Latest / The Dan Show / Training Autonomous Delivery in Simulation

Transcript

Training Autonomous Delivery in Simulation · The Dan Show

• 0:04
  SPEAKER_02: Hi everyone, welcome to another Dan Show.
• 0:06
  SPEAKER_01: Hey, here we are.
• 0:07
  SPEAKER_02: Here we are.
• 0:08
  SPEAKER_02: We have a special guest.
• 0:09
  SPEAKER_01: This guy.
• 0:10
  SPEAKER_02: We do Tan May Havelhar.
• 0:11
  SPEAKER_02: Did I say Arrive?
• 0:12
  SPEAKER_02: Yes, that's working.
• 0:13
  SPEAKER_02: Great.
• 0:14
  SPEAKER_02: Um robotics engineer here at Arrive AI.
• 0:16
  SPEAKER_01: Hey, real quick.
• 0:22
  SPEAKER_02: I can't see you on Talking.
• 0:24
  SPEAKER_01: All right, that's my that's my Easter egg.
• 0:26
  SPEAKER_01: That was a good entrance.
• 0:27
  SPEAKER_01: All right, there you go.
• 0:28
  SPEAKER_01: Joy Coast, everybody.
• 0:30
  SPEAKER_01: Love Joy Coast.
• 0:31
  SPEAKER_01: Okay.
• 0:31
  SPEAKER_02: Yes, we do love Joe.
• 0:32
  SPEAKER_01: Future so bright.
• 0:33
  SPEAKER_01: You gotta wear joints.
• 0:34
  SPEAKER_01: I don't want to take them off for now.
• 0:36
  SPEAKER_02: I didn't know that that was coming.
• 0:37
  SPEAKER_02: So um Rainhaldekar, thanks for joining us.
• 0:41
  SPEAKER_02: Anything else before we get started, Dan?
• 0:43
  SPEAKER_01: Yeah, you'll see in a little bit.
• 0:44
  SPEAKER_02: Okay.
• 0:45
  SPEAKER_02: Let's go.
• 0:45
  SPEAKER_02: Okay.
• 0:46
  SPEAKER_02: Um we have some exciting things to talk about with you.
• 0:50
  SPEAKER_02: Um, the brain behind a lot of our operations right here.
• 0:54
  SPEAKER_02: We're lucky to have you.
• 0:55
  SPEAKER_02: First.
• 0:55
  SPEAKER_01: Are you the brain behind a lot of the?
• 0:57
  SPEAKER_01: I thought I was.
• 0:58
  SPEAKER_01: Some of it.
• 0:58
  SPEAKER_01: What about me?
• 1:02
  SPEAKER_01: You gotta have the the okay, okay.
• 1:04
  SPEAKER_01: I'm not the brand, obviously.
• 1:05
  SPEAKER_01: That's okay.
• 1:06
  SPEAKER_02: We need visionaries just as much as we need to.
• 1:09
  SPEAKER_01: I like to think I'm both.
• 1:10
  SPEAKER_01: Am I not both?
• 1:11
  SPEAKER_01: You think I'm both?
• 1:12
  SPEAKER_01: Uh you know, yes.
• 1:12
  SPEAKER_01: Okay.
• 1:14
  SPEAKER_01: Good answer.
• 1:14
  SPEAKER_01: Good answer.
• 1:15
  SPEAKER_01: Okay.
• 1:15
  SPEAKER_01: All right.
• 1:16
  SPEAKER_01: What do you got?
• 1:16
  SPEAKER_01: Go ahead.
• 1:17
  SPEAKER_02: Um, so just tell people who are listening and watching a little bit about yourself.
• 1:21
  SPEAKER_00: Sure.
• 1:21
  SPEAKER_00: Uh so I am Tanmah Aldenkar.
• 1:24
  SPEAKER_00: I work as a robotics and AI engineer at Adifai.
• 1:27
  SPEAKER_00: And a lot of the work that I do here um involves robots, simulations, cameras, um, and AI inferencing and all the applications that we're building around the AI that we're leveraging in at IPAI.
• 1:38
  SPEAKER_02: So Tambe was brought on to work on our next gen of the AP3.
• 1:43
  SPEAKER_02: And then recently you might have seen the press release that um we brought all of our engineers together.
• 1:48
  SPEAKER_02: And so now he is actually working on some improvements within the AP3, um, just kind of bringing it all full circle, a lot of which he've been doing in the past week or so.
• 1:58
  SPEAKER_02: I know a little bit you can't say, but some of it you can, right?
• 2:01
  SPEAKER_01: Well, say say it all, man.
• 2:03
  SPEAKER_01: Come on.
• 2:04
  SPEAKER_00: Sure.
• 2:04
  SPEAKER_00: Um, so this past week I have been working on getting the 3D files for our AP3 model and adding them into our simulation.
• 2:13
  SPEAKER_00: Um, most of our simulation work is in Nvidia's Isaac Sim, which is Nvidia's simulation platform.
• 2:19
  SPEAKER_00: And what Nvidia does really well with Isaac Sim is that Isaac Sim simulation environments are photorealistic.
• 2:25
  SPEAKER_00: So through ray tracing, which is executing the math behind light and simulating light, um, Nvidia has figured out a way of providing us with a simulation um tool that is not only physics accurate, we can simulate collisions, friction, but we can also simulate light accurately.
• 2:42
  SPEAKER_00: And what this means is that it allows us to collect visually photorealistic data from simulation.
• 2:47
  SPEAKER_00: And so the performance of various AI models that we train that use vision as a modality of data is much better tuned to be deployed to real systems already out of the box, even though it's only been trained on simulation data.
• 3:00
  SPEAKER_01: One of the things that blows me away is how granular you get in these simulations, right?
• 3:05
  SPEAKER_01: Things that honestly, even as the visionary of the product, things I never thought about, how different things would be impacted.
• 3:13
  SPEAKER_01: And you guys are back there doing all this, and the way you're able to shortcut things and rapidize through the omniverse or the simulation.
• 3:21
  SPEAKER_01: Talk about that.
• 3:22
  SPEAKER_00: Yes.
• 3:23
  SPEAKER_00: So um Nvidia provides a lot of inbuilt um tools, and they have pretty good documentation about um all of the different tools like Omniverse and Isaac Sim and all of their APIs.
• 3:34
  SPEAKER_00: So we're able to leverage a lot of what they provide from out of the box, um, a lot of the textures that they provide, 3D files.
• 3:40
  SPEAKER_00: They have really cool environments that um replicate hospital scenes or outdoor scenes with simulating um people walking on the road or cars and vehicles.
• 3:51
  SPEAKER_00: And like you said, it's as granular as simulating friction or contact between bodies.
• 3:56
  SPEAKER_00: And you can get very granular with these simulations.
• 3:58
  SPEAKER_00: Um, and that helps us capture motion artifacts and make sure that whatever we're doing in simulation is as close to the real world as we can get.
• 4:07
  SPEAKER_00: Um, and that affords us to not make as many mistakes in the real world because what we're training in simulation is that much better before it gets deployed to the real world.
• 4:15
  SPEAKER_02: Yeah, you're not um crashing drones into the nightmare.
• 4:22
  SPEAKER_02: That would be a nightmare.
• 4:23
  SPEAKER_02: So problem solving before it's actually deployed is uh the number one biggest value in it, I would imagine.
• 4:30
  SPEAKER_00: Correct.
• 4:31
  SPEAKER_02: And it's also a quicker process to market, right?
• 4:33
  SPEAKER_00: Yes.
• 4:34
  SPEAKER_00: Much it it lets us prototype very quickly too, right?
• 4:37
  SPEAKER_00: Um so all of the interactions that we have, let's say, between the arrive point and um a mobile robot like the autonomy robot, we can actually simulate those and through simulation understand if there are any failure cases that we may not have foreseen before the models that we want to deploy have been deployed.
• 4:54
  SPEAKER_00: So that helps a lot.
• 4:55
  SPEAKER_01: We're not just building a mailbox 2.0, right?
• 4:58
  SPEAKER_01: We're building a whole platform network company, right?
• 5:00
  SPEAKER_01: And there's a lot involved.
• 5:02
  SPEAKER_01: Why don't you just talk about, give some behind the scenes without too much and just talk about the depth of what we're really doing here so people have a sense.
• 5:10
  SPEAKER_00: I mean, we um, you know, I think we have a lot of very talented engineers here, and we have simultaneous work going.
• 5:18
  SPEAKER_00: You know, my my team primarily focuses on cameras, vision, um, robotics, and our applications in that space.
• 5:25
  SPEAKER_00: And a lot of that is AI.
• 5:27
  SPEAKER_00: But we also have brilliant teams and engineers that are working on, like you said, building a network of arrive points, um, trying to show the um you know the benefit that this network will bring to last mile logistics and being able to simulate that as well.
• 5:40
  SPEAKER_00: And that's a whole different simulation scenario.
• 5:42
  SPEAKER_00: Um, and then we have you know a very talented engineering team that's actually putting the ideas onto paper, modeling the you know, the system that we want to build and deploy and making sure that it's reliable.
• 5:53
  SPEAKER_00: Um and overall, our team we have test engineers, we we really have um a lot of work going towards the correct direction, and it's very comprehensive and it's very nice to see.
• 6:03
  SPEAKER_01: What about the compute capacity to be running all these different simulations in different areas?
• 6:07
  SPEAKER_01: Why don't you speak about that?
• 6:09
  SPEAKER_00: Sure.
• 6:09
  SPEAKER_00: So um we have three um workstations, um, and the workstations each house multiple Nvidia RTX 6000 Pro Blackwell GPUs.
• 6:21
  SPEAKER_00: Um I know the name is a mouthful.
• 6:22
  SPEAKER_00: Sure.
• 6:23
  SPEAKER_00: Um, but the Blackwell is just the name for their latest generation of GPUs.
• 6:28
  SPEAKER_00: Um so what the GPU gives us, and the Blackwell GPU specifically, is a lot of high-speed memory and CUDA cores and tensor cores.
• 6:37
  SPEAKER_00: So um what CUDA and Tensor Cores do is they um allow us to run AI models very efficiently and uh run the compute that is required by the AI model in its inference.
• 6:49
  SPEAKER_00: And um, having a lot of high-speed memory means we can load larger models because large language models or generative models, vision language models tend to be very big these days.
• 6:57
  SPEAKER_00: They have billions of parameters.
• 6:59
  SPEAKER_00: And so it's a challenge to even be able to load those models onto older generation of GPUs.
• 7:04
  SPEAKER_00: Um, so the Blackwell architecture, which has fifth generation Tensor cores and fourth generation ray tracing cores, is provides, I think, three times, close to three times performance over their previous architecture, their ADA architecture.
• 7:19
  SPEAKER_00: Um, and the ray tracing cores, like I mentioned, also allow us to run bigger simulation environments that are photorealistic even more efficiently.
• 7:28
  SPEAKER_00: So combined the high-speed memory, the extra GPU cores, and the ray tracing cores really allow us to leverage Nvidia's tools to be able to build large simulation environments that can model certain parts of our network, like you mentioned, and be able to do it efficiently and um not drop any performance while running it.
• 7:48
  SPEAKER_01: What's your previous experience prior to working here at Arrive?
• 7:51
  SPEAKER_01: What tell us about that?
• 7:52
  SPEAKER_01: Sure.
• 7:53
  SPEAKER_00: So um you'll be interested to know.
• 7:54
  SPEAKER_00: I before working at Arrive AI, I was working as an international researcher at the National Institute of Standards and Technology, NIST, in Gaithersburg, Maryland.
• 8:04
  SPEAKER_00: And surprisingly, um, the research that I was working uh on primarily there was focused on robotic systems in recycling.
• 8:12
  SPEAKER_00: So um at materials recovery facilities where all the recyclables that we collect in our curbside bins go, um, it's quite a challenge for the workers there to be able to reach out on the conveyor belt and pick out the recyclables from the conveyor belt.
• 8:25
  SPEAKER_00: And there's always, you know, because um we we don't necessarily do a good job of educating people about recycling, there's always, you know, um needles and sharp objects and all other like dangerous and hazardous items on the conveyor belts.
• 8:40
  SPEAKER_00: And it's it's not a great facility for a person to be working in either because there's a smell to the place and it's hot or cold.
• 8:47
  SPEAKER_00: So we got them B.
• 8:49
  SPEAKER_00: Yes.
• 8:50
  SPEAKER_00: It doesn't smell that bad here.
• 8:51
  SPEAKER_00: Yes.
• 8:52
  SPEAKER_00: Oh, it's it's much nicer here.
• 8:53
  SPEAKER_00: Well, yeah, I don't think you've worked in there.
• 8:55
  SPEAKER_00: I know I'm just giving them all.
• 8:56
  SPEAKER_00: I have been to a fair few um just to see what they're like and talk to the workers, yeah.
• 9:01
  SPEAKER_00: But uh we try to keep it light here.
• 9:04
  SPEAKER_00: Yes.
• 9:04
  SPEAKER_00: Oh, this this place smells really nice.
• 9:06
  SPEAKER_00: The TV rolls with a punches.
• 9:07
  SPEAKER_02: I could see in his eyes he was trying to figure out where Dan was going in real time.
• 9:11
  SPEAKER_01: So was I, so was I.
• 9:12
  SPEAKER_01: Good job.
• 9:13
  SPEAKER_01: If I don't know where I'm going, it's hard for you.
• 9:17
  SPEAKER_00: Yes.
• 9:18
  SPEAKER_00: So so at recycling facilities, the robotic systems that they're trying to deploy are uh they're not generic robotic arms, but they're these spider-like robotic arms that sit on top of the conveyor belt with a camera that try to find the valuable recyclables um that actually have some value for the uh the recycling facility itself to separate and try to catch them from the conveyor and separate them.
• 9:38
  SPEAKER_00: So while a lot of our research was not yet very applied because these systems are quite new and we were trying to get some funding for this project, it was a lot of surveying the research field, talking to folks, understanding robotic systems.
• 9:50
  SPEAKER_00: Sure.
• 9:51
  SPEAKER_00: Um and then on the side, I was also working on uh uh a physical AI project where we were using robotic arms and training them in Nvidia's Isaac Sim tools to do different things like picking place or assembly tasks.
• 10:04
  SPEAKER_02: So a lot of that you can apply to what you're working at.
• 10:08
  SPEAKER_02: Yes.
• 10:08
  SPEAKER_00: Hey, where are you from originally?
• 10:10
  SPEAKER_00: I am from Mumbai in India.
• 10:12
  SPEAKER_01: Why don't you tell the exciting news?
• 10:16
  unknown: Oh.
• 10:17
  SPEAKER_00: About your Oh, yes, okay.
• 10:20
  SPEAKER_00: It took me a second.
• 10:21
  SPEAKER_01: I guess it wasn't that exciting.
• 10:22
  SPEAKER_01: Not that exciting.
• 10:23
  unknown: Of course.
• 10:24
  SPEAKER_00: Well, uh the news that Dan's referring to is that uh me and a few of my other colleagues that work here, we actually got selected for our H1Bs this year.
• 10:32
  SPEAKER_01: That's right.
• 10:32
  SPEAKER_00: Sponsored by sponsored by Live AI, because that IV AI wants to keep us around.
• 10:36
  SPEAKER_02: Yes.
• 10:37
  SPEAKER_01: And I'm glad.
• 10:38
  SPEAKER_01: Now he's stuck here and we're stuck with you, man.
• 10:40
  SPEAKER_01: Uh-huh.
• 10:40
  SPEAKER_02: So kind of going off of that when you were saying your colleagues, it just seems like a great camaraderie.
• 10:45
  SPEAKER_02: That at lunchtime, the conversation and the laughs and everything, it's real.
• 10:49
  SPEAKER_02: It's really fun energy all the time.
• 10:51
  SPEAKER_00: Yes, we also have a uh very competitive um Jess scene.
• 10:55
  SPEAKER_00: Yes.
• 10:56
  SPEAKER_01: So I'm bringing a back gammon board in.
• 10:58
  SPEAKER_01: I want to take you guys.
• 10:59
  SPEAKER_01: Anybody play back gammon back there?
• 11:01
  SPEAKER_01: Uh I don't think so.
• 11:02
  SPEAKER_01: Well, they're gonna learn.
• 11:03
  SPEAKER_01: They're gonna learn.
• 11:04
  SPEAKER_02: It's easier to learn than chess.
• 11:06
  SPEAKER_02: So I think that's a good thing.
• 11:07
  SPEAKER_01: I think it's more fun too.
• 11:09
  SPEAKER_01: Anyway, we digress.
• 11:10
  SPEAKER_01: Sorry.
• 11:10
  SPEAKER_01: What else do we got?
• 11:11
  SPEAKER_02: Something we do have a bunch of chess players here.
• 11:14
  SPEAKER_02: And you do have so do you have it?
• 11:16
  SPEAKER_02: You I know that there was a bracket at some point.
• 11:19
  SPEAKER_02: Is that bracket still continuing, or is that recent?
• 11:21
  SPEAKER_00: Uh you know, in in recent times, uh the work has increased.
• 11:24
  SPEAKER_00: So I see fewer chess games.
• 11:26
  SPEAKER_00: Um, but I I I could ask a few people.
• 11:28
  SPEAKER_00: Some people, especially the ones that were uh on top and winning, have kept the record.
• 11:32
  SPEAKER_00: Of course.
• 11:33
  SPEAKER_01: We'd all see Neuro's name and keeping it going.
• 11:35
  SPEAKER_01: Neuro's fine checkers.
• 11:37
  SPEAKER_01: Right.
• 11:38
  SPEAKER_01: Just kidding.
• 11:41
  SPEAKER_02: Um but so our COO, Mark Cam, always talks about storming, forming, and norming.
• 11:47
  SPEAKER_02: And you've been a part of this.
• 11:48
  SPEAKER_02: And it's a path to go to market um as efficiently as possible.
• 11:53
  SPEAKER_02: So, where in that phase are you guys?
• 11:55
  SPEAKER_02: And what does that look like for you and the team?
• 11:58
  SPEAKER_00: Good question.
• 11:59
  SPEAKER_00: So um I think we're it's been a while since we've been storming and we're finally um, it feels like we're forming now.
• 12:07
  SPEAKER_00: And um I think we're still in the prototyping phase.
• 12:11
  SPEAKER_00: So we're working on different ideas and different, you know, we have a vision for what the system needs to be, and we're validating the different aspects of it.
• 12:18
  SPEAKER_00: Um, but I'm sure with how our work has been so far, um, it won't take us too long to form and be able to um put that form into an actual physical representation of this unit.
• 12:29
  SPEAKER_00: And I'm super excited to see what we actually end up building and like the first version of it um deployed.
• 12:34
  SPEAKER_01: Yeah, it's really important to build that strong foundation, right?
• 12:38
  SPEAKER_01: Ahead of revenue or profits or anything like that, get in that baseline.
• 12:42
  SPEAKER_01: And that's what we're really focused on, right?
• 12:44
  SPEAKER_01: Yes, yeah.
• 12:44
  SPEAKER_01: And and you're seeing some great movement.
• 12:46
  SPEAKER_00: Yes, absolutely.
• 12:47
  SPEAKER_00: Have you been blown away?
• 12:48
  SPEAKER_00: Is anything blown you away?
• 12:50
  SPEAKER_00: Oh my god.
• 12:50
  SPEAKER_00: I mean, the the engineers that we have on the team always blow me away.
• 12:54
  SPEAKER_00: Because um, you know, I'm I'm so focused and local to the work that I do.
• 12:59
  SPEAKER_00: But when I step out and actually see all of the other people around me working on the different things, and if you walk by our product room um and talk to the different people, everyone will always have like a small demo to show you.
• 13:09
  SPEAKER_00: And it's crazy how fast we're progressing too.
• 13:12
  SPEAKER_00: So I'm always blown away by the work that the other teams are doing.
• 13:16
  SPEAKER_00: Always.
• 13:16
  SPEAKER_00: And if I just want some time off, I'll step out of my room out of the you know, the room where we have the workstations and go talk to anyone.
• 13:22
  SPEAKER_01: You're not taking time off, are you?
• 13:24
  SPEAKER_01: No, no, no.
• 13:25
  SPEAKER_00: Not in the logic sense, no.
• 13:26
  SPEAKER_01: Okay, don't let that happen.
• 13:27
  SPEAKER_01: Okay.
• 13:28
  SPEAKER_00: I take I take time off to learn more about what's happening with the other teams.
• 13:32
  SPEAKER_01: Yeah, no, I'm I'm giving you a hard time.
• 13:33
  SPEAKER_01: Hey, I know.
• 13:33
  SPEAKER_01: Proud to have you on the team.
• 13:34
  SPEAKER_01: Go ahead, Kyle.
• 13:35
  SPEAKER_01: What were you gonna say?
• 13:36
  SPEAKER_02: Um, no, I was just wondering, kind of in the same vein.
• 13:40
  SPEAKER_02: I see you all plugging away with code and simulations and whatnot.
• 13:45
  SPEAKER_02: When was the last time you got this moment of like, oh, we did it?
• 13:49
  SPEAKER_02: There was a breakthrough.
• 13:50
  SPEAKER_02: Because I can imagine those.
• 13:52
  SPEAKER_02: You you you storm in your own little universe.
• 13:55
  SPEAKER_01: Don't say it was at the place with the conveyor belt either.
• 13:58
  SPEAKER_01: Say it was your okay.
• 14:00
  SPEAKER_02: In your work here, it arrives.
• 14:02
  SPEAKER_02: Yeah.
• 14:03
  SPEAKER_00: Um, I think there's there's been a few of those.
• 14:05
  SPEAKER_00: Um ever since we started working here, um, me and my team especially.
• 14:10
  SPEAKER_00: Um, I work with James and Troof, and we've all learned so much.
• 14:14
  SPEAKER_00: Um, and I'm very proud of them too, because they they have also like grown um and developed as engineers massively.
• 14:20
  SPEAKER_00: So we've had a few moments where um you know we're working on the different architectural designs for you know our AI applications, and every once in a while we have this aha moment where you know what we were researching since December and trying to put into like an application or a system actually makes sense now.
• 14:39
  SPEAKER_00: And it's a culmination of all of the work that you know our team has done since then.
• 14:44
  SPEAKER_00: So we've had a few of those moments with you know um testing out some cameras or testing out some of the you know new hardware that we're getting or running some of the AI models that we've been testing currently.
• 14:53
  SPEAKER_01: Yeah, just because it's never been done doesn't mean it won't be done, right?
• 14:57
  SPEAKER_01: Yes.
• 14:58
  SPEAKER_01: That's kind of our mantra, right?
• 14:59
  SPEAKER_01: Yes.
• 15:00
  SPEAKER_01: So so we're building new technology in a new industry with new people here in the US, right?
• 15:06
  SPEAKER_01: And that's great, right?
• 15:08
  SPEAKER_00: Yes, and it's a very exciting time to be building new technology and working in this space also, because we really are at the forefront of physical AI and you know uh last mile logistics.
• 15:19
  SPEAKER_00: And there never has been a better time, I think, than now, with all the different tools that we can leverage and with how far the industry has come.
• 15:26
  SPEAKER_00: So I think it's it's it really has set us um in a position for success.
• 15:30
  SPEAKER_00: And I think with all the different people that we have working on this, I can definitely see it's it's our game to win.
• 15:36
  SPEAKER_00: That would be my take on it.
• 15:37
  SPEAKER_00: Thank you.
• 15:38
  SPEAKER_01: That's awesome.
• 15:38
  SPEAKER_01: I feel the same way.
• 15:39
  SPEAKER_01: Kyle, how do you feel?
• 15:41
  SPEAKER_02: The same way.
• 15:42
  SPEAKER_01: Okay, I'll get an answer.
• 15:43
  SPEAKER_01: You'll probably see her on the next episode.
• 15:46
  SPEAKER_01: You're probably gonna come back.
• 15:47
  SPEAKER_01: Uh-huh.
• 15:47
  SPEAKER_01: Anything else we want to talk about?
• 15:49
  SPEAKER_02: Yeah, anything, anything else that you would like to add from just your time here, some of the development process, the workstations, anything at all?
• 15:57
  SPEAKER_00: Um, well, I'm just super glad to be working here.
• 16:01
  SPEAKER_00: And I really appreciate the opportunity.
• 16:03
  SPEAKER_00: Um, and you know, I I can't complain because I get to work in the field that I love to work in with the latest and greatest toys every day.
• 16:10
  SPEAKER_00: Um, and so I have a smile on my face every time that I come into this place.
• 16:15
  SPEAKER_00: You really do.
• 16:15
  SPEAKER_00: Yes.
• 16:16
  SPEAKER_00: I love working with my co-workers.
• 16:18
  SPEAKER_01: Yeah, no, I guess that's a good segue.
• 16:20
  SPEAKER_01: Why don't you kind of gauge Arrive AI and how you think it fits against other peer companies that are creating new technology, not only the level of talent, but the tools that we've made available in the general morale of the company.
• 16:33
  SPEAKER_01: What's your take on that?
• 16:34
  SPEAKER_00: Yeah, um, I think, you know, uh, like I said, I'm I'm a big fan of everyone that I get to work with.
• 16:40
  SPEAKER_00: And never have I like, you know, I've always seen a positive um environment and like a positive mood around the office.
• 16:48
  SPEAKER_00: Everyone's always super excited about everything that they're working on.
• 16:51
  SPEAKER_00: Um and we've been moving so fast, also.
• 16:53
  SPEAKER_00: So, like I said, like whenever you walk by and talk to someone about what they're working, it's always like, oh my god, I didn't even know that could happen.
• 16:59
  SPEAKER_00: Right moment.
• 17:01
  SPEAKER_00: And uh being being like a small um uh team that's growing and growing together, I think it sets us up for success because um you know, we we're not really tied to any um, you know, a certain way of working.
• 17:15
  SPEAKER_00: And we can be quite agile, right, and be able to pivot as we require.
• 17:18
  SPEAKER_00: And with how far technology has come and how our tools are and the experience that we have on board with all of our employees, I think you know it it allows us to be able to like navigate this space quite efficiently and be able to work in this very state-of-the-art, very quickly changing field and not be left behind.
• 17:36
  SPEAKER_00: So I think as compared to other competitors that might have been in the industry for longer, you know, with how quickly Genitive AI and the tools today uh seem to be improving and progressing.
• 17:45
  SPEAKER_00: Um, actually, I think it's to our advantage that we have such excellent engineers and we're able to pivot so quickly and be able to leverage the technology.
• 17:54
  SPEAKER_00: Awesome.
• 17:55
  SPEAKER_01: Hey man, thanks for being on the team.
• 17:58
  SPEAKER_01: Doing a great job.
• 17:59
  SPEAKER_01: Thanks so much for having me.
• 18:00
  SPEAKER_02: All right, well, may have the car, everybody.
• 18:02
  SPEAKER_01: Thanks for joining, man.
• 18:03
  SPEAKER_01: Yes, have a great day.
• 18:04
  SPEAKER_01: Anytime.
• 18:05
  SPEAKER_00: I hope I can make a comeback appearance.
• 18:06
  SPEAKER_00: Yeah.
• 18:07
  SPEAKER_00: And the next time we can talk about what we actually have already.
• 18:09
  SPEAKER_00: You know what?
• 18:10
  SPEAKER_01: We're gonna check the ratings and see show us how it was good.
• 18:14
  SPEAKER_01: I hope please get me good ratings, guys.
• 18:15
  SPEAKER_02: I think it means it's our best performing one ever, and Dan doesn't want you to get back.
• 18:19
  SPEAKER_01: Yeah, that's right.
• 18:20
  SPEAKER_01: You're either too good to come back or you're too bad.
• 18:22
  SPEAKER_01: So you better hope for the middle.
• 18:24
  SPEAKER_01: Just kidding.
• 18:24
  SPEAKER_01: Okay.
• 18:25
  SPEAKER_00: Well, we'll we'll do something so cool that you're forced to get me back.
• 18:27
  SPEAKER_01: Yeah, let's let's get it back and let's show them next time.
• 18:29
  SPEAKER_01: Absolutely.
• 18:30
  SPEAKER_01: Okay, thanks, man.
• 18:31
  SPEAKER_02: And I think that's gonna do it for today, dude.
• 18:32
  SPEAKER_02: Is that it?
• 18:32
  SPEAKER_02: Is that it?
• 18:33
  SPEAKER_01: Okay, we don't want to give everything up in one episode, right?
• 18:36
  SPEAKER_02: That's right.
• 18:37
  SPEAKER_01: Okay, all right, guys.
• 18:38
  SPEAKER_01: Well, thanks for joining.
• 18:39
  SPEAKER_01: Time, thanks, Kyle.
• 18:40
  SPEAKER_01: Thanks so much.
• 18:41
  SPEAKER_01: Great job.
• 18:41
  SPEAKER_01: Thanks, guys.
• 18:42
  SPEAKER_02: Bye next time.
• 18:44
  SPEAKER_01: Bye, guys.