Latest / Logixly Live Unsupervised Learning Podcast / Why Forward Deployed Engineers Turn AI Into ROI

Transcript

Why Forward Deployed Engineers Turn AI Into ROI · Logixly Live Unsupervised Learning Podcast

• 0:04
  SPEAKER_00: I gotta say, um he's my favorite train.
• 0:08
  SPEAKER_00: Point him in a direction.
• 0:12
  SPEAKER_02: But I I love that you didn't force them in a direction.
• 0:17
  SPEAKER_02: You might point him, but like if they don't you know, I I will always go back to the little, you know, caven kids and I'm just like I hate them.
• 0:28
  SPEAKER_02: That's not how I would do it.
• 0:30
  SPEAKER_02: Like I think your kids they find the right way.
• 0:33
  SPEAKER_00: We'll see.
• 0:33
  SPEAKER_00: Right?
• 0:34
  SPEAKER_02: I well you said he wants to do math now, right?
• 0:37
  SPEAKER_02: There you go.
• 0:38
  SPEAKER_02: Heck yeah.
• 0:39
  SPEAKER_00: I think the ultimate um measure of success is they all come back home.
• 0:44
  SPEAKER_02: That's what that's what not just checking the mom box.
• 0:47
  SPEAKER_01: I think his mom babysits.
• 0:49
  SPEAKER_01: Yeah.
• 0:50
  SPEAKER_00: But uh that was uh a good turnout last night.
• 0:54
  SPEAKER_00: A good Was it?
• 0:56
  SPEAKER_00: It really was.
• 0:56
  SPEAKER_00: It was yeah, it it I really uh really good content, really good conversation.
• 1:03
  SPEAKER_00: I uh yeah, it was it was another good, another good evening.
• 1:06
  SPEAKER_02: Yeah, I'd sneak out early to go to my little class, but the uh the guy just we talked beforehand on like very similar page to where we're at on what he's seeing in industry.
• 1:16
  SPEAKER_02: It's awesome.
• 1:17
  SPEAKER_00: Yeah.
• 1:18
  SPEAKER_00: Yeah.
• 1:19
  SPEAKER_00: Speaking about industry, I I gotta say, I feel like we jump into this a little earlier today.
• 1:23
  SPEAKER_00: Um, because I I feel like this is this is a hot topic.
• 1:26
  SPEAKER_00: So again, welcome everybody to Logic Flee Live, where we cover everything that's happening in AI.
• 1:32
  SPEAKER_00: And as typical with every uh session, it's like this last week, you could pick any number of topics, right?
• 1:40
  SPEAKER_00: I I my opinion is this a topic.
• 1:42
  SPEAKER_00: So the topic today is a forward deployed engineer.
• 1:44
  SPEAKER_00: I kind of feel like this is it's in the vernacular, it always has been, but to me, it feels like it's it's kind of rising to the top.
• 1:55
  SPEAKER_00: My data point is actually, and we'll do we'll define it, give a little history of it, but I to me it feels like another version of what data scientist was maybe a decade ago.
• 2:08
  SPEAKER_00: Yep.
• 2:08
  SPEAKER_02: Yep.
• 2:08
  SPEAKER_00: Do you guys okay?
• 2:10
  SPEAKER_02: My my data point is from 25 to 26 in Google Trends, it's a thousand percent growth.
• 2:17
  SPEAKER_02: For board deployed engineering, okay.
• 2:19
  SPEAKER_02: Yeah, okay.
• 2:20
  SPEAKER_02: It's like it went from nothing to like, oh my god, it's huge.
• 2:23
  SPEAKER_02: Yes.
• 2:24
  SPEAKER_02: And it's yeah.
• 2:25
  SPEAKER_00: And and I'll put my and I I do want to define it and give some history, but um like my like I feel like it feels like data scientists because you know, going in the Wayback Machine, uh, I think everybody had data scientist as the thing they pop in and solve all the problems.
• 2:44
  SPEAKER_00: You can't really define the role really well, and everybody defines it maybe a little bit differently.
• 2:50
  SPEAKER_00: And then when you start talking about skills for it, or if you try to hire one, you're like, well, we need this, but we need a lot of like for data science, right?
• 2:57
  SPEAKER_00: It was like, well, we need you have to be able to program, you got to be able to understand the business, you need the statistics, understanding.
• 3:03
  SPEAKER_00: But if you don't have the math, I mean it's just you just keep stacking the skills, right?
• 3:08
  SPEAKER_02: Right.
• 3:08
  SPEAKER_02: It was that you could not find a person who could do all of those things.
• 3:12
  SPEAKER_02: And oh, by the way, they need to be able to talk and storytell as well.
• 3:15
  SPEAKER_02: Yeah, that I'm sorry, that person doesn't exist.
• 3:17
  SPEAKER_00: 100% they have to communicate.
• 3:19
  SPEAKER_02: Yeah, forward deployed engineer, you're getting the same thing where everybody's got a bit of a different definition.
• 3:23
  SPEAKER_00: And yeah, so a bit of history, just to uh at least give some context.
• 3:28
  SPEAKER_00: So I pulled uh a definition from Palantir.
• 3:31
  SPEAKER_00: So Palantir, I think by everybody's uh, you know, everybody agrees that they they kind of created defined what a forward deployed engineer is.
• 3:40
  SPEAKER_00: So I'm gonna borrow their definition of it and give a little bit of history of it.
• 3:44
  SPEAKER_00: So they define a forward-deployed engineer as an engineer who works alongside a customer's employees to understand a business problem, build or adapt a solution, and help put it into everyday use.
• 3:58
  SPEAKER_00: And I think the thing that, you know, from what I gather, what they would say was uh, you know, Paneltier has software platforms that are, I won't say uh, you know, a little complicated, if you will, to connect to the business problems.
• 4:14
  SPEAKER_00: And so their solution was we're gonna take an engineer and we're gonna put them, embed them with the customer on site, and they're gonna go through the implementation of connecting the software to the business problem and to stay with the business until the outcomes were achieved.
• 4:32
  SPEAKER_01: Right.
• 4:32
  SPEAKER_00: Uh, so and and you know, you know, to go a little deeper, they they actually had a um a their model had kind of two types of engineers, and Ted they did refer to these as software engineers.
• 4:45
  SPEAKER_00: So their their definition of forward deployed engineer from 10 years ago included the engineer type was a software engineer, but they kind of carved out devs and deltas.
• 4:54
  SPEAKER_00: So devs were this this is their terminology.
• 4:57
  SPEAKER_00: Okay, their their devs were the developers and engineers kind of back at the home base that are building the Palantir software platforms.
• 5:06
  SPEAKER_00: And their mindset for that group is just like all SaaS products, they're looking to add one capability but deliver it to many customers, right?
• 5:17
  SPEAKER_00: That classic one-to-many.
• 5:19
  SPEAKER_00: Deltas deploy, which I feel like is a keyword now.
• 5:23
  SPEAKER_00: The Deltas deploy software platforms to customers, and their mindset is one customer, so they're embedded at one customer, but they're trying to deliver mini capabilities to that customer from from you know from the software.
• 5:40
  SPEAKER_00: So that's how they described it.
• 5:41
  SPEAKER_00: I looked at it, put it in our terminology.
• 5:43
  SPEAKER_00: I would say deltas are last mile.
• 5:46
  SPEAKER_00: Yeah, you know, that's how we always talk about it, and deltas are kind of focused on the outcomes.
• 5:51
  SPEAKER_00: Uh any thoughts on on kind of what you guys heard something different, different understanding of what for deployed engineers are.
• 5:59
  SPEAKER_01: Yeah, I think it frames it up perfectly, but I think the obviously biased here because of our business model for the last 15 years, but the it feels like the technology and things are changing to the point where it's it's really no longer about having the best models or the best tech.
• 6:19
  SPEAKER_01: It's really on the execution phase.
• 6:22
  SPEAKER_01: You know, you used to always say a good model deployed is better than the perfect model that just sits on the shelf.
• 6:28
  SPEAKER_01: I always remember you saying that.
• 6:30
  SPEAKER_01: And that is more true now than than ever, I think.
• 6:33
  SPEAKER_00: Right.
• 6:34
  SPEAKER_00: 100%.
• 6:34
  SPEAKER_00: So and I know Katie remembers this era well, but the the mindset that, you know, we used to always talk about the lab in the field and how it was, I won't never never was easy to create models, but if you create a great great model in the field that or a great model in the lab that you can't get to the field is is no business value.
• 6:55
  SPEAKER_00: But a but even just a lot of times like a pretty good model that you can get into the field, you're gonna drive some business outcomes with that.
• 7:04
  SPEAKER_00: And uh so 100% agree with with what you're saying there, Ted.
• 7:08
  SPEAKER_02: Yeah, very, very much so.
• 7:09
  SPEAKER_02: I mean, I think there is uh a weird terminal.
• 7:13
  SPEAKER_02: I like the term Delta.
• 7:14
  SPEAKER_02: I didn't know that they used that because it really sort of aligns in two ways.
• 7:18
  SPEAKER_02: One, you've got kind of the brave new world, we're gonna define everyone by their Greek letter, but then on the other side, you have like they are the change makers, right?
• 7:25
  SPEAKER_02: Like math-wise, delta is change, and those are the people that are going in and seeing that change through.
• 7:31
  SPEAKER_02: And I love that.
• 7:33
  SPEAKER_00: Yeah, super interesting.
• 7:34
  SPEAKER_00: So I gotta say, the question that I have now kind of like linking this to other current uh you know happenings in the AI world.
• 7:44
  SPEAKER_00: So how do you guys how do you guys think about this?
• 7:47
  SPEAKER_00: You have you have pretty much all the big frontier labs saying the same thing this last week, which is we've we've at least OpenAI is saying, look, we've hit AGI, but but even the other labs, um you know, um X AI as well as Anthropic are saying, hey, we need to slow down.
• 8:06
  SPEAKER_00: So you've got one side saying, hey, we really need to slow down, but then on the flip side, you have those same companies investing and building services organizations that are focused on delivering this last mile.
• 8:20
  SPEAKER_00: Like, how how how do you guys think about that?
• 8:22
  SPEAKER_00: Like to me, it's like I'm not saying it's exactly pointing in opposite directions, but like there is some amount of wait a second, if both those are true, then I don't really understand exactly what's happening.
• 8:34
  SPEAKER_00: What do you guys think about that?
• 8:36
  SPEAKER_02: From from a business side, I love the slowdown.
• 8:39
  SPEAKER_02: Okay, we need to go and we need to figure out what are the ways that we deploy this, right?
• 8:44
  SPEAKER_02: And it is, is it always AI, right?
• 8:46
  SPEAKER_02: Is it always the most complicated thing?
• 8:48
  SPEAKER_02: No, like find the most simple solution to solve the problem, and that's awesome.
• 8:54
  SPEAKER_02: So we need that side of it, but then we do need RD and we need to keep making progress.
• 9:00
  SPEAKER_02: So, you know, stop with the slowdown, like slow down in some ways, but we have to keep making progress.
• 9:07
  SPEAKER_02: Other people are going to continue making progress.
• 9:09
  SPEAKER_02: We have to as well.
• 9:10
  SPEAKER_02: Yeah, but from a business side, I love the idea of slowdown.
• 9:14
  SPEAKER_00: Like, take the time, do the service when you say that, you mean conceptually, like it maybe I don't want to put words in your mouth, but like slow down from the standpoint of hey, typical business.
• 9:24
  SPEAKER_00: You actually don't need to wait for a new iteration of a frontier model to go take today's goodness and apply it to your business.
• 9:32
  SPEAKER_00: Like, slow down as in like take what's here, think about your business, right, decompose it, figure out where AI fits and it doesn't, and and run in that direction.
• 9:41
  SPEAKER_02: Right.
• 9:41
  SPEAKER_02: And I'm saying, like, with these new iterations of frontier models or whatever we're changing, and whether we call it AGI or not, right?
• 9:49
  SPEAKER_02: Those things we do need to continue making progress.
• 9:52
  SPEAKER_02: Does that mean it needs to be progress that's immediately released to the public and driving business value?
• 9:56
  SPEAKER_02: No.
• 9:56
  SPEAKER_02: We just need the research to happen and we'll the business will catch up.
• 10:01
  SPEAKER_02: But 100% from a forward-deployed engineer standpoint, like you have to be able to go and say, What can we do with what we have today?
• 10:10
  SPEAKER_02: Yeah, just waiting for it's gonna be a little bit better tomorrow, a little bit better tomorrow.
• 10:13
  SPEAKER_02: You're never gonna get there.
• 10:15
  SPEAKER_02: Get the value today while you can.
• 10:17
  SPEAKER_01: Yeah.
• 10:18
  SPEAKER_01: Yeah.
• 10:18
  SPEAKER_01: I think it's a recognition of the chasm between business value and actually deploying this stuff day to day and the research that's happening in the lab, kind of going back to that perfect model, it's smarter, it's AGI, it's all this stuff.
• 10:35
  SPEAKER_01: That's great.
• 10:36
  SPEAKER_01: But in the end, for business, who cares?
• 10:38
  SPEAKER_01: Like until you can get that done.
• 10:40
  SPEAKER_01: And I think that, you know, anthropic um standing up ode and and there's uh the deployment company.
• 10:46
  SPEAKER_01: Deployment company.
• 10:47
  SPEAKER_01: Yep.
• 10:47
  SPEAKER_01: Absolutely.
• 10:48
  SPEAKER_01: And it's like, yeah, okay, that's where the gap is right now.
• 10:51
  SPEAKER_01: The gap is actually deploying this stuff, sitting down with a customer, understanding their their operational needs, where are we gonna get the ROI?
• 10:59
  SPEAKER_01: How do we get to that faster?
• 11:00
  SPEAKER_01: I think it's just it's the market catching up to it.
• 11:04
  SPEAKER_01: Um, now, whether we all die in a year because AI is gonna kill us all is probably a different topic.
• 11:10
  SPEAKER_01: But it's it's I think to me, it's a recognition that um services now, services is probably king as far as when it comes to deploying this stuff.
• 11:20
  SPEAKER_02: For sure.
• 11:20
  SPEAKER_02: But from the research standpoint, like go back to the stories about Hinton, right?
• 11:24
  SPEAKER_02: How long did Hinton sit in an office where nobody cared about his research at all until all of a sudden convolutional neural nets are the thing to talk about, right?
• 11:33
  SPEAKER_02: Like that's 20 years ago.
• 11:35
  SPEAKER_02: We have to keep doing these pushes now, but those the benefits are gonna be realized in years, not tomorrow.
• 11:42
  SPEAKER_02: It's yeah, we've we've got to be able to go and say, What do we have today and what great can we do?
• 11:48
  SPEAKER_02: Like, how can we change today?
• 11:50
  SPEAKER_00: 100%, 100%.
• 11:51
  SPEAKER_00: And the reference there to Jeff Hinton, who is kind of the quote unquote one of the the godfathers of AI, I think most people believe, researched at University of Toronto and um, you know, spent 20, 30 years working on neural nets and back propagate back propagation, excuse me, really before it was very clear at all that this technology had had really any, this is crazy to say, but really had any value.
• 12:18
  SPEAKER_00: Right.
• 12:19
  SPEAKER_00: It was it was neat, it was good on a whiteboard, it was kind of toy-ish.
• 12:23
  SPEAKER_00: And then, and then yeah, that flipped.
• 12:25
  SPEAKER_00: But I I mean to me, it almost comes back to that that thing of like we're and maybe now it's just even more apparent, but you have the frontier labs.
• 12:35
  SPEAKER_00: We clearly have amazing technology, and it's getting more and more amazing.
• 12:39
  SPEAKER_00: And then we have all of the businesses in the world that are clearly not at the pace of the frontier labs, and they really don't have to be, but that gap, that gap to me, we called it the last mile forever.
• 12:53
  SPEAKER_00: That is the gap that I feel like is the opportunity because it's not just Ode, it's not just the deployment company, it's all of the big consulting companies.
• 13:01
  SPEAKER_00: They're all they're all having this model of, yeah, the real opportunity is how do we get people in the business that understand the business?
• 13:09
  SPEAKER_00: Ted, and your point I think you were making was yeah, great, LMs are clearly the killer app for coding assistance.
• 13:17
  SPEAKER_00: But these four deployed engineers, like maybe they used to be software engineers, but but probably, probably what?
• 13:24
  SPEAKER_00: I mean, how does that change?
• 13:25
  SPEAKER_01: Yeah, we were talking about that earlier at a uh Phal and I had a meeting earlier this week with a gentleman who was a pilot in the military, got out, got his law degree, and then was an operations guy for like for like five years at a uh manufacturer.
• 13:45
  SPEAKER_01: And now as we were talking, he was walking through and scoping out um some things for a mutual client.
• 13:53
  SPEAKER_01: And it's like to me, I'm like, that's a forward-deployed engineer.
• 13:57
  SPEAKER_01: He's looking at change management, he's mapping out the workflows.
• 14:00
  SPEAKER_01: Does he have to know everything about AI?
• 14:03
  SPEAKER_01: No, but if it at least he understands the tools.
• 14:06
  SPEAKER_01: The guy was a pilot, right?
• 14:08
  SPEAKER_01: He couldn't build an airplane, but he could fly it.
• 14:10
  SPEAKER_01: He understood the systems and things behind it.
• 14:13
  SPEAKER_01: So I think that that chasm is is should be filled by people that have operational experience and then have the at least the capacity to be able to understand the tech.
• 14:25
  SPEAKER_01: Right.
• 14:26
  SPEAKER_02: Really hard to teach somebody business.
• 14:29
  SPEAKER_02: You know, these are things that you learn by doing business.
• 14:32
  SPEAKER_02: Like, you know, you can go to business school and learn about that stuff, and you know, I think I always tell people I think maybe the best business lesson you can learn is like, do you guys remember Dino Park Tycoon?
• 14:44
  SPEAKER_02: That might be really specific to like my age group.
• 14:47
  SPEAKER_02: Sorry, I don't.
• 14:48
  SPEAKER_02: It was a game we used to play in school.
• 14:50
  SPEAKER_02: And it was like you would go, it was on the computer, and you'd go and you'd build a you build a dino park.
• 14:54
  SPEAKER_02: Okay, and you would go and you'd you know, you got a cooler dinosaur, more people would come, but if more people come, you gotta feed more people, and then you need more bathrooms.
• 15:00
  SPEAKER_02: Like, I kind of feel like we should probably play this in business school because like at the end of the day, it's just deliver value to your customers, right?
• 15:06
  SPEAKER_02: And it's these are hard lessons to learn, and you learn them through going and doing the work and learning about business and sitting and listening to people's challenges.
• 15:16
  SPEAKER_02: And you gotta take those people and like you can learn the tech.
• 15:19
  SPEAKER_02: Like you can most people can learn the tech, but you need somebody that can go and hear somebody talk about their business and think that's a challenge that I could solve.
• 15:30
  SPEAKER_02: And like that is a skill that we're we're certainly not historically teaching software developers, right?
• 15:36
  SPEAKER_02: Like, this is a group of folks that it's really gonna change their role.
• 15:40
  SPEAKER_00: Well, okay, so if I could double down on that, or at least give it uh uh A16Z had a really good newsletter this last week, and they were talking about the old loop and the new loop.
• 15:49
  SPEAKER_00: So, so you know, I'll just speak for myself.
• 15:52
  SPEAKER_00: You know, trained software engineer, did plenty of school, spent plenty of time executing an industry on it.
• 15:58
  SPEAKER_00: But it's like the loop being there's an idea that gets translated to a spec that gets translated to a design of what you know, what are we gonna build, the UML, that kind of stuff.
• 16:10
  SPEAKER_00: That gets built, tested, and then you ship something, and then you learn at the very last step is you get some feedback from the field.
• 16:17
  SPEAKER_00: Yeah.
• 16:18
  SPEAKER_00: And the flip of like now, it's like you take idea, so you're with the business.
• 16:23
  SPEAKER_00: You see, there's an idea in the business, you're sitting with them embedded, you go right to build.
• 16:28
  SPEAKER_00: Hey, hey, you know, we can actually create something right now that we can show you.
• 16:34
  SPEAKER_00: Maybe it doesn't solve the problem, it certainly isn't production ready.
• 16:38
  SPEAKER_00: But what if, what could we do?
• 16:40
  SPEAKER_00: And we can do it right here on our desk with you, sitting at your desk with you, and and we can show what can be done almost an art of a possible, like next to you.
• 16:51
  SPEAKER_00: And and and the change in that loop of like idea to build something and play with it and iterate there with the business, and then and then when you have something, get it into the enterprise or you know, select the right route to get it deployed into the enterprise, but just how much of a shift that is.
• 17:08
  SPEAKER_02: Well, it's a huge shift, but it's so important, right?
• 17:11
  SPEAKER_02: Like it is so incredibly valuable.
• 17:14
  SPEAKER_02: If you go and try and build it without the people there, you know, I I always go with the analogy of like if I go and ask, like, what are the steps it's gonna take to make a cake, right?
• 17:24
  SPEAKER_02: I bake a lot of cakes.
• 17:25
  SPEAKER_02: I used to make all of our birthday cakes, and then we got too many people, and I'm not baking more than 40 cakes a week.
• 17:30
  SPEAKER_01: So full-time job, and then you forgot someone's birthday, and oh god, oh my god, no, that HR issue.
• 17:35
  SPEAKER_02: Yeah, we can't.
• 17:36
  SPEAKER_02: So, you know, but I could I could tell you how to bake a cake, but the steps I'm gonna write down, they're not gonna be enough for you to be able to bake a cake because there's like things that in my head I'm like, oh, that's obvious, right?
• 17:46
  SPEAKER_02: You just that's the next step, but you don't write those down because you think that they're obvious.
• 17:52
  SPEAKER_02: If you watch somebody do their job, you're like, oh, why'd you do that?
• 17:57
  SPEAKER_02: Oh, it's this thing, and then you understand like what's actually happening.
• 18:00
  SPEAKER_02: If you just try and go off of here's the five steps they told me about, like you're gonna build the wrong thing, it's not gonna solve the problem.
• 18:06
  SPEAKER_02: And that's where we go back to the MIT report that's like, oh, five percent of things ever succeed.
• 18:12
  SPEAKER_02: And it's because we're not taking the time to build out like what's actually solving the problem and what is the whole scope of that problem.
• 18:19
  SPEAKER_01: So obviously, Katie is our star forward deployed resource slash engineer.
• 18:25
  SPEAKER_01: You've been on dozens, 20, 30, 40, however many, getting with the customer solutioning from the standpoint of listening and then bringing it back and and then putting putting things into production.
• 18:39
  SPEAKER_01: We haven't even talked about maintenance yet, right?
• 18:41
  SPEAKER_01: Support and maintenance.
• 18:42
  SPEAKER_01: We haven't gotten there yet, but different topic.
• 18:44
  SPEAKER_01: Yeah.
• 18:45
  SPEAKER_01: But maybe um share a recent opportunity.
• 18:49
  SPEAKER_01: That's the other thing is she likes to get so forward deployed, and we have such cool clients that she couldn't get a ride in an F-35.
• 18:58
  SPEAKER_01: But maybe I know that was your that was heartbreaking.
• 19:01
  SPEAKER_01: I think there's only one like like one seat.
• 19:02
  SPEAKER_01: I told them I'll hang off the wings, but they said apparently you're not allowed to do that.
• 19:07
  SPEAKER_01: But what's uh anecdotally, or or maybe kind of yeah, give get an idea of what a forward deployed resource would look like in a day-to-day.
• 19:15
  SPEAKER_02: So okay, I'm gonna give you one that is gonna make me sound like a real, you know, idiot because I don't always have the common sense.
• 19:23
  SPEAKER_02: And so somebody comes to us and they say, Hey, I'm a I'm a floor broker.
• 19:27
  SPEAKER_02: I would love your help.
• 19:29
  SPEAKER_02: And I'm like, Oh, they must sell carpets, something like that.
• 19:32
  SPEAKER_02: Uh no, they they do financial trading.
• 19:34
  SPEAKER_02: Um, you know, but that's not my world.
• 19:37
  SPEAKER_02: You know, I understand that stocks are a thing.
• 19:39
  SPEAKER_02: I don't know anything more than you buy and sell and shout.
• 19:43
  SPEAKER_02: Um, so they said, Hey, come up for a day.
• 19:46
  SPEAKER_02: Just come and sit with us and watch.
• 19:49
  SPEAKER_02: And I tried to read about it beforehand.
• 19:52
  SPEAKER_02: I tried to go up there and like I was like, oh, I'm gonna know all about that.
• 19:55
  SPEAKER_02: I know nothing, right?
• 19:56
  SPEAKER_02: And you go up there and like by the end of the day, I'm talking to people about jelly rolls and you know, all this trading terms, and like it's this is great.
• 20:04
  SPEAKER_02: You learn so much.
• 20:05
  SPEAKER_02: And like while we're there, one of the problems they have is like a math problem.
• 20:08
  SPEAKER_02: And I think the next day I sent them over.
• 20:10
  SPEAKER_02: I'm like, can we check if this number would match?
• 20:12
  SPEAKER_02: Because uh, you know, it's for the love of the game, like math side, it's awesome.
• 20:17
  SPEAKER_02: But you know, from that changing skill side, I Ted, you can tell the story if you want of like what what did I say when I started here?
• 20:28
  SPEAKER_01: You said that you would nut all you would work with us on the condition that all you would do is sit and do math.
• 20:35
  SPEAKER_01: And you didn't want to talk to anybody.
• 20:36
  SPEAKER_02: Ever make me talk to anybody and I'll come work for you.
• 20:38
  SPEAKER_01: That's right.
• 20:39
  SPEAKER_01: And then the next day we made you talk to a client and you said, Hey, what's up with that?
• 20:44
  SPEAKER_01: And we said, Yeah, that's you gotta talk to people.
• 20:46
  SPEAKER_02: But we're gonna have to do that with software developers.
• 20:49
  SPEAKER_02: Like, you're going to have to get to this point where honestly, it's my favorite part of my job, right?
• 20:54
  SPEAKER_02: I get to go and I get to see the amazing things that people do, right?
• 20:58
  SPEAKER_02: Because like I I talk to software developers every day.
• 21:01
  SPEAKER_02: It's not really that interesting to me.
• 21:03
  SPEAKER_02: Like, I go places and people want to talk because they're like, you know about this AI thing.
• 21:07
  SPEAKER_02: And I'm like, doesn't everybody like everybody I work with knows about AI.
• 21:12
  SPEAKER_02: But when you go to other places, I have no idea.
• 21:14
  SPEAKER_02: We learned about furnaces and heat treating two weeks ago.
• 21:17
  SPEAKER_02: It's like this stuff is incredible.
• 21:19
  SPEAKER_00: Yeah.
• 21:20
  SPEAKER_00: And I'll I'll throw I mean it's sort of related, but I'll go back, like, you know, I I mean to me, if I could just summarize as part of like embedding in the business is so much so much um, you know, able to to perceive so much information that you're not gonna get from reading or hearing or maybe getting a spec given to you.
• 21:39
  SPEAKER_00: I don't know if you remember this one, Ted, but even you know, we went to a um rebar manufacturing, right?
• 21:48
  SPEAKER_01: Columbia, South Carolina.
• 21:49
  SPEAKER_00: Right.
• 21:49
  SPEAKER_00: And they had a real legitimate issue, which was they make rebar and it's a very it's dangerous, honestly.
• 21:56
  SPEAKER_00: It's it's its own process.
• 21:57
  SPEAKER_00: And they had a really legitimate issue of counting the rebar and making sure they're delivering the right amount, which which when I heard that, I'm like, hey, this isn't that hard, trust me.
• 22:08
  SPEAKER_00: Counting.
• 22:09
  SPEAKER_00: But you see the operation of the environment and it's dangerous, and there's time constraints, and you got hot rebar right behind you, and you got trucks waiting for you, and it's like I and you have a very harsh environment.
• 22:20
  SPEAKER_00: Like, we can't stop and put this high-res camera right here and this and that.
• 22:24
  SPEAKER_00: And it's like I wouldn't like how many times have we run into this?
• 22:28
  SPEAKER_00: You sit in this, you sit with the customer and you see it, and it's like, wow, I didn't like all this context.
• 22:36
  SPEAKER_00: I we would have been over here when we needed to be over here.
• 22:39
  SPEAKER_00: Yeah, and that's just the moral of the story.
• 22:41
  SPEAKER_02: Okay, my favorite one of those by far is do you guys remember snake vision?
• 22:48
  SPEAKER_02: We were working with the poison control center, and it was like, what is cool stuff we could do for them?
• 22:53
  SPEAKER_02: So they have like an awesome new tech.
• 22:54
  SPEAKER_02: And it's like people call in and they go, I got bitten by a snake, and they've got to figure out what snake it is, and why don't you just take a picture of the snake and then send it into the poison control center?
• 23:04
  SPEAKER_02: They can use computer vision and it will tell you what kind of snake it is.
• 23:07
  SPEAKER_02: It's like, well, that's great, except for we really encourage people to not get up close to poisonous snakes.
• 23:12
  SPEAKER_02: Like, you know, don't take a picture of it, leave the snake alone and go away.
• 23:16
  SPEAKER_00: And it's like no snake selfies.
• 23:18
  SPEAKER_02: Yeah, from a but from a comp size standpoint, you go, oh, obviously, this is solving the problem, and it's totally not solving the problem.
• 23:24
  SPEAKER_00: Right.
• 23:25
  SPEAKER_00: But it's a cool use of technology.
• 23:26
  SPEAKER_00: Like that that's another that's another trap, right?
• 23:28
  SPEAKER_02: That's one of those 95% where you build something awesome that doesn't go anywhere because it doesn't solve a business problem.
• 23:34
  SPEAKER_00: That's exactly right.
• 23:35
  SPEAKER_00: That's exactly right.
• 23:37
  SPEAKER_00: Well, let's let's wrap it there.
• 23:38
  SPEAKER_00: I will say we had a few folks reach out at the event last night, but then also through LinkedIn to kind of connect outside of this venue.
• 23:45
  SPEAKER_00: Happy to happy to make that connection, so please keep those coming.
• 23:49
  SPEAKER_00: We're if you can't tell, we're a very social bunch, love to love to hang out and and just hear what other groups are doing or seeing.
• 23:56
  SPEAKER_00: Um but until next time, keep learning, even when it's unsupervised.