I wouldn't actually
00:06:13
when I'm elected because I don't want to
00:06:15
lose their money I thought it was like
00:06:16
budgeted four or three flights I mean
00:06:20
technically I did have a plan where I
00:06:22
had to have the money from PayPal I had
00:06:24
like 180 million from PayPal that should
00:06:26
be fine I'll have 90 million like so
00:06:28
slots you know but yeah 2008 we had the
00:06:32
third consecutive failure of the Falcon
00:06:35
rocket for SpaceX I tried very hard to
00:06:37
get the right expertise in for for
00:06:40
SpaceX I tried hard to to find a great
00:06:43
chief engineer for the rocket I don't
00:06:45
really have a business plan of SpaceX I
00:06:50
think Tesla's actually been probably
00:06:53
two-thirds of my tone a problem on those
00:06:55
over time practices drama magnet it's
00:07:00
crazy I think probably one the biggest
00:07:03
of a search tenet is that I'm actually
00:07:05
not an investor I'm so much people think
00:07:07
I'm weird best ROI invested things I
00:07:08
don't actually dirty missing anything
00:07:09
back the only public security that I
00:07:12
would have any kind is Tesla I wanted to
00:07:14
start the tunnel from where I could see
00:07:18
it from my office at SpaceX it so I said
00:07:20
well let's just call before part of the
00:07:21
parking lot across the road now we're
00:07:23
making good progress and we're finding
00:07:28
the company from merchandise sales just
00:07:30
have fun with the boring company but my
00:07:33
time allocation is about literally about
00:07:36
two percent we are all much smarter than
00:07:38
we think we are much less smart dumber
00:07:41
than we think you are they defined
00:07:42
themselves by their intelligence and
00:07:44
they they don't like the idea that a
00:07:47
machine can be way smarter than them so
00:07:48
they discount the idea which is
00:07:51
fundamentally flawed
00:07:52
that's the wishful thinking that a
00:07:53
situation I'm really quite close to I'm
00:07:56
very close to the cutting edge in AI and
00:07:59
it scares the hell out of me you can see
00:08:03
this in things like alphago which went
00:08:05
from in the span of maybe six to nine
00:08:09
months it went from being unable to beat
00:08:10
even a reasonably good go player so then
00:08:12
beating the European world champion who
00:08:14
was ranked 600 then beating Lisa doll
00:08:17
for five then beating the current world
00:08:19
champion then beating everyone while
00:08:21
playing simultaneously
00:08:23
then then those alpha 0 and alpha zero
00:08:26
just learnt by playing itself and it can
00:08:30
play basically any game that you put the
00:08:32
rules in for if you whatever rules you
00:08:34
give it literally read the rules play
00:08:36
the game any superhuman nobody expected
00:08:38
that great of improvement to gas those
00:08:39
so those same experts who think AI is
00:08:43
not progressing at the rate that I'm
00:08:45
saying I think you'll find that their
00:08:46
predictions that we'll see this also
00:08:48
with with self-driving I I think
00:08:54
probably by intermixture self-driving
00:08:58
will be well encompass essentially all
00:09:01
modes of driving and the Natsuko study
00:09:05
on on Tesla's autopilot version one
00:09:07
which is relatively primitive and found
00:09:10
that it was a forty five percent
00:09:11
reduction in highway accidents the
00:09:14
advent of digital super intelligence is
00:09:16
one which is semiotics with humanity I
00:09:19
think that's the single biggest
00:09:21
existential crisis that we face and the
00:09:23
most pressing one I'm not normally an
00:09:26
advocate of regulation and oversight I
00:09:29
mean I think it's once you generally or
00:09:31
on the side of minimizing those things
00:09:32
but this is a case where you have a very
00:09:35
serious danger to the public a danger of
00:09:37
AI is much greater than the danger of
00:09:40
nuclear warheads by a lot and nobody
00:09:44
would suggest that we allow anyone to
00:09:47
just build nuclear warheads if they want
00:09:49
that would be insane and mark my words
00:09:52
AI is far more dangerous than nukes far
00:09:57
so why do we have no regulatory
00:09:59
oversight this is insane
00:10:01
I'm not really all that worried about
00:10:03
the short-term stuff things that are
00:10:05
other but like narrow AI is not a
00:10:08
species level risk but it is not a
00:10:10
fundamental species level risk whereas
00:10:14
digital super intelligence is so it's
00:10:18
really all about laying the groundwork
00:10:20
to make sure that and obviously the
00:10:22
things that I believe in like extending
00:10:24
life beyond Earth making life
00:10:25
multiplanetary and I'm a big believer in
00:10:28
sort of Asimov's Foundation series or
00:10:32
the principle that you
00:10:33
you really want to we want to make sure
00:10:36
that there's enough of us of a seed of
00:10:38
human civilization somewhere else to
00:10:42
bring civilization back and perhaps
00:10:46
shorten the length of the dark ages the
00:10:49
Mars base might survive is more likely
00:10:51
to survive than a moon base but I think
00:10:53
a moon base and a Mars base that that
00:10:58
could perhaps help regenerate life back
00:11:00
here on earth it would be really
00:11:02
important and to get that done before a
00:11:04
possible world war 3 again I'm not
00:11:06
predicting this seems like well if you
00:11:09
say given enough time will it be most
00:11:11
likely given enough time this is this is
00:11:14
has been our pattern in the past
00:11:16
so sustainable energy is also obviously
00:11:20
really important as tautological if it's
00:11:22
not sustainable its unsustainable yeah
00:11:25
how close are we to solving that problem
00:11:26
well I think that the core technologies
00:11:28
are are there with wind solar the
00:11:32
fundamental problem is that there's an
00:11:34
unprecedented the cost of of co2 you
00:11:39
start off with a low price but then that
00:11:43
price and then depending upon whether
00:11:45
that price has any effect on the past
00:11:47
million a possibility of co2 in the
00:11:49
atmosphere you can adjust that price up
00:11:51
or down so it's really up to people and
00:11:54
governments to put to put a price on on
00:11:57
carbon and and then automatically the
00:12:00
right thing happens it's really
00:12:02
straightforward more carbon in the
00:12:04
atmosphere is actually good but up to a
00:12:08
point well right now the only things
00:12:10
that are really stressing me out in a
00:12:12
big way or AI obviously that's like
00:12:14
where's there and and working really
00:12:18
hard on Tesla Model 3 production and
00:12:22
we're making good progress but it's
00:12:24
usually hard work but those are the two
00:12:26
most stressful things my life right now
00:12:28
but we do want to close coupling between
00:12:31
collective human intelligence and
00:12:33
digital intelligence and a newer link is
00:12:39
trying to help in that regard by
00:12:42
creating a an interface between
00:12:46
the high bandwidth interface between AI
00:12:48
and your human brain so that there's
00:12:51
going to be essentially vast numbers of
00:12:54
tiny electrodes that are able to read
00:12:58
right from your brain of course you know
00:13:00
security is pretty important in the
00:13:01
situation say the least I was they say
00:13:05
I'm not coming with I'm keeping my brain
00:13:06
air-gapped yeah
00:13:08
well I think a lot of people will choose
00:13:10
to do that that is an AI the AI
00:13:13
extension of you a tertiary layer of
00:13:15
intelligence so you've got your limbic
00:13:18
system your cortex and and the tertiary
00:13:21
layer which is the digital AI extension
00:13:23
of you and that high bandwidth
00:13:24
connection is what achieves the tights
00:13:27
and meiosis I I think that's the best
00:13:30
outcome I we don't talk that much about
00:13:32
StarLink but essentially it's intended
00:13:35
to provide low latency high bandwidth
00:13:38
Internet connectivity throughout the
00:13:39
world that actually will not be enough
00:13:42
cognitive processing power onboard the
00:13:45
satellite system to to in any way be a
00:13:48
Skynet thing like it's the Starling
00:13:52
system will be important in providing
00:13:54
the funding necessary for SpaceX to
00:13:56
develop interplanetary spacecraft and at
00:14:01
the same time yeah helping people who
00:14:04
have even no or super expensive
00:14:06
connectivity and giving people in urban
00:14:08
areas more of a competitive choice the
00:14:11
form of government on Mars would be
00:14:12
somewhat of a direct democracy where you
00:14:16
vote on issues where were people vote
00:14:18
directly on issues instead of going
00:14:20
through representative government in in
00:14:22
you know when the United States was
00:14:24
formed representative government was the
00:14:26
only thing that was logistically
00:14:27
feasible because people there's no way
00:14:29
it was all your people to communicate
00:14:31
instantly a lot of people's doesn't even
00:14:33
have really access to but I think I'm
00:14:36
washed most likely it's gonna be people
00:14:38
at everyone votes on every issue and
00:14:41
that's how it goes
00:14:42
I've a few things I'd recommend which is
00:14:45
keep blowing
00:14:47
long Louis it's like that's that's
00:14:50
something suspicious is going on if
00:14:52
there's long mole you know if if you if
00:14:54
the size of low exceeds the word count
00:14:57
of Lord of the Rings
00:14:58
signs well put the eggs on hold and do
00:15:04
something to help out the internet or do
00:15:07
something useful on the Internet
00:15:09
and that's when I talked to Kimball and
00:15:12
you're working in Canada at the time and
00:15:17
said hey why don't we try to do this
00:15:18
this company and Silicon Valley I think
00:15:21
that's that's where decide you really
00:15:22
want to go to the end consumer if you've
00:15:24
got great technology you want to go all
00:15:26
the way to the end consumer don't tell
00:15:28
this to some bonehead legacy company
00:15:31
that doesn't understand how to use it
00:15:33
either
00:15:34
solar electric car or space
00:15:37
I thought space was like the least
00:15:40
likely to have somebody at least likely
00:15:44
to attract entrepreneurial times I'd
00:15:47
look like like nobody is gonna be crazy
00:15:49
enough to do space so I better do space
00:15:52
so I started off with with space first
00:15:56
how old have done something called Rosen
00:15:58
Motors which was like an attempt to DV
00:16:00
startup I said well I've always been
00:16:02
super interested in electric vehicles I
00:16:03
was gonna do my PhD on an advanced and
00:16:07
reduced energy storage I was gonna do
00:16:09
grad studies on on advanced energy
00:16:12
storage techniques for electric vehicles
00:16:14
and and so JB said well have you heard
00:16:18
of this company called AC propulsion
00:16:20
because they had created it the T zero
00:16:23
electric sports car as a prototype
00:16:26
lithium-ion batteries had really
00:16:27
achieved a level of energy density that
00:16:31
for the first time could allow you to
00:16:33
have significant range in an electric
00:16:36
car and they had a sports car that had
00:16:39
zero to 60 in under four seconds at 250
00:16:42
mile range so I got the test drive from
00:16:44
AC repulsion and I was like wow you guys
00:16:47
should really commercialize this this
00:16:49
would show people what electric cars can
00:16:50
do and I tried for months to get AC
00:16:54
propulsion to go into production with
00:16:57
the t zero but the problem is like the
00:16:59
electric saundra could cost $70,000 or
00:17:03
you could build a sports car for
00:17:06
$100,000 okay but like nobody's gonna
00:17:08
buy the electric silent
00:17:10
if you're gonna try to park ties t0
00:17:12
there's some other teams you should talk
00:17:13
to that also interested in doing that so
00:17:16
that's where why never heart Mach
00:17:19
topping and Ian Wright came in and no I
00:17:25
think that was probably the biggest
00:17:26
mistake of my career quite frankly I'll
00:17:30
dedicate 20 percent of my time to Tesla
00:17:32
and that'll be fine but actually it
00:17:37
didn't things really melted down went
00:17:42
through hell where to recapitalize the
00:17:43
company the Kimmel was there sing in
00:17:45
real time but the truth is stranger than
00:17:49
fiction all the crazy stuff you see in
00:17:51
that just look valley
00:17:52
the reality is way crazier than that
00:17:57
[Applause]
