Welcome to another episode of Adventures in dev Ops. It feels like we were just here, Warren. That was to you. Yeah, I know, you know. I just thought that you had something else lined up after that, because sometimes you just keep on going, right. Oh so maybe I tend to ramble. No, that's fair. Yeah, it does happen to even the best of us once we get onto a topic that really meets close to our hearts. And I feel like we have a good one lined up
for today. Dude, I am so excited about today because we'll just cut to the chase because I'm sure everyone on the podcast feed has already seen the title. And joining us today is Kelsey high Tower. Kelsey, Welcome to the show. Happy to be here. I'm so excited to have you here, so just to I feel like you don't need an introduction, but I want to make sure that everyone who's listening who may not know who you are,
is familiar with you. Your software engineer, developer, advocate, and speaker, most well known for Kubernetes and your work for open source and your time spent at Google. Is that a pretty fair assessment. Yeah, I think that's a good assessment of the last mile the last eight years of a twenty five year career. You know, before that, I used to work in enterprise it financial services, praying to God that the Apache server does at fall over in the middle of the night. I've worked in web hosting,
tech support. I've contributed to Python and cible terror form, and so throughout my career I've had many of the roles that I'm guessing a lot of your listeners have had. And I think maybe the thing that maybe some of the listeners haven't done is that part where you start to contribute to the tools that you use every day, the part where you actually join the community of people thinking about what comes next, thinking about not just reporting the bugs, but
fixing the bugs. And I think that just brings balance to someone's career where there becomes a lot of outward facing work and you see that in form of conferences, open source contributions, tweets on Twitter, those kind of things.
Yeah, I think I talk with a lot of people getting started in their tech career, and I think that's one of the things that a lot of them have they struggle with, is they see someone like you in the last eight years of your career and they're like, how do I get there, you know, and they're looking for you know, they're searching Google job postings, looking for Kelsey High Tower's replacement. And I'm like, no, you have to. You got to start at the entry level and you you sort
of have to grind your way up. And I think that's I think that's somewhere along the way we've missed communicating that message over the last generation. Yeah, I mean it's hard. I mean when we watch things that are typically televised professional sports. You know, they don't televise the middle school games,
right, they televise the pro games. And so when you're watching the pros, you know, you're seeing twenty five year training be perfected, and then you're only seeing the one zero zero one percent of all the people who've tried this do it. So what you're watching is the absolute pinnacle of that particular thing. And I think it's okay to kind of observe those things and then people say, hey, what happens when you meet someone like that? Try
to ask, like, how how they got to that point? Because it's these little things where I always encourage people. If you're on year one, year five, most people are trying to get to what they see on the screen that they don't pay attention to the nuance and the detail, like learning how to work with other people, learning how to communicate to your manager that you don't feel challenged and you would like to do something more, learning how
the tools you use work. I'm just writing PHP. PHP is used by Facebook in production. There are ways to make it faster, there are ways to build things like Shopify. Pay attention to the things that are in front of you. Don't race past that, because I think what people end up with is those first five to ten years almost throw away, and they didn't really learn a lot. I met a guy that reached out for mentoring. He's like, hey, I'm trying to get to the next level, and
he blamed everything on the job. Oh, I'm only doing this. These people are slow, they don't know what's going on. And I'm sitting here and I'm like, what do you do. He's like, well, I do a little bit of this, and he mentioned I managed the VMware environment. I said, oh, oh this is great. You can definitely carry that particular thing onto the next thing. Why are they using vms instead of bare metal can answer the question? I was like, oh, you haven't
been learning. You've just been doing. You've gotten the instructions, and you've mastered the instructions so well that you think you know how it works. I was like, what I want you to do is go back on Monday and just ask a question. Why are we using vms instead of bare metal? Is there a performance overhead? When should you not use the VM? Is there a security related clause? How does Intel assist in this? There are so many levels to this that I promise you that you have not necessarily learned
everything you can to the current job. You just haven't been asking the right questions. It's always encouraged people slow down a little bit, go deep before you move on. Yeah, and I think the attitude that you have makes a huge difference there. And instead of finding faults outside Jocko Willing talks about this a lot, you know, instead of finding faults outside, internalize it and say what can I do to change this? Because really I'm the only
variable I have control over. And it seems like that seems to be very very effective for people long term. I always ask myself, what do you attribute that to throughout like your childhood when you grow up, you know, I was lucky. I was born in eighty one, so I was part of a generation that played outside, you know, pre internet, pre smartphones and going outside you kind of do have to fit for yourself. You can't call mom, right you go play basketball at the park and someone gets you
a hard file. There's no referee, no replay, there's no challenge. You just keep playing. You figure out what the boundaries are and you try to stay in those particular lanes. My first job was at McDonald's. Now you can complain about that and it doesn't make all that money, but you can't actually go sweep the lobby with some integrity and do it to the highest level that you're capable of. And so and I also played sports, and also like it's a team, but the team expects you to do your part.
And so I think there's a lot of this part of what am I supposed to do? And have I done what's necessary in terms of preparation showing up with the right attitude. I've learned that it's not always about being the best. Sometimes the best person with the wrong attitude is the person you don't want right like you think about it, you need a surgery, and your surgeon has a bad attitude today, I don't know of being the best is
all you need here. I need someone to stay focused. I want to make sure that I am the most important person in that room at that moment. That's the type of attitude I think. And it's hard to do this every day. If you've been at a job for six years, is it easy to have their best attitude in the world on six years and five months? It's hard to come in like it's day one. But if you find yourself kind of too far off, then you've got to just ask yourself,
like, hey, is my attitude where it needs to be? And if not, maybe I do need that take a vacation to hit the reset button.
Yeah, for sure. And I think that's where like a solid morning routine is helpful long term for that, you know, making sure that right off the bat, you're doing something that puts your mind in that framework that you want to be because you can just kind of ride the waves and end up however you whatever mood you're in for the day, or if you have a routine like oh, I'm going to listen to you know, this particular piece of music or this motivational speech, or do you know, there's a
thousand ways to do it, but you can actually direct what kind of attitude you're going to have during the day. And I notice personally when I fall out of that routine that it definitely shows up. And then I am really fortunate to have a good set of friends who call me out on it and help me stay on track. I needed a reminder last week. I was given a keynote, and I mean, I don't think I put enough attention
into the preparation. You know, I have a lot of tricks in my bag, I have lots of stories, I have a lot of go tos. But I was meaning to do some new stuff and so I ended up holding off doing the new stuff until two days before. And so now I'm like, ah, man, I'm not putting my best foot forward. I'm trying to compensate by doing unnatural things, like why am I adding all these slides? You don't even use slides? And I just have to remind myself.
And it was a little different than like the motivational thing. I had to remind myself that I have put in the effort. This is not a two days of preparation. This has been a lifetime of preparation. And I remember saying to myself that morning, as I was deleting all this extra stuff that I did need, don't afraid to be great. You're good at this. Why are you so afraid of being great? Why are you overcompensating trying to plan the whole thing out. Leave a room, leave a little room
for just being great. Go on stage and have that confidence you need. And what I needed at that moment was just a confidence component. You think, Hey, you've done lots of these keynotes, you should no longer be fearful of the stage. I am, and sometimes I just need that little confidence boost, because that's the difference between a good talk for me and a great talk for me, is how I'm feeling. And so once I said
that, I walked to the conference. I didn't know how I was going to open it, but I wasn't afraid of being great, and it turned out great, right on, right on. So, speaking of conferences and your career, you are recently retired, So congratulations. Does this mean that you're trading in the family car for a golf cart moving out to Sun City, Arizona. The phrase I've heard the best is I'm retired not tired. So it doesn't mean that I just want to go off and do nothing and
sit by the beach. Just not it. But I realized is that there were lots of things that I like doing. I like building stuff. When it comes the technology, I love the delta between a project and a product. And then when it comes to public speaking, it's how I think out loud. Right, So I'm thinking about a bunch of things all the time.
There are things that catch my eye. I have my own opinion, hopefully it's an informed opinion, and then when I get a chance to go on stage, I get to think out loud and sometimes it causes other people to think. I love those components. So when I retire in July last year, one thing I kept were those things. So how do you do projects to products? Well? Advising companies like Docker and for sale companies behind RGCD. These are all teams trying to go and say, hey, we
got a great project, people like it. Can we build a business out of this so it can be sustainable? And so those elements I'm pretty good
at. So when I can join the team and exist in that in any way, I can keep doing that and the same goes for building things that these days way less software, a bit more dy home improvements and people like do you just buy like a brand new house, Like, yes, it's a brand new house, first person ever live in it, but there are parts of it that I think needs to change, And so I started learning all the things, just like you learn software a bunch of YouTube videos.
But there's one big difference about software, and like physical things, there are codes. There are compliance things you have to do. Like in software, you can take all the shortcuts you want if it builds, so be it and the thing you live in that you expect to keep you safe, there are codes. And as someone who considers themselves a software developer, some people
will call it an engineer, having guidelines really helps. Like when you're writing in a type language, it tells you when you're passing around the wrong type between functions, or you have a compiler that really understands what you're trying to do and it yells at you when you're not doing the right thing. Having building codes, if you can find them, is really helpful way of guiding you into Oh, I should staple this row mix cable twelve inches before it
goes into my electrical panel. Those things are important because there's a group of people who do this professionally. They have taken time to write down these codes so you don't burn your house down. So that's been an interesting kind of
pivot, but lots of carryover right on. So that one thing you mentioned there that I thought was really cool was the difference between a project and a product, because I've spent most of my career working with early stage startups, and I found that, like, that's the difference, Like there's always a difference, Like in successful startups that make it, there's always the product that you launched with and the product that your customers thought you launched with, and
successful companies are able to identify that gap and close it as quickly as possible. And I think that sort of resonates with the same thing that you were saying, the difference between a project and a product. You launch with a project and you use that to build the product that your customers want. The hard part about this is that a spoon is a very useful tool. There's lots of value in a spoon. Almost every person in the world he knows
what a spoon is or has used one. And so when you think about it. This should be an easy way to start a business around spoons, at least the first spoon. But you'll realize that a spoon is a feature. It is a feature of a kitchen, or it requires all the other things plates, forks, knives, So you gotta sell a set. But everyone can sell a set, and so like, there's not a lot of
margin. Maybe the materials you use. How fancy it is that some people just want a spoon, and so spoons do well when sold alongside other things. Right, You're probably better off selling your spoon in Nikea than starting a standalone brick and mortar spoon shop. It's not going to work. There's a lot of value in a spoon. You can go and do research and say thirty four percent of the world uses a spoon. Doesn't matter. You will
not be a standalone business selling spoons. And I think startups have the same thing. You give away software the lowest price point possible on open source and it's pretty good. It solves a need, that's why you built it, especially the ones that come from like someone used in production at a real company and then they decide to open source it. All the kings have worked out they've done all the right things to get into production, but then they give
it away. Lots of things sell quote unquote if you gave them away, like even cars you don't like, if they gave them away, they would be the best selling cars in the world because they're free. And a lot of people confuse is the ability to give something away for the ability to sell something, And they're not remotely the same thing. And the teams look very different. So when you just have a great project, you can over index
on technical people. Ideally those technical people know how to interact with the community. But even then that's not required if it just works well, like REP. I don't know who maintains right REP just works and so I don't I
don't know, like who the product manager is if there is one. But the moment you want to start charging money and you expect people to pay, now you got to start talking to customers, and customers usually want things like integrations, fab REP, compliance and says so all these things that technically your project doesn't need but your customers do. So helping people go from project to product is that kind of delta building real rounded teams. The one thing I
noticed a lot is a lot of that to struggle with the story. I have a cool new ide, I say, great, look at all the things that can do. I say, oh, that's beautiful. But here's the thing. People use Vim, emacs and Vius code. What is your story to them? Oh no, we're better than them emacs and vias code. So you're not listening. Those things already work, They've been here, they have a great track record, and they have communities. Your thing is new. What is your story? And what I mean by story is do
you support Vim bindings? Can I import my Vim config that I spent three years getting just right? Or do I have to start over learning all your settings? These are the little nuances that go from project to product. Will is totally sold already on your ide and he already wants to buy it if he can import his Vim bindings right away. It's all he's been talking about for like the last six weeks is how much he wants the right code directly
in production using his ide. See I see baby. Actually I'm stuck on the spoon staying I'm gonna I'm gonna start as SaaS spoons as a service. Well, it's really interesting because I think there was a very important embedded wisdom here that I feel like you just slightly highlighted and then moved on to the
next thing. And that's I bet. There's a great desire to actually have spoons as a use case, right, you know, I want a spoon, and so your customers maybe even giving you this feedback about how great a spoon is. But if you doesn't, if you aren't able to turn it into an actual product the everything around it, that feedback is actually what you shouldn't be listening to because there's no money to be made in that specific feature
that you're offering. And so there's a difference of whether or not if you're listening to the users that are able to help you make your company great or your product great and deliver it to market, and the ones that are just telling you the one little annoyance thing that they have and getting stuck on that. I think we're also getting to a point where everything doesn't need to be
a product. And then people struggle with this, right because most people believe if you create some value, there should be a path of getting rich from that value. And as much as I'm a capitalist just like anyone else, don't get confused but not everything is a product. I think about like fruits and vegetables. Should all of those be products? Or should we still reserve the ability to grow them in your backyard if you needed to, Should nature
still be able to grow them where they naturally grow? Or should we just get rid of all of that and everything is only accessible from the grocery store. That just wouldn't be good. And when it comes to open source, it's the same situation. Some people will go and solve a small problem and they'll put it on get hub, and sometimes that's as far as it should go. It should just be a thing on GitHub. It's a project. If someone really wants some features, they can pay a contractor to add the
feature. Maybe they make donations. I think let's encrypt is a great example. You'll probably remember our time where s CL certificates were high value products, like you could pay VeriSign four hundred to a twelve hundred dollars a year for
an extended validate. I don't even know what extended validate it means, but they would give an s sale certificate that would expire like clockwork, and then you'll give them twelve hundred dollars again for the same certificate, and the difference between the one you make and the one that they make is that they signed it and people trust them. So in many ways, they were sharing their reputation with you. Zoom two. Where we are now you have let's encrypt
generating what almost half of all SSL certificates. The reputation is high, they're built into every browser and is zero dollars cost to you, and it's a nonprofit. So if you consider donating to something, Let's encrypt is a good example of a project that doesn't need to become a product, and we should still try to figure out how to fund those things, but it doesn't mean that they have to IPO. And I think that's a little bit more sustainable
for most of the projects that are out there. I mean, I think that's a really good point, actually, because there's something very close to that in the same view. There's a BIMI certificates to get your logo to show up in emails and has almost no other value, and it's been trying to be sold like old certificates of old for twelve hundred and fifteen hundred dollars just
for your logo to show up there. And there's only two companies that offer it, and people are complaining that Let's encrypt, you know, isn't offering it yet. It's so ridiculous, honestly that some of these things have been productized and there is no way around that. I mean, look, if you can sell it, more power to you. But I don't think we should prohibit Let's encrypt likes so that shows up, competes with it and then becomes the standard. And I think a lot of people are afraid of that,
you know, And people say you've copied my project. It's like, dude, you open sourced it, you chose a license that said I could do that. I don't know how this is stealing or copy Like I thought that's what you wanted me to do. So one of the things I think would be really interesting to talk about is how you've obviously been very successful in your career, but that's only like one part of who you are as a
human. So what what did you do throughout your career to balance the professional side of Kelsey Hitar with the personal side, and specifically around how you maintain sanity and prevent burnout? Yeah, I mean I think I walked the burnout line most of my career, But I think it feels different. When you put a lot of effort into something and you get the results you want, it's the same effort. It feels like, Man, that was good, I want to do it again. So you probably need some rest, but
you're probably still working nights and weekends. You probably on call a little too much. But if you do the same amount of effort and you don't get the outcomes, like you don't get that promotion, you don't get the raise, it doesn't work. I think that's when all the other emotions creep in, right. That feeling sucks, and when you're tired and you have that feeling, it feels like burnout, and they're dangerous because they're the same.
And so for me, one thing I've done in my career is I kind of learned how to set like a goal that meant something to me and not necessarily the company. So if burnout would be I'm going to do all of this and hopefully I get promoted, and it turns out maybe because of politics, you don't get the promotion, and so you felt like the whole last three to four months was a waste of time, and you want to quit
something of y'all rage, quit, unplug the keyboard. I don't know what when you get to the other you drive all the way to work, so you can quit in front of everybody, right, because that's how you feel. But when you kind of think about yourself, I started, I don't know, maybe it's around year seven of my career. I started thinking like, yo, this job can fire me for any reason. Right, this is not my family, and even they can kick me out for any reason.
And so what skills do I need for the next gig? So I'm going to always just for my attitude and my ethics. I'm going to do as best of a job as I can with the current one. But if they're not teaching me the next set of skills I need to be successful, that's on me. And I'll give you a good example. On twenty twelve, I was working at Puppet Labs. Right, so I was in a financial institution contributing to open source. I learned Ruby just so I can contribute
to Puppet the configuration management tool. I gave a talk at puppet cough. This is why public speaking is great. All those open source contributions. You would never know my name given a talk at puppet cough you know who I am, and so now people can see me. And while I was in town, I did an interview and I got a job at puppet Lapse, and I let Ruby and those skills I was able to cash in on.
Around a year, maybe a year and a half, in a new programming language emerges Go Link, and I'm looking at him, like, why do we need another programming language we got? I thought we got Ruby, we have all of these things, and it was like, hey, this thing's in the middle between dynamic languages and things like C plus plus. And at the time I'm working on Puppet, and Puppet uses Ruby, and Ruby has
a global lock interpreter. It can really start to have performance problems at a certain scale, even for something as simple as like pulling configuration files down and trying to sink them across the set of notes. And so we started moving towards the g using tools like closure. But I saw this go a thing and I was like, h let me try it on one of our projects back there. Back there's a thing in puppet land that gets all the information about a node and sends it over to the main server, so it can
generate your configs. I was like, why is that in Ruby? Wouldn't it be faster and a single binary? No, nothing to install and not even been a Ruby and I got it to work. I was like, Yo, we might be able to solve our performance problem and we can ease installation. And I remember being like, Yo, you can't use go here because it doesn't run on SELAIS or AIX and I was like, what are you talking about? Like they'll get around to that, but I don't know
if that's where the industry is going in the enterprise company. And then I remember Docker coming out and it was written and go and I'm sitting here like I got a feeling that this is going somewhere. And then Terrorform comes out and I'm sitting here like what are we doing? We're not paying attention.
Things are changing, And so I left puppet Labs. I created a config management tool called comb using Go and ETZD, and I went to the next space like, hey, this is the type of thing I want to do, and Go just pay it off for me so much, because then it became my on ramp to coros and I had all the skills to participate and developed and open source my own project. And then the rose lead to Google, and so just that little bit of planning of saying I know where I
am, but just investing on what's next and just making that personal. That became a big part of my career. So the balance was not there just to be honest and transparent. I didn't figure out the family and work situation, but I did believe that these skills would contribute to the future that I want. And so last year I cashed in on all of those big bets and me and my wife retired at forty two, and so it kind of feels like, yes, I didn't have perfect balance the whole way through.
There were times where I did take good vacations and got off right at five, but there were times where it was lots of nights and weekends. But when I turned to my wife in July of last year and said was it worth it, she said, yes, we're both retired now, and so that's the way I thought about it. There's a long game component to this too, for sure. Yeah, have you found since retiring that you have
to struggle or that you're undriven in how you spend your time? Because it's like I think a lot of people look at retirement as a wake up play some golf, go hang out in the lounge, you know that, that type of thing. And I don't think that's realistic. I don't think that's sustainable. So have you found it difficult to occupy your time and keep yourself active mentally and physically? No, that's not at all, because you know how we are. Hey, babe, I would like a bidet. What
is the best bidet in the world? Right? That's thirty hours right there, just researching comments, Reddit post, Oh, Toto, Oh? People like those? One was Toto invented? I mean, have they sold? What's the history? Now I'm watching a total documentary on the daisis and how they were created and how they were tested in the engineering that goes into those. I was like, well, I can't just buy a bidacy without imagine toilet. So now I got to learn how to swap out all the toilets.
What's the difference between one point six gallons per flesh versus one point two? Oh, Washington State has a law that all new toilets must be one point two eight gallons of flesh for the climate efforts. So I'm now deep in this rabbit hole. I got three new toilets and three Beidet seats, but they draw fourteen forty watts of power. You cannot put that on a
normal electrical socket with your lights because the lights will flicker. If you do the math fifteen times point twenty, that gives you about eighteen hundred lots of power. Didn't know that. So when you turn on the Bidet seat, the lights will ficker because of the draw. And some of these lights are just sensitive the break it doesn't trip. So what do you need? You
need more power. More importantly, you need dedicated power. So now I have holes in my garage ceiling because I'm running twenty amp circuits only after I learned how to pull a permit from the county. I have no short of things to do. And I would say one other thing. I just started saying yes to everything. Do you want to go to my friend's birthday? Probably they'll be dancing and eating. I don't like to dance, but now yes, and I'm going to dance anyway. And so it's a lot more
yes to the people's stuff. It's a lot more patient in who are these people? Talk to them being interested. There are other stories out there that are not tech related, and so I think I just started opening up myself more to just more things, because you know how it is when you work nine to five, that means someone else finds something to do with your time from nine to five. They figure out how to maximize their investment, and all you have to do is can you not do the same, And so
it's more of it's not about not working. It's more about working for yourself and having more time to work at being a good human, being a good father, being a good husband, being a good all these other things. They're kind of on the hook for as well. What's been the most surprising thing that you said yes to that you were like, Oh, I didn't know I was going to like that. Honestly, it's going to be those parties where there's lots of dancing and getting I hate getting dressed up. I
don't like shirts with collars. I bought my first blazer a month ago. Dressing up hate my thing. It's like, oh, you got to go through this, What do you mean it has to match? Oh my god, we're not going to the met Gala? Like why You're just supposed to go as a birthday party? Check e cheese, some coins for the machines, pizza were done. What what's all this other stuff? And so you just say yes and then you go and all these people have different backgrounds,
you listen to them. I like dancing. I used to dance a lot when I was younger. You can still have fun dancing. You don't need to be the best dancer. You can just be there and just jul and just be in the moment and not rush so much. So what I found is I like just going to these random events as long as I have the time to just be there and not worried about some email poling up thin or about getting power to the days. I like those two. I feel like
the flush followed by the lights flickering. That's sort of a special sort of experience. Yeah, you're like, what's wrong, and then you start doing the research. But I've been so Yeah, I think it's being important to know how these subsystems in your house work, from the plumbing to the AC to the electrical to how the Floyd Joyce align. And so my toolbag is
getting pretty massive now as I'm collecting these random tools. But I don't know, I just kind of see it like my software and my tech career. The more tools you have, the more problems you can solve when they come up. So this is just another investment for me. Yeah, I mean it's pretty related from my experience having worked with and hired through a lot of people that how you operate in a professional environment is very similar to maybe your
core values and how they come out in your personal life as well. So I mean, obviously you've transitioned and you're still going out and looking into how to solve these problems that you're generating yourself, but important to you in a similar way in which you did in your career. Absolutely, Home Depot, Lows or other Amazon, mainly because of the ability to select and choose.
Maybe I just like the home Depot color scheme and their fall to home Depot, and I think Lows if they carry the thing I want, I'm just going to buy it there. But honestly, you know how this works. There's always like A and Home Default across the street from each other, and my guess is where I get off the freeway exit. Whoever's first, you're winning. That's one of the things I do like about having the physical store
close. Whenever it comes to tools, as you can go in, you can pick it up, feel it and kind of get an idea for it that you don't really get that experience when shopping online for it. Now, I do try my best to go to my local hardware store, like those little you know in your neighborhood. They have some obscure things that you need last minute, and they're close by. I try to give them as much money as possible when I'm buying this stuff, because it's really convenient to have
one closed. You know, how to drive, you know, five or ten minutes down the road and some people further. And so those small businesses they tend to be helpful. They tend to answer questions that other people don't have time to answer. So I do try to support those local businesses as well. Yeah, there's an A Hardware just near us here, and all the ACE hardwarees I guess are independently owned and operated. But you're right, you know, whenever we go in there, they're always like, hey,
how did that last project turn out? You know, and you get this rapport with them where they're kind of they're kind of emotionally invested in whatever you're working on along with you, which is cool. Yeah, one of the biggest surprises that I know a lot of people probably think this should be normal. When I was going to like those trade stores, like where plumbers buy their equipment and you know, wholesale various things, I walked in, I
told the person my name once. Then I came back, and as soon as I walked into store, it's like, hey, Kelsey, and I look around, like, you can't possibly know me from that. You just gave me a quote lest time, and he made it a pride to say my name. And I just thought that was amazing that they pay that level attention to detail, because I think when you're selling what is arguably a huge commodity, they don't necessarily make any of this stuff. Any other store could
sell this stuff. The level of customer service and the way they treat people, it's just a good reminder is that sometimes that makes all the difference. Yeah, It's amazing how much something just just remembering your name leaves such a favorable impression. I had a similar scenario about a month ago. I went to an irrigation supply store. Didn't even know those things existed, but it's just like warehouse place that just had irrigation, you know, the sprinklers and
the valves and the manifolds and stuff. And I go in and the lady that's working there, I kind of but butchered out in my best English what
I was trying to do, and she's like, oh yeah. She hands me this empty cardboard box and it's like follow me, you know, and takes me through these aisles and she's throwing shit in the box, you know, and just completely hooks me up. Then I get back and start piecing all this stuff together, and through that little bit of a description, she gave me everything that I needed to complete the project and nothing I didn't need. And I was like, Wow, that is that is so cool.
I mean, the DeVos folks listening, I don't I don't think people understand you're also in the customer service business. Yeah. A lot of times I think people get very confused. You work at this company, this is how we do things. Look, maybe you're even open to a new ideas, maybe you're like doing this DevOps thing. But truth is, I don't think
people realize that you're in the customer service business. When someone's trying to get something accomplished, typically they're doing it because it's the work they didn't wake up one morning it was like, you know what, I want to do a thing we don't need to do like they typically need to have it done. And if you could be like that person that says, here's everything you need because you learn your job so well that you know that even though they didn't
ask you, you're just making sure that they're set up for success. And that's where those kind of I don't know, those getting started guides, these one button deploys, all of these little things we do just to make sure that people don't struggle or get in away of the big job because in your situation, you have a bigger problem that you need to solve. I'm only coming to the store to get the tools so I can go back to the
big problem. That's why I like referring to go ahead, Warren. No, I'm just saying, I'm so glad you didn't say, like just so they could do their job, like just so that they could they didn't have to worry about this, because I feel like it's so often in the industry that I hear anyone who's outside of doing on feature related work that oh, I'm just here to make it so that no one has to focus about this
area at all. But I think realistically that's not what's super effective. And I think the you know, the story about pulling out all of the parts for water reclamation, et cetera. It's you have to know what's going on there. You have to be able to have a conversation and collaborate and give them what they need. But it goes both ways, right, you know they have the project idea in mind, you don't just hand them a one
click deploy. You know, that's probably the right answer, but there is some conversation about what you're working on to actually even make that recommendation in the
first place. Yeah, I wanted to add that. That's why I like referring to DevOps as a trade so much, because it mentally I associate that with you know, electrician or a plumber or a carpenter, and it's a you know, a skills a skilled trade, and it makes me think of someone who's gone through the ranks of apprentice to journey on and taking the time and effort to master that trade so that they know the ins and outs of it and they know which tools to use to help you successful in whatever it
is you're trying to accomplish. I like this analogy a lot because that was kind of always my hang up with just the word DevOps. You can just use the title whether you're a trades person with the right attitude or not.
You just said, we're just doing DevOps. How Like, when you become an electrician, there's a lot of stages, like you got to be like a journeyman for like a decade before you consider like a master electrician, and every job you do there's a knowledgeable inspection like this doesn't meet code, and you have this sometimes it feels like a rigorous feedback loop, but there is
a feedback loop saying that's not quite right. And then you have a group of people that can look at each other's toolbags and like, yo, that's what this thing is for. This is this particular procedure. A lot of times what I've seen, and look, it's not everyone. There are some people who do this immaculately well, whether they use the word DevOps or not,
they're just going to be good at their jobs. But there are some people who don't really understand the job, but they've adopted the title, and so in those scenarios, just like yo, I mean, we've just been running this raggedy ass script for seven years, but it's DevOps now because we had a flags to it. It's like, what are you talking about? There is this part where you have to pay attention to how things are changing.
If you're going to be a professional electrician, the standards change, and your job is to have I think they require so many hours of continuous education to make sure that you're up to date on the various specs and codes. I've watched a lot of people not learn anything like even if you're not going to use there for you shouldn't know what it does. You shouldn't know how
it compares to the thing you're currently doing. I'm not saying you got it all cargo cult the next new thing, but you have to understand why the new thing exists. You have to at least be able to compare it to what you're currently doing so you can make an informed decision not to use it.
And so I've seen a lot of people just holding back themselves, their team, or the whole organization because they forget that professional development component that, Okay, are my skills where they need to be, Are my tools where they need to be. I'm gonna like, let's be clear about what I mean. Are your tools where they need to be, Like those professional ideas
of like why would people ever use those? And I remember I was trying to get something done and the language is never used, and I pulled down one of these, like jet brains ideas the auto complete, the debugging, some of the things they do, because as a full time software developer, I don't have time to be switching between seven five. It's gonna cost me too much time. And you see this where it trades people some of the tools they have, Like why do they make a wrench just to do this
one thing? It's like because without this wrench, you're gonna be back and forth with the wrong tools. You're gonna break something, you're gonna cut yourself. That's what this tool is for. And so I think, as you're gonna be a person a practitioner in the debos movement, nothing wrong with being where you are, it's just that that can't be the case five years from now. Yeah, for sure. It's not the type of career that you're going to land your first job and you can ride that thing out til you
hit sixty five and start collecting social security. Well, I think the mindset is sort of the whole point there. That the mindset is that you have to automate your job away almost you want to understand what's an improvement for that even that space. You know, does that mindset translate to the team that you're working on or the organization making the shift? If that's not what's happening, you know who is who is responsible for that? And you know,
Kelsey, you said something here and I feel like I got triggers. Home may have heard it and be like, oh, you know what, there's no like DevOps certification yet, right, maybe we should go out and get one of those. You know, you know there's some company that can start selling this for sure. These are the tools you need to be have in your and your toolbox in order to be an effective DevOps person. You know what you know? I might I might say maybe if it was from a
good place, if it came from a right kind of moment. For example, you said the mindset, it's hard to have a good mindset without accountability. I'm good at what I do. How do you know? You don't know? Right? Because most professional endeavors, like when I say professional, meaning they're judged like, yo, you are not good, you don't get to play on the team. You're not good, or you're good, but you're no longer competitive. See that's different. Like you can have ten years
of experience, you've done good for us. But the new people coming into the game, the game is changed. They move faster, their tools are better. As much as we want to keep you around, we cannot. And so that's when a mind mindset shift has to occur because you just won't have a job. But imagine scenario where the company doesn't hire those type of people. They don't want to pay for those types of people. So you're in your little bubble, still doing stuff from ninety eight, like I'm killing
it. You see this corn shell. The spaces are perfect. It's like look the Roman corn shell. I'm just saying, for what you're trying to achieve, there might be a better way of doing it. Like, no, it doesn't corn surre. It works just fine. See that's the dangerous component that I think is the trap that a lot of these you know, whether it's devopsis on their movement, tries to address. And I do think
awareness, so I'm not necessarily the biggest fan. But it's really valuable having things like the Devous Report because when you get to see what other companies are doing, that's extremely valuable because it's hard to get better if you don't know where the bar is. And so I do think people have to go and just at least do a little bit of self reflection and look at these things and say, hmm, has the goalpost moved and should I move towards them?
Yeah. I mean there's an important aspect here which would be a huge miss otherwise realistically if they're not paying attention to that. I think an important aspect here is if your company isn't changing, it's really because they may not even know, right If you're in the position of doing the in what you call DevOps, you know what, if that's what you're doing, if that's even your title, or maybe you're doing something that helps and just supports the
teams by building some tools, it's you're not innovating in that space. If you're not going out and learning those additional what other companies are doing or what new tools are available, even how the mindset is changing, there may not be anyone doing that else in the company. You know, that that means your company won't be effective, your company won't be innovating if you're not the
one doing that. And I feel like if you're just focusing on the tools you've got and what has been working for the last ten to twenty years, you're definitely gonna be falling behind. Now. One thing I will say, though, there are some things that don't want to change. And so I know we say innovation a lot, and so much should be innovad for sure.
Somewhere there should always be some form of innovation happening. It's harder than most people think, and sometimes you do just need companies to be in execution mode. Like some of these microphones like this r A twenty that I'm using from electro Voice. These things are roughly the same for decades, and when you change them because you're trying to do innovation, like at a USBC port on it and it throws off the sound because of the circuitry that you put
in. Don't want that innovation. Leave the XLR cable and I will plug it into my own and don't change it. Now. They've innovated to get to this point, no doubt. But sometimes people say, hey, so it is a It's such a hard thing to do. It's like when do you change and when do you not? And I think that's the part where a skillful person will constantly be aware and have a reason. We do not
need to use this new thing in this context. So I just want you to be aware, have the ability to But boys judgment so important to keep you from chasing new and shiny versus things focus on the task at hand. I use a personal rule here. I call it it is my rule of three Warren's rule of three. If I hear about a product or a tool, or an area, mindset, whatever it is, three times, I know that I have to invest in understanding what that is more either because if
I don't, no one else will do it. Maybe it really will become my problem. It will land on my lap in some way, and I'll regret not having done it earlier. Or I may just be the only person who is at the intersection of where that keeps happening. I am the right person to actually take that next step. So if I say beetle Juice, Beetle Juice, Beetle Juice, that's three or eight the second movie coming out there, well, right, And I cannot tell you how excited my wife
is to see it. It's like, oh, I didn't know this was a thing, but okay, let's go. Peter Juice is a classic. I mean I went back and watched it. I was like, yo, I mean it's you got to be into that kind of thing. But it is a classic for its time. And even the rewatchability was pretty good. So the rewatchability was good on it. I mean part one. I mean it's it's a quirky movie. So yeah, sure, I thought were amazing. I go back and watch it, like, yo, this is pretty
terrible. But Peter Juice has its own unique style and story, format and visuals. I think it's worth a rewatch right on because I remember seeing it whenever I was younger. But but like you said, so many Like there's so many movies I watched when I was younger that I just have built up in my mind and then I go back and rewatch them. I was like, oh, I was I should have just stuck with the memory on that
one. I'm lucky if I remember the titles because I saw so many movies where you missed the first couple of minutes And this is back of course where if you don't see the title screen, that's it. You know, you'll never know again. If I'm lucky, like years down the road, I'd see a rerun of it, and I'll watch the beginning of see the name of the title, and then wait, I've seen this before, but I have no idea what it is, and then I'd put it together. Should
get entertained twice. I mean, there's movies I probably watched ten times and still don't get the title, right, Lie, I think it's The Last Dragon. Yeah, the Last Dragon, I believe, And I don't know the title. But it's like a spoof on karate movies. But it's a serious movie, Bruce Lee Roy. They got this glow thing going and it's like in like the eighties settings and you got black folks doing kumfu and karate and Harlem. It is inside this movie. If you haven't seen The Last
Dragon, you will. You don't know whether to laugh or take it serious because it's that type of movie. So have you've ever seen like Big Trouble, Little China. It's now of those. So if you've seen Big Trouble, Little China, then you have to add The Last Dragon to the collection just to round things out, all right, I'm I am. It's got to be a black guy doing this thing. If you see that Bruce Lee Roy, that's how you know I'm not talking along with Bruce Lee. No,
this is Bruce Lee Roy. Seriously, and you watch it, there's the bad Guys showgun like. Trust me, I think you're going to appreciate it. I know what I'm doing tonight. So do you spend a lot of time watching movies? When I do? But when I do, I try to really lean into that one movie. I try not to judge whether this movie is better than that movie. You just you're gonna watch a movie that's like two and a half hours of your life. I just try to
just really appreciate that movie. And I used to have this. I don't think it was a good rule at all. If I've seen it already, I'm done. I don't want to see it again. I already saw it. But now what I'm realizing is that second and third watch you're starting to notice and pick up on things like now that you got the plot, and you go back and you just watch how they build up to it, sometimes you can respect it more. Sometimes instead of you because I used to get
nervous when you're watching a movie. I don't want to pay attention too much of this one character because they may not be around the whole movie. But once you go through it once, then you can go back to the character development. Why do they make these flaws decisions or good decisions? So yeah, multiple times, but I'm a little bit more patient now. I for sure had the same problem very early on. I was like, I don't need to watch anything again. It's already ruined if there's a spoiler or a
twist. And I was a prettyularly a fan of film noir. But I think what broke me was actually rewatching Futurama, where it's such a high level of intellect gone into building every single episode, and the story is so continuous over the original seven season run that there are some little jokes that you don't get. You for sure will never get that you only know because you know
the later content. And you come back and there's an episode where someone says, oh, yeah, that's ridiculous, that could never happen, but you don't know why that's the case until later, and I feel like, you know, you get double entertainment out of some of your favorite movies though, for sure. But you know, it's odd though we don't typically do that
with music. We can listen to the same song over and over and over again, but movies, for some reason, we feel like, you know, maybe if the plot is ruined or we know how it's going to end, but we know how songs end, and we still listen to them again. And the one thing I'd done on the music side was listen to it sometimes with the lyrics, Like on Apple Music, you can bring up the
lyrics and so as you're listening, you can also read the words. And sometimes I would just go read the words without listening to the song, because number one, I'm not I can't sing, And it turns out I don't know the lyrics to the majority of the songs I listen to, and people like, you know what that song is about, right, I'm like,
no, I don't know what that song's about. And so now I found it really important to understand what the artist was thinking at the time, who the artist is, and riot they write the song the way that they wrote it, because it's so important. It brings so much depth to music, like In the Air of the Night, Phil Collins, like that song is super deep. Most people don't really understand what it's about. But when you read the lyrics and then your mind starts to like fill in the gaps,
You're like, oh, this is a very serious song. And then you learn the context of the song of inviting the person to the concert and then singing the song, It's like this is much deeper. So then it finds its place in kind of my top top songs of all time. I see that there's a huge pattern with how you approach just whatever you're focusing on,
whether or not it's music, movies, personal life, or professionally. This idea to really have the discipline to go really deep in whatever that thing is that you're working on, not there's not just one more thing or whatever I'm doing at this moment, but real one hundred percent focus probably more on that area so that you can fully understand and appreciate whatever that is. And like, I feel like there's a lot of aspect where I do that, but
it really does require intentional discipline to pull that off. It's not something that happens accidentally. I mean, I think I'm afraid of things I don't understand, you know, what I mean, like, what does this thing do and why is it doing it? And so whenever that question comes up, I feel like I got to solve it. You know. It's like when people like I'm one of these people that if I see a puzzle, it's over for me. And my mom knows this, Like I'm forty three years
old. When I go visit, she has a new puzzle on the table and she opens the box, puts it next to it, and there'd be the pieces scattered out, and I'm like, oh my god, this must be solved now, and you're no way out, say, I gotta do at these a few pieces before we leave, and before the trip is over, I am not satisfied to that puzzle issaulved. And so my real life is like this too. It's like why did someone do this? Why is the stock its current price? Does my phone listen to me? If so?
How? Like I just need to know these things. So it also keeps me from I think a lot of stress people have. You're trying to pay attention to everything and then you don't know how anything works, and so then you're just you're not really growing or improving. And these people are very succeptible to chatter propaganda. They don't know what's true or what's not because they don't have time to process, so everything is just streamed at them, and
then everything just happens to them. And so I think it's like when you when you first get into tech and you log into your first server s A stage X, Y Z, and you're in and you're like, oh, and so you can actually go really far in your career just run in commands over and over and it doesn't work. One what do I do here to say? Run it this way? And you just run it this way.
You're an operator. But the moment you write that first shelf script and you make the other scripts do something different, that's when you feel like, hey, I'm not just here to be a consumer or an operator. I can create stuff too. But the key though is you have to know how things actually work and then you can rearrange the pieces. And so when it comes to anything in life, it's like, hey, you can't learn everything, Kelsey, Like, you're right, I'm just saying learn the things you're dealing
with in that moment. So if you're gonna be listening to a song, just read the lyrics who is the artist, were they in the band previously? What other music did they do? What genre is this? Do you like that genre? Who else is in this genre? So then it just gives you a chance to go super deep on this thing that's in front of you. And then hopefully by the time you're also forty three, you've done this so many times that it does feel like you're a knowledgeable person because you've
taken the time to go deep. What do we call these people, We call them like t shaped engineers. They have the breath, but the depth because instead of just like learning a bunch of things, they went deep. And so, yeah, it does take a long time, but when you're done, you have both. Do you have a favorite genre of music? I grew up on hip hop? And the reason why that's so nostalgic. As a kid, good things would happen in your life right like you would
go outside, you would win the game. You just got a new Nintendo game for your birthday, So you remember the song that was playing right like you're listening to your back of my day to walk Man, and you played this game. You had a really good experience playing the game, and this is a song you heard first, and that song at just the memory and
links everything together. And so for me, that was hip hop. And then when I'm reading the lyrics around hip hop and studying the origins of hip hop, you think about this group of people in the earliest forms of that genre. You can't afford instruments and creating music from scratch. A lot of people didn't go to school for music theory and knew how to read or write sheet music. But what you didn't know is like what music you like and what music other people like, and so what do you do? You take
records and you find the part that everyone likes. How do you know it's a good part because you're watching people dance and when it gets to that break, you're like, oh, people seem to like this part. So you take that part and you loop it. Not with some fancy computer or screen. You literally got two turntables and you're marking on the record where the good part is and you loop it right, and then the next person is like,
yo, I can rhyme over the top of the good part. But you just got to keep it going long enough for me to do this. And so you just think about how technical you need to be to produce hip hop, you can't play any instruments, you can't necessarily read sheet music. I'm not saying everyone's like that, there's so many people in this genre then and now, but to still be able to create something that has such a
global impact. And also when you look at the history of a lot of those people, they're coming from some really rough environments where there's a lot of politics, things that are happening to them, various movements. And for anyone that writes blog posts, journals, anything writing is thinking and a lot of those people when they wrote their rhymes, they were showing you or expressing their situation. They couldn't get media coverage, they're not on the movies, they're
not on the TV shows. They got their music, and so hip hop to me is one of those things where it documented in my life. You're hearing other people document theirs, and it's just one of these genres that resonates so much. But then the bleeding and crossover because when I go listen to other genres of music, the drum patterns, the rhyme schemes, the collaborations. Right, So the first time I remember listening to like other gramas of music, when I heard, like a rap artist and an R and B
artist get on a record together, I'm like, what's that? When I started to understand where the samples came from, Oh, what is this disco thing? What is this jazz thing? And so then it just kind of branched out into these other gramres. So that's the nucleus for me, but it branched me into all these other ones. So I love shade. Sometimes I just need to be calm and relaxed, so I'll jump into shade. I grew up in southern California and I remember hearing Nirvana teen Spirit for the
first time. Right, I'm all in my hip hop, but when you hear that song, I don't care what genre of music you like, that one's going to jump in. And then last thing I would say is when I played guitar hero, I was never into some of the rock and metal stuff. I just I never branched out. But man, playing guitar hero and you get introduced to all of these bands, and there's something about paying attention to a song. I've listened to rock in the past, but it
didn't jive. But it just didn't make sense. But then when you play guitar hero. You're not reading sheet music, but you kind of pay attention to the song so you can play the notes. And at that moment I was like this, I can't even push the color buttons fast enough. How is Slash doing this in real life? And so then I became respectful of
the instrument and then the people who could play that way. So yeah, that's been my kind of overtime expanding into the other genres, but all stuff with the respect for the instruments and the artists that write the music that goes
with them right on. Mine was almost the exact opposite of that, because I had a like, growing up, I heard a lot of country music and like what's now called classic rock, and then later whenever I started hearing hip hop, I picked up on the sampling like wait a minute, that's a led Zeppelin riff or wait a minute, that's thirty eight special, which kind of then just forced me deeper down to that to see, well,
if they're sampling that, what else did they sample? You know what, when run DMC, I forget what band they collaborated with, Aerosmith Aerosmith, and you know, I should have known who Aerosmith was by then, but I didn't and just see them at the same time. I was like, I like the way the sounds, and when I started listening to Rick Rubin and how he made music, I think what made Rick Rubin is so unique.
He brought in those elements into hip hop and so for a lot of hip hop artists, that whole brand of music from def jam in those days, you were hearing this ultimate collaboration and crossover for sure. Yeah, Rick
Rubin was. He was a very i don't know if you want to say, forward thinking, but innovative, like the way he just listened to music, and then I think he was very instrumental in bringing those different genres together and letting them capitalize on what was successful from the others and bring it into their own genre. All Right, Well, I kind of feel like we covered a few topics here. You know what we should cover AI. Oh,
let's talk about AI. I've been seeing you post a little bit on X Yeah, so which thinking so like a lot of you all when I saw early AI, like the stuff that's inside of your Google Maps, the stuff that's been in your Gmail, the stuff that has been in a lot of your sensors and cameras and all of these things. Like you've lived with this so much so that you don't even talk about AI anymore, just like we don't talk about syntax and the products that get shipped, right, It
just it becomes an ingredient of the whole. And so this resurgence of large language models, right, this would be your chat GPTs, you know, Lama from meta and these things are interesting because people are trying to build this general purpose tool that can interact largely with the language driven interface. And for a lot of people they haven't made the connection on where this fits on the lineage of AI mL, But it's the first time that you're seeing something that
looks like you do. Think about it. When we say prompt engineering, this idea of asking a thing what is is my birthday? It doesn't know because it has no context, doesn't know who you are. Now you could say what's my birthday? What is Kelsey high Tower's birthday? And it's going to guess from what from whatever it found on the internet. So that's a training set. But when you interact with something like that, it feels a lot how we interact with each other. I was reading somewhere earlier that it
takes twenty watts of power for your brain. And your brain, look, you don't know everything and you can't memorize everything, so that's your disadvantage. But your advantage is that you can generalize human with new problem on twenty wats of power. You can do with a room, a building full of the most expensive GPUs. We're talking killer wats of power times in number of those units, right, so you're a factor of multiples of what you can accomplish.
What's really intelligence and the things we don't know yet. And so when I see people look at this new paradigm, and it's scary because the only thing we've seen interact this way are people. And for some reason, we put people at the highest part of the hierarchy in the galaxy. Right, people are at the top, the most intelligent. They can be dangerous, they can be safe, they can be kind, they could be mean,
they can be whatever. And we're usually only limited by our physical abilities or the machines we tend to operate, and you can fill in the gaps. And so when we say that we're going to generate a thing that's a void or void of ethics, reason, purpose, and give it access to all the power and let it operate the same machines. You all get nervous. So, in simpler terms, if you're in the DevOps or you write code, you're thinking to yourself, what do I do? And at a high
level you kind of get driven the same way. So one opens a JY or ticket, I would like X. That's the prompt. You convert that prompt to the best of your ability. Remember that thing where you can't remember everything and you haven't seen everything, or those lms a little different. Usually they've seen it all, they've seen every repository on GitHub, and they're not biased. They don't care if it's a Ruby or Java. I've seen a lot of people who say I will never touch anything of written in Java.
Fair enough, that's your preference, but that's not what that thing does. This thing learns everything with very little biased except for the weights we create, and so those biases become the model. Right. You have a model, too, and your model tends to go under invested. Right, So think
about it. When we talk about like the new version of these llms, what we get excited about is the amount of maybe net new training data or unique training data plus all the work we put into making the model and adjusting those little weights like dipswitches to try to tune that model hopefully towards the correct answer. Or we're working towards the truth. So you put a natural language in front of this truth model. Hopefully it's truthful. This is what you
end up with. Now, Look, they're far away from that. So this one thing I want you to take comfort in. What you do on twenty wats of power is much more efficient than what that thing does. But your biases are probably holding you back. So when people think that a LLLM like CHATGPT is going to replace your job, it could if you think of your job as simply taking prompts from others and just doing things. But what is your job. Your job isn't to just do what people tell you and
run commands. That's the thing we were talking about earlier. Your job is to form a better model of the world, of your team, of your company, and your problem space. See that model is hard to compete with what you currently have. As long as you keep tuning that model, you're going to be hyper competitive. And then here's the punchline. That thing is
just a tool to compliment your model. That's the trick here. So a lot of people on DevOps are really nervous, at least the ones I've talked to number one, because I don't think they understand how this thing works. They haven't really walked through the depths of what it's good at and what it's bad at. So there's just a bit of fear. Remember that thing we talked about, just the stream of stuff hating people all the time and things
are just happening to them. So I really want people to rethink this AI thing. So when I was growing up, the last thing you wanted was the artificial version of anything. The real version costs more. It was real for a reason. You're still the real version, And I want you to just to think about what the real version can do. And I guarantee you
it's undertrained. You haven't thought about walking around with one of the world's most sophisticated computing devices and also the world's most efficient computing devices if you thought about it that way, right, The same effort that Nvidia puts into the next GPU release, the same effort open ai puts into the next model release. Imagine if you put in a fraction of that effort to training your model and then releasing that on your colleagues, the world, your family, and yourself.
So that's kind of my thoughts on AI is so important, not because you need to figure out what's the best model, just remind yourself that you
got one too, and it's also worth investing in. Right on fair point, I think, yeah, I think for me, For me, I've been using it as almost just like yeah, a colleague to bounce ideas off of, you know, whenever I'm whenever i get the JIRA ticket and I'm like, oh, I'm thinking about doing this, you know, I'll use it as something to balance ideas off of because it has such a larger set to work with and see what other ideas it has for solving that problem that
I hadn't thought of. But I tend to agree with you that I don't think it's at a human replacement capability for most scenarios. I think there are some where it probably could. But for me personally, which is all I can speak to. My job is to solve problems, and even whenever it starts solving the problems that I'm currently tessked with solving, now that just means I'm free to go find other problems to solve. And the other thing I try to think about is a lot of the tools we have don't have good
APIs or interfaces of interacting with them. Let's just think about this for a second. In the cloud, there are thousands of resources with thousands of parameters. It's impossible to find the right permentation of those things together. So what are we given? Polumi, terraform, puppet, chef antable and these are fine, but they're almost one to one leaking all of those details. They just give you a way to express your choices in their language, but you
still have to be exposed to all of this detail. And so that's just a very static way of solving problems. It works, you know, we solve the problem once we go through all the pain. We still don't know if we're making the right decision, but we at least will make a decision form consensus, and then we write it down in HBL as say terraform apply. Nothing wrong with that, but the interface is not good. I used
to work at a cloud provider. What should the interface be? The interface should be I need a database that can hold this amount of data, and I need it in this particular reason because of these particular compliance laws. And I also need a set of credentials that only let these apps. That's how
we would think about this. So imagine I take that and give it to the system that's enough information to set up a VPC in the right zone or region with the right set of credentials, lock everything else out, generate the IM components, make them available, and then return back to someone. Your database is ready. Here are the IM accounts that can log in the IP addresses over here. You're at this version and it meets all the compliance things.
So what does that tell you? It tells you that the systems that we currently target don't have great APIs. And so when I think about what L Andm's do, and what fascinates me the most is they tend to bring a much better API to the systems we interact with, at least in that
specific domain. And the thing that made me less scared of what it does is the machine that it targets is still roughly the same, right because it takes your prompt and it translates that into something that it can understand and if I can use that word, and then it still has to manipulate the same underlying machine. The cloud provider is still the same, the low balancer is
still the same, the config is still the same. It's just that the model for gluing these things together is refined, or can be refined to the point where we can get people a more natural interface, hence large language models. And so that's the part where I'm like now going back and looking at all my tools and saying, man, what would a better API be? We've been so locked into it has to have a RESTful API, get put, post delete, Jason object, and what is that I have to leak
all the details about how the machine works versus a high level API. And I think that new high level API the thing we've always been afraid to do because of composability. A high level API, whether it's using a large language model or not, you can still take the same expression and just really combine things using the best practice. And I think that's what a lot of people
in devov do with our golden paths. Right. If you do things with these inputs, we already know how to wire up a better machine, so you don't have to piece it together yourself. I think a specific example that resonates with me on that is least privileged security with cloud providers, because like everyone is like, oh, least privileged security, least privileged security. But every piece of documentation, every being a relative word, of course, every
piece of documentation references admin privileges when implementing it. So you're in this iterative loop that lasts for hours of trying to figure out which permissions you need. And I think something like a well trained LLM, you can tell it which permissions you want it to have and then it returns the im roles that you need to implement without spending hours beating your head against the wall. Yeah, we're going to get a So what MS are teaching me is that when you
design these systems, you can't design them in isolation. You have to be aware of contexts. And this is why I think when you say neural nets and the network that goes is that share context that these things have and their ability to try things until they find the perfect combination. But again I can't struss enough. A lot of these systems are tuned, so they're guessing and when they're wrong, we can tune them and then we have a much more
dynamic, expressive system and framework. So if you learn anything from this whole chapter of Lams and this movement number one, we can build better systems. We can give them better APIs and interfaces, and it will feel like sky nets, even though it won't be. He wants to intentionally scare people. I mean, I think a lot packed in there, realistically. I mean, I think there's something very on point about how we've been developing maybe not
even an API effectively, but really the protocol. We're just doing a better job exposing what you have access to and never really sort of abstracted it away in a way which actually really makes sense. Yeah, I think that's a really good way of looking at it. It's just an abstraction layer that is more suited towards talking to humans than a rest API is. Well, I
don't know. I don't know if I agree with that part. It's sometimes very challenging for me to get allance to spit back out what I'm looking for in a reasonable way. I know other people are much better than myself, but there is something challenging about the natural way in which humans speak and then get that converted into what's been seen on the internet into a relative common form. Yeah, we're making progress, you know, I think that's the way
I kind of look at it. We're making progress. Like SQL was a huge progress on structured data and how we get at it and insert it, and then trying to do that with natural language is a little more challenging, especially when these things aren't tuned. But we're making progress and in the highlights
an area of opportunity. Yeah, I mean, I think we're right now in the sort of map phase where we're seeing everything it can do, and very quickly we're also going to see the reduced cycle where very specific use cases have a much better interface for interacting with them as we identify what it's optimized for. And we've already gone that far with whether it's good or bad.
You know, it was still a question of content generation for words, audio and video and images, like we know that that's a pretty good use case, so we've improved the interface. They're much more on top rather than interacting with just text you know what I want, Like I'm I'm personally less excited about the generation. Maybe for things I can't do, like I can't draw,
so something that can help me draw, no problem. But for things like this where We're talking into these mics and degrees of background noise, and it's hard to get the levels just perfect. Just like you know how sometimes you use a really good plugin or audio plugin or filter and just makes it sound better. Just give everyone MPR studio level quality from any mic. Right,
I'm already generating the content. Don't waste cycles doing that. Just make the audio sound perfect, right, the thing that I've already generated, Just make it perfect. So anyone enhancement is probably what I'm really looking forward to. So that's the highest thing on my radar. I do have a meeting in two minutes. I just realized that for my calendar, so fair enough. Thank you so much for your time, man, I really really appreciate you taking the time to talk with us. I've enjoyed it a lot,
and you are more than welcome to come back at any point. Would love to have you back on Awesome. Well, thanks for having me. This is dope, and I also like your all interview style and format. As you know, the hosts bring out the best in the particular interview or a person being interviewed, So I just wanted to give a shout out to you all. It felt really nice. Awesome. Thank you, Kelsey, have
a great day, and thank you everyone for listening. I hope you all enjoyed the episode and we will see y'all next week.