Latest / Automated with Brian Heater / When AI Leaves the Demo Stage: Kence Anderson on What Lasts

Transcript

When AI Leaves the Demo Stage: Kence Anderson on What Lasts · Automated with Brian Heater

• 0:00
  SPEAKER_00: I love the fact that startups just live and die by the demo.
• 0:04
  SPEAKER_00: You know, Steve Jobs was an absolute master at this.
• 0:07
  SPEAKER_00: I mean, just absolute master at this.
• 0:09
  SPEAKER_00: He wasn't just talking, like he was showing, you know, the thousand songs in your pocket.
• 0:13
  SPEAKER_00: Bam, there's the iPod.
• 0:15
  SPEAKER_00: I remember the demo for the MacBook Air.
• 0:18
  SPEAKER_00: Pulled it out of a vanilla envelope.
• 0:21
  SPEAKER_00: I love that kind of stuff where it's like a real demonstration of a technology and proving how it can drive business value.
• 0:28
  SPEAKER_00: Oh yeah, I love it.
• 0:38
  SPEAKER_01: Hello everyone, and welcome back to another episode of Automated.
• 0:42
  SPEAKER_01: I am Brian Heater, the managing editor of the Association for Advancing Automation.
• 0:47
  SPEAKER_01: Uh, first off, thanks everyone who attended our first live podcast recording with Michael Taylor at last week's Business Forum in Orlando.
• 0:58
  SPEAKER_01: For those of you who missed it or just want to uh relive the thing with fancy video editing, we will be posting that up here in a few days.
• 1:07
  SPEAKER_01: This week's main event is Ken's Anderson.
• 1:10
  SPEAKER_01: And a quick note to say that uh one of the downsides of recording podcast episodes well in advance is that sometimes uh a company will go and and rebrand, will change its name before you edit and actually post the thing.
• 1:26
  SPEAKER_00: And now we're launching a Mesa, Advanced, Modular, Enterprise Systems for Autonomy.
• 1:32
  SPEAKER_01: Thanks to Kent.
• 1:33
  SPEAKER_01: Always a joy to speak with.
• 1:35
  SPEAKER_01: Uh, this is actually probably the first time I managed to stop myself from asking him where he gets those wooden sunglasses.
• 1:41
  SPEAKER_01: Um, always, always a pleasure speaking with Kent.
• 1:44
  SPEAKER_01: Uh, thanks to you for tuning in.
• 1:46
  SPEAKER_01: As always, please like and subscribe, and we will see you in about 45 minutes.
• 1:52
  SPEAKER_01: One of the biggest challenges in industrial AI right now isn't building a single smart system, it's getting a bunch of them to work together.
• 2:00
  SPEAKER_01: Ken Sanderson has spent years doing exactly that.
• 2:03
  SPEAKER_01: He's helped companies like Shell and Delta design AI agents that actually collaborate, learn, and adapt.
• 2:09
  SPEAKER_01: And now he's teaching others how to do it in a new on-demand course called Designing Industrial AI Agents from the Association for Advancing Automation.
• 2:19
  SPEAKER_01: It's just a few hours long, and you can take it at your own pace.
• 2:23
  SPEAKER_01: And by the end, you will know how to build multi-agent systems that think and act like a real team.
• 2:29
  SPEAKER_01: Check it out at automate.org slash AI agents.
• 2:33
  SPEAKER_01: That's automate.org slash AI agents.
• 2:38
  SPEAKER_01: What one of the things that I really appreciate about you is we were doing the programming for um or uh a focus event recently.
• 2:48
  SPEAKER_01: And um I I think it was Robert who said, we can only give Kent two hours.
• 2:55
  SPEAKER_01: And he's the word only.
• 2:57
  SPEAKER_01: Like we we've got to we can only slot him in two hours, which would be an extraordinary amount of time for just about any other person who we could possibly book.
• 3:06
  SPEAKER_01: I mean, you're able to speak for an extraordinary amount of time.
• 3:11
  SPEAKER_01: You could have you could have been a professor.
• 3:13
  SPEAKER_00: Well, I get I gained a lot of teaching, thank you.
• 3:15
  SPEAKER_00: I gained a lot of teaching sensibilities for my uh from my father, who was a teacher.
• 3:20
  SPEAKER_00: He was a he was a professor, he taught second grade all the way through college at different points of his career.
• 3:26
  SPEAKER_00: Um and so I grew up with a teacher, um, understanding what it's like to be around someone who can really teach.
• 3:33
  SPEAKER_00: I'm amazed with the A3 courses that we've done, how many people sit uh for so long.
• 3:39
  SPEAKER_00: I think part of it is it's a topic, you know, AI, AI agents, um AI that can actually do things in the physical world um or in the enterprise.
• 3:52
  SPEAKER_00: I think it's it's interesting enough to people because of the value that it could potentially drive.
• 3:57
  SPEAKER_00: Then it's a matter of can you explain things in a way that people can you know actually understand, which is most of what I try to do.
• 4:05
  SPEAKER_01: But again, just the pure stamina.
• 4:07
  SPEAKER_01: I mean, is is that something that you kind of is that you learned from osmosis from from watching your father?
• 4:12
  SPEAKER_01: Did he actually like impart anything specific on you?
• 4:16
  SPEAKER_00: The stamina is just passion about the subject.
• 4:18
  SPEAKER_00: I love engaging with people, and our courses are always very engaging.
• 4:22
  SPEAKER_00: Uh, you know, people are sharing about what they're doing in their business, they're sharing about the problems.
• 4:27
  SPEAKER_00: I had one of my favorite moments actually at Automate.
• 4:30
  SPEAKER_00: We did the two four-hour classes.
• 4:33
  SPEAKER_00: That was exhausting, to be fair.
• 4:34
  SPEAKER_00: But at the end of the second four-hour class, this um this uh senior vice president from a major manufacturer stood up and basically said, I asked people what it, you know, give me some feedback and some impressions of the class, and he stood up in front of everyone.
• 4:48
  SPEAKER_00: I said, Yeah, I thought this AI stuff was you know a bunch of baloney, use slightly different words, but you provided a framework that an engineer can use to understand AI.
• 5:03
  SPEAKER_00: And I took that as a huge compliment, as a really huge compliment.
• 5:08
  SPEAKER_00: But I would say it's really driven by passion for seeing people use this in real life, you know, in their businesses, and just passion for the subject.
• 5:18
  SPEAKER_01: I don't know how much more information you got from that specific person, but like what is it that could make an engineer think specifically that AI is a bunch of balode or whatever other word that person used, you know, in this current era?
• 5:32
  SPEAKER_00: I have gotten to speak with that executive quite a bit.
• 5:36
  SPEAKER_00: I think one is there's a lot of philosophy without practicality, a lot of abstraction people jumping straight to philosophy, but an engineer is about is about building things that work in real life.
• 5:46
  SPEAKER_00: Like, remember how AI has always worked.
• 5:49
  SPEAKER_00: Like, and I think I might have said this at the session that we were in together.
• 5:53
  SPEAKER_00: AI since the beginning, I mean, the term AI was coined in 1951 at a famous conference at Dartmouth, artificial intelligence being the term.
• 6:02
  SPEAKER_00: Um, but ever since then, there have been these very promising technologies that AI researchers have developed, and then the public has said, okay, this is it.
• 6:12
  SPEAKER_00: This is gonna change everything.
• 6:14
  SPEAKER_00: In the 60s and 70s and 80s, there's a technology called expert systems.
• 6:18
  SPEAKER_00: We now call them rule-based systems, but they were called expert systems, and people thought of them the exact same way that we think about generative AI today.
• 6:26
  SPEAKER_00: It's gonna replace doctors, definitely, gotta replace lawyers, definitely.
• 6:30
  SPEAKER_00: Um, you know, it's it'll become conscious and it'll you know lead the robot revolution.
• 6:35
  SPEAKER_00: And then when there's that trough of disillusionment and the hype dissipates, the bubble bursts, however you think of it, then it's engineers that go and pick up the pieces and say, yeah, but there's valuable technology here.
• 6:51
  SPEAKER_00: And then they build those technologies into everything we use, which is why now to us, rule-based systems aren't even considered AI or anything special.
• 7:00
  SPEAKER_00: They're just built into the fabric of everything we use.
• 7:03
  SPEAKER_00: And so if you break down technologies in terms of their strengths, their weaknesses, you know, if it's a tool in a toolbox, which tool is it and uh what kind of task should you use it for, then an engineer can digest it and pick it up and go, oh yeah, that's it, just like the nail or the screw or the bolt or the gear or the lever.
• 7:25
  SPEAKER_00: It's a building block for me to build these engineering systems.
• 7:28
  SPEAKER_00: Thank you for presenting it that way, and now I can go use it in my business.
• 7:32
  SPEAKER_01: This is really interesting because I think that you've you've framed this in a way that I haven't really approached it from.
• 7:39
  SPEAKER_01: Like usually when I have this conversation and I and I've been speaking to somebody who has been at this a lot longer than even I, you know, either you or I have, I ask, you know, is this this time different?
• 7:50
  SPEAKER_01: You know, which means is is the outcome going to be different?
• 7:52
  SPEAKER_01: Usually I'm talking about robotics, but uh the the question is like, are the robots themselves going to be more useful?
• 8:00
  SPEAKER_01: Or is the AI itself going to be more useful?
• 8:02
  SPEAKER_01: But maybe the way I should be framing it, and kind of what you're getting at, is um is the penetration of the that it's had in the public deeper?
• 8:13
  SPEAKER_01: And are more people going to be disillusioned by seeing what the technology can't actually do?
• 8:20
  SPEAKER_00: I will answer the question directly.
• 8:22
  SPEAKER_00: The outcome will be exactly the same, which will be net positive, which is there will be a trough of disillusionment.
• 8:29
  SPEAKER_01: But it'll be a larger trough, right?
• 8:31
  SPEAKER_01: I mean, it's just like there are more people.
• 8:33
  SPEAKER_00: But the engineers, there's so there will be deeper penetration, but then engineers will pick it up and build it into everything we do.
• 8:40
  SPEAKER_00: I mean, expert systems are still used today to run the entire banking system, you know, it's entire sections of the banking system.
• 8:49
  SPEAKER_00: They're used in engineering, you know, significantly.
• 8:51
  SPEAKER_00: And so what we'll find in I don't know how many years, you know, two, five, ten years, is people won't talk about generative AI or LLMs that will just be built into um what we do, and you'll use LLMs when you need it, and you'll use an expert system when you need it, and you'll you'll build these uh sophisticated hybrid systems, just like a car.
• 9:14
  SPEAKER_00: I mean, imagine a car or a plane.
• 9:15
  SPEAKER_00: I mean, imagine if you walk into a car and it was just one piece.
• 9:18
  SPEAKER_00: Like, no way.
• 9:20
  SPEAKER_00: Um, you know, there's wheels and there's and there's gears and there's there's a transmission, and there's all these different things that serve different parts.
• 9:27
  SPEAKER_00: And that's what was really fascinating to me back in 2017, 18, 19, when I was traveling all over the world, going to all sorts of factories and mines and steel mills, things like that, and talking to engineers about AI, and I realized, and I should have known this, I was trained as a mechanical engineer, I should have known this, that all the engineer is going to see is the tools.
• 9:47
  SPEAKER_00: And so if you uh eventually it all boils down to is this a tool, this is a tool, is it useful?
• 9:56
  SPEAKER_00: What is it useful for?
• 9:57
  SPEAKER_00: And then it'll be combined into sophisticated systems.
• 10:00
  SPEAKER_01: It almost seems like this really inefficient system, though.
• 10:02
  SPEAKER_01: It's like we we have to let it become a big bubble, we have to let it break, we have to let it collapse, and then somebody's there to pick up all the pieces and put them back together.
• 10:11
  SPEAKER_00: That is an extremely inefficient system.
• 10:13
  SPEAKER_00: I was thinking about this, uh, Ryan, um, and I saw an article um online that where someone was positing a very similar statement that you just made.
• 10:24
  SPEAKER_00: And I was wondering to myself, for like the 10 seconds that I thought about it, why that is.
• 10:29
  SPEAKER_00: And maybe it's because in the past, it was either governments or large laboratories that are designed specifically for research.
• 10:39
  SPEAKER_00: So they didn't have to make anything practical.
• 10:41
  SPEAKER_00: So um they they were they were developing scientific inventions and they did well, they did the research and then they did the development, which is at least bringing it to a place where it could start being tested in the real world.
• 10:55
  SPEAKER_00: Um, and maybe as there's less and less of those, then there has to be a substitute.
• 11:01
  SPEAKER_00: And you know, venture capital picks up the um the slack and goes, okay, well, then we're gonna have to develop technology this way.
• 11:09
  SPEAKER_00: And we see in the public sphere this new mode of technology development.
• 11:14
  SPEAKER_00: I call it the research to PR pipeline, where you have these AI labs, which are now mostly part of large corporations, but in a different way than like a Bell Labs.
• 11:25
  SPEAKER_00: So, you know, like a Bell Labs or Xerox Park, which they were Xerox and Bell were certainly large corporations, but it really was almost that government lab model where no, those guys are just there to do research.
• 11:38
  SPEAKER_00: And then you see these, now you see these papers being published, and there's an expectation that value has been generated already just by the fact that this research has been done, but that's not actually how it works.
• 11:52
  SPEAKER_01: There are a lot of VCs out there.
• 11:54
  SPEAKER_01: I I won't name any names, but there are a few that I can think of who had one really big bet payoff 20 to 15 years ago and are still able to coast on that.
• 12:04
  SPEAKER_01: And and that that that kind of keeps us going, right?
• 12:07
  SPEAKER_01: I mean, they're able to continue to throw meth funds a lot of things.
• 12:11
  SPEAKER_01: Yeah, they can keep throwing money at at long shots, uh, with you know, which is good.
• 12:15
  SPEAKER_01: I I like I like supporting long shots, but that just keeps you know the bubbles inflating on and on.
• 12:21
  SPEAKER_00: But maybe the devil's advocate position is that when something hits, it does change everything.
• 12:27
  SPEAKER_00: I mean, you know, you think about Google or even Uber, you know, for example.
• 12:33
  SPEAKER_00: Yeah, we saw something, I think, similar for self-driving cars.
• 12:37
  SPEAKER_00: I remember in 2017 the company that I was working for, and we were working on autonomy, but we specifically decided not to go into self-driving cars because we felt like there's just too much froth and too much, you know, hype here.
• 12:51
  SPEAKER_00: But also we decided, and this was the real reason, that you got to prove out autonomy first in a more controlled space.
• 12:58
  SPEAKER_00: We figured, well, factories have more controlled space than out of the open streets of San Francisco.
• 13:03
  SPEAKER_00: And then we saw you know huge amounts of money go in, and we saw a lot of companies, even some of the companies that I thought were going to be the winners fail.
• 13:11
  SPEAKER_00: And then a couple emerge.
• 13:14
  SPEAKER_00: And now, if you go into San Francisco, you will see driverless Waymo taxis everywhere, absolutely everywhere.
• 13:21
  SPEAKER_00: It's not even considered out of the ordinary.
• 13:23
  SPEAKER_01: It's great to see certain institutions like TRI, like RAI, these uh these research institutes that can exist and be um you know supported by large corporations that are really, at least for the time being, do exist to do to do pure research.
• 13:39
  SPEAKER_01: Because as somebody whose job it has been to be a reporter, to be a you know consumer electronics reporter, I've often gone into these labs, seen this really interesting research, and then asked asked the question, okay, like when what when's the product coming out?
• 13:53
  SPEAKER_01: And that really is, thankfully, in a lot of these cases, the wrong question to be asking.
• 13:58
  SPEAKER_01: I mean, it's great that there are places that can really do research for research's sake.
• 14:03
  SPEAKER_00: So you like the independence of it, which I understand, especially coming from a journalistic background, independence is important.
• 14:10
  SPEAKER_01: I like that not everything is necessarily being created um specifically with the end goal of a specific product at some point.
• 14:18
  SPEAKER_00: Yeah, that makes perfect sense.
• 14:19
  SPEAKER_01: But obviously, you know, it's probably not a uh a super self-sustaining model, I would say.
• 14:26
  SPEAKER_00: It might not be, but we'll see.
• 14:28
  SPEAKER_00: There's a balance, there's there's a there's a pendulum because it's there are trade-offs to having your the research product, your innovation driven by the need to turn that into a profitable product quickly, versus you know, you know, longer horizon innovation that may take decades, and it will be really interesting to see.
• 14:53
  SPEAKER_00: And I think different countries have different models of that.
• 14:55
  SPEAKER_00: You know, so some countries have more state-sponsored innovation, some countries have more corporate-sponsored innovation.
• 15:01
  SPEAKER_01: And maybe this is what you were getting at a little bit before as well, because the flip side of what we're seeing of that is the attempted productization of everything and the over-productization of everything, and the attempt to kind of I was watching some uh I can't remember what talk you were giving, but but you know, it was kind of around co-pilots, and it was okay, we have this really what we we've all sat here and agreed that Dreadden of Ray, that LLMs are really incredible technology, that they're doing these things that uh you know are are really on the face of them unbelievable.
• 15:33
  SPEAKER_01: So let's try to jam as many square pegs into as many round holes as humanly possible.
• 15:39
  SPEAKER_00: I hear a lot of railing on the problem, but not a lot of discussion of the solution, not by you, but uh at large.
• 15:46
  SPEAKER_00: You know, folks saying, well, everyone's you know looking for a you know, has a hammer and is looking for nails with a particular technology.
• 15:53
  SPEAKER_00: Yeah, but what's missing again is that engineering approach.
• 15:56
  SPEAKER_00: The engineering approach says, okay, you know, the analogy that I'd love to give is the the nail, the screw, and the bolt.
• 16:03
  SPEAKER_00: They all do similar things that fasten two pieces of things, the two pieces of material together.
• 16:08
  SPEAKER_00: Um when the screw came out, it's not like the nail was rendered obsolete.
• 16:13
  SPEAKER_00: They have different, they're they're different.
• 16:16
  SPEAKER_00: A nail is cheap and easy to drive.
• 16:19
  SPEAKER_00: So in your house, in your home, where you need 10,000 fasteners, it's the obvious choice.
• 16:25
  SPEAKER_00: Like it would be a complete waste of money to use screws because you don't need that kind of force to join the things together.
• 16:32
  SPEAKER_01: And they're more time efficient.
• 16:34
  SPEAKER_00: The bolt is the extreme.
• 16:35
  SPEAKER_00: That's when you need an extreme amount of force with a bolt and a nut.
• 16:38
  SPEAKER_00: Um, but that doesn't render the screw up or the nail obsolete.
• 16:42
  SPEAKER_00: They're different things for different for different tasks.
• 16:45
  SPEAKER_00: So I think when you, you know, again, I think it boils back down to that kind of engineering mindset of what you know, what is it and what and what should it be used for.
• 16:55
  SPEAKER_01: So so you're a kid growing up in Bangor, Maine.
• 16:58
  SPEAKER_01: You go to school to become an engineer.
• 17:00
  SPEAKER_01: What was your conception of what an engineer is or does going into university?
• 17:07
  SPEAKER_00: I loved planes and space growing up.
• 17:10
  SPEAKER_00: So my dad used to take us, me and my sister, to air shows all the time.
• 17:13
  SPEAKER_00: I got to see all the great airplanes, you know, that were around in my lifetime.
• 17:17
  SPEAKER_00: I got to see the SR-71, I got to see the U-2, I got to see the B-2, the F-117, you know, all these great uh planes, you know, the great fighter planes, and I was just fascinated.
• 17:27
  SPEAKER_00: I was like, oh my gosh.
• 17:28
  SPEAKER_00: And so I loved, you know, things that move.
• 17:31
  SPEAKER_00: Now I wish some someone would have just told me that's what mechanical engineering is.
• 17:36
  SPEAKER_00: Uh, you know, the engineering of things, you know, how to design and build things that that move.
• 17:41
  SPEAKER_00: But what actually happened was my guidance counselor, the junior year of high school, said, Well, you're good at math, you should do in science, you should do engineering.
• 17:48
  SPEAKER_00: So I was like, oh, sure.
• 17:50
  SPEAKER_00: My first innovation experience as an engineer, I was just so fortunate.
• 17:56
  SPEAKER_00: I landed an internship at Ford Motor Company as a freshman engineer.
• 18:00
  SPEAKER_00: So I've essentially taken hardly any coursework.
• 18:03
  SPEAKER_00: You know, I'm worth nothing as a true engineer.
• 18:06
  SPEAKER_00: And in Ford and other companies, they really invest in folks.
• 18:11
  SPEAKER_00: And I went out to Ford Motor Company and they paid me, you know, to spend a whole summer.
• 18:15
  SPEAKER_00: At first, I get there and you know, they're busy.
• 18:17
  SPEAKER_00: This was a group that had just invented this new, you know, car park, this new thing, so air conditioning compressor.
• 18:22
  SPEAKER_00: And they're busy working, and I and I didn't have much to do for the first two weeks.
• 18:26
  SPEAKER_00: And I said to my boss, I said, if you don't give me something to do, I'm going home.
• 18:30
  SPEAKER_00: And he said, sure, why don't you work with these two MIT grad students?
• 18:35
  SPEAKER_00: And it was me and these two MIT grad students inventing a new process for how to manufacture things.
• 18:44
  SPEAKER_00: So this air conditioning compressor used to have to make a mold out of sand.
• 18:49
  SPEAKER_00: It's called sand casting.
• 18:51
  SPEAKER_00: And then if you wanted to make a prototype part, like it's not ready for the factory yet because you're still testing things, and so you'd make a mold out of sand, it would take like six weeks to get a part to test.
• 19:00
  SPEAKER_00: And what we did, the three of us, the two MIT grad students and freshman engineer from Howard University, we were the first people at Ford Motor Company to use CAD to design how a tool could cut the metal, and we sent the part over a phone line to the CNC machine on the floor in the factory in Connorsville, Indiana, and we get a part in six hours instead of six weeks.
• 19:27
  SPEAKER_00: And I got to write a paper about it, and we're the first people to ever do it at Ford.
• 19:32
  SPEAKER_00: And I learned something, I was like, oh my gosh, like one engineering is about experimentation.
• 19:38
  SPEAKER_00: Because the next summer I was working with uh, I was actually supervising some PhDs, which is amazing.
• 19:44
  SPEAKER_00: I was sophomore, a mechanical engineer supervising PhDs on how to design that same compressor, and it was very experimental.
• 19:51
  SPEAKER_00: The second is you don't have to have any particular license to innovate.
• 19:56
  SPEAKER_00: Like you can engineer and innovate without you know, quote unquote credentials.
• 20:01
  SPEAKER_00: I wasn't the MIT grad student, you know, I wasn't the PhD from you know the University of Michigan, but yet, especially in the second summer, I was making these design decisions and we did some great work.
• 20:12
  SPEAKER_00: Engineering was challenging, but I was very uh fortunate and grateful that I ended up choosing mechanical engineering because it essentially teaches you how to think and teaches you how to thaw solve problems, teaches you how to think in a systems way.
• 20:25
  SPEAKER_00: Everything can be decomposed into pieces, everything.
• 20:29
  SPEAKER_00: And that's the best way to design things.
• 20:31
  SPEAKER_00: And if you think of things as a system with component parts, you know, I found this has been so helpful as I've been in AI for this last decade that not everyone thinks that way.
• 20:44
  SPEAKER_00: And I remember working at a large tech company, and some of the computer scientists definitely didn't think that way.
• 20:50
  SPEAKER_00: I remember this one particular conversation where I was like, um, you know, it's just like in um it's just like in algebra when you know you you learn how to break a problem into pieces.
• 20:59
  SPEAKER_00: That's exactly what we should do for this AI.
• 21:01
  SPEAKER_00: And not only did the person disagree, the person looked at me and talked to me like, are you kidding?
• 21:09
  SPEAKER_00: Why would you ever do that with with code?
• 21:12
  SPEAKER_00: Because there's a very, there's a very different approach sometimes in software engineering.
• 21:16
  SPEAKER_00: It's an algorithmic approach can be very different.
• 21:18
  SPEAKER_00: Claude Shannon talked about this.
• 21:20
  SPEAKER_00: Claude Shannon was one of the godfathers of AI that was at that 1951 conference, where they that they coined the term AI.
• 21:27
  SPEAKER_00: He wrote this paper, it was either in 1950 or 1951, and he predicted this.
• 21:31
  SPEAKER_00: He said people are going to use AI to learn how to play chess, which is so uncanny because there's been so many breakthroughs in AI have been demonstrated on chess.
• 21:40
  SPEAKER_00: Alpha chess, which I think is 2016, 2015, what was the chess playing program that beat Gary Kasperov from um from IBM?
• 21:48
  SPEAKER_00: Deep Blue.
• 21:49
  SPEAKER_00: Deep Blue, exactly.
• 21:50
  SPEAKER_00: And you know, and even Alan Turing worked on uh worked on it.
• 21:53
  SPEAKER_00: But his point was he said some folks are going to develop this rule based system that's gonna try to capture codify.
• 22:00
  SPEAKER_00: The rules of chess, the strategies of chess as rules.
• 22:04
  SPEAKER_00: And the benefit of that is that the system is going to understand something about chess because there's expertise built into it.
• 22:10
  SPEAKER_00: The problem with that is you have to manage the programming of it, you're going to have to manage the rules, the exceptions of the rules, and that's what the expert system is, really.
• 22:17
  SPEAKER_00: But then he said, but there's going to be this other approach.
• 22:20
  SPEAKER_00: And the computer programmer is going to love it because they're going to just program one algorithm, and that algorithm is going to search options and then choose the best move.
• 22:28
  SPEAKER_00: And he said, the problem with that is that algorithm is going to have zero understanding of chess, but it's going to appeal to the programmers because it's one thing that you program.
• 22:38
  SPEAKER_00: And so you see how there's this kind of trade-off.
• 22:40
  SPEAKER_00: Um, and I'm generalizing, but I've seen it a lot where when you're doing sometimes physical engineering, you can have this, it demands this more systems and component approach, where computer science can have a more sometimes demand a more algorithmic approach.
• 22:58
  SPEAKER_00: It just depends on your perspective and it's useful in different ways.
• 23:02
  SPEAKER_01: It sounds like you're not easily intimidated.
• 23:05
  SPEAKER_01: I mean, I I'm trying to put myself in that situation where I'm a like a first-year undergrad.
• 23:10
  SPEAKER_01: There are these two MIT grad students, and I, you know, I'm just gonna kind of stand there and like watch them do their thing.
• 23:16
  SPEAKER_01: But you you you sort of inserted yourself in there and you felt like you were an equal party to the work they were doing.
• 23:22
  SPEAKER_00: My father, he he was an amazing man.
• 23:25
  SPEAKER_00: He was born in 1916.
• 23:27
  SPEAKER_00: And uh, you know, he he went through a lot of he had to become a teacher and an academic.
• 23:34
  SPEAKER_00: He got his uh he fought in World War II and then he got his uh master's degree from New York University.
• 23:39
  SPEAKER_00: And he talks about the different opposition and racial adversity that he faced, and I can't even believe it.
• 23:46
  SPEAKER_01: Like you know 1916 is not that far removed.
• 23:50
  SPEAKER_00: I got a lot of that determination from him, and he always he always would impress upon me.
• 23:55
  SPEAKER_00: He said, You should you should watch everything, you should learn everything you can.
• 23:59
  SPEAKER_00: He showed me how to experiment and how to he was he was a very innovative person, very innovative person.
• 24:04
  SPEAKER_00: It was also honestly those MIT gracious uh grassuits were very gracious.
• 24:08
  SPEAKER_00: Like they could have easily, easily treated me like get out of here.
• 24:12
  SPEAKER_00: Yeah, but they didn't.
• 24:13
  SPEAKER_00: So I think innovation it depends on both sides, you know, working.
• 24:17
  SPEAKER_00: But I'll tell you, when I was at that tech company, there were multiple times where I thought, you know, I'm just gonna leave.
• 24:23
  SPEAKER_00: Uh this isn't this isn't really working.
• 24:25
  SPEAKER_00: Uh no one's really listening.
• 24:27
  SPEAKER_00: That wasn't true, but seeing things from a different perspective can lead to great innovations, but it can also be a tough road to travel.
• 24:37
  SPEAKER_00: You know what I mean?
• 24:38
  SPEAKER_01: And you keep ending up at startups, you keep starting startups, and and that doesn't seem like a coincidence.
• 24:45
  SPEAKER_00: I love the fast pace of startups.
• 24:47
  SPEAKER_00: I love the fact that you get to work on so many different things, but it's really the it's Missouri's the show me state.
• 24:54
  SPEAKER_00: It's really the show me part of startups that I love, which is I see something, but I believe it.
• 25:00
  SPEAKER_00: I think there's legs to it, as I said.
• 25:02
  SPEAKER_00: I think there's some merit to this, and it's okay.
• 25:05
  SPEAKER_00: You don't have to see it right now.
• 25:08
  SPEAKER_00: You know, when you work at a much larger company, you're working on things that people have seen for a long time, and then they've handed down to different layers of the organization to execute.
• 25:18
  SPEAKER_00: Sometimes can see things a little bit farther, you know, ahead.
• 25:22
  SPEAKER_00: And I want to work on those things and I want the opportunity to show.
• 25:25
  SPEAKER_00: I love the fact that startups just live and die by the demo.
• 25:30
  SPEAKER_00: You know, Steve Jobs was an absolute master at this.
• 25:32
  SPEAKER_00: I mean, just absolute master at this.
• 25:34
  SPEAKER_00: He wasn't just talking, like he was showing, you know, the thousand songs in your pocket.
• 25:38
  SPEAKER_00: Bam, there's the iPod.
• 25:40
  SPEAKER_00: I remember the demo for the MacBook Air.
• 25:44
  SPEAKER_00: Pulled it out of a vanilla envelope.
• 25:46
  SPEAKER_00: I love that kind of stuff.
• 25:48
  SPEAKER_00: Where it's like, oh no, no, no, I'll show you.
• 25:50
  SPEAKER_00: Now, I think in AI there could be a little bit too much of a kind of manufactured demo.
• 25:54
  SPEAKER_00: And there's been some famous instances, you know, even in humanoid robotics, about completely manufactured demos, like demos that weren't even real at all.
• 26:01
  SPEAKER_00: Yeah.
• 26:02
  SPEAKER_00: But those, you know, disingenuous things aside, like a real a real demonstration of a technology and proving how it can drive business value.
• 26:11
  SPEAKER_00: Oh yeah, I love it.
• 26:13
  SPEAKER_01: This is something that's super interesting to me.
• 26:15
  SPEAKER_01: And you know, as I've gotten older and worked with more people and interacted with more people and just sort of read up more on, you know, just the brain and neurodivergence, um, you know, I've come to appreciate the value of the way, as you say, people approach uh ideas differently and the value that that uh people bring to different problems.
• 26:37
  SPEAKER_01: I'm wondering, is there something that you can see even like during the hiring process that you feel like give some insight into that kind of value?
• 26:47
  SPEAKER_01: Obviously, you're juggling a lot of things.
• 26:48
  SPEAKER_01: You know, there's personality, there's ego, but can you really tell at that early stage the first time you meet somebody, like, oh, this person's gonna bring something very unique to this role?
• 26:58
  SPEAKER_00: I remember the first time I ever hired someone, it was at the second startup I worked for, and there were many, many candidates that folks said, Oh, you should totally hire them because they had specific expertise for the job category and the thing that our startups do.
• 27:16
  SPEAKER_00: And I didn't choose any of them.
• 27:17
  SPEAKER_00: I chose a person that just graduated uh with a master's in psychology, and I was like, they were like, Well, why'd you choose her?
• 27:25
  SPEAKER_00: I was like, because she learns fast and she speaks up about what she thinks.
• 27:30
  SPEAKER_00: I was like, I'd rather have that any day at a startup, every single day.
• 27:34
  SPEAKER_00: And the earlier stage your startup is, the more that's important.
• 27:36
  SPEAKER_00: It's just it's a it's a bunch of smart, you know, smart people sitting there trying to figure things out.
• 27:42
  SPEAKER_00: So the ability to speak up about what you think and and which is to say the thing that's that that's different, the ability to articulate that difference, because I'm looking, I'm looking for the difference.
• 27:53
  SPEAKER_00: Um, I'm looking for the different opinion, I'm looking to match different things together.
• 27:58
  SPEAKER_00: If there are innovations about AI that you can find in my book, they come from me pulling from teaching social science, like actual studying social scientists talking about how to educate and um help humans acquire skills and the intersection of artificial intelligence and machine learning.
• 28:17
  SPEAKER_00: Um, it's almost like never the twin shall meet.
• 28:20
  SPEAKER_00: But when I talk to this people in this field, and I was like, oh, I could use that.
• 28:25
  SPEAKER_00: And then I talked to people in this field, oh, I could use that.
• 28:27
  SPEAKER_00: But that person has to articulate clearly what their different perspective is.
• 28:32
  SPEAKER_00: I would much rather take a uh kind of a fast learner, uh, an experimenter and a fast learner than a person that's good at a very particular thing.
• 28:42
  SPEAKER_01: It's interesting because I think to a certain extent, you and I are good at our jobs for the same reason, but again, we're approaching them from very different angles in that, as you said at the beginning of the conversation, as an educator, you're very good at distilling complicated concepts into simple language.
• 28:59
  SPEAKER_01: And that that's a lot of you know what I do, but I I largely do that through asking people questions and and then writing it.
• 29:06
  SPEAKER_01: And I'm coming, you know, you're you're coming at this through uh uh being uh you know an engineer at Howard, and I'm coming at this being a uh creative writing major at UC Santa Cruz, right?
• 29:17
  SPEAKER_01: Um but you know, these these are both these are both uh proven, you know, we're sitting here talking to each other to some extent, these have both proven to be um valuable, valuable routes.
• 29:28
  SPEAKER_01: Uh, you know, what is it do you feel in in your background that has made that has given you this ability to, as an engineer, as somebody who does, you know, I assume think complexly complexly?
• 29:42
  SPEAKER_01: Who does think about things in complex ways?
• 29:45
  SPEAKER_01: Um how are you able to break it down into concepts that are really digestible?
• 29:52
  SPEAKER_00: I do think journalists and teachers have one similar superpower, which is asking great questions.
• 29:57
  SPEAKER_00: The greatest teachers and the great greatest journalists ask great questions.
• 30:00
  SPEAKER_00: I think I remember seeing a Charlie Rose interview, and he said, and someone asked him, How are you so good at this?
• 30:05
  SPEAKER_00: He said, I'm the best question, you know, askers in the world have kind of saying is what he said.
• 30:10
  SPEAKER_00: Yeah.
• 30:11
  SPEAKER_00: So learning to ask really provocative questions, I think, is super important.
• 30:18
  SPEAKER_00: But also part of the way my brain works is I'm a categorizer.
• 30:23
  SPEAKER_00: You know what, Brian?
• 30:25
  SPEAKER_00: It took me a long time to make peace with what my superpowers were because they're so different from others.
• 30:31
  SPEAKER_00: People used to say to me all the time, why are you constantly using analogies?
• 30:37
  SPEAKER_00: Why?
• 30:37
  SPEAKER_00: Like ever since I was in, like, not a little kid, but ever since I was in middle school or high school, I was constantly using analogies.
• 30:45
  SPEAKER_00: But that's a huge part of teaching.
• 30:47
  SPEAKER_00: And I have an analogy for everything.
• 30:49
  SPEAKER_00: That's one of the ways that I can explain things in plain language without using a bunch of jargon, is you compare it to something that somebody else knows.
• 30:56
  SPEAKER_00: But I'm telling you, I almost stopped doing it when I was a young adult because people would ridicule me.
• 31:01
  SPEAKER_00: Enough of the analogies do.
• 31:03
  SPEAKER_00: Now, people don't do that anymore.
• 31:04
  SPEAKER_00: I don't know why.
• 31:06
  SPEAKER_00: Maybe it's because I became more comfortable with it and you know, used it, or maybe use it in a less obnoxious way.
• 31:11
  SPEAKER_00: I don't know.
• 31:12
  SPEAKER_00: That's one way my brain works is I'm constantly comparing things with analogies.
• 31:16
  SPEAKER_00: I'm looking, I'm I try to understand the world by going, oh, with models.
• 31:20
  SPEAKER_00: This is like this, this is like this, this is like this.
• 31:23
  SPEAKER_00: And also I'm a categorizer.
• 31:25
  SPEAKER_00: Like, oh my gosh, I'm a ferocious categorizer.
• 31:28
  SPEAKER_00: Everything belongs in a category.
• 31:30
  SPEAKER_00: And again, people used, even some of the scientists used to say, Oh, this categorization thing is you know, false, it's you know, a crutch for the human mind.
• 31:38
  SPEAKER_00: And no, no, no, no, no, no, no, no, no.
• 31:40
  SPEAKER_00: It's a it's a way of understanding the world.
• 31:42
  SPEAKER_00: And there are fuzzy boundaries between categories and gradients between categories and all that.
• 31:47
  SPEAKER_00: I get that.
• 31:48
  SPEAKER_00: The categories still exist.
• 31:49
  SPEAKER_00: The best engineers I know, the best teachers I know do the analogy thing.
• 31:54
  SPEAKER_00: The best engineers I know, they're they're compartmentalizers and categorizers, like you know, like nobody's business.
• 32:01
  SPEAKER_01: Was the pushback, do you think, because there's a certain degree of gatekeeping?
• 32:06
  SPEAKER_00: I think so.
• 32:06
  SPEAKER_00: Earlier in my career, I mean, I've been working startups for 20 years at this point.
• 32:10
  SPEAKER_00: Earlier in my career, I did have this feeling like, oh my gosh, if someone would just tell me what the rules of the game are, I see some game as being played, but I don't know what the rules are.
• 32:20
  SPEAKER_00: And if someone would just tell me what the rules were, I think that uh I would I would do a lot better.
• 32:24
  SPEAKER_01: In baseball, they call them the unwritten rules, which is all the kind of like the old guy stuff, yeah.
• 32:28
  SPEAKER_00: I think that might be.
• 32:29
  SPEAKER_00: And I also think that thank goodness that this didn't happen to me at Ford, but I think people sometimes reject innovation when it comes, innovation or leadership when it comes from an unexpected source.
• 32:44
  SPEAKER_00: Yeah.
• 32:45
  SPEAKER_00: I think it's part of the predictability in our life as we go, oh, these are the people that you know teach us things, and these are the people that are leading us somewhere, and these are the people that are innovating, and that that's the beauty of startups.
• 32:55
  SPEAKER_00: People pop up that you've never heard of ever before.
• 32:58
  SPEAKER_01: They don't even look like the other people that you relied on for innovation, and then but it's also like this person paid their dues, this person did their time, you know, it's it's this person's time now in line, things like that.
• 33:10
  SPEAKER_00: Exactly.
• 33:11
  SPEAKER_00: That's the other thing I love about startups, they're irreverent.
• 33:14
  SPEAKER_00: It's like, you know, when when Uber comes along, it's like dues, what dues?
• 33:19
  SPEAKER_00: Like, because it's not just a one-sided argument where someone's saying, You need you need to want this.
• 33:26
  SPEAKER_00: Like, people wanted it.
• 33:28
  SPEAKER_00: People wanted to drive, you know, and people wanted to to give the rides and have this interesting source of income.
• 33:35
  SPEAKER_00: And people, and there were a million things.
• 33:37
  SPEAKER_00: I'm I'm not commenting on the morals of how it was executed or anything like that.
• 33:41
  SPEAKER_00: I'm just commenting on the innovation.
• 33:42
  SPEAKER_00: Sure.
• 33:43
  SPEAKER_00: There were people that wanted alternate sources of income, and there were people that wanted rides, and I remember being one of them.
• 33:49
  SPEAKER_00: Like I'm tired of not being able to hail a cab or calling a cab and then the the calling the dispatcher going, where's my cab?
• 33:58
  SPEAKER_00: and getting no customer service.
• 34:00
  SPEAKER_00: And I was so happy when there was an app.
• 34:02
  SPEAKER_00: That's just it's it's irreverent.
• 34:04
  SPEAKER_00: It's irreverent to the establishment.
• 34:07
  SPEAKER_00: And sometimes it goes very, very, very wrong.
• 34:09
  SPEAKER_00: And there's a million examples, but I like the part of innovation that's willing to break the rules for paying your dues for the sake of innovation.
• 34:19
  SPEAKER_01: So it's something else I think you were maybe maybe kind of edging at here.
• 34:23
  SPEAKER_01: And and I I read a really great piece that you wrote, really kind of at the height of the pandemic was about biases, right?
• 34:30
  SPEAKER_01: Was about and and you said this people not looking the way you expect them to look.
• 34:35
  SPEAKER_01: Was it was that was that part of what you were getting at there?
• 34:39
  SPEAKER_01: Yeah.
• 34:39
  SPEAKER_00: That was, you know, when Black Lives Matter movement, you know, was in full swing and a lot of people were talking about race and thinking about race.
• 34:47
  SPEAKER_00: And I realized, you know what, we can get into really dogmatic positions uh and and take sides instead of going, you know what, everyone has racial biases.
• 34:57
  SPEAKER_00: And I admit it, I have racial biases.
• 34:59
  SPEAKER_00: I remember moving to California and having almost zero experience uh with people of Asian descent and Asian culture.
• 35:09
  SPEAKER_00: And I had biases.
• 35:11
  SPEAKER_00: I had stereotypes, and I talk about it in the post I wrote, I had stereotypes that I had to admit to my Korean friend that I had, and they were wrong.
• 35:21
  SPEAKER_00: Maybe we'd all be a little bit better off if we would admit our biases instead of like get on either side of these entrenched lines and pointing fingers.
• 35:28
  SPEAKER_00: It's like, no, biases and and and stereotypes happen person by person, and they'll be torn down person by person.
• 35:37
  SPEAKER_01: Listen, I again I'm not going to get into the to macros of politics and everything else, but the the this seems like biases in um artificial intelligence seems like something that we were really talking about five to ten years ago, but isn't really spoken about that much anymore.
• 35:52
  SPEAKER_00: That is a great point.
• 35:54
  SPEAKER_00: The pursuit of business results and financial gain will trump a lot of things.
• 35:59
  SPEAKER_00: There's a lot of talk about AI governance, but there's not a lot of talk about the bias anymore.
• 36:05
  SPEAKER_00: And honestly, I'm just gonna state my opinion and view on this.
• 36:08
  SPEAKER_00: I'm not sure the way people are talking about and executing AI governance, I'm not sure it's of much use.
• 36:15
  SPEAKER_00: Like it's it's more, it's what the way I've experienced it, and I don't work at a large company, so it's me looking into large companies, is it seems to be this like governing body or gatekeeping organization that is, you know, mostly talking about kind of philosophies.
• 36:31
  SPEAKER_00: And I'm not saying there shouldn't be ethics.
• 36:34
  SPEAKER_00: There's actually some really tough ethical aspects of AI.
• 36:38
  SPEAKER_00: Like I remember working for a company and having to determine, and thank goodness, that there were there was a governing body inside the company that said, no, we're not gonna use it for that particular military application.
• 36:52
  SPEAKER_00: There were two different military applications, like we're not gonna do that.
• 36:56
  SPEAKER_00: But here's the thing I actually don't believe that AI is conscious and ever going to be conscious.
• 37:04
  SPEAKER_00: I don't um so at the end of the day, it's all a tool, it's all a piece of engineering.
• 37:10
  SPEAKER_00: That's true for everything.
• 37:12
  SPEAKER_00: Like, how are we going to use automated control systems?
• 37:15
  SPEAKER_00: How are we going to use weapons and certain, how are we going to use certain types of fertilizers and certain types of seeds?
• 37:23
  SPEAKER_00: That's some thorny things to think about.
• 37:25
  SPEAKER_00: One of the arguments is the capability of AI is so, so different.
• 37:30
  SPEAKER_00: I think we'd say that about every technology revolution.
• 37:33
  SPEAKER_00: The capability of the automobile is so, so much different.
• 37:36
  SPEAKER_00: So, how are you going to use the automobile?
• 37:38
  SPEAKER_00: Are you going to use it, you know, for military applications?
• 37:41
  SPEAKER_00: If you do, you you have obvious advantages over the horse.
• 37:45
  SPEAKER_00: But there's nothing about the automobile that's or the combustion engine that makes it fundamentally different than any other advancement in technology.
• 37:54
  SPEAKER_01: Are you talking about people being overly cautious about implementing it?
• 37:57
  SPEAKER_00: No, I'm actually not.
• 37:59
  SPEAKER_00: I'm talking about the tough ethical debate about how about the specific types of applications that AI should be used for.
• 38:08
  SPEAKER_01: Yeah.
• 38:09
  SPEAKER_00: Surveillance, you know, automated weapons, drones.
• 38:12
  SPEAKER_00: I mean, there's there's so many, so many things.
• 38:16
  SPEAKER_00: Um, even things like therapy and dating.
• 38:20
  SPEAKER_00: I mean, that's every day I see an article telling really sad stories about what happened to human beings, when they used AI as a therapist or as a dating relationship.
• 38:30
  SPEAKER_00: I'm not commenting on whether it's good or bad.
• 38:32
  SPEAKER_00: I'm saying that's a very thoughty issue.
• 38:34
  SPEAKER_01: When you're on the inside of things, obviously you're seeing a lot of these companies work on these for a very long amount of time.
• 38:40
  SPEAKER_01: And and then obviously, you know, the ramp up to the creation of all of these models and LLMs are happening in research institutes over the course of decades.
• 38:51
  SPEAKER_01: But certainly from the outside, it it sure seems like everything just crops up overnight.
• 38:56
  SPEAKER_01: And and I I can understand why there's a lot of fear around suddenly there being uh AI therapist and then reading, you know, there are stories, and and and that it is truly tragic about um people self-harming and harming others after a poor relationship with a chat bot.
• 39:13
  SPEAKER_00: You bring up a good point about also about you know the gatekeeping.
• 39:16
  SPEAKER_00: I do see in corporations, AI governance committees, which are essentially just slowing down innovation.
• 39:22
  SPEAKER_00: They're not don't seem to me to be.
• 39:24
  SPEAKER_00: There's others that are more sophisticated, but I do see some that seem to mostly just be slowing down the innovation process.
• 39:30
  SPEAKER_00: But there's also the perspective that in certain industries, things move slowly on purpose.
• 39:36
  SPEAKER_00: When you're making things, physical things, and it's taken 30 or 40 years to figure out how to make that properly, you're not going to just change it on the drop of hat because someone uh wrote a research paper.
• 39:47
  SPEAKER_00: Nor should you.
• 39:49
  SPEAKER_01: It's frustrating in a lot of ways that uh, you know, I talk about this a lot, that the kind of the the boiled frog or that the the goalposts keep getting pushback for self-driving cars.
• 40:01
  SPEAKER_01: But it's really good that we're not seeing them on highways right now.
• 40:04
  SPEAKER_00: We do have a different standard when we talk about a judge AI, like for example, self-driving cars, than we do for people.
• 40:11
  SPEAKER_00: There was a time, I think it was a couple of years ago at this point, where pretty much every accident that a Tesla got into for self, if you had set well in self-driving mode, was like national news practically.
• 40:24
  SPEAKER_00: Definitely local, almost national.
• 40:27
  SPEAKER_01: Yeah.
• 40:27
  SPEAKER_00: And I thought, what if we published every crash that a human got into?
• 40:33
  SPEAKER_00: Like, oh my gosh, we have a different standard.
• 40:35
  SPEAKER_00: There were some studies done that said, you know, for apples to apples comparison, the self-driving cars actually get into fewer accidents than humans do, which I get.
• 40:44
  SPEAKER_00: Um, which I get that.
• 40:47
  SPEAKER_00: But maybe there should be a different standard.
• 40:49
  SPEAKER_00: I the way I think about it is not that there should be a different standard.
• 40:53
  SPEAKER_00: The way I think about it is AI is a tool.
• 40:56
  SPEAKER_00: And so you should think about it as a tool.
• 40:58
  SPEAKER_00: If you think of it as a person or a conscious human being or or something like that, then you're going to think of it the wrong way.
• 41:06
  SPEAKER_00: Like I drive a Tesla, I enjoy it.
• 41:09
  SPEAKER_00: I don't use it at self-driving.
• 41:10
  SPEAKER_00: I personally like to drive a lot, but there are two things for which I very much value its abilities.
• 41:16
  SPEAKER_00: One is um it does a lot of kind of menial tasks for me that I don't that I have to do in my other car.
• 41:23
  SPEAKER_00: Like I never turn on and on the with off the windshield wipers because of the perceptive capabilities, it can do a lot of things for me that I normally have to do.
• 41:31
  SPEAKER_00: Um, but the thing I value the most is specific expert behaviors.
• 41:35
  SPEAKER_00: I saw uh YouTube video, someone had a go-crow camera on and uh something fell off the truck, like I beams or cement pipes or something, sewer pipes.
• 41:45
  SPEAKER_00: And this Tesla performed a maneuver that I couldn't perform.
• 41:50
  SPEAKER_00: You know, you'd have to practically be a professional driver, you know, a race car driver or something to perform that maneuver.
• 41:57
  SPEAKER_00: And it's like, okay, for specialized maneuvers and V DL tests, that's the tool for the job.
• 42:03
  SPEAKER_00: For my everyday driving, I personally enjoy driving.
• 42:06
  SPEAKER_00: For other people, for their everyday driving, that's a good tool for them.
• 42:08
  SPEAKER_01: That's interesting.
• 42:09
  SPEAKER_01: It's it's it's uh it's almost as if it performs best and worst when it comes to edge cases.
• 42:14
  SPEAKER_00: Yes, that's actually I love the way you said that.
• 42:17
  SPEAKER_01: So we're at about time, and and I and we can end on this.
• 42:20
  SPEAKER_01: I kind of want to loop a little bit back to something that you were talking about before.
• 42:24
  SPEAKER_01: You were discussing this woman that you had hired who had been a psychiatry or psychology major.
• 42:30
  SPEAKER_01: Part of the reason why you hired her was because she she really spoke up.
• 42:33
  SPEAKER_01: And I I like, I like, I want to pull on this thread.
• 42:36
  SPEAKER_01: I like this idea of uh hiring someone who, in some senses, is an outsider in that you know they hadn't been in that job for a long time, but clearly is is an intelligent person.
• 42:46
  SPEAKER_01: And I want to go back even further and get back to the story about you as an intern at Ford, as an undergrad, these two uh MIT graduate students.
• 43:00
  SPEAKER_01: What is it specifically as an undergrad without that training, without that expertise, do you think you brought to the table?
• 43:10
  SPEAKER_00: I think it was the willingness to try things.
• 43:12
  SPEAKER_00: I didn't have a lot of preconceived notions.
• 43:14
  SPEAKER_00: I didn't have notions about what quote unquote should be done and shouldn't be done.
• 43:19
  SPEAKER_00: And that's what happens in startups a lot.
• 43:21
  SPEAKER_00: Some of the best conversations in startups are when someone goes, Well, why can't we do this?
• 43:26
  SPEAKER_00: Well, that's not normal practice.
• 43:27
  SPEAKER_00: There's a lot of things that aren't normal practice.
• 43:29
  SPEAKER_00: Like if you look at a chess game, some of the best, most brilliant moves are completely, completely unorthodox and out of the ordinary.
• 43:36
  SPEAKER_00: That's the whole point.
• 43:37
  SPEAKER_00: So sometimes the thing that's going to unlock the innovation, it requires someone to not be steeped in 20 years of, but this is how it's supposed to be done.
• 43:48
  SPEAKER_00: Even if things are done that way to keep things safe or for a million good reasons, you might not get to the breakthrough unless there's someone there that can say, well, you could do it this unorthodox way and it could still work.
• 43:59
  SPEAKER_01: Well, Kent's always a pleasure.
• 44:01
  SPEAKER_01: Thank you so much.
• 44:02
  SPEAKER_00: Yeah, absolutely.
• 44:03
  SPEAKER_00: Thanks for having me.
• 44:04
  SPEAKER_00: It was really nice.
• 44:06
  SPEAKER_01: Thanks so much to Kent.
• 44:08
  SPEAKER_01: Uh, always a good sport and always super insightful.
• 44:13
  SPEAKER_01: Uh, thanks to you for making it to the end of the show.
• 44:15
  SPEAKER_01: Please like and subscribe.
• 44:18
  SPEAKER_01: Don't forget to check out our sister newsletter over at automated.fm.
• 44:22
  SPEAKER_01: And with that, we will see you next week for another episode of Automated.