WEBVTT

NOTE Sentence-level transcript of https://www.youtube.com/watch?v=2wgPHvW0mG8

NOTE One cue per sentence. Cue ids are the line anchors on /transcripts/2wgPHvW0mG8.html. A cue ends where the next begins, or 2 s after its last word.

s1
00:00:12.480 --> 00:00:13.599
Thanks everyone for coming.

s2
00:00:13.599 --> 00:00:23.840
So we'll be talking about how um to use Agentic sort of orchestration to command many many different drones to achieve the different tasks that we have ahead of us.

s3
00:00:23.840 --> 00:00:30.880
and just going to rethink this presentation a bit and not just jump straight into slides, but instead we're going to fly.

s4
00:00:30.880 --> 00:00:36.079
Um, so what we have here on the left hand side is a real time view.

s5
00:00:36.079 --> 00:00:40.239
What we're going to be doing right now, am I showing anything?

s6
00:00:41.600 --> 00:00:42.879
Well, I'm flying.

s7
00:00:42.879 --> 00:00:44.160
There you go.

s8
00:00:44.160 --> 00:00:44.719
Sweet.

s9
00:00:44.719 --> 00:00:45.120
Okay.

s10
00:00:45.120 --> 00:00:50.800
Um, we are somewhere here in the city down south, which where where our headquarters is.

s11
00:00:50.800 --> 00:00:55.360
Uh we have a few drones up and I'm going to hit launch.

s12
00:00:55.360 --> 00:00:59.120
Uh and these are drones that are in docked stations.

s13
00:00:59.120 --> 00:01:11.439
So we calling this drones as infrastructure uh where we actually have thousands of these drones now planted around the country with power utilities with uh public safety uh with construction companies.

s14
00:01:11.439 --> 00:01:16.400
And what I'm doing right now is using my keyboard to simply fly here.

s15
00:01:16.400 --> 00:01:19.439
Uh for those of you that know the area, this is San Mateo.

s16
00:01:19.439 --> 00:01:25.759
And perhaps if I kind of zoom in here, we may see a very faint render of the uh SF skyline.

s17
00:01:25.759 --> 00:01:28.799
Although the uh the fog city is always there.

s18
00:01:28.799 --> 00:01:33.759
So we might just say hello to um the uh SFO airport here.

s19
00:01:33.759 --> 00:01:34.799
Some planes are launching.

s20
00:01:34.799 --> 00:01:39.680
If anyone has a plane tracker app, they can sort of see this is uh realtime stuff there.

s21
00:01:39.680 --> 00:01:45.600
So what we have been building is the full autonomy stack behind like how does a vehicle operate autonomously?

s22
00:01:45.600 --> 00:01:47.360
How does the cloud system operate here?

s23
00:01:47.360 --> 00:01:49.040
How do the cloud servers work?

s24
00:01:49.040 --> 00:02:00.880
uh and where does the uh different levels of intelligence and automation needs to happen for us to make this happen robustly such that when I'm here and saying hey oh there's an incident happening here I need to go respond

s25
00:02:00.880 --> 00:02:10.800
I could at the same time go back here and say hey what about that other thing that's happening on the other side of the country what if we launch that instead so while that's happening

s26
00:02:10.800 --> 00:02:21.680
I'm now going to launch something in Colorado so this is a imagine some sort of fire incident has happen on your on the power lines and we wanted to go out and

s27
00:02:21.680 --> 00:02:22.959
inspect those things.

s28
00:02:22.959 --> 00:02:28.480
So, a dock is now opening up in Colorado while the first drone is still flying safely.

s29
00:02:28.480 --> 00:02:33.519
Um, and I'm just going to uh let it do all the safety checks that it needs to do before the launch.

s30
00:02:33.519 --> 00:02:36.400
And this is all happening in sort of conference Wi-Fi traffic.

s31
00:02:36.400 --> 00:02:42.480
So, you can imagine I can shut down my laptop right now and everything needs to safely happen behind the scenes.

s32
00:02:42.480 --> 00:02:47.599
Um, so I'm going to do this while the first part is still happening.

s33
00:02:47.599 --> 00:02:51.120
Let me let's go back to the first drone.

s34
00:02:52.800 --> 00:02:54.480
Might give it another set of instruction.

s35
00:02:54.480 --> 00:02:56.720
Let's have a look around here on the first drone.

s36
00:02:56.720 --> 00:03:04.800
The second drone has sort of kick started off and maybe there's some cars sort of going around there um that we can perhaps track.

s37
00:03:04.800 --> 00:03:09.200
So, let's track and what's this car doing here?

s38
00:03:09.200 --> 00:03:09.599
All right.

s39
00:03:09.599 --> 00:03:10.959
So, we're now tracking this car.

s40
00:03:10.959 --> 00:03:13.519
Maybe this is a runaway car that we needed to follow.

s41
00:03:13.519 --> 00:03:14.959
And I'm hands off right now.

s42
00:03:14.959 --> 00:03:22.159
This is uh the autonomous system kind of taking over um through the various interactions that I've given uh for it to do.

s43
00:03:22.159 --> 00:03:34.080
And at the same time while this is happening we can do the final thing which is yet another sort of drone in the system uh back at our HQ

s44
00:03:34.560 --> 00:03:38.400
and we can say hey let's why don't we run a third drone.

s45
00:03:38.400 --> 00:03:50.319
So what we're building towards is how do we enable autonomy at scale where traditionally how it started off historically 15 years ago that you might have a drone at home.

s46
00:03:50.319 --> 00:03:52.560
It's a hobbyist drone and you might play around with it.

s47
00:03:52.560 --> 00:03:53.440
You will tinker with it.

s48
00:03:53.440 --> 00:03:55.680
You will work with the controlling software.

s49
00:03:55.680 --> 00:04:01.760
And then about 10 years ago drones started to become a lot more available and they started to become a tool.

s50
00:04:01.760 --> 00:04:03.840
A lot of industries out there started to use them.

s51
00:04:03.840 --> 00:04:07.439
They will carry them with them in the truck, go out there, deploy it.

s52
00:04:07.439 --> 00:04:10.959
And the next era that we're working towards is drones as infrastructure.

s53
00:04:10.959 --> 00:04:12.239
Imagine these systems.

s54
00:04:12.239 --> 00:04:14.959
And for the sake of this conference, I'm going to call them agents.

s55
00:04:14.959 --> 00:04:27.120
These are physical emboded agents that are kind of available at all times at anywhere for the different sort of use cases that we're interested in that can automatically uh uh launch, execute, and do their tasks.

s56
00:04:27.120 --> 00:04:30.160
And what is the minimum amount of autonomy that needs to be begged in?

s57
00:04:30.160 --> 00:04:32.960
And what does the future interface look like today?

s58
00:04:32.960 --> 00:04:35.040
Everyone needs to be a dedicated pilot.

s59
00:04:35.040 --> 00:04:36.960
I go went through a certification exercise.

s60
00:04:36.960 --> 00:04:38.960
I need to think about the safety standards here.

s61
00:04:38.960 --> 00:04:47.600
But you can imagine in a few years time the safety is going to be determined by the autonomous system and the interface becomes really high level.

s62
00:04:47.600 --> 00:04:50.880
It could be a little slack bot that says, "Hey, something's happened here.

s63
00:04:50.880 --> 00:04:52.000
Why don't we go send a drone?"

s64
00:04:52.000 --> 00:04:55.280
And you might not even know that the drone launched.

s65
00:04:55.280 --> 00:05:01.600
So, while this is happening, I'm going to tell all the drones to pause and return to doc.

s66
00:05:03.600 --> 00:05:05.440
So, we'll write this.

s67
00:05:05.440 --> 00:05:06.880
They're all going to start returning to doc.

s68
00:05:06.880 --> 00:05:10.720
And while that's happening, I'm going to start on the presentation.

s69
00:05:14.320 --> 00:05:17.199
So, what I've shown you here is not just concept.

s70
00:05:17.199 --> 00:05:20.400
These are literally systems that are being used in production.

s71
00:05:20.400 --> 00:05:28.400
Uh as was mentioned we are the largest manufacturer of drones in the US and we want to give people superpowers uh through this technology.

s72
00:05:28.720 --> 00:05:31.120
Let's take a quick look at what some of the people are doing here.

s73
00:05:31.120 --> 00:05:42.000
This is um in the northeast coast of the US uh where our client has set up uh the system next to a power station and they kind of used this to do normal patrols and as they flew around

s74
00:05:42.000 --> 00:05:52.080
they found that some of the there was a the pole was burning from the inside and it was really starting to show up here and this could have fallen at any time and created a fire risk that they would have otherwise

s75
00:05:52.080 --> 00:05:57.919
not caught without having to send people there which itself is quite uh expensive.

s76
00:05:59.919 --> 00:06:03.120
Moving on to the sort of next use case is on the public safety side.

s77
00:06:03.120 --> 00:06:04.800
This is San Francisco.

s78
00:06:04.800 --> 00:06:07.120
Uh we work very closely with SFPD.

s79
00:06:07.120 --> 00:06:11.520
Um normally when a car gets stolen, you'll see a high-speed chase happening in the city.

s80
00:06:11.520 --> 00:06:12.560
Quite dangerous.

s81
00:06:12.560 --> 00:06:18.960
But what if you could deploy a drone instead where the uh people in the car don't even know that there's a drone following them.

s82
00:06:18.960 --> 00:06:26.240
Here's a person who has stolen the car on the right hand side and they're about to change their license plates uh on that.

s83
00:06:26.240 --> 00:06:29.280
So they go here, get out their tools.

s84
00:06:29.280 --> 00:06:34.240
They come back, luckily they point the license plate up so the drone can see it and we know exactly what they're doing.

s85
00:06:34.240 --> 00:06:38.479
But they're now replacing the plate uh in the car uh with a new one.

s86
00:06:38.479 --> 00:06:41.360
And all this time they don't know that there's they're being chased.

s87
00:06:41.360 --> 00:06:49.840
The the the how they behave in the uh public, how they behave out there is very different and it's a lot safer uh in how the operations are done.

s88
00:06:49.840 --> 00:06:52.639
They're now going to go ahead and tint the windows.

s89
00:06:52.639 --> 00:07:03.680
uh but this allows the police to like strategically position themselves in the most safest way form to intervene at the right time rather than doing a high-speed chase outside.

s90
00:07:04.639 --> 00:07:16.160
So these are being used across many different industries and we are really starting to treat this as infrastructure that can operate day in day out nighttime rain sunshine.

s91
00:07:16.160 --> 00:07:19.280
We've got a few of these docs deployed in Alaska, so very cold weathers.

s92
00:07:19.280 --> 00:07:22.560
Few of these docks deployed in uh Texas, so extreme heat.

s93
00:07:22.560 --> 00:07:29.840
And these need to be reliable down to 99.9999% uh where we do our sort of simulation and testing to be able to prove that.

s94
00:07:29.840 --> 00:07:45.440
And today, uh we have about 16 million people living within 2 mi radius of this infrastructure that the uh um public safety, the power companies can kind of use this technology to be able to respond to such incidents without having to travel there.

s95
00:07:45.440 --> 00:07:48.720
So, what really happens when this happens at scale?

s96
00:07:48.720 --> 00:07:52.000
Can I get a hands up of people that have flown a drone before?

s97
00:07:52.000 --> 00:07:52.880
A couple of hands up.

s98
00:07:52.880 --> 00:07:57.440
When I started to fly it, it was like kind It took me a few hours to like really figure it out.

s99
00:07:57.440 --> 00:08:04.560
Then I put on the FPV, that was even more tricky, but it felt good to get get that expertise up, but it's kind of like a skill that you develop

s100
00:08:04.560 --> 00:08:12.720
and you develop that skill over time and you say, "Okay, for each skilled person, we're going to put them next to a drone and they're going to start working."

s101
00:08:12.720 --> 00:08:14.479
But now you want more of these.

s102
00:08:14.479 --> 00:08:18.960
So, uh, police companies, infrastructure companies need to start hiring these people more.

s103
00:08:18.960 --> 00:08:20.879
And at some stage, this really starts to break down.

s104
00:08:20.879 --> 00:08:26.080
The more 911 calls come that come in, the more alerts that can come in, it doesn't really scale.

s105
00:08:26.080 --> 00:08:35.279
So, we kind of rethinking as to what this means in terms of this like initial firsterson view engagement with these systems to how do we convert this to a more strategic

s106
00:08:35.279 --> 00:08:42.719
uh, multi- aent view that you can kind of command the entire fleet with an objective in mind without having to worry about the flight.

s107
00:08:42.719 --> 00:08:44.880
Let me just go back and see if that was all working well.

s108
00:08:44.880 --> 00:08:45.120
Great.

s109
00:08:45.120 --> 00:08:46.160
They all landed.

s110
00:08:46.160 --> 00:08:48.800
I am still pleased when that happens successfully.

s111
00:08:48.800 --> 00:08:51.920
Although it's meant to happen all the time.

s112
00:08:51.920 --> 00:08:55.200
All right, let's go back to this.

s113
00:08:55.200 --> 00:08:58.320
Oh, all right.

s114
00:08:58.320 --> 00:09:00.959
We're just going to carry on.

s115
00:09:02.640 --> 00:09:05.440
So, what does that mean when we start to launch different things?

s116
00:09:05.440 --> 00:09:06.320
You saw me launch them.

s117
00:09:06.320 --> 00:09:08.080
I was still kind of thinking about it.

s118
00:09:08.080 --> 00:09:11.040
Okay, I need to launch this one, that one, that one.

s119
00:09:11.040 --> 00:09:21.200
And even myself in this like uh who's used to this, I'm going to have some cognitive challenges where our vision is to be able to launch uh many many of them uh in a potentially unsupervised

s120
00:09:21.200 --> 00:09:21.440
way.

s121
00:09:21.440 --> 00:09:23.040
So what sort of commands these things?

s122
00:09:23.040 --> 00:09:32.560
How do we get it to like just hey get out there, launch, search in this area, find a missing person or look out for this type of car and hold your position there.

s123
00:09:32.560 --> 00:09:41.839
And in order to do that, we really need to think about how do we get like really like the many nines of reliability that we need in autonomous flight.

s124
00:09:41.839 --> 00:09:52.640
And that's where we get an edge in the industry because at Sky we're sort of controlling the hardware, the software, the cloud, the user interface to be able to manage all that and specifically the autonomy.

s125
00:09:52.640 --> 00:10:02.720
uh allowing us to think about how our underlying vision system should work to see the environment to behave in the environment correctly whether it's at high altitudes whether it's cloudy

s126
00:10:02.720 --> 00:10:16.800
whether it's at high speeds how do we deal in the rain on the bottom left we're showing how do we navigate in cities how do we plan large scale and be able to do that and for anyone that has worked with any sort of GPS device in the city even our phones they kind of suck

s127
00:10:16.800 --> 00:10:28.240
uh so how do we robustly do that and how do we also like do tracking when there's a lot of occlusion These are the all the different places where we're thinking about how to train AI systems quote unquote models.

s128
00:10:28.240 --> 00:10:35.600
The word model itself is uh has different meanings in different places and I'll discuss a little bit on what that means for us.

s129
00:10:37.040 --> 00:10:41.839
But in order for us to you like really harness this is we are learning on the go.

s130
00:10:41.839 --> 00:10:50.480
This is a learning flywheel that we're getting out there, we're collecting data, we're operating, and we're coming back and doing that so that each flight we can log the data.

s131
00:10:50.480 --> 00:10:56.560
Uh kind of like Google Street View where we need to think about sanitizing that data, make sure there's no private information left there.

s132
00:10:56.560 --> 00:11:01.519
Uh and make sure we don't sort of like uh uh the customers know exactly what they're sharing with us.

s133
00:11:01.519 --> 00:11:08.959
But if we can once we do that, we have access to a huge amounts of data that we can kind of learn from every single time we instructed this but the drone did this.

s134
00:11:08.959 --> 00:11:09.600
every single time.

s135
00:11:09.600 --> 00:11:11.680
We thought this was going to happen, but this happened.

s136
00:11:11.680 --> 00:11:21.440
That can come back to our uh learning agents, our reinforcement learning ecosystems to be able to retrain, evaluate, and send it back out there and kind of continue that flywheel

s137
00:11:21.440 --> 00:11:25.519
that allows us to uh have that robust framework.

s138
00:11:25.519 --> 00:11:30.000
And the other added advantage that we have is it's not just about having autonomy on the drone.

s139
00:11:30.000 --> 00:11:37.279
As I mentioned earlier, we have the luxury now to have autonomy on the edge device, but also have autonomy on the cloud.

s140
00:11:37.279 --> 00:11:41.920
what I was showing you earlier, all that video feed, all the telemetry that's going through a cloud server.

s141
00:11:41.920 --> 00:11:57.839
We could set up GPUs and we could set up inference engines there to be able to have that heavier lifting, maybe that longerterm planning there, whereas the immediate autonomous actions happen on the drone and kind of always thinking about the trade-off that we need to make uh to make that successful.

s142
00:11:57.839 --> 00:12:06.959
One thing is true, however, that uh once you do start thinking about having your agents from the cloud, this the amount of data coming to the cloud really matters.

s143
00:12:06.959 --> 00:12:18.560
Uh here's us sort of investing in how we think about enabling the best um uh video quality coming up through to the servers uh in uh low lower low sort of bandwidth areas.

s144
00:12:18.560 --> 00:12:32.720
Uh being able to sort of optimize that being able to un uh um encode that information into smaller sizes and be able to decode into something clear that allows us to have much higher quality exact same sort of network conditions here.

s145
00:12:32.720 --> 00:12:41.600
uh and it's the some of the areas of investments that we make so that we can have this more cloud-based infrastructure uh to be able to manage this at scale.

s146
00:12:41.600 --> 00:12:53.839
So once we do that I want to sort of explore at a high level some of the uh models that we um uh have in our system uh that allows us to orchestrate all of this.

s147
00:12:53.920 --> 00:12:56.880
Firstly a model this is a section about world models.

s148
00:12:56.880 --> 00:12:59.680
I want to sort of talk about world models from a context of maps.

s149
00:12:59.680 --> 00:13:02.959
Not too dissimilar to how uh Whimo works.

s150
00:13:02.959 --> 00:13:05.600
They have a map of the world and they kind of navigate in that world.

s151
00:13:05.600 --> 00:13:08.480
They do local perception, but also think about global planning.

s152
00:13:08.480 --> 00:13:14.800
If I want to go from part A of the city to part B, I can't just like keep hitting every building and kind of navigating around them.

s153
00:13:14.800 --> 00:13:18.240
I need to think about what's the optimal path uh along the way.

s154
00:13:18.240 --> 00:13:20.959
So, we start off with a lot of prior information.

s155
00:13:20.959 --> 00:13:32.079
uh um and we merge that with uh not just sort of building data, but maybe there's uh vector data such as where the power lines are, where the roads are, if you want different behavior

s156
00:13:32.079 --> 00:13:32.959
in these areas.

s157
00:13:32.959 --> 00:13:39.200
And we think about how to combine these resources to ultimately build a map that we can plan and navigate around.

s158
00:13:39.200 --> 00:13:44.880
And the drone has knowledge of this map uh at all times to be able to go around that.

s159
00:13:45.040 --> 00:13:48.639
But like any map, maps can go out of date.

s160
00:13:48.639 --> 00:13:56.639
Uh luckily we have so many eyes in the sky to think about how to maintain and update these maps along the way.

s161
00:13:56.639 --> 00:14:05.920
On the left hand side uh we are rendering our knowledge of the world in the points onto our video feed.

s162
00:14:05.920 --> 00:14:08.320
However, that doesn't line up perfectly everywhere.

s163
00:14:08.320 --> 00:14:18.959
Uh there's some sections here that um if I sort of zoom out here, there's some sections here that a new construction site had set up uh that we did not know about.

s164
00:14:18.959 --> 00:14:27.519
Uh however as a drone now starts to fly and not this is not just one drone but your fleet of drone they're now observing these things that we can feed back into our

s165
00:14:27.519 --> 00:14:41.120
uh map syncing process that can come back land give the data and now once uh in the next iteration all the drones in the fleet have this most updated map of the world that they can uh do all the planning in.

s166
00:14:44.959 --> 00:15:02.000
A second type of model is uh perhaps today a more conventional sort of uh machine learning uh inference model uh which is the ability to uh be able to track understand objects in the scene but be able to track them be able to track them behind occlusions.

s167
00:15:02.000 --> 00:15:11.279
So the uh this implicit representation behind the scene is some sort of world representation of the object that hey it's gone behind this building and it might come out on the other side.

s168
00:15:11.279 --> 00:15:18.160
So I should navigate myself so I can kind of follow it there or I should move myself in a different direction to be able to do that.

s169
00:15:18.160 --> 00:15:21.920
Maybe 5 years ago this would be more done uh in a more conventional way.

s170
00:15:21.920 --> 00:15:24.000
You have to like manually think about how to move.

s171
00:15:24.000 --> 00:15:33.839
But now you can think about uh more reinforcement learning style techniques or more learned end to-end approaches that can really help out here without you having to engineer all the edge cases

s172
00:15:33.839 --> 00:15:35.040
uh that can go in.

s173
00:15:35.040 --> 00:15:43.839
And then obviously like how do we do this robustly rain, snow, day, night um uh using um vision only.

s174
00:15:44.880 --> 00:15:54.000
One other thing that we're sort of doing here on the tracking side is we have some minimal set of tracking that's happened on device on the edge but we can do some higher level tracking that happens on the cloud.

s175
00:15:54.000 --> 00:15:56.639
So perhaps it can reason more about your entire map.

s176
00:15:56.639 --> 00:15:59.440
Perhaps you can reason more about use heavier models.

s177
00:15:59.440 --> 00:16:10.800
Use VLMs uh uh with lower which have u which don't respond as quickly uh at a rate of like let's say 7 to 10 hertz but can give you feedback at a 1 to two second uh latency

s178
00:16:10.800 --> 00:16:16.560
but that's good enough for us to make uh broad decisions about um where to move.

s179
00:16:16.800 --> 00:16:22.399
For a lot of our um infrastructure customers uh we're doing a lot of semantic reasoning.

s180
00:16:22.399 --> 00:16:24.160
So what is there in the scene?

s181
00:16:24.160 --> 00:16:40.320
Uh here is an illustration of us thinking about um uh utility poles and we want to uh when a instruction comes in hey go look at this line there's something gone wrong the drone needs to go there it needs to understand the scene then it needs to take actions within that scene and we're constantly looking at how

s182
00:16:40.320 --> 00:16:53.759
to build these primitives these tools ultimately uh that uh we today code but uh at any time an agent can access these tools to better understand uh what the drone is seeing and what it could do about

s183
00:16:54.959 --> 00:17:01.440
So an example of uh the agentic sort of system in action is a visual language model on the top left side.

s184
00:17:01.440 --> 00:17:07.919
The user here is typing in look for a white jeep uh and it's doing a sort of find and follow.

s185
00:17:07.919 --> 00:17:13.439
It in it uh accesses a drone API to command a certain sort of trajectory that it should take.

s186
00:17:13.439 --> 00:17:21.360
While that is happening the detection head is kind of uh the V uh the VLM is kind of running here to say hey what's what in the scene am I looking for?

s187
00:17:21.360 --> 00:17:26.559
it finds something, then it has access to the tools that allow the drone to track and follow.

s188
00:17:26.559 --> 00:17:43.760
And that's without any specific coding of that law rules, but instead uh having a more sort of agentic uh giving the agents the all the tools that it needs to be able to um understand the drone state and make decisions given the uh information and the context that's available there.

s189
00:17:45.600 --> 00:17:53.360
And then there's uh always the sort of long-term vision uh that's often there in the self-driving community here right now or any sort of robotic system.

s190
00:17:53.360 --> 00:18:01.440
What if uh we could just give it raw sensor data and out comes uh the perfect uh results.

s191
00:18:01.440 --> 00:18:03.679
Uh perhaps that's the actuation that happens.

s192
00:18:03.679 --> 00:18:05.360
Perhaps it's where the drone is pointing.

s193
00:18:05.360 --> 00:18:06.960
Perhaps it's where the drone goes.

s194
00:18:06.960 --> 00:18:16.880
And we're definitely sort of doing a lot of testing and triing with reinforcement learning on what that looks like if we have multiple uh instantiations of uh this um behavior.

s195
00:18:16.880 --> 00:18:19.760
Does it get to the right um end spectrum?

s196
00:18:19.760 --> 00:18:30.960
The main consideration whenever we work with a physical system is that you're often looking at really high volumes of reliability as I was saying many nines of reliability.

s197
00:18:30.960 --> 00:18:41.760
uh and the um doing a completely end to-end system uh does have its challenges in the sense that the observability of what's going wrong and the guarantees on reliability is very difficult today.

s198
00:18:41.760 --> 00:18:51.360
So while it's a direction that we're continuously taking and exploring, it's kind of figuring out which segments of your end to-end chunk need to move to a more uh um

s199
00:18:51.360 --> 00:18:54.320
like sort of world model representation of it.

s200
00:18:54.320 --> 00:18:59.440
Uh, and it's specifically the things where we always find ourselves, okay, I need to handgineer this.

s201
00:18:59.440 --> 00:19:00.880
I need to code this in specifically.

s202
00:19:00.880 --> 00:19:01.919
I need to look at these rules.

s203
00:19:01.919 --> 00:19:10.960
Like search and rescue is one of those things like, oh, go look for these trees, but if you don't find the trees, look under here or then turn on thermal or but then if you didn't find it here, look there.

s204
00:19:10.960 --> 00:19:18.960
We're trying to get away from having to have all these if statements and the code and these branching strategies and kind of let the agent have some highle tools

s205
00:19:18.960 --> 00:19:24.080
uh to be able to instruct these very high level commands uh for the drone.

s206
00:19:26.000 --> 00:19:36.559
So um I talked about a quadcopter today, but we're kind of doing this similar with a much smaller from form factor quadcopter and we're now starting to look at this what this looks like from a fixed wing as well.

s207
00:19:36.559 --> 00:19:38.559
all in the world of infrastructure.

s208
00:19:38.559 --> 00:19:47.840
Uh so they can be launched from anywhere, recovered from anywhere and ultimately the sweet spot where we can kind of start to very quickly build on the cloud uh to orchestrate these things.

s209
00:19:47.840 --> 00:19:58.160
And the way we thinking about it is allowing these systems to have basic APIs of interactivity that the cloud agents can come in and uh tap into and make decisions on

s210
00:19:58.160 --> 00:20:04.960
and ultimately allow for very high level thinking when we uh uh work with these drones.

s211
00:20:04.960 --> 00:20:15.440
So like many talks here we are hiring uh as as I said we have full stack sort of uh we do the end toend thing hardware software autonomy full stack front end back end

s212
00:20:15.440 --> 00:20:27.600
uh wireless networking everything all the technologies on our mobile phones we're now making them fly um so uh come say hi or look at our website uh would love to talk more

s213
00:20:27.600 --> 00:20:29.600
thank