0:00 Brad It's a completely legitimate observation to say, that's a different job. Welcome everybody to another episode of the How Many CTOs Does It Take podcast. My name is Brad Hefta-Gaub, and this is my trusty co-host.
0:29 Scott Scott Porad, how are you doing, Brad?
0:30 Brad I'm doing really well. I've been pondering, pondering all about the state of the industry, and I can't wait to talk about what we didn't finish talking about last time we met.
0:43 Scott Yeah, we tend to go long. You know, we'll get about 20, 30 minutes in and then we'll go to like an hour.
0:49 Brad Plus, we were talking about last episode we were talking about time travel. Like how did we we got we got way off. We got way off way off base before we even got into the thick of it.
0:59 Scott I would love to know from all of our listeners who are going to rate and review this podcast if they love when we get way off or if they just want us to stay focused on business.
1:10 Brad Maybe they only listen to the intro banter and then they don't even hear the all business.
1:18 Scott Before we get into the back to business, let me ask you a quick question.
1:21 Brad Yeah.
1:22 Scott For those of you who are listening who don't know, Brad here is an avid skier. He is a ski patroller. He is a search and rescue guy. He teaches avalanche training. He probably can do like 360s and daffies and spread
1:35 Brad No,
1:35 Scott eagles
1:35 Brad I don't
1:35 Scott and all do the
1:35 Brad any
1:36 Scott cool
1:36 Brad of
1:36 Scott tricks,
1:36 Brad that. I
1:36 Scott you
1:36 Brad don't
1:37 Scott know,
1:37 Brad do any of the tricks.
1:37 Scott all the stuff the kids are doing these days. What do you think about the El Nino?
1:44 Brad Oh, man, dude, I don't know. It's going to be. I don't know what to say. I mean, we're recording this on October 16th, you know, surprise to the listeners when this comes out in November. But just three days ago, they got their first big dump of snow in the Wasatch, the Cottonwood Canyons. Alta, people were hiking up to the top of Alta and skiing in like 18 inches of fresh powder.
2:11 Scott Yeah.
2:11 Brad So, yeah,
2:12 Scott I saw some videos of Park City and they were. And they were just, it was great.
2:16 Brad Yeah. Now, I am thinking here in the Pacific Northwest, you know, it's either going to be it's going to be a coin toss. It's either going to be the worst year that in that I know of, like that I remember, or it's going to be unbelievably epic with some because we're going to get I think we're going to get a lot of rain. I think we're getting a lot of precip.
2:40 Scott Mm
2:40 Brad But
2:40 Scott -hmm.
2:40 Brad it all depends on whether or not the polar It's vortex
2:43 Scott cold.
2:43 Brad comes down or not.
2:44 Scott And that's what we had last winter was it got a lot of rain and it didn't get cold.
2:48 Brad Yeah.
2:49 Scott Well, I am going to pray for cold.
2:52 Brad Pray for cold. Feel
2:54 Scott All right.
2:54 Brad our cold dances. Okay.
2:56 Scott Okay, last week, just to recap for folks who weren't listening last week, there was a presentation given by Gergely Oroes, I think that's how he pronounces his name, he goes by the private a pragmatic engineer at a big conference a few weeks ago called about probably about two months ago now or maybe six weeks ago called the state of the tech industry and he talked about what's changed what's not changed what's broke and what is next and last week we talked about some other things that have changed and would have not changed And then we took a little bit break and now we're back talking about what broke and what is next.
3:34 Brad Yeah.
3:34 Scott And for those people who are listening, we're just going through bit by bit. I was reading a summary of what. was said, and then we were just sort of seeing what resonated with us. So under what broke? Code maintenance and quality control, the massive influx of machine-generated code has overwhelmed or request and code pipelines and code review standards. That's definitely broke. Hardware and compute constraints, honestly, that hasn't broke that much in my world, but maybe it's broken in other people's worlds. Workforce morale and fatigue, high
4:12 Brad Yeah.
4:14 Scott expectations around AI. An AI output of heightened engineering burnout created widespread security concerns in a rapidly auto-generated code, exacerbated feelings of job instability across rank and file developers. What do you think? What stands out to you there?
4:34 Brad Well, like you, I think the code maintenance, code quality one and then the workforce morale fatigue, I think those are the two that jump out to your point about hardware and compute constraints. I'm not seeing that as much in the in the teams that I work with, but but I you know, when I look at industry trends, I think that is out there. But yeah, the code quality. Yeah. Like, I mean, isn't like it's almost like 98% has become the new normal of 98% uptime is now somehow acceptable. Like we're not even at two nines anymore.
5:08 Scott not in my world not in my world uh well there used to be an expression code is craft i saw and i might have mentioned it already on the podcast as i saw it earlier this year where somebody said guardrails are the craft and
5:21 Brad Yeah.
5:22 Scott it really is that is becoming the craft of how do you create the boundaries around the agents so that they do the right thing and
5:30 Brad And
5:31 Scott i think that's a really it's actually come to think of it it's kind of an interesting problem
5:35 Brad I, so let's just stipulate that the point that like. Code maintenance has broken down and code quality has broken. Let's just stipulate that. Like, I don't think anybody could possibly argue with that. Now, I think the interesting question, and I think as I, you know where I'm going to land on this one because I've been like a broken record on it. I think the real reason this has broken is because a lot of organizations. thought they had quality gates and by the way organizations i've worked for will fall into this category right they thought they had quality gates but they were really predicated those quality gates were really predicated on human speed and
6:23 Scott Right
6:24 Brad now that we're not going human speed those old style of quality gates that depended on human speed And we're really kind of just getting lucky. I don't think they I don't really think they were the quality gates we thought they were. I think we were just kind of getting lucky and we moved slow enough that when something broke, we could fix it, you know.
6:47 Scott Mm-hmm.
6:49 Brad But now things are moving so fast that if it gets out, it's 12 other things have already piled on top of it.
6:56 Scott Mm-hmm.
6:57 Brad So that's my sense is that it really, and this goes back to the one of the principles of what had not changed is core software principles, right? Like that was one of the things that the pragmatic engineer talked about hadn't changed. And so to me, these two are these two things of core software principles and code quality code maintenance is broken. These two things are intertwined with each other.
7:19 Scott Agree. I think it gets into a little bit of what's changed. though which is or what's going to have to change is like the mindset of our industry the way we invest in terms of how we allocate resources and what we spend time on like i'm just going to say like before it was 20 percent I feel like this is generous. I was going to say it was 20% figuring out the requirements, 60% coding, and 20% verifying that you did it right. And I think that's probably generous to the requirement. It's probably going to go something like 30-20-30.
8:04 Scott Actually,
8:04 Brad Yeah.
8:04 Scott that doesn't add up to 100.
8:07 Brad 30-20-50.
8:10 Scott Yeah. I mean, something like that.
8:12 Brad By the way, it's almost like you read ahead or you watched the whole video because that's literally the first thing that he talked about in the what's next is AI coding safety layer.
8:22 Scott Yeah, that's
8:23 Brad Widespread
8:23 Scott right.
8:24 Brad adoption of, I mean, by the way, I mean, I will admit I'm so vain. I probably think that video was about me because I feel like I was watching it and I was just getting this great sense of validation because he was talking about all the things I've been talking about for a year now, right? About like you got to have high quality SDLC, you got to have you got to put energy into building the quality gates. So how much have I talked about static code analysis, guardrails, verification tools? I've talked about. about you know ai generated tests and then ai generated mutation testing of those like just like all these guardrails that are good quality standard software engineering best practices that we use it was really hard to convince our business owners and our in our our engineers that we should be doing this investment in these quality gates it's it was so i feel like my whole career it's been that's the hard
9:24 Brad That's like pulling teeth is convincing people to do these quality gates. But now I feel like this is fine. It's finally practical for us to build the kind of quality gates that we always imagined we were getting by having people read the code. Now we got to put the true guardrails behind it.
9:45 Scott Hang on a second. I'm not disagreeing with you that we have to do it. I'm also not, but what I don't know that it's going to be any easier to convince our stakeholders, convince the CFOs of the world and the CEOs of the world, the product management, chief product officer of the world to want to invest more in all this.
10:07 Brad Well, that's fair. And I guess I will just add, I. have been in the room with a product owner who's saying i don't i really want engineers to build new features and i was saying i really think that you should be demanding that we fix the an authentication provider that keeps going out, which causes our customers and our users to not be able to log into the system, which has caused us to miss SLA five quarters in a row. We need to fix that. That is quality. And eventually they came to see the wisdom of my ways that it didn't matter if we had all the features if the people can't use the software.
10:55 Brad And we're at 98% uptime now. GitHub is at 98%. It's ridiculous, Scott. Like this quality has broken down so bad that we've lost like. It's just like this stuff is unreliable. We have to fix it or else we're not really building software that people can use.
11:14 Scott Yeah. So like, is the prescription you're offering to say, hey, engineering teams start to ship crappy code and start to have your systems go down so that your business stakeholders will
11:25 Brad No,
11:25 Scott value
11:25 Brad I'm definitely not saying that. We're not saying that, but I'm saying it's already it's too late, man. That particular that particular horse is out of the barn.
11:35 Scott Well, I generally agree with that statement. Maybe it is, but I can tell you the systems we run don't have those problems right now. And I don't think it's because we've got the most amazing quality gates.
11:45 Brad I'm really glad that that is true for your your venture. And I'm simply saying that we're talking about what the pragmatic engineer. came to the conclusion after surveying the entire industry. And he listed as the number one thing that broke is code quality is so bad that 98% is the new normal.
12:05 Scott I agree. But I guess I still okay. So yeah, so I guess what you're saying is, I guess we're asking, you know, we're talking past each other. Yes, if you're running at 98%, it's going to be pretty easy to convince your stakeholders to invest in quality. I think the question is for teams that are not having that challenge yet either because they haven't adopted AI so they're still going at human speed or they've gotten lucky or who knows what right the question yeah
12:31 Brad What do we say to our business owners?
12:32 Scott as CTOs how do we tell this story of like if you want AI if you want the benefits of AI it comes it comes with a cost and a benefit right and you got to take them both you can't just get one well you can't just get one Good one. And then you get 98%. I
12:50 Brad I think
12:50 Scott don't
12:50 Brad that's
12:50 Scott think it's
12:50 Brad a great
12:50 Scott all that's all that's the story. I want to
12:52 Brad that's the story you tell now workforce morale and fatigue do you want to pull on that thread
12:57 Scott I want to pull out for one more second on the quality controls and the guardrails and stuff like that. It feels not this is not the getting the stakeholders. It's actually building those guardrails to me feels really hard. And my hypothesis for why it feels so hard is because since the day I started in this industry as like a 21 or 22 year old, I've been building new features and capabilities. So like it's a thing that I'm comfortable with and familiar with and like, you know, yes, we built tests and we build quality controls and things like that. that was always it wasn't the thing right it was the it was the thing behind the thing and so like i guess for i guess i just it seems like for me and i guess for a lot of people that that having the quality be the thing is like it's new it's like clothes that don't quite fit yet if that makes
13:46 Brad Yeah.
13:46 Scott sense certain parts are going to get a whole lot more emphasis than i used to and that's just it's just different i guess is what it boils down to I
13:55 Brad Yeah.
13:55 Scott think that's a big change. That's a big change that as an industry we're kind of going to be going through. Like if we were having a prediction show, right, my prediction would be like, you know, this is obvious, but just like how this all feels is going to be so different in five years.
14:11 Brad Yeah. Well, and so I agree with what you just said. You're not going to get an argument from me on that. And I think that it dovetails nicely into the other. So one thing that's interesting is in his talk, he talked about what broke and he talked about what's next. And
14:25 Scott Mm
14:25 Brad in
14:25 Scott -hmm.
14:25 Brad the case of the code quality and code maintenance, like the other side of that coin was this notion of what's next is this AI coding safety layer, right? And that's, you know, kind of. what i've been saying you're saying that maybe that we need to change our mindset around that in the other side of the coin of the workforce morale fatigue thing which we let's start pulling out that thread is
14:48 Scott Yes.
14:48 Brad this evolution of what software engineering role looks like and i think that's what you're starting to hint to this this and so i think all four of these things are really kind of like they're pretty coupled you know like what we used to identify so why why are we seeing workforce and morale fatigue like why is that happening
15:10 Scott Is that is a rhetorical question? I
15:12 Brad i
15:12 Scott think
15:13 Brad think that yeah i would love to hear your take on it i mean i i have opinions
15:16 Scott a bunch of reasons. I'm gonna just throw a few things out there. I don't know if it's comprehensive. One is the pace of change. It's exhausting to just try to keep up like you know you go on vacation for a week and you come back and all of a sudden it's like what's new and so it's hard to develop your craft if you're changing your tools every five minutes right
15:36 Brad Yep
15:36 Scott I think that you know learning I was thinking of like wood carving or something but learning how to use that tool and really make it do what you want it to do you got it takes time it takes the 10,000 hours and when you're where they only get about 45 hours before you got to do when it's real tough. I think that's exhausting. I think the expectations that non-technical people have for what we can do now. they know we've been granted some superpowers that we didn't have before and so their expectations are really high and that puts a lot of pressure on but the superpower i feel like probably use this analogy on here before i feel like we've been given this fire we had a garden hose and now we have a fire hose and a lot of water's coming out but
16:21 Brad Ha,
16:22 Scott it's also
16:22 Brad ha,
16:22 Scott really hard to control it
16:23 Brad ha, ha, ha, ha, ha, ha,
16:25 Scott and they're just saying like there's just they're just saying like why can't you get that fire out we're like i can't even get this hose it's following the aim in the right direction you
16:31 Brad Yeah.
16:32 Scott know so i think that's putting that's adding a lot of pressure so there's just like trying to keep up there's the pressure from stakeholders um the the career job uncertainty i think is a whole thing also i think there's certainly segments out there for whom look i know how to get it done this way Right. And this way works and I'm getting. I guess it's a flavor of career job uncertainty. Maybe they don't feel uncertainty about their career, but it's just like there's change happening that feels uncomfortable. I think those are all leading to it.
17:12 Brad By the way, well, those are I think the only thing that I think you didn't touch on because it's a little bit redundant to what we already talked about is I think there's real morale impact and fatigue around seeing that number go from. five nines down to nine eight like
17:28 Scott Yeah,
17:29 Brad i
17:29 Scott sure, sure,
17:30 Brad don't think the engineers that um have been have been in this pressurized system and then all of a sudden now they're getting incidents every other day that's definitely putting like the number one like in my career like the thing that always panicked engineers or or was the most distressful thing for engineers was is dealing with incidents right like we we we we invent We invented the whole notion of SREs, you know, or reinvented. Like how many times has there been this like these movements to like somehow shield the engineer who wrote the code from the having to be woken up at three in the morning, right?
18:07 Scott Yeah, that's exactly right. I mean, I have been preaching to my teams and to my stakeholders, the CEOs and others, our number one job is to keep this system running. I know you want a lot of features and you want things to go faster and so on and so forth, but that is actually our number one job. If the system is not running, if the store is not open, people can't buy stuff. The golden goose is laying eggs. We've got to protect the golden goose.
18:33 Brad Exactly.
18:33 Scott And they just expect that as table stakes. I think there's another one that I thought of as you were talking. And you've talked about this on this show, and we've talked about this in other places as well. I'm sorry, we have talked about many people have talked about this other choices. The qualitative or aesthetic aspect of the job is different. You've talked about going from a programmer to a manager. We're now managing agents and just the flow of that work and how it feels and what's gratifying about it is very different and I think is maybe. ungratifying isn't a word what would be the opposite of gratifying
19:10 Brad Less gratifying.
19:11 Scott less well the opposite of gratifying and i think that's i think that's emotionally hard for people it's just like i'm doing this thing and it doesn't it isn't fun it just isn't enjoyable yeah
19:22 Brad Well, and that is exactly what, you know, the pragmatic engineer says about what's next is evolving the way we think about being software engineers, right? Transitioning from developers who manually, you know, type out syntax to People who think about systems and think about orchestrating that systems and think about designing the software factory and thinking about modeling the various threats that can be come against, you know, the code built by a software factory. And and, you know, I think it's it's a it's a completely legitimate. observation to say that's a different job that's
20:02 Scott Mm-hmm.
20:03 Brad a different job and it not and it's a job that it's a job that i have enjoyed doing for a big part of my career like I as a as a as the you know a junior engineer I am I wanted to go and build bigger things and think more about the architecture of not just the software but also the team or the software factory and and that's what drew me towards engineering management and ultimately you know being a CTO right but it does pull you away from the coding elements of it. This might be the biggest change right there that we're talking about. If you think about the shift from assembly to higher level languages, you are still doing the craft person's job, right?
20:51 Brad You are still operating the wood lathe and spinning the wood to make the bowl, right? But this shift... is different you're no longer the craft person you're the supervisor you're a machine supervisor and that's a fundamentally different job now you're just a guy with a clipboard checking Yeah. to make sure the gauges are Find a point in the right direction. Well, OK, that's fair, but that's I think that's underselling the creative part of the higher level functions. Right. Like I think that, you know, I don't know if you know that much about sort of like. How art studios work, you know, the great masters of the Renaissance or even like modern art masters like Dale Chihuly and his glass blowing studio.
21:45 Brad But Dale doesn't, Dale Chihuly doesn't touch the glass and he doesn't touch the metal skewers and he's got all these other people working around him, but he is the visionary and he's. And he's there watching and overseeing it, even Right. if his hands aren't physically doing the work. And I think he probably feels like he created something in that process. He feels that creative energy from that. And now this is an interesting. Question. Can we just really quickly sidebar and on tangent for a second, Sure. going to our audience? Have you been to the Chihuly Museum in Seattle? Yeah.
22:30 Brad Yeah. For our audience, it is worth going. Super cool. Back to business. Okay. So Dale is doing that. I think he might have passed away though. Is Dale truly still alive? I don't know. Nevertheless, he was doing that with other people. And so there was a teamwork there, right, that maybe might not exist when it's just you and a bunch By of robots. the way, Dale's still alive and kicking it as of the recording of this podcast goes. This episode has been brought to you by the Chihuly Glass Museum in Seattle, Washington. Okay, so I hear what you're saying. Yeah, absolutely, there's still creativity to it. You have to have the vision for what you're building,
23:17 Brad but there is this sort of you don't get this experience that I think so many engineers love of having their hands-on actually making it and getting into that flow state and hitting F5 compile or whatever and seeing it work, right? Yeah, and I think that's what Kent was talking about last week when we talked to him, or a couple weeks ago when we talked to him. Yeah, that's right. That is a really big change. But now what's interesting, here's an interesting question. Will the colleges start teaching it different? Or are we going to be graduating people for another decade that are doing it the old way? Well, let me say that I believe. The computer science degree that I got from the University of Washington,
24:03 Brad you know, from 1987 to 1991, I Mm-hmm. think it prepared me perfectly for the new job of software engineering in 2026. That program was a computer science program and computer science. I mean, yes, I took a languages class and yes, I took a data structures class and I, you know, in those classes I wrote code and I also took an operating systems class and I had to write code, but those classes weren't about. They weren't about the language that you were writing. Like, I did not take a C++ class.
24:49 Brad I did not take a, you know, prologue class. I took classes and I was forced to or I was, you know, the classes taught me about all of these concepts, these system level concepts. And they were practical, like you practically did things with those concepts. This is really interesting because it really, to me, fits with what you say a lot that we're in the golden age of making, which is if you're a person who really enjoys making things. And you want to go make computer things, going and getting a degree in computer science to learn how the computers work will be very beneficial to you because when you go to make things with the robots that use the computers,
25:41 Brad you'll know what's happening there. Just like I'm sure Dale knows Dale Chihuly, now we're on a first name basis, but yeah, you know, just how Dale Chihuly, I'm sure know how. of the properties of glass and how it melted Yeah. and I don't know I don't know a lot about glass but but he was an artist who was thinking how can I then put it together to make something visually or visually appealing or functionally useful Yeah. and appealing etc Yeah. Absolutely. And also, by the way, he, I'm sure he studied art. He studied art history. Yeah. He studied visual arts. He studied the concepts of art before he even decided that glass was his medium.
26:29 Brad I'm sure he took painting classes. I'm sure he did clay sculpture classes. I'm sure that his 10,000 hours included just... learning about the systems, right? And then he found his own voice and he found his own vision and now he's able to deliver that, you know, without necessarily putting his hands on the medium. Right. And so for the young people out there going who want to make things and use technology to make things in addition to the computer science, it's going to be learning about. business and learning about the humanities i mean this is steve jobs talked about this all the time i think he had like a degree in calligraphy or something like that i mean something ridiculous but like um yeah it's going to be about learning about all these other things so that you can actually figure out what you want to do with it this is actually very inspiring to me because i've talked to so many parents you know my kids i've talked about this you have a kid in college a kid out of college so i've got a lot of friends
27:30 Brad of kids who are college age and many of them are worried about you they have kids who or in computer science programs or maybe have graduated recently and they're very worried about like what's the future for my kid I just was talking to a friend of ours the other day you know she's a fashion designer she's her son is studying computer science he's in his senior year of college she's she's like what's going to happen with this guy he doesn't she doesn't know and you're making me feel very hopeful I'm going to pull when we release this I'm going to go send her this episode and say look Don't worry, you're Well, still going to be fine. I'm glad I'm glad that that's inspiring. I would say that like I will I will I will. pivot this a little bit to say the following there has been for the last several years there's been this before the kind of the ai started you know supposedly eating all white collar jobs there was this narrative that somehow liberal arts educations those educations in the in in in literature or languages or art or whatever that those were
28:36 Brad fluffy, useless, not very meaningful career academic pursuits. And instead, we should be focusing on STEM and skills and focusing sending young people to those types of programs. And Mm-hmm. I. I have always said that's wrongheaded and the wrong way to think about it, that what a liberal arts education, I mean, a computer science degree is in fact a liberal arts education because it's not politics. The liberal part of that doesn't mean politics. It just means the notion of open mindedness and curiosity and exploration and learning how to learn and learning how to think and learning how systems work. work and that's why steve jobs had a calligraphy degree or an art history degree and still was one of the most influential technologists of in all history right is because he understood how systems work and how people work and how to think and how to be curious and how to learn new ideas right like i used to i say this to every engineer i ever have hired on a team is like
29:43 Brad It's like, you know, if they're nervous about like, well, I've never coded this type of thing. Well, of course you've never coded the thing that we're building right now because we're trying to build something new that no one has ever built. So I don't need you to have done the thing before. I need you to know how to do new things. Yeah, that cross-disciplinary. learning about all sorts of things will be essential for students young people who want to build things I think that's an interesting pivot too right we had these years where the tech industry was in such demand we saw the boot camps became a big thing and it became a little bit of like programmers were the with a hammer swinging hard hats right Exactly. and the hammer swinging hard hats kind of just got put out of job not not entirely but kind of and now it's more the architects and the engineers who
30:34 Brad are suddenly in demand. They're Right. going to be in demand. Right. People And I who would say, want to yeah, build things. exactly. So the boot camp thing that if the only thing you learned out of the boot camp was how to place the semicolons in the brackets and the curly brackets, then. And that didn't that didn't that's not a persistent skill that you're going to be able to take away into 2026 2027. But if your boot camp, if you were the type of curious person, you already had a learning open mind, you had a growth mindset, and you just and you use the boot camp to imagine other things that you would build. Great. Mm hmm. Then your boot camp probably served you very well. But if you just learned the syntax, it's going to be it's going to be rough.
31:22 Brad Yeah. You know, it's like. I the analogy I would write and write well writing is going to be part of the analogy the analogy I would use is like being an author you went and learned how to write a language and the grammatical the grammar and the syntax but you still had to decide what story you wanted to write did you want to write short stories did you want to write novels did you want to write movies what did you want to write and when you chose one of those formats then what story did you want to tell did you want to tell romance stories or did you want to tell non-fiction right and that's kind of i think the part that's giving me hope here is like for a young person who's a cs student it's like sweet you're learning how to write a novel now like what's the now what story do you want you're learning how to write now what do you want to do with that you want to go write a novel like what do you want to do right
32:12 Brad Yeah, totally. um you want to go make cool experiences for people or do you want to go save money for big banks or do you want to go figure out or get And back to the end of the moon on that whole notion of computer science. So if the computer science is this area you want to be, I think it's perfect because that's what's going to give you those unchanged core software principles, the core computer science principles that matter. That's not changing. That's not going to change. That still matters. And that's what's going to help you build the AI coding safety layers so that you're not in the code quality and code maintenance nightmare that seemed to have happened in 2026. And so to me, that's the path forward here.
32:58 Brad Or we kind of always talk about consumer or business applications, right? I watched a video yesterday. I don't actually know how old it is, but it's like the chief scientist at Google or something like that. He was giving a talk at Princeton about sort of the history of machine learning and LLMs and just sort of how we got this spot. And it was very, very deeply technical. He was there at Google. say 2012 is when he started there so but how long ago was that 2020 that's 15 years ago right um they were doing deep research and he even talked about earlier in undergrad where it was like he had seen these neural networks this was in the mid 90s you'd seen these neural networks and thought that was cool maybe I could do this thing with it right and so there's this other path we often talk about applications consumer and business applications there's this other path for computer scientists which is you often talk about the k-shaped economy I think about it like the k-shaped technology you want to go the application direction or do you want to go I'm not sure what the other direction is but it's like in
34:00 Brad Systems. that direction we're like the system's direction yeah do i go the system's direction and i thought it was interesting for me to watch that video because so much it was like man i don't quite know what he's talking about because i've never worked with the system layer like that Um, I mean, it was fascinating. A lot went through a lot went over my head, but it was really interesting to see how this evolved and so much of it evolved through really on a super tangent here, but just these kind of researchers that were curious, and then they figured something out and that created another problem. I was watching, I stopped watching about 45 minutes in because I ran out of time, but he just got into the part in the story where like, oh, we figured out we needed to make new chips. because to do the thing we wanted to do would have meant having twice as much compute as all of Google had at the time.
34:48 Brad And that wasn't practical. So we had to go build a chip team and start designing specialized chips for doing this thing. They're called, I can't remember what the chips are called, I but the think it's the TPU probably. TPU, that's right, the TPU. Yeah, so it's interesting. It's a different style of work. You're kind of doing research and then solving problems in a different way. way which is maybe another another route this goes for people these days yeah There you go. I think that's a nice way to wrap this segment up is what's next is there's a lot that's available to you. There's a lot that's available to you. well i want to wrap it up with this last thought if we could it is times of turmoil that create opportunity right Yep. because everything's moving around and you can wiggle your way into a spot where you know you can wiggle your way into a lot of opportunist spots and so whereas a lot of people can see this as chaotic time and uncertain time and unnerving for them it's also especially for the young people out there and I don't know how many young people listen to this but it's a time of opportunity
35:50 Brad Yep, there you go. All you young people out there, what do we want to tell them to do? Smash that subscribe button. Yeah. And then if you're more mature age such as ourselves. Also smash that subscribe Smash button. that subscribe button right in review and introduce your mom to the podcast because moms are the fastest growing demographic of listeners of the how many CTOs All right, does it take podcast? Brad, this has been fun. We'll talk again next week. Talk in next week, Scott. Bye, Brad. Bye. This episode of the How Many CTOs Does It Take podcast was produced and edited by Safran Heptacow. Our theme song was AI generated by Brad. Please follow us on all the social media platforms at How Many CTOs Pod.
36:36 Brad You can find us on YouTube, LinkedIn, Instagram, Facebook, Threads, X, TikTok, and Patreon, or at our website at howmanyctospod.com. And
36:46 Scott There's a challenge
36:46 Brad remember to rate, review, and subscribe on the podcast platform of your choice. Thanks for listening.
