Episode 003: Jason Mele

Adam Pippert (00:02)
Awesome, and we're live. This is great. So for those of you who may not be aware, the first time Jason and I recorded this particular episode, I thought that everything was all set up. I was actually out in the middle of a park. I went to Tryon Creek State Park and recorded my bit of the episode because I was out in that area. We had a really great conversation, and then I went back to go upload the recording for the episode.

and it was gone. It was not there. So Jason and I had to unfortunately get back together and do this again. I'm actually at my house now, so the background is a little bit more stable than having just tons of rain pouring on my head while we're talking, but hopefully that translates to a good conversation.

Jason Mele (00:51)
I really enjoyed the rain vibe and where you were just like field recording. You're like the weatherman out in, in a hurricane reporting on, you know, news from the tech world.

Adam Pippert (00:53)
I know. Yeah.

Yeah, and because I travel so much, I have had a number of episodes that have been in various different spots, but I try to record as many of these at my house as I can because it means that I know I have a stable internet connection and everything works out appropriately. So Jason, let's go ahead and start by having you introduce yourself and tell everybody who you are, how we know each other, and what you do now.

Jason Mele (01:29)
Yeah, hi, I'm Jason Melee. I have worked in the tech market for like 16, 17 years now. Currently, I lead the technical account management and professional services group for streaming media and investor relations company called Notified. But I know Adam because we worked

kind of a dawn of streaming back when like, I think if I recall, they were just starting to launch like streaming on devices, which was like one of these newfangled, like who would wanna watch a movie on their phone? And now people are like, who owns a TV? I think we had, I think I got sat with you my first day to shadow and I think we ended up talking about music instead of like,

you know, context of what I was supposed to be doing for my day job, which is what I end up doing to everybody. Cause like, and that's the context of what, why we started, like the idea of me coming on this podcast is that like all of the best people I've worked with this, like the most creative, smart people are like, Oh, I don't have a computer science background like me, like I have a music degree and, uh, or they've got a, um, Oh, they got a BFA of fine arts or a philosophy degree. And, you know, um,

think we start at this conversation because it's like in the AI era where the idea of just sitting around writing code maybe going away our creatively minded people maybe better suit it.

Adam Pippert (03:11)
Yeah, and I made a post on LinkedIn where I commented that the best thing about LLMs is it actually makes my liberal arts degree relevant all of a sudden. Because we do see a lot of people that are going back and harkening to their roots of critical thinking and putting together thoughts and structures and sentences in a way that's not just like, let's go bang outlines of Python, right? So that skill set combined with

the understanding of tech and the understanding of its limitations really makes artificial intelligence in 2024 very, very powerful for the right kind of people. I think that we're in a unique spot for sure.

Jason Mele (03:53)
Yeah, I mean, as we build the core databases that these LLMs are based on, really, as we map up all these data sets, there's gonna come a point where to really move AI forward, it's really gonna take someone capable of coming up with an idea that no one had previously thought of, right? Because as much as people think

which is going to naturally come up with ideas. It does hallucinate, but really like articulating a wholly new idea is going to be like where you're going to see the most growth, I think.

Adam Pippert (04:36)
Yeah. And by its very nature, an LLM can't, because all it is a fancy language translator. If you think about it, you know, it's taking pieces of data that it's already received and recombining it in new ways. But at the same time, those building blocks are already built. You know, an LLM that's going and grabbing something with retrieval log meta generation is going and retrieving nearest neighbors. It's not.

generating something brand new from that existing piece of content. It's pulling it from some vector database, bringing it in, and then coming up with an insight, but it's really just pieced together out of stuff it already knows. So an LLM's limitation is it can't be original, 100% original, right?

Jason Mele (05:18)
Right. You know what's a, you know, this just made me think of like, so this fall, I did, one of the volunteering things I do is I volunteer for the Mark Cuban Foundation. I do their AI bootcamp. We do like a bootcamp for high schoolers teaching them AI. And as we were like, for those familiar with like the concept of tokenization, which is like breaking apart words into like their elemental functions. And

I think I use the analogy of music theory, right? So you've got like your quarter, your half, your whole notes, right? Where you're indicating like segmentation and like pieces and stuff like that. And then, you know, where that sits on the like treble or bass staff will tell you kind of what note to play. So there's like, when you really map it out, really like the building blocks of an LLM is like a music theory, right? Where you've got to, it kind of masters all this knowledge over here so that it's able to do.

like you can't just come in day one and play like Jaco's Donnelly, right? You need to have an established idea of like, oh, this isn't just like a bass guitar falling down a set of stairs and playing all the notes. This is actually very complex to really understand all these things and where they're supposed to be and kind of like, oh, there's like a chord section here that it isn't being played, but the people

playing these very complex arpeggios, understand that that's what they're playing over, even if no one's actually playing that piece.

Adam Pippert (06:52)
Well, it's interesting you mentioned Jaco because I always use him as an example of a way that you can look at a tool and have a different interpretation of the capabilities of that tool. Because up until Jaco, there really weren't a lot of players that were playing bass fretless electric. So you would think in terms of an instrument that does not have frets, oh, this is this big thing with a 42 inch scale length.

is this tall, you have to pluck it like this, you only can use three hands. I don't know if you're aware of this, but string bass and electric bass actually have different fingerings despite the fact that they're the same notes.

Jason Mele (07:30)
I didn't know that.

Adam Pippert (07:31)
Yeah. And part of the reason is because the scale length is so long on a string bass that you literally can't stretch your hands far enough to be able to have a four finger position. So so the sound that comes out of a string bass is obviously not fretted, you've got the ability to do slides and those kinds of things. So there's a certain set of functionality that comes with string bass. And then there's a certain set of functionality that comes with an electric bass like the ability to use all four fingers to be able to make notes.

Well, up until Jaco, nobody had combined those two capabilities together. So with Jaco now, you know, taking a pair of pliers and pulling all those, uh, frets out, suddenly you've got the shorter string length, you know, the 34 inch string length on a jazz space. But with that ability to slide into notes and get really precise tunings and the other advantages or disadvantages depending on how you look at it of fretless.

So he took two separate tools, combined them together and created a new set of functionality out of it. And we're, you know, I think we're starting to see that in the AI space too, where we've got these LLMs and their capability is basically to go against a fixed body of text, break them down into elements, into tokens, and then recombine those with some prompt. Then we have databases and they're traditionally the place where you store data outside of.

Jason Mele (08:28)
So he took two separate tools, combined them.

Then we have databases. They're traditionally the ones where you store data inside of other stuff.

Adam Pippert (08:55)
other stuff and have a retrieval, right? And then with RAG, that's like those two worlds getting collided together. And suddenly it increases the scope of what kinds of problems large language models can solve because you have the advantages of both tied together in a new capability.

Jason Mele (09:01)
and collided together and suddenly increases the scope.

that analogy. And it's crazy to think because if you think it took 20 years to kind of come back around to the concept of understanding intonation in that way, right? It's all out there and then recontextualizing it that way took... So the precision bass came out in 50? I think it's 49 or 50.

Adam Pippert (09:39)
Yeah, the broadcaster base was 50, yeah.

Jason Mele (09:40)
Which ironically, it's called the precision, if you don't know the history, it's called the precision bass because the frets allowed you to precisely play a note. And it took, what, 20, 30 years, 33 years of experimenting for someone to say, like, why haven't we like taken the sort of like really nuanced intonation of a string bass and kind of applied it this way?

And a lot of stuff happened in those 33 years, 32 years or so. Um, and I may be getting my dates wrong, but like, I thought it was like 72, 73, but like it's, it's fascinating because I don't know if you've ever experienced that working in tech, right? Where you've been operating off of, like I said, a core principle so long. And, you know, you naturally make the assumption that like,

Yeah, we must be doing this for a reason. And you come to find out that these things that have been often articulated to you as rigid rules are often established guidelines for the sake of speed to market or just ability to replicate it on an operational scale. But when you're able to find a more sophisticated path to solve a technical problem.

And all of a sudden, that sort of gives you that same sort of satisfaction. Like I figured something out like a puzzle. And I'm curious to see, like, as these tools get more sophisticated, right? Again, they're not going to do, like, any sort of, like, significant thought work for you. But your ability to sort of take, like, raw ideas and then really kind of see those through, see those down the line.

like faster and more effective is really going to amplify what you're able to come up with. Because you won't have the same like time to like ramp up prototyping that you used to have in the past. And it'll give people a lot more ability to be innovative faster. Because faster equals like not lost budget. Like I had to say, put it in those terms. But that's frankly where we get the opportunity to innovate or not is like do...

get to do it in our budget, you know? And just, it's gonna increase the ability to do experimentation and try new things.

Adam Pippert (12:18)
Yeah, I mean, that's definitely true in my space, right? Cause I work in automation. So every bit of extra capability that I get to do as an automator, it only comes from the ability for automation to save time on the front end, to be able to create that set of resources back. Whether it's saving developer time for being able to do a more with fewer people, or it's just simply saving that team that already is in place a certain amount of hours because of the thing that they don't have to do over and over again.

Jason Mele (12:49)
So yeah, so like when you see like, I often this, I have the option, because I ended up staffing a lot of my own teams. So I do have this tendency to like gravitate towards creative types. And like I, you know, the creative mind, right? And then I'll take the opportunity to arm them with technical skills and the right tools to do the job. We're in adjacent kind of roles. We're in like,

technical account management is what my team is and you're an essay, you're principal essay, and that's sort of the same thing, a customer facing technical implementation. So I don't know, just gate, just gut checking this. Is this like, have I surrounded myself with creative folks and musicians? Because of course those are the people I like chatting with. Or is that something you see around the space? Like that really talented tech folks have a creative mind.

Adam Pippert (13:32)
Thank you.

Yeah.

Honestly, in my space, it's more so than in your space. And it's not because, and it's because I'm in pre-sales and not post, so we can kind of get away with being a little more creative, whereas you have to actually do the work, as mean as that sounds, but it's kind of true. I mean, when you're coming up with trying to match a solution to a problem, as opposed to having to actually go through the nuts and bolts of solving that problem, sometimes it's a little easier to have that liberty of.

Jason Mele (13:50)
Right.

Adam Pippert (14:14)
being able to come up with a creative way to solve something. And then ultimately it ends up being the post sales, the TAMs, the delivery folks, whomever, that are responsible for understanding the real limitations of the product and figuring out how it actually works. But that still requires creativity, right? That's definitely still important.

Jason Mele (14:37)
I guess, yeah. That makes sense. But as someone is on the other side of that dialogue, I kind of, I love it when I get something interesting. And you'll get those sort of reactions like, oh, my God, how are we going to figure this out? But that's infinitely more interesting to me than just doing repetitive implementation tasks. Oh, we're going to get another, you know, Marketto integration or another, let's get

Adam Pippert (14:42)
Mm-hmm.

Jason Mele (15:05)
get something set up to, you know, with some data visualization to Power BI or to have, like that stuff is important and it's good to do it well, but like I am absolutely the chaos monster that will like, I love being on a customer call and I can just, and I've, and I'll just tend to like sit back and listen at the beginning of like client calls and just sort of listen and like trying to figure out what's the one thing that they probably would love to have, but they don't know how to ask for it.

And then I'll love blurting out, like, what about this? And then you can just see someone on my team eyes bug out, like, why would you say that in front of them? And I'm like, because I, from just like keeping myself interested standpoint, I want to see what we can build because I'd love to see what's possible. And a lot of times I think...

Adam Pippert (15:43)
Right.

Jason Mele (15:59)
pushing yourself to see what you can come up with. You'll find, like you don't always have the right tools to do the job. That's always an element of like what your company has. But I find that a lot of times it's a matter of effort and like creative thought on how to solve for like, not necessarily this like particular like function, but like think in terms of like, how can I achieve this core goal with the tools I have, like a set of Legos that maybe don't all match. And,

it's sometimes a hard path, but like when you get there, it's always like, wow, that's great. You know, I wouldn't want to do that again. And I'm like, well, we definitely can do it again. Just wait till I come up with my next idea. But it's what keeps me interested, right? You know, we work in, I guess, I mean, I think like we're both in that semi, like we're working in, you know, in between different systems and different businesses.

Adam Pippert (16:38)
Yeah.

Jason Mele (16:59)
it's easy to feel like, you know, you don't have the same creative enjoyment. Like again, we worked for like back in the day, a company that was decidedly a cool company to work for. But like, as you develop your skill sets, you know, you maybe, the core premise as you describe it at like a cocktail party isn't as fun, but you can find enjoyment. Like I can find the kind of work that it's an interesting problem to solve. I think that's what makes me want to do stuff.

Adam Pippert (17:30)
Yeah, I find it really interesting that I see a lot of content that people create on YouTube or in podcasts or whatever that is the dark side of fang. And it's in general, folks that work for these very large, big tech companies, like even bigger than what we're working for, oftentimes have the problem of the larger the company, the

greater, they can break down a particular problem into smaller segments. And so you get stuck on a responsibility that is very focused. And sometimes that's good, and sometimes that's not so great, because you may feel like you don't have a good understanding of the holistic context of where you fit in your role. And so a lot of times people that go and work as developers at Netflix or Facebook or another

larger company are like, oh yeah, my day job is just doing x, y, z widget and just doing this very small niche thing and I don't understand where it fits in the big picture. And the smaller the team you work with or the smaller the company you work with, the greater view you have of the holistic puzzle and you get an understanding of where you fit in.

Jason Mele (18:43)
That's a good way of articulating it. I was just describing this to someone else and I didn't have the best words for it. When I think about the kind of companies I want to work for, that is a big factor for me. I love to work on the whole picture. I love to think of it in terms of like, oh, yeah, all the technical integrations in the ecosystem. But I also, the reason I've got some solutions consultants, I've got some...

Tams, I've got some pro server folks, but I really am big on the program. Like I don't want to, I can do an integration, but I really want to think about like, what's your whole program. And let's think about the training too. And let's think about like the end user experience, like, you know, like think about like, oh, you're trying to like generate leads or like, how do you make that fun and interesting to people? And you know, you get some looks like, why does that matter? It just needs to perform the function. I'm like, yeah, but.

at the end of the day, if you're driving traffic or you're trying to do something, it's like, why wouldn't you put in the effort to make things as engaging as possible? And it's sort of like, not expressly a must, right, in the job, but I think that it provides a superior end result for everybody if you think about the whole ecosystem. You may be familiar, there's a...

There's a methodology that came out at MIT, and I think like the 60s, called System Dynamics. There's an author, Danella Meadows. She's got a book that I always recommend to people when they join the TAM team. And it's like, it's thinking in systems. Let's think about how everybody interacts with all the components. You map out things like.

like the things themselves, the functions and the action as it goes to other things, and then how they all interact and affect each other. And it absolutely can get out of hand in terms of like how much of a rabbit hole you get down in terms of, but it's really good to at least do the exercise and understand if you're building a program or you're building any sort of technology systems, think about everybody that interacts with from

the user that might join a marketing thing or might join a conference to kind of do an educational piece, to the program lead, to the IT folks, to the executive who's going to look at it at the end of the day and decide, like, is this thing profitable? And maybe any piece of that isn't specifically interesting to you. But if you're like,

You don't have to prioritize one or the other, but you should be thinking about how it affects everybody because it'll allow you to spare a moment to think about, have I really thought about why I'm doing everything I'm doing, you know?

Adam Pippert (21:41)
Well, yeah, because if you don't, then there's long term implications to that. So maybe solving your small technical problem has short term implications because you get the thing done, you get the thing out to production. But whether or not that thing works in production, you aren't going to know until you've got an iteration of it out there, right? So the more you can think about the thing that you're building first and get it out there, the better. There is some argument.

against that too for certain kinds of applications where it's just like, you're not going to know. So just go ahead and put something out there that sucks and then fix it later. But we don't oftentimes have the luxury of being able to do that in the kinds of work that we do. So we have to think about things first.

Jason Mele (22:27)
I am, I'm overtly like resistant to that thought. Like of let's just put it out there and this get feedback is I always, I always want to put out the best version of what I'm going to do. Like this should be, we should do the best version of whatever we're able to do. Do we always have the right tools? No, can we always, maybe we can justify the budget or like the use of it later. But like, I always want to have the best version I can possibly do.

and then just be open to like, there's still gonna be like five or six things that like someone's not gonna like, and you're gonna get feedback. I always say like, you know, let's design a system that won't need us, that makes us obsolete. And I guarantee if we try to hit that target, we're never going to be like out of work. If you're worried that you're gonna like,

design perfection in a way that you're not going to be needed to do more updates and fixing. It's like, don't worry about it. It's always going to be something you didn't imagine. So always do the best version and then just accept that, you know, be prepared to find out that you did the best possible version and you were still good.

Adam Pippert (23:32)
Yeah.

Yeah, a corollary of that is if you're afraid that your job is going to get taken by AI, then your job probably could have been taken by not AI.

Jason Mele (23:55)
Yeah, absolutely.

Adam Pippert (23:56)
You know what I mean? Like, if you think that whatever thing that you do, some LLM is going to go and replace it, well guess what? That LLM just crowdsourced a bunch of information from somewhere else. I mean, it has memetic value, right? It's taking information from literally everything on the internet and condensing it down for you. But somebody with a good sense of Google foo that could go out there and search for the things that you need would be able to outwork you.

Jason Mele (24:09)
I mean it's

Oh, you beat me to it.

You beat me too. I was like, well, like, okay, how many people here, like, you know, raise your hand, you know, in the group if you've ever googled your way out of a complex problem, you know.

Adam Pippert (24:35)
We all, that's what our job is, or mine is anyways.

Jason Mele (24:37)
I was at a job and I absolutely like in a like technical like training manual like instruction number two was once have you tried googling that you know before you get into like escalating a ticket to a subject matter expert or like going over like try googling it first like and don't be afraid that maybe like if it's they have that it can't be that simple it's like it might be that simple

You know?

I think like you but you have a computer science background like a proper like I went to school for computer science at some point Right. I looked it up the other day

Adam Pippert (25:14)
So in a roundabout way I do. So I started off taking CS classes at Roanoke College in the 90s into 2000s to be a dual major. So my original plan was actually to be a dual computer science and music major. So I did like the first year and a half I think of my CS program and went through music composition. My senior piece.

for music composition was actually a computer composition with MSP Max and a prepared piano. So like I was doing stuff with computers the whole time I worked for the IT department. I was doing, you know, like troubleshooting all the Mac systems for the College of Arts section. Well, I say College of Arts, but it was a small school, like there weren't like separate colleges was just a department, right. But I was doing all of that.

Jason Mele (25:48)
Okay.

Sure.

Adam Pippert (26:10)
I was still doing technical things even though I had dropped my CS major. Because my thought was, well if I went back to school, it would be easier to go back to school for CS, and be able to get an internship and be in computer programs, than it would be for me to want to go back to school to study music, because who's going to hire you to do anything as a music intern? Like you may be able to go work for a studio or something, but that's not realistic. So my thought was, I know you did! I know! I know!

Jason Mele (26:10)
I was still doing the technical thing.

music

Hey, I worked for a studio. I worked for a commercial studio for years. I mean, until like, 08 happened and then yeah, I had to... I only developed technical skills, like actual IT-based skills because music kind of fell off a cliff. Which is really... So, like, can you spare me a segue for a second?

Adam Pippert (27:01)
Absolutely.

Jason Mele (27:02)
So I am really impressed with your vision of your journey, because I didn't entirely see it. I let the river take me down with the current. So I got into school. It was still ADAT and reel-to-reel tape. I actually, after the commoditizing of the digital recording market where everyone had a Mac with GarageBand and you can't see it,

commercial studios were imploding, I still, I got more work in the late aughts to early tens doing recording because I knew how to record and edit it on real to real tape. And Razor Blade editing is like, it's like those people that can code in COBOL. It's just sort of like, they'll call you off the bench, even if you've been out of the game a while and you can charge a premium. But I remember, I think it was maybe,

Adam Pippert (27:55)
Right.

Jason Mele (28:00)
sophomore year of college, the Digio 2 came out. Maybe it was the Digio 1, but it was the Pro Tools, but like Pro Tools existed, it was a $10,000 avid system with proprietary hardware, and then they're like, right. There was the mixing console, and then there was the rack unit that everybody had, and all of a sudden, it was great, because you're in college, and you're like, I can do pretty good recording projects.

Adam Pippert (28:08)
and Pro Tools.

Yeah. Yeah, but it wasn't like a little box that's like this big, like the DigiWords. Yeah.

Jason Mele (28:30)
This is great for a college student, but none of us saw around the corner, which is you don't need to be somebody with this sort of fundamental proper education in this, like understanding how to like wire a patch bay and how to like calibrate a Studer 24 track, like you don't need these technical skills in the same way with this new technology, but I remember that I was right there in the shift and the analog studio got outfitted, everything got a Pro Tools rig.

automation in the mixing board which I was just completely resistant to.

Adam Pippert (29:05)
Yeah. What do you mean my fader can move without me?

Jason Mele (29:13)
I see what you mean. You're mimicking me off the air. Yeah, it's like, I don't want that. I was being a retro grouch. And I, thoughts, things sounded good, but in like retrospect, right? That was me, you know, resisting the tides because ultimately what I've really discovered over the years, I'm back to using the rig I used in college. I've still got my 2007 MacBook Pro, my Digio 2. I still use the same like SM.

Adam Pippert (29:14)
Yeah.

Jason Mele (29:41)
SM7 and SM57s and a Rode NT1. Same exact setup because I can still make music on it and I've got quick shorthand with it. And what I really discovered is that digital's fine, right? Like I don't think the analog has all the sort of like magic, like sparkle dust on it that everyone likes to attribute it to.

What it is, is deciding to do two things. It is capture natural human interaction in a room, which is just a choice, right? And it is, it is.

just like, oh, I just lost my train of thought. I had, I was like, so.

I just said it is not about the actual sound of analog gear, but it was really that choice of how you capture. Oh, the other one, sorry, that came to me, which was the choice of I'm going to get the sound in the room and not fix it in post with plug-in, right? Is let me move the microphone. Let me make sure I dial in a quality sound on everything. Get your drum sound dialed in, and then capture a natural performance with human interaction.

And those are the only two important factors because everything else is really my new sort of nuance that ultimately you're probably going to lose once it's like mixed down to be put on Spotify.

Adam Pippert (31:10)
There's also the zero latency issue of the only way that you can get actual true zero latency is to do so through an analog input.

Jason Mele (31:19)
Yes, that is absolutely true. However, if you are capturing a dialogue between musicians in the room.

Adam Pippert (31:21)
Yeah.

Jason Mele (31:32)
You don't have to worry about the latency because it should be theoretically consistent through everything going through. I guess that's what I mean.

Adam Pippert (31:37)
Yeah, yeah, if you're capturing a live environment and the interactive and you're recording an interactivity that's outside of the scope, then that's great. If you're overdubbing, it can take a little bit of time. Like I remember hearing a story about Billy Corgan doing tracks for some Smashing Pumpkins album. And it was when digital first started. And he was talking about how he tried to track a guitar bend over

Jason Mele (31:48)
Yeah.

Adam Pippert (32:05)
over like a new track, doing it on a digital recording and he couldn't get the bend. But as soon as they went to them, they went to like a mastering house or something that had tape and he just, and he like re-recorded it in an analog way. He was able to hit the bend. And I know it sounds really crazy, but like.

Jason Mele (32:27)
I agree. I think we're kind of, we're like making points and we're coming to the same point in the middle, I think. I guess what I'm saying is like for those who think that there's, that have thoughts on like computerized technology and how that's changing, right? Things like Beat Detective and like working off of a like a really like rigid grid and like being able to auto-tune vocals. Those aren't overtly...

bad things, those are tools. The question of analog versus digital, those old records you loved, is all about how you chose to record and capture. So I think that the technology moves forward, I think the vibe you're trying to capture is really in my head up to choices that you can make, and the tools are largely there to aid in that.

function. And I guess that's true. That's really true of what everything we're saying is like, really, these are all tools that will aid you towards a goal. And the question is how you make choices and how you like articulate these things.

Adam Pippert (33:36)
Right, and you have to understand the constraints that exist in each of those tools and whether you're adding constraints as you combine tools or you're subtracting constraints as you combine tools. And that's true with programming too. I work with Ansible, right? That's like my primary tool that I work with. And I do work with Kubernetes and some other stuff, but the thing about Ansible is it's a Python application. And

Jason Mele (33:51)
Very well put.

Adam Pippert (34:04)
One of the constraints of Python is that it's crazy slow. So we oftentimes will have to over engineer on the infrastructure side to make up for Python's lack of performance because I'm working with telcos that are trying to automate systems that are like all across the United States in order to make sure that you have telephone service and make sure you have 5G internet and all that stuff. So we oftentimes will have to over engineer.

Jason Mele (34:22)
across the United States. Make sure you have telephone service, make sure you have 5G, internet.

Adam Pippert (34:31)
the number of nodes that we have in our automation mesh to make that work. Whereas if that thing had been written in Rust, we might not have to have as much stuff, but the, on the flip side, the product would not be as popular as it is today if it hadn't been written in Python, because it means that it's an environment that's easier to develop in. And so more of our ecosystem partners were able to say, Oh, we're already using Python anyway, let's go and write the sensible module or let's

Jason Mele (34:36)
Whereas if that thing is in the rest, we do not have to have as much. But.

as popular as it is today.

Adam Pippert (34:58)
contribute code to somebody who can write this a handsable module. We have other automation technologies like Puppet, that's our local Portland favorite, and Chef up in Seattle, that had a large amount of market share five, 10 years ago that have just sort of faltered. And part of the reason for that is a technical constraint. Those applications are written in Ruby. How many Ruby developers do you go and meet on the street today compared to 2008? Not as many. So the fact that

Jason Mele (35:02)
We have other automation technologies like Puppet, that's UnivArt local core, and the favorite, Chef, in Seattle.

and the reason for that is a technical constraint. Those applications are written in Ruby. How do you Ruby develop?

Adam Pippert (35:27)
the ecosystem has moved largely towards Python and JavaScript. Let's be honest. That's where a lot of sort of public facing applications are going these days. And the fact that Ansible was already a native Python application has just made its adoption that much easier. So sometimes the constraints of what you choose to put together to make something work, you can either work with it or work against it. But ultimately, those set of constraints will lead you down a specific opinion.

Jason Mele (35:36)
applications are roaming these days. And the fact that Ansible was already a native Python application has just made its adoption that much easier. So sometimes they do.

what you choose to put together to make something work. You can either work with it or work against it, but ultimately those set of constraints will lead you down a specific opinion.

Adam Pippert (35:57)
of how to do the thing.

Jason Mele (36:00)
Yeah, I mean, and I think it all goes into, you know, you need, but that I think it's a great analogy too, because it actually illustrates both the need like you like the need for very specific technology, what is the best is not always necessarily the best for this use case. I think what you're describing goes back to that point of like trying to understand everything as a holistic system. Think of in terms of not just like

Adam Pippert (36:17)
For sure.

Jason Mele (36:26)
You're not thinking of just, what's the best tool for us to build this, but like, what's the best tool to make this usable for the market so that people that are not me can get the most value out of it.

Adam Pippert (36:38)
Mm-hmm. Yeah, like this right here, that is a Blue Yeti. That's a USB mic. Like, I would love for it to be a Neumann U87. That would be awesome. But would it be the right tool for the job? Absolutely not. Like, can you imagine with all this like, no, you know, no sound deadening, I've got all these weird angles in this room, the sound would sound like crap. Who cares if it's a $10,000 microphone? It would sound terrible.

Jason Mele (36:44)
Yeah.

I agree. I think that's a great point. However, just complete segue just for the sake. So like, I get it. Like I'm not going to buy a U87 either. But if you want what is fundamentally like the shell and basically like the basis of a U87 but like 300 bucks, I shouldn't say this. I should go buy this stuff up before I make this recommendation just in case it takes off. But the, you know, I'm a fan of like the old 90s road stuff.

Adam Pippert (37:08)
Yeah.

Mm, yeah.

Jason Mele (37:35)
was way better stuff than it should have been. But the original NT2 was basically like an Australian made U87 clone. And they're still like super inexpensive and it's still one of my favorite mics for getting the job done in a way that you can be, it's like, I'm not really gonna sweat it if someone's like using the capsule a little bit. Sorry for people that are not into recording tech, but like.

Adam Pippert (37:39)
Mm-hmm.

Mm-hmm.

Yeah, it looks like... No, no worries.

Jason Mele (38:04)
If you are, like, you're welcome. They're great underrated mics. I love Rhodes stuff. Blue stuff's great as well, but I'm pretty impressed too when you think about like, yeah, like, what is the purpose? Like noise canceling, like these headsets. I know it's not that great a headset, but we have, in addition to my dog, a 70 pound Australian Shepherd who occasionally decides that he wants to.

yell at me that he would like to go outside, please. And I'm surprised, and there's like an Aussie right here off screen, like yelling at the top of his lungs. I'm like, I'm sorry about the dog barking. People are like, oh, I can't hear it. I'm like, it's amazing.

Adam Pippert (38:47)
Yeah, noise canceling is wonderful. It's wonderful.

Jason Mele (38:52)
But, like, what tool do you need for the job? Though, just know that I support you getting a U87 if you really want a U87, if it brings you joy.

Adam Pippert (39:03)
Well, if I told my wife I brought home a U87, I'm not going to tell her how many figures I paid for it because that is...

Jason Mele (39:12)
I think, I think, yeah, the common knowledge is like, if you feel like you would need to make that move, perhaps, if you think you would need to make that move, perhaps just like, don't make a purchase if you think you need to lie about it. But I don't know, I feel like you could probably like justify 299 in a normal conversation and not have to lie and be like, this is not bad for just for home recording use, going back to the road.

Adam Pippert (39:28)
Right, right.

Yeah, no, I agree. I agree. I mean, one day I would like to go back to my studio at Bongo Fury just so that I could have an isolated space that is outside of the house. Because I used to have a space over there and all the, when I worked at Intel, like the interns and all the other folks, we would all get together and play on Sundays and stuff. And now I've got a family, so I can't really get away and do that much. But when the kids get a little bit older, I'll probably set all my stuff back up over there and do that.

Jason Mele (39:51)
Yeah.

That's cool.

That's super cool. Yeah, I do kind of miss the external space. Like that was the great thing about running a recording studio is that if I wanted to go even just make loud music, I could just go there and like it's on the door, right? I'm renting the building on the premises recording. I put in a sub floor for it. I ran all the wiring. You knew what this was. That I didn't book a session is frankly, none of your business.

Adam Pippert (40:25)
Mm-hmm.

Yeah.

Jason Mele (40:38)
If I'm just looking to be loud. I mean, we're all, everyone gets into music because they're like loud children who want to continue being loud children as an adult. Okay, so segue. I use this analogy for somebody when it comes to growth. Cause like I didn't go to computer science school but I've developed.

like a lot of the coding and system architecture skills on the job as I found the need arose for, you know, to move forward as we just sort of developed business and like grew our line of business. And I always applied this rule going back to college when you're playing with, especially when you're playing jazz, because I was in the jazz ensemble for a bit. And I always applied this rule of

Always play with people that are better than you, right? If you are the best person in the group, you have zero chance of personal growth or like skill-based growth. And it'll do two things. A, it's an opportunity to push you to actually grow, but also like gives you a bit of like emotional that sort of like fear instance of like, oh, I need to be better if I don't want to be.

Adam Pippert (41:37)
Oh yeah.

Jason Mele (42:00)
embarrassing myself in front of people. Has that been your experience as well, just sort of developing both like musical and technical skills?

Adam Pippert (42:08)
Yeah, I'd say even outside of those scopes, like in business, in relationships, in whatever, like always try to be the dumbest person in the room whenever possible.

Jason Mele (42:20)
Jen finished her engineering degree, so absolutely, that is true of me and my personal life too. I've got half of an electrical engineering degree. She has the whole of an environmental engineering master's, which means it makes her more qualified than me at a whole bunch of stuff. So yeah, it's a good way to live and it doesn't always have to be credentialed per se, but I, and I think this, you know, thinking about like saying this,

Adam Pippert (42:25)
Yeah, yeah, yeah.

Right.

Jason Mele (42:50)
to an audience of people, if they're people that are younger in their career and that they maybe decided, for better or worse, on a philosophy degree or a music composition and sound recording degree, and then they're in the tech space, is that developing skills aren't per se directly correlated to.

to leave and go back to school. That is absolutely valuable if you hit the point where you've got a very particular thing you'd like to do that requires a certification or some growth. But let that be targeted. Have an idea. But taking time to try to work with the folks that are smarter than you on a particular thing and interact with them and try to...

bring ideas to the table is absolutely an excellent way to sort of get into the conversation and be open to feedback and be willing to sort of be pushed and, you know, pushed back on maybe sometimes in your ideas. But I've always found like pushing things forward, I'm probably wrong about like the proper way to do things, right, particularly when it comes to like systems architecture stuff. But I'm able to contextualize a vision

and just be open to that feedback. And the 40% of the time I get pushback, I'll always have this dialogue, like, what would you do instead? And really try to develop relationships. And that goes to the same thing about who you interact with being smarter than you. But also, you think about working at a tech company, whether you're in sales, whether you're in like,

customer facing tech or like backend development. Like you're part of a team, right? And you're part of the band, like, you know, and trying to think about how you interact as an interconnected unit is more important than showing off for your own personal sort of glory, right?

Adam Pippert (44:58)
for sure. And I think a lot of it too is the fact that I don't know if this is necessarily true of all subjects in school, but I feel like mathematics and music are the two primary subjects where the entire curriculum is based on first principles. Like I don't find that to necessarily be true in history. I don't find that to necessarily be true in English, although it is to a certain extent, but not really because you pick up linguistic skills before you actually go to school. But with math

Jason Mele (45:26)
with masks and what we use in particular.

Adam Pippert (45:27)
And with music in particular, in general, for however long those curricula have been around for hundreds of years, those subjects are always taught from first principles, no matter what. Like, when you go take piano lessons, it's usually you learn essentially like how to sit at the keyboard, you learn what white keys are, you learn very simplistic rhythms first on a single note, and then you move on to other notes, and then you move on to other rhythms, and it expands from this very first principles.

point of view and with math too. It's like you learn to count numbers one through X and then you learn how to add them, subtract them and everything is built from these very small building blocks and combined together. And obviously technology is built that way too, right? Like we put things together from these very simple building blocks and then combine them together and learning how to combine that stuff seems to come more natural to people who study either math or music.

Jason Mele (45:58)
It was navigate, it's like you learned to count.

small building.

and obviously technology.

very simple building blocks and putting them together. And then you can kind of define that stuff. It seems to come more natural to people who study either math or music because they already are trained in the first principles of music. They're so related. I've always equated music theory to another form of math. It just, it follows completely different principles. And it can be in some ways more complex because different modes will have different

Adam Pippert (46:25)
because they already are trained in first principles thinking.

Jason Mele (46:45)
rules, but I read somewhere recently, and I think you'd appreciate this based on that statement. I was looking for the article before I joined today, and I couldn't quite find it, but it was talking about some of the earliest...

post-primary education, like college, sort of, I wouldn't call them degrees at the time, because these are like, I think like 15, 16 hundreds or so. But like the earliest curriculums, religion, right, of course, that's common, because that's probably funding quite a bit of it, but also music and mathematics, and there was a fourth, and I couldn't quite remember, that's another reason I was looking for the article, but it was talking about these are the core things that like, oh yeah, you're going to do continuing education.

These are the things that you need, like that stronger continued education basis to kind of build on that structure beyond what you learn in primary and high school or whatever the equivalent was at the time. And it's funny to think about like, you know, the idea of going to school for me is like so marginalized these days as this absurd thing. But it's one of the oldest sort of like, like

continued education, things you could do if you go back hundreds and hundreds of years.

Adam Pippert (48:03)
Yeah, well, the interesting thing is it tends to get lumped with other arts, like visual arts and film and other, other arts. And in some ways it has some commonality with those things. But I guess the big difference is that to get to a modern competency in music, you still have to start from first principles to get to a modern competency in art. You can fudge your way through and still become an artist. Like look at Mark Rothko.

Jason Mele (48:17)
I guess the big difference is that to get to...

It's like you're waiting for something.

Adam Pippert (48:33)
You know what I mean? So yeah, just being honest. Like, and I think a lot of the pushback against the arts really is that people view music with that lens without really understanding exactly how your mind has to work in order to be able to put musical compositions together. And some of it may just be the types of music that ends up getting created by people who don't have that background. Sometimes they get lucky.

Jason Mele (48:34)
Brutal.

arts really for people.

Adam Pippert (49:01)
Sometimes they create hip hop. Sometimes they create these other types of music. But fundamentally, people get tired of it. People get bored with it. Like, yeah.

Jason Mele (49:07)
Yeah. Two examples. Great examples to show you that like, it isn't about proper Western music theory. Two examples from two opposite ends of the spectrum.

prove that it is about fundamentals and math. Number one, Nashville number system. I can't sight read, but I'm gonna write out a chart this way I can transpose if the key changes. I can, I just, I know the chord and the sequence and if it changes to B minor, cool, right? The other one's J. Dilla, thinking about like a very non-traditional way of looking

at timing, right? And these are both not rooted in the traditional western canon of music theory, but are also like very related to the those two polar ends of structure and core fundamentals to understand how this stuff comes together. It's two great examples of finding your way to kind of map

out and articulate that structure even if you weren't given that basis. It's so endemic to music as a form.

Adam Pippert (50:24)
Mm-hmm. Yeah. Well, and Dilla is interesting because a lot of people may not be aware that the Lindrum has those swing settings, but they're fixed. There's like eight of them that are built into that Lindrum. It's not as if he just randomly arbitrarily flipped to like 72% on the swing or something. It doesn't really work that way. It sounds like that because it's just like, I mean, but they're still discreet.

time points, but you can still kind of fudge with it a little bit and make something awesome out of it. And now everybody talks about Dilla as being an influence, you know, when really it's Alindrum was the influence, he was just the operator of the tool, you know.

Jason Mele (51:07)
Absolutely. But there's a great book on that called Dillatime, if people are looking to really dig in. I think it's Dillatime, right? But it talks all about the invention of this. And if you look back, then you can really see how like.

such a massive influence off of a moment that maybe isn't fully understood for its influence. But, you know, and a great example of like that's only because that was he was using an MPC, right? That given the era or was he pre-MPC?

Adam Pippert (51:48)
Let's see, when did he die? 2000... I think it was in NPC. I think it's about that time.

Jason Mele (51:56)
Um, yeah, I remember the MPC was the, was the, uh, the Lindrum of my generation. Um, obviously I'm like massively influenced. There's plenty of musicians like Prince and, and many others that made amazing use of the Lindrum to do things that were just, um, that were massively different than they were before. But it's amazing how like, you'll see, just bringing it back to AI, how a new tool, like

There's plenty of drummers in what 80, 81, Roger Lynn comes out with the Lynn Drum, and there was this fear of, oh, people are just gonna use this and they're not gonna need real drummers. Like, well, that hasn't happened, right? The technology moves forward. There was people like Phil Collins, who himself is a great drummer, leveraging the Lynn Drum to good effect. It's merely a tool and a mechanism to apply your ideas to. And I think

Adam Pippert (52:37)
Right.

Jason Mele (52:55)
you know, I hear that overuse here, I'm like, oh, jobs that jobs are not going to be replaced by AI, but people using AI replace people that aren't using AI. And I'm like, yeah, I, but I think that can be kind of reductive and a bit discouraging, as to say that like, you need to look at it like a tool, like a graphing calculator or something else where it's, it's allowing you to do more.

and experiment more and it just sort of moves you forward.

Adam Pippert (53:29)
Yeah, it's like the buggy operators. Maybe the buggy whip manufacturer didn't get to, you know, extend the life of his company, but you couldn't really say, um, cars aren't going to replace people. Drivers will because then people ended up becoming drivers. Cars got easier to use and then, you know, tons of people drive. And then some people choose to take public transit, which is like a car, but modified for, you know, mass consumption. And.

There are certain trade-offs taking public transit, but it's a thing that exists. I think the same thing will be true with AI. There are going to be some people who are the developers, they're still going to be developing with AI. There are going to be some people who consume tools that are built with AI, and then there will be other kinds of tools that are created that will allow for mass consumption, but there will be some trade-offs in terms of customization, et cetera, that they'll have to deal with in order to be able to use those tools, but they'll be massively available to everybody.

Jason Mele (54:26)
dynamic hasn't fundamentally changed that much, if I'm totally honest. So it is still a matter of, in a pre-AI world, the programs I was putting together are still based on taking what somebody thought from an engineering side this was due and try to have a better understanding of insights and how people are going to use it.

Adam Pippert (54:31)
Mm-mm. Yeah.

Jason Mele (54:55)
And thinking about like, well, how can I prompt this to get the best result? And it's not fundamentally that different than kind of having a dialogue with an engineer and trying to understand, trying to get down to core objectives as opposed to, you know, there's maybe a commercial marketing or revenue audience versus an engineering audience and navigating the conversation, sort of you're bringing everyone together, that AI is a part of that dialogue, but that's still going to happen.

Adam Pippert (55:07)
Yeah.

Jason Mele (55:23)
your ability to sort of make sure you're asking the right questions, right, is at the core of it. So it's just a matter of it's a tool that can amplify what you do, but just understanding where your skills have become more valuable and maybe which elements of your skills are going to start getting commoditized. You meant?

Adam Pippert (55:46)
Right. And that's where prompt engineering really comes into play because that is your ability to take natural language and turn it into something that you're asking the tool to do. But there are a lot of nuances to get the thing that you want out of the AI. It's not just simply like make me X and then X pops out. Like it's if it were that simple that it would have already taken your job.

Jason Mele (56:07)
If it were that simple, then it would have...

Absolutely, and it's not like solidified yet, right? There's still like that market and that role and responsibility is still materializing. So it's a great chance for thought leadership. And the concern around artificial intelligence is really something that the users can craft. And I think those who are really focused on like creative insights

actually delivering interesting, cool products are still going to succeed better because having the same information regurgitated back out to the market, it's gonna start getting staler faster if you're not putting in and really moving the conversation forward and using AI as a bicycle, right?

Steve Jobs said a long time ago, the humans are by far the least efficient species on the planet for the consumption of their energy, but give us a bicycle and all of a sudden we blow everyone else away. And it's, he for the year of the computers is a bicycle for the mind. And it's like, does it take what your creative mind is capable of doing and amplify it and make you capable of doing more? And I think the fear comes from the idea that like,

You don't know what the more is yet because you haven't come up with it because you've maybe been stuck in repetitive tasks. But letting yourself push the line and drive a little harder on new ideas, I think that fear will kind of turn into excitement if people apply it correctly.

Adam Pippert (57:58)
Well, we've already been through it before with the rise of the Internet. The same thing applied, right? Like suddenly we have this mimetic body of knowledge that everyone can contribute to and you can pull from readily at any time. So people had the fear that like the Internet was going to be taking away their ability, people's ability to think clearly, like who's going to want to go to Wikipedia and make that a source for a paper and you know, obviously Wikipedia has its, has its limits, right? And some of those limits are because of the

Jason Mele (58:23)
you're obviously just seeing the limits, right? And some of those limits are because of...

Adam Pippert (58:28)
nature of having this open tool that anyone has access to, right? And then these clicks develop and then Doug Barr's number rears its ugly head again. Suddenly the world is kind of destroying itself. But that's going to happen with or without AI.

Jason Mele (58:39)
suddenly the world is kind of destroying itself. I still think Wikipedia. But I still think Wikipedia is possibly one of the best examples of an open source solution not going off the rails. Not that there's not inaccurate stuff on there, but that it has not gone off the rails more than it has is really a testament to how much care gets put into that as a solution.

take it for granted, right? Like, like there's a omnipotent entity managing Wikipedia, but there's not.

Adam Pippert (59:19)
I mean, it's still there, right? Like how old is Wikipedia? No, I mean, I did, but it was more of, I wanted to see what year it was created because I can't remember what year Jimmy made Wikipedia. How old Wikipedia? 20, so 20 years. So it was founded in 2003. Well, it says here 2001. Oh, that's the foundation. So the foundation was 2003. So the...

Jason Mele (59:21)
Did you just check to see if Wikipedia was still there?

Okay, so I'm not looking, I'm not typing, so I'm gonna guess. I'm gonna say.

in 2000. That late? Oh yeah, that's the foundation. So the foundation...

Adam Pippert (59:47)
project is from 2001. So yeah, so I was still, yeah, that's, I think so too. I was kind of surprised. But how many applications are out there that have been around for 20 years that fundamentally do the same thing that they did 20 years ago?

Jason Mele (59:50)
Still, that's later than I thought it was.

Oh boy.

Adam Pippert (1:00:06)
Other than Emacs and Vim, you know, like text editors, most applications, even if they do the same fundamental thing, have a whole bunch of extra features that they didn't have 20 years ago. There are a couple that are still, that still do the same fundamental thing like Word and Excel, like the Office suite, that stuff is still around. But if you think about Google in 2001 versus Google today, like that company is fundamentally just totally different.

Jason Mele (1:00:18)
Yeah.

Oh, yeah. Yeah, I had formulated a couple of examples in my head, but you just convinced me out of it by mentioning just sort of, like, oh, no, that invalidated every point I was about to make. I think, though, it's interesting, though, we're in a unique time where innovation's going to be really fast at the same time. Funding and you know.

Adam Pippert (1:00:37)
You know?

Yeah, yeah.

Jason Mele (1:01:05)
gathering financing for projects is challenging. So it's gonna be an interesting time where like, I think like companies getting out of like their core, with their core value offering. I think the smart companies are really gonna stick to like, we know what our ideal customers are and we're gonna lean into that and developing towards that. So we'd rather be a stable.

Adam Pippert (1:01:21)
Hmm.

Jason Mele (1:01:34)
like mid-market company than some company that's over diversified and is now collapsing on several fronts because we've like we've stretched ourselves thin we cannot make payroll because we were dependent on like venture capital dollars to come in and save us so

Adam Pippert (1:01:48)
Yeah. Right. And I think those market forces are going to force more people to have smaller companies rather than everything aggregating into larger enterprises. Yeah.

Jason Mele (1:01:59)
It's already happening. So there's a lot in the market about the tech space and layoffs. And it is true. And like I've talked to a lot of people because it's extremely stressful for people, right? But there are a lot of companies hiring. They're just not the giant over extended companies. And so I think my prediction, right? Is that I think that we're gonna start seeing

really specialized, really focused, like mid-level companies are really going to have a good era because they're going to be able to target their markets better. And then the talent's going to start shifting there because they'll be able to provide a lot more for people. Previously, you know, you were around in that era where it was like, and you know, we know some people at a certain

twice what I make. But I also know that the expectations are also maybe not the most reasonable for some of those folks. And at the end of the day, I think that it's going to start flattening out, because if you're doing too much, they're going to have to start making really tough choices. Whereas a really targeted specialized company may not have as many people, but they might be able to target their market really effectively and adapt in a way that those companies cannot.

And I think when I talk to people at those companies, they're a lot happier, right? And so I think those companies are gonna have a great time, which is great, because I think from employee happiness and the way people are being treated, I think that like those kinds of businesses, right? That are creating a good environment for people are gonna create really interesting and innovative projects. And I think that might be excited to see what comes out in the next 10 years. Because we didn't...

Again, we were there at Netflix when the idea of streaming TV directly to your living room over the internet was a thing you still needed to explain to people. But it's ubiquitous now. It's somewhere south of a utility to have a streaming service these days. So it'll be interesting to see the thing that my brain can't imagine that's going to be.

Adam Pippert (1:04:02)
Yeah.

Mm-hmm.

Jason Mele (1:04:19)
ubiquitous in 10 years, right?

Adam Pippert (1:04:23)
Yeah, and we'll just have to wait and see. We'll have to come back and visit this podcast episode in 10 years and see what that thing was that we couldn't figure out. But we are at the top of the hour, Jason. So I just wanted to thank you for coming back on and re-recording, this has been awesome. I'd imagine we could probably go on for, you know, Lex and the Lodgy length of time discussing. So, you know, and maybe in the future, once I record a couple of these, maybe we'll do another.

Jason Mele (1:04:34)
We are at the top of the hour, Jason. So I wanted to thank you for coming. Of course.

notorious for it so yeah absolutely

Adam Pippert (1:04:52)
recap and we'll come up with some other topics. So I'd love to have you back on again some other time. Cool.

Jason Mele (1:04:57)
Absolutely. You know how to find me. If requested, we can do some elaborate field recording version of this, like at the Portland Zoo or something like that.

Adam Pippert (1:05:06)
Yeah, could be fun. So we'll figure that out. Awesome. Well, thanks so much, Jason. It's been a pleasure. All right. Yeah, take care.

Jason Mele (1:05:09)
Awesome. Well, thanks so much, Jason. Yeah, it was great chatting with you again, Adam. Bye.