- 6 hours ago
Europe's security future will be won or lost by the companies it builds today. From AI and autonomy to industrial transformation, what does it takes to build sovereign, software-defined capabilities for Europe's defence?
Category
🤖
TechTranscript
00:20good afternoon it's nice to see you all and we're looking forward to this discussion i think modi
00:26really sets the stage in very interesting ways and if we really take a step back and looking at
00:34what's happening in the world we really have a very interesting tension going on on the one hand
00:39we see a revertant to i'd say hard power a you know time we thought we'd overcome a world that
00:48is very much defined by who is the strongest and who do we feel most credible in the way that
00:55they
00:55that they can deter and then on the other side you have a phenomenon of civilizational acceleration
01:02which is ai and so that's kind of a tension that i feel like it's very interesting and it's a
01:07good
01:08key up to our discussion that we're now about to have with with antoine because helzing is a company
01:14that very much sits at the intersection of both and so maybe throwing the ball to you on one why
01:20don't
01:20you walk us a little bit through what helzing actually does and is yeah thank you uh hello
01:25everyone uh thank you janet uh glad to be here to be back at vivatech so helzing was created five
01:32years
01:32ago uh exactly with this premise but at the time this was a premise the premise was like
01:36sovereignty and defense for europe who are going to be very important for the future of the continent
01:42maybe at the time where not everybody was realizing that you were one of the first to realize that
01:46um but also at the time like europe the the second part of this question will be that a new
01:52type of
01:53technology is going to come is going to change the game and this technology was ai was already software
01:58as well software and ai and now it's even more robotics which we didn't even know five years ago
02:03so helzing was created there and now five years later what are we doing well we're doing autonomous
02:09systems uh systems uh for defense which means that we try to build robotic system flying and some of
02:16them are going to sea that can carry out effect and help us protect uh our continent by the usage
02:23of
02:24software and autonomy stack which been made by ai that allows the system to have bigger effect more
02:29autonomy be resistant to jamming be able to be deployed on the most complex situation um for for now and
02:37so
02:37if you have two keywords that you need to remember from this talk about helsing it's autonomous mass
02:42right this is really how we're thinking about it so we are in an era where you can produce much
02:48more
02:48type of hardware okay much more type of hardware namely drones but not only which are not very
02:55uh expensive to build but of course they become effective when you have software inside that brings
03:00them autonomy function so that's why autonomy and mass are really the two-parting that are driving
03:05defense today and of course helsing yeah and i think people actually underestimate what the
03:10complexity is and on one i mean you've been one of the most prolific researchers or are one of the
03:16most prolific researchers in the ai domain you led the meta lab together with yan le koon um on ai
03:22and so
03:22you are one of the pioneers of the domain and if you now think of what you just described to
03:27us
03:27walk us a little bit to the complexity of what it takes to actually build
03:31a software and the ai defined hardware product which is very different to a hardware the way you
03:37just slam some software on top yeah yeah so the so you need to realize that uh when you work
03:44in defense
03:45of course you work on very very critical topics and you work in very difficult conditions and um
03:51so that's that's one hand on the second hand when you're building ai systems what you care about first
03:56and foremost is data and compute of course the talent as well but it's mostly data and compute
04:01in the defense field data is very very hard to build very hard to get very very hard to acquire
04:07it's either classified it's segregated in some places or in the case of drones data doesn't even exist
04:13anywhere because the drones have never been deployed at a scale at which you can look up and actually get
04:18the data from so you need from day zero almost to think of your data strategy pretty early on
04:23with different types of assets the second aspect when you build robotic systems is that
04:29that you should not forget that you need to have a production muscle very early on as well
04:33because if you can build one or two drones first off it's not really the mass producible asset that
04:39i was talking about but also you won't have enough to be able to test them at scale and to
04:44gather the
04:44data so something that we've done at lc quite early on is that think about mass manufacturing
04:48all devices even at the time where we're still doing the iterations and what it gives you is that
04:55at the point where you start to have the good idea you can produce hundreds of them test hundreds of
05:00them
05:00get the data from them and then you start this very virtuous circle and i think that's something that
05:06the advice that i give as well if you think about going into physical ai you need to think in
05:11terms of how
05:11can i acquire data fast and it usually goes through scaling up the hardware and the last lesson on that
05:17it's very hard to collect data for a specific hardware for a specific drone or a specific quadruped or
05:23whatever using another platform so you say oh i can buy some drones of the shelf collect data from them
05:29i'm going to use this data put it on my device it makes the problem way harder when you can
05:34actually
05:34collect the data with your own platform you're actually way better so this complexity are really
05:40coherent to the fact about physical ai systems of course when you're deploying uh on server etc you
05:46don't have the same kind of constraint and usually that's a big price to pay when you move in this
05:50range yeah and so housing is actually the only company that is really battlefield proven the only
05:55company that you would argue is a modern neo prime which basically defines the next category of of a
06:01defense prime and so you know you've been live in ukraine and there's also some learnings um from from
06:08from other regions i would love to understand um when you think about the modern deterrence stack
06:14right like drones fit in the picture software fits into the picture but maybe walk all of us through
06:19how you think through reconnaissance where that starts and how that actually leads to um to activity in the field
06:25yeah so i think i mean ukraine of course uh if you're following is is unfolding very quickly um i
06:34think the
06:34and the ukrainian are have i mean have a lot of qualities but i think there are two that i
06:39think are
06:39very important for us first the bravery right they've been standing up for themselves alone first and now we
06:46support since the beginning but also they've been very innovative and very fast to adopt innovation
06:52very quickly and sometimes we forget about this right now we talk about the drones the the drones
06:57the drone war that's coming up okay four years ago none of these existed right none of these existed
07:03there were no drones over there and of course now it seems obvious that oh yeah of course if you
07:07have
07:07drones you can do that but the way that it came to friction is that the ukrainians were able to
07:13change the
07:14way they procure systems so say oh drones is interesting how can i buy them quickly to put them to
07:19my forces which is
07:20something that if you go to western europe we are very very bad at doing we have much slower so
07:25they
07:26were able to adapt new system quickly to to purchase them and two they were also very good at training
07:31their troops to use the new system very quickly as well which is also something that sometimes we don't
07:36realize if you bring a lot of new assets to the field to the front line and your people know
07:42all to use
07:42the the older hardware it doesn't work and they've been very good at doing that so the lessons for us
07:48when
07:48you you came out of you came out of ukraine of course there are the technical ones how you can
07:52resist to jamming and to autonomy maybe we'll get to that but for me the meta lessons as well is
07:57that
07:57as a society how you are able to actually bring technology and bring innovation at the core of
08:03your critical processes how the government are able to purchase to test to try new system quickly so that
08:09they can actually follow the innovation scale and on your hand also can you upskill and train people
08:14quickly so that they can use that i think this is the ukrainian are very good at that but i
08:19think
08:19this lesson is actually true for all of us and i know that you janet uh you've been updated your
08:24your sparta paper and i think these topics also from your eye as a uh investor as well has been
08:29very
08:30key for some of your vision and so maybe you also have some lessons from ukraine i would be curious
08:34to
08:34hear about them yeah so i actually started my own from la familia 10 years ago very naive not having
08:40any clue on venture capital but my mission was always how do we make sure europe doesn't fall
08:44behind in technology and so started out and the initial goal was let's bring established industry
08:49players startups together and then over time realized that you know you needed governments to really lean
08:55and especially in hello i think we're losing the we lost the mic
09:05janet are you back i don't know i'm back you're back great um so if we sort of look at
09:11the way you
09:13need to orchestrate an ecosystem now it becomes a lot more complex and so you do need to think through
09:19a systems design approach on how companies like helsing and others can become successful in
09:26breaking through different layers of inertia and different layers also of just lack of comprehension
09:32on how to actually grapple with the technology i think one thing we have to acknowledge is ai is
09:36moving incredibly fast physical autonomy systems are moving incredibly fast in their development and
09:42governments are very oriented towards you know 70 80 years of like a an existing and ongoing
09:49procurement system that hasn't hasn't changed and so what we've put together is sparta which is
09:54essentially a um paper really that lays out the critical capability gaps that europe needs to um
10:00needs to go after in order to really achieve strategic autonomy and what we mean by that is
10:06when you look back in in into january and you have the remember the whole greenland debate
10:12then you really understand that europe has a very unique value system that needs to get protected and i
10:17think it's on us to actually protect our values we cannot rely on others to do that because others
10:22just have different opinions and that's okay but in order for us to become effective and building
10:26a credible deterrent stack we need to ensure that companies like like helsing and others
10:34are getting procurement pull from the government customer and so sparta is an attempt of educating and
10:40understanding what are actually the strategic capability gaps because it's actually not it wasn't clear to
10:45everyone and then thinking through what are the mechanisms and how much capital is actually needed
10:50to get that done and the number is actually quite low it says we came up with a 500 billion
10:56number over
10:56the next 10 years it needs to be invested for europe to become fully strategically autonomous which isn't a
11:01lot and like it could actually be covered by the german defense spending alone another question is how do you
11:06shift the mindset away from you know buying more tanks and more vessels and more like heavy systems that
11:13essentially as we can all now see in ukraine and then and in iran it takes only one drone to
11:18come
11:18through and actually hit hit a given target for that target to become invalid and destroyed and so
11:24that sort of mass advantage is something that we are now trying to translate into something that
11:29governments can work with which is something that we as a venture capital firm actually really
11:35understand is our job is it's to kind of meet insanely powerful founders like like antoine and his team
11:41where they are and then really try to orchestrate these what we call insanely powerful advantages like
11:46how can we how can we create these linkages that's something we we cared about from the beginning
11:52excellent um so you're german we are i'm french we are we are friends uh i'm a strong believer in
12:00the in the
12:01fact that europe to actually uh succeed as well and you mentioned the 500 billion etc needs to have
12:07uh a very strong partnership between france and germany we're in paris germany is actually
12:12the owner the guest of honor of vivatec helsing has been created in germany with a strong pool in
12:16france so we try to to do what we can how do you think we can do even better so
12:21that we can actually
12:21together uh can really try to drag as much as possible this vision yeah so i think the best example
12:27would be airbus right it's it's it's one of the the the sort of really prominent examples of french
12:33german cooperation and if you now zoom forward you would argue your partnership with mistral is
12:39actually a great example where that actually that spirits get reignited at the at the young company
12:45stage i do think that um europe as a whole still acts way too nationalistic i think germany france
12:52there's actually there's two examples but literally every country in europe would much
12:56rather buy american technology than buy from each other so this is something i think we need to
13:00change um because the problem is we're not going to find and we're not going to fund 10 helsings in
13:06europe right there's going to be one housing and there's going to be a couple of other um a couple
13:09of other companies in different domains but we need more concentrated capital and concentrated
13:14procurement pool on capabilities because it doesn't make sense to just fragment the landscape further so
13:19i think this is something we're paying a lot of attention to and i think the ship is moving in
13:23the right direction um but i think there's definitely more more legway to go but i think
13:28speaking of our friendship and the and the spirit in which housing was built i think one thing that i
13:34find extraordinary and this is kind of you know i think you joining the the company was a big signal
13:40for us to really double down in major ways onto the company at the series b walk us a little
13:45bit
13:45through your journey on one because you were you know you had all doors open like what made you
13:49decide to quit your job um which was very prolific and you and many others on the team and actually
13:55come back to to to build for european sovereignty walk us a little bit through your logic and what are
14:00you seeing and in the in your ability to attract talent like you have an enormously high high talent
14:06pool at housing yeah so yeah i mean it's a it's a good question i feel that so before joining
14:12uh housing uh
14:14i've been working at at fair i'm in the meta ai laboratory for almost 10 years uh i joined it
14:20almost at the beginning in 2013 where basically ai was still something that people thought was
14:25an idea from the 70s that would never work but when jan look and called me and say oh i'm
14:31starting this
14:31stuff said do you want to join oh that looks interesting and so i said 10 years at the end
14:36i was
14:36actually leading the the whole lab um and the latest project that i saw was the small project called llama
14:42uh llama one um and uh and so it's true that the direction was clear but when the helsing founder
14:50called me a bit like cold call right it it was not really planned like that um they told me
14:56two things
14:57that i think uh resonated a lot with me and still do and resonate that i use as well and
15:01resonate with
15:02a lot of people are joining um it's one don't you think that in europe we can also prove that
15:09we can
15:09build technology that is at the level of technological dominance on the world scale defense or not
15:16right it could be healthcare it could be automotive uh it could be space it could be foundation models
15:21you know there was this idea of maybe pushing a little bit my guys say hey you've been working for
15:26an american company great but don't you think in europe we can also uh build something that is of
15:31that ambition i said oh that's uh you know a bit uh are you up for the challenge a little
15:36bit right
15:36um and you know when i talk to arthur or the missile founder there's always a bit this idea as
15:41well in
15:41terms of okay can we show them also what we've got a little bit so that was clearly one trigger
15:46and the
15:46second trigger was of course they didn't join any company it was a defense company in the context where
15:52the ukraine full-scale invasion had just started okay it was uh it was at the time and so they
15:57had the idea
15:58well they quickly showed me the data say okay arthur look at this and this and that and look at
16:03basically the state of software and ai and it was even before robotics uh in the military compared to
16:09what you know or what you have and i saw such a gap such a gap uh that i was
16:14like okay we need to do
16:15something because of course a lot of the signals were there that ai and and data data processing at the
16:22edge etc was going to to be more and more important we didn't know exactly how much more important
16:27and i looked at basically the gap between what was deployed and what i knew from what was done
16:32at meta or at google or at open ai at the time was like okay we need to do something
16:37to try to close
16:38the gap as much as we can because this is just going to get wider okay and this is before
16:43the
16:43the arrival of claude this is before the arrival of all these models that you see now say holy holy
16:49crap
16:49we need to have um we need to have a real solution there so i would say this is the
16:53two is one
16:53one i saw something that needs to be fixed uh and when they came to say hey if you don't
16:58do it who
16:59will right and then say okay let's let's try to do this together and to your question about when i
17:05try to bring people uh and i hire a lot you know when i joined helsing we have 100 people
17:09now we are
17:10almost like uh 1500 right so this is way well a lot of these people are here yeah a lot
17:16of people are
17:16coming from american company from abroad etc say yeah i want to be able to join a european um
17:23operation i want to do a re-opern project that has the ambition to be technology leading you know
17:28i don't want to join a project i just want to be the best in my courtyard right i don't
17:32want to
17:33join something that's going to be the best ai model in france or the best drones in germany i want
17:38to be something that is leading technology but of course that has a big purpose for the
17:42sovereignty of our continent 100 and i think like the what what is interesting is people aren't
17:50pausing enough in my view on how strategic and incredibly relevant it is to control your own ai
17:57systems that's why mistral matters that's why housing matters and if you think about what happened
18:02last week with fable you kind of get a taste on what it could look like if that if that
18:08was happening
18:08to us and so one of the key elements that that we are looking for as a firm is how
18:13do we activate
18:14more people like antoine and the housing folks to really kind of go and like go for ambitiously
18:20crazy big ideas and i think the thing that for example right now i'm pausing on and i'd love to
18:25get
18:25your your sense of that as well um antoine is if you really think agi is around the corner right
18:31and
18:31you would argue it's maybe one or two years out then you know what what does it actually unlock you
18:37know and i'm i'm currently thinking a lot about space as a domain and like what if we had abundance
18:42of rocket fuel for example you know you can just you can you can imagine abundance through ai and
18:46like what would that do to our capability to actually explore the universe and actually kind of move
18:52beyond things that um that that we all all know and understand and so that's just one one area but
18:59on one i mean you're thinking a lot through the application of ai the application of ai into um into
19:04physical systems and the sort of the frontier that that that can happen there walk us a little
19:09bit through what you're seeing more broadly yeah i mean this is this is uh this is an excellent
19:14question maybe this is the mother of all questions right now you know when it's there what is going
19:18to change i would argue that it's already changing a lot uh and even changing a lot like months by
19:24months and it's true you said something very very true and you've been also backing a lot of company
19:28doing that the dependency that we have with regard to these technologies is becoming greater every
19:33week right uh we saw that when we started dancing but now i don't think we were even foreseeing
19:38how important that would be today right i think uh uh and that's something that we try to contribute
19:44other companies and you're as well trying to to work in terms of bringing that but we need more
19:50of that we need more investment we need more talent we need more people i think we need more
19:54coordinations as well more consolidation uh because once again the train is not waiting for us
20:00and sometime when i talk to sorry this was not your question i wanted to make my point um sometime
20:05uh
20:06when i talk to politicians or decider say oh but Antoine well you know look at what we're investing
20:10it's way more than before you know we're getting faster and like it's true we're getting faster than we
20:16were so this is the glass glass alpha but if you look both sides usually like more on the on
20:22the east
20:23toward china if you go to the west uh toward the us they're not waiting for us so they're also
20:28investing way more than they were using before right so it's not like the acceleration is everywhere
20:33and so if you look at the rate of acceleration you know we're accelerating maybe not even as much as
20:38the other one so we always need to keep that in mind that yes if i do better than before
20:41but they're
20:42also doing through traces better so then to your question about the the agi and the application so um
20:49in helsing we're really about embedded ai models into physical systems it can be drones it can be
20:54it can be satellites uh as you said it can be also autonomous uh fighter jets uh that that we
21:00try to
21:00build and i think there there is something that we are seeing more and more and that i think is
21:05i think
21:05an opportunity for europe and i um i try to to to explain that that the when you go in
21:11a physical
21:11system that was my point earlier on this talk is that when you collect the data for one system usually
21:16it's hard to make it work for all the other systems as well so there's really a tight dependency
21:20between the system you're building and the data even the system you build you can use some components
21:24that were pre-trained but there's this application and then you get into something that's much more
21:29specialized much more distributed and there having a big um fable model that you have access on your
21:35api to do that will not make the cuts it can help you design that and i think we should
21:39have access
21:40to this but it will not make the cut and then how you can actually insert that how you build
21:44that
21:44this hardware and software and ai co-design so that you can build it at scale is something that
21:49i feel the game there is really not done it's really not done and then we still have a lot
21:55of
21:55assets we have the people but we also have an industrial base that's been pretty good and very
22:00good at being adapting etc that is also looking for purpose so i think we should also looking at that
22:05through this angle as well the physical ai system change a little bit the technological rules
22:09and there some of the recipe that work for big tech are not as easy to bring right google has
22:14been
22:14trained to do robotics for like 15 years they they never they never really break through right
22:19this is not really the way you do it they have elements but it's not there so there are i
22:23think
22:24elements for us and in europe we've been doing robotics for decades right a lot of the you talk a
22:31lot of godfather of ai but if you look at the godfather robotics a lot of them are actually europeans
22:35it's just like we maybe forgotten about this and now we try to bring to bring this back in so
22:40so
22:41that's that's uh that's also i think a call for hope a call for action as well and and i
22:46know that you
22:47you're looking also at these companies you're looking at these founders you try to unlock that also
22:50to bring also the industry all the big industrial partners together with these new founders to try
22:56to connect this tissue right so what do you think about this and do you think we have we have
23:00a shot
23:00of doing that yeah it's actually quite interesting when you look at china you would argue they really
23:05care about driving ai into physical systems you don't see them spending as much time talking about
23:11agi the us is clearly on the exponential on the race to to owning agi and being being at the
23:18frontier
23:18of developing the the software-based technology and then the question is what could what role can europe
23:24play and i think europe has a unique industrial base so one thing that i'm really excited
23:30about is thinking through the next level of value capture from ai within product development so
23:37think like the example of mistral and asml like asml needs to also you know innovate on their core
23:44product stack and so will many many other companies in europe think of semen think of many of these
23:49hidden world market leaders that we have in europe all of which own very sensitive data very unique
23:55data sets that can really sort of power a whole next generation of of sort of ai enabled systems so
24:01i think that is an area we're spending a lot of time and i think where europe has a unique
24:06has a unique edge
24:07the other thing we're actually seeing in europe as well is the whole sort of component stack and the
24:13ability to potentially lean into robotics more just because when you look about think about what is a
24:17robot right like essentially consists of a lot of many many different parts like motors and actuators and
24:23some of the most prominent companies and technologies for that stack actually sit in germany for example
24:29because they're all used to serve the automotive industry and i know that you know helsing you are
24:33spending a bunch of time like in the supplier base in in in europe for the same reason so i
24:38think there's
24:39an interesting unique mode that europe can actually shape in this in this next generation around physical
24:45systems and then i think the last thing i'd say is there's a very interesting i think shift in mindset
24:52within europe when it comes to how do we think about on-sharing services right like so for the
24:57longest time we have really taken us our we've taken our physical manufacturing to china we've taken our
25:02digital services to the us and then for for for like other other service industries we've kind of
25:07outsourced that the philippines to morocco so ai gives us now a unique capability to actually onshore
25:13services back back into europe which is essentially a really important productivity uplift that we that we so
25:18urgently need and that can actually um really serve our gdp growth going forward and that's actually
25:24something that i'm pausing on the most right now is thinking through how do we ensure that in this
25:29next you know of over these next two decades we don't erode our economy even further right if we
25:35think about the past 12 years we basically saw that you know the us has essentially increased the gap in
25:42gdp to europe by roughly 12 trillion and that is essentially the combined market capitalization of the big
25:47five right and these were only services if we're now not hyper careful and like leaning into these
25:53opportunities in europe then we will be seeing the largest involuntary transfer of wealth from europe
25:59and potentially other rest of world countries to the us um for for for for the reason of basically
26:05labor getting transferred into compute and then and then getting taxed in the us and not not in europe so
26:11i think this is just something really worth pausing on and something that we are really on a mission to
26:16mobilize um all stakeholders in europe to to understand that this is a strategic imperative and not a not an
26:22option really yeah is it so you've been with supporting housing for a long time since uh almost the
26:30beginning i mean i would argue we've enjoyed housing at the almost at the same time um from your
26:35role and mine is it were you already seeing just the same kind of picture at that time you know
26:40if you
26:41backtracked four four years ago you are you are seeing that nice and folding even further i think
26:46it was very visible so torsten who's one of the co-founders and co-seos of of of housing and
26:52i actually go
26:53back a pretty long time so we've actually been talking about this idea before before housing even
26:57started and i think the thing that resonated with me at the time was the fact that um you know
27:03you
27:03saw trump won you saw brexit on the horizon you saw russia and crimea was already was it was already
27:10happening and so there was just an increasing amount of geopolitical tension and there was literally
27:14nothing that europe put up against it um and you already saw the emergence of autonomous systems so
27:20i think that was just something that was ambiguous but i think at the early stages it's always about
27:23people do you feel like this person is actually able to do incredible things and do i trust him enough
27:29to
27:30um to go and achieve that and bring in amazing people like you and and and hire an incredible
27:35team and do that especially in this domain with the right ethical framework right because if you think
27:40about what this is like we are we're now talking about this in a way that feels rather casual if
27:46you
27:46spoke about this four years ago there was a lot of hesitation there was a lot of pushback i think
27:51nobody really um really wanted to wanted wanted to embrace defense technology as something that
27:57was actually an integral part of society um and i think now fast forward i don't think we often
28:01pause enough on the question of of what the right sort of ethical underpinning actually needs to be
28:07and i think for for us the most important lever there is to you know you have people like you
28:12people
28:12like like like tourists and your entire team you think about this a lot you have an ethics council
28:17maybe you want to talk about that yeah yeah uh i think this is this is very important that you
28:23mentioned that thank you janet uh i mentioned that we want to boost about mass and autonomy to build
28:28these new systems we do it because we really feel it has to be done it's important for our sovereignty
28:33we're not really happy to be doing that right we know that the systems actually can be dangerous
28:38they are carry actually so there's a lot of gravitas of the responsibility even within the team
28:43even before we engage with the stakeholders that are of course our government the french or the german
28:47government or ukrainian government so there is that in the team it's true that in the dna
28:51of the company from the founder torsted that talks even openly a little bit about this but i think
28:56a lot of our engineers they come from technology company they come from academia they don't come
29:01from defense for a lot of them this journey as well to come from actually i was working on something
29:06that's much more like normal to come to something that defense is very abnormal in a way uh was a
29:13journey as well and so the journey that's all the community done we've done it internally and we
29:17feel when you ask the the team right now and say okay uh how do you feel about the responsibility
29:22of what you're building they feel the responsibility for themselves even though they are not going to
29:26be the soldier using it they feel the responsibility of doing it with the right principle testing it
29:31correctly etc it's actually very very important because we know that this mission is important but it's
29:37it's a big one and it's also something that you don't want to take like for granted as well and
29:41we talk a lot about
29:42the man and the machine uh collaboration now the human in the loop uh testing for this kind of
29:48decision ai decision making this stuff for us are core also to the reflection we have we don't take that
29:54lightly say oh yeah we are going to be a builder tech and you can take faster decision whatever no
29:58no
29:58a big question that's in my head all the time is like how do we put the human in the
30:02best conditions
30:03with this system more and more complex so that they have all the elements to take the best decision
30:07at the best time i would say this is for us if you're asking me what's the biggest challenge
30:12for other things down the road i think this is this one yeah i think this is this one how
30:15you
30:15scale up the effect while you keep the human at the common with the understanding of the situation
30:20100 and you know dario from anthropics had it once and i think it's one of the most important lines
30:25ai
30:25diffuses at the rate of trust and that's the way capital flows to people like you it's the way you
30:30can
30:30build distribution it's the way you can um you know leverage and build the teams in in the best way
30:37so i
30:37think trust actually becomes the ultimate hard currency in this tension and this world we we
30:42just kind of looked at anyway it was really nice to see you all um more founders in the room
30:48please
30:48go for ambitious projects like helsing and reach out we're looking forward thank you so much
30:53thank you
30:58so
30:58you
Comments