Become a Creator today!Start creating today - Share your story with the world!
Start for free
00:00:00
00:00:01
/API sprawl: herding the chaos
30 Plays5 days ago
Wrangling APIs? Dreaming about AI agents that actually get your specs? This episode delivers. We chat with Dave Shanley (aka quobix)—a musician-turned-engineer with a sci-fi streak—about cutting through the noise of OpenAPI standards, building tools developers actually want to use, and how to stay grounded while navigating the fast-moving world of agentic AI.
Recommended
Transcript
The Problem with Standards
00:00:02
Speaker
There's a great XKCD strip that does things explains this exactly. think it begins with something like there's 15 standards and we we need to solve this problem, but none of the standards fit.
00:00:14
Speaker
You know what? We'll create one new standard to rule them all. We'll have one overarching standard that fits all the other standards inside of it. And then the next plan is now we have 16 standards.
Meet Dave Shanley
00:00:33
Speaker
Welcome to the Forward Slash, where we lean into the future IT by inviting fellow thought leaders, innovators, and problem solvers to slash through its complexity. Today, we have with us Dave Shanley, who also goes by Quobix. Did I pronounce that right, Dave?
00:00:48
Speaker
That is correct. Yes, Quobix. Quobix. He's a hacker, product thinker, software architect, inventor, and technical leader with nearly three decades of experience designing and building software across a wide range of industries.
00:01:01
Speaker
He specializes in what he calls experience engineering, creating technology that people interact with directly. By day, he's a distinguished engineer at Splunk. By night, well, he's also a failed musician.
00:01:15
Speaker
Those are his words, not ours. Um... The reason we wanted to have you on, you you have an interesting ah kind of ah take on kind of an AI twist on something that's been around
Dave's Musical Journey
00:01:27
Speaker
for a little bit. We wanted to talk about that open API. ah there's ah There's a little bit of a connection there. So I thought that was that sounded very interesting.
00:01:33
Speaker
um But I got to get into this like, This failed musician thing. My morbid curiosity when I saw that on your bio, I'm like, okay, how bad could it be? You know what I mean? Like how bad could this really be? So I clicked on it. I went to your Spotify and I'm, I'm expecting like step brothers prestige worldwide kind of, you know, you're wrecking your dad's boat into the shore kind of a thing.
00:01:55
Speaker
But i i was i was surprised. was pretty awesome. um but I'm going to have to listen to this. how is this failed How are you a failed musician? That's good stuff, man. Thank you. i really appreciate that. i always wanted to be a musician. wanted to be rock star, like 99% of other young children out there, and tried it. And we you know had some progress. you know When I was but maybe 15 years ago, 20 years ago now,
00:02:19
Speaker
I it's a long time ago, long, long, long time ago, but um when I was in England originally, we you know actually had ah so a bit of a following. We had a couple of record deals. We actually had a music video on MTV2, if you know you can believe that.
00:02:33
Speaker
um And we have we did some tours. Actually, we did a tour with My Chemical Romance right kind of just before they blew up.
Engineer at Splunk
00:02:39
Speaker
um And it was a good time. But you know we never quite cracked it, never cracked the nut, never quite made it.
00:02:45
Speaker
And it all feels fizzled out. and By that point, it was just like, well, you know, there's no there was no money anyway. the only The only money we ever made was for for recording. So ah kind of fell back to my second passion in life, which is software engineering, which is always there. was always doing it.
00:03:01
Speaker
But I just kind of doubled down on that and gave up the life of trying to be a professional musician. Oh, man. ah I enjoyed it. I was listening. It was great. Oh, thank you. so I wasn't expecting that. But say it it would have been my choice and in life if I could have a musician. would have been.
00:03:17
Speaker
yeah But i see I see a lot of similarities in coding and engineering software with music and art. and For me, it is a form of art. you know it's not just Sure, it's algorithms, but it's algorithms that combine together that can create something beautiful and unique.
00:03:30
Speaker
Absolutely. No, i love it. I love it. So, well, first I want to talk to to Dave. Like you, you, you have a, this, this alter ego, but I want to talk to Dave because you, you got a kind of a cool gig at, at Splunk and Cisco.
00:03:44
Speaker
Tell us a little bit about what, what do you do there? This distinguished engineer sounds, sounds pretty cool. Yeah, thank you. um So, you know, my my day job is I'm the i' a Distinguished Engineer for Splunk, but now we're owned by Cisco. So we were acquired by Cisco just over a year ago. And um nothing's really changed, in fact, in in in the sense of what I'm doing is just a part of a much bigger company now.
00:04:07
Speaker
But you know the the nuts and bolts of it is vision, leadership. I work in security. So that in in Splunk, there's the Splunk platform. I'm not sure if you know much about Splunk, but and it's a huge big log aggregator for feeding any any kind of unstructured data.
00:04:22
Speaker
And it gives you a way to search, query, and build applications on top of that data. Splunk Security, which is what umm one of the distinguished engineers for, is all of the premium security products that sit on top of Splunk.
00:04:35
Speaker
So enterprise security being our cash cow, which is you know the main the main product which a lot of businesses use. In fact, most businesses that use Splunk use it for security and use one of our security products. And my role basically is to set the vision of how we're going to get from where we are today to where we want to be and ah how all the problems is going to solve and why we need to do that. So it's really kind of setting that ground vision and then selling that vision, not only need to the engineers, but the product managers, the UX team, the technical product managers, the executives, um yeah you know, the
00:05:11
Speaker
the whole the whole organization needs to be brought along on that journey. But anyway, i um yeah, that's kind of what I do on ah my day-to-day is leadership, thought leadership, vision leadership, technical leadership strategy, um and taking us on that long journey of solving the big problems.
Mentorship in Tech
00:05:26
Speaker
That's cool. As as technology you know leaders and you know, as we kind of grow up in the industry, it's hard to to take that backseat and realize that, you know, we're we're there to mentor and and let others shine. at at a certain point, you need to take a step back away from, you know, getting, you know, like, move, get out of my way.
00:05:46
Speaker
Stop doing that. Like, you don't need to be the smartest person in the room. The other people are really smart too. So, yeah. Yeah, it's it's a really hard skill to learn. um and And part of kind of growing into that role is the ability to let go of the reins and say,
00:05:59
Speaker
What would you do? How are you going to do it? How are you going to make it better? How would you design it? And then supporting them and helping mentor them through that versus saying, well you should do it this way. You should do it that way. Now that's really, really hard to let go of that, that, that desire.
00:06:14
Speaker
I was mentored once by an engineer. um senior engineer at VMware, and he gave me the question. It was, how can you take these two hands and turn them into 100 pairs of hands? That's your goal.
00:06:26
Speaker
Not to do the work of 100 people, but to have 100 people do your work for you in a way that they feel like they own it and they enjoy it and they want to do it. To do that, to encourage people to do it that, is really, really hard. It's a really, really hard skill to learn.
00:06:40
Speaker
um Yeah, so that's basically what do on the 9to5.
The Quobix Persona
00:06:45
Speaker
So all of you senior engineers out there, if you want to learn how you you get to be a better and better engineer in this industry, you need that mindset. I love it. I love that. That's awesome. So and enough talking to this Dave Shanley guy. Okay, we've got an alter ego, Quobix. Now I want to talk to Quobix. Is he available right now? Is he around?
00:07:03
Speaker
Absolutely. I've got my cowboy hat on. cowboy I was wondering, like I don't know that I've known anyone that has an alter ego. like You see like Superman, you know it seems like it's real easy to just change between. He just parts his hair a different way, puts on some glasses, and nobody knows. is that Do you find that to be the case for you? Is it is it easy to just go back and forth? her what what do you ah How's your experience there?
00:07:24
Speaker
you know That's a great question. um And I think that if you kind of roll it ah back a little bit further and you think about like people who have alter egos, like Superman or Spider-Man, Batman, and they're all this person that they have to put on a mask to become their alter ego.
00:07:42
Speaker
But Superman's different. um And this is actually a quote from Kill Bill, I believe, volume two. But Superman is different because you know he is Superman. The cape, the red boots, the the blue outfit, it's all that's his. that's Those are his clothes.
00:07:59
Speaker
He has to become Clark Kent. to hide who he is. And I'm a little bit like that. Not that I'm super mad, not nothing like him, but, um, I am in the sense that that I always feel like Quobix versus Dave Shanley. Dave Shanley is who I've become to operate in, in the world.
00:08:15
Speaker
But you know, if I had unlimited money, time and resources, I'd never take my cowboy hat off. I'd be hacking code all day long and making music somehow trying to blend them together, which is really hard to do well.
00:08:26
Speaker
Um, But yeah, that's that's how I would operate. So I kind of see myself as Quobix all the time. And I have to become Dave Shanley to make money and have a career. All right. So you're you're you're an enigma. I said I have more curiosity. Okay. So you said you live talked about you live in Virginia.
00:08:45
Speaker
You don't sound like you live in Virginia. You mentioned earlier you're you're from England. Um, but you also, you play music that sounds to me like Lincoln Park and Green Day. And then you mentioned cowboy hat. There's, there's a whole lot of worlds here that are combined that don't make any sense.
00:09:00
Speaker
You know what i mean? And yeah explain all of this connection. How did the cowboy hat come to be? So I've always had an affinity for anything Western. i love I love the idea of the cowboy and what the cowboy stood for, you know, protecting the herd, driving it home, being out there on the front lines, you know, sole survivor. And if something went wrong out there on a drive, you are on your own and you still have to make sure that you got the job done.
00:09:26
Speaker
Everything about that just rings to me. And plus the fashion is is awesome. And, you know, I can't believe, I can't, You know, it's just uniquely American. I know it's influenced heavy hair, very heavy by Spanish and Mexican culture, but there's something about the cowboy in itself is, is uniquely American, you know, whether or not it's kind of composed of different cultures over the years.
00:09:47
Speaker
It's something just always sung to me. So I've i've always wanted cowboy, uh, buckle, like a big belt buckle. you know, even when I was in England, I come over here and you know, the only place I could find it was like some like hot topic and I'd find some kind of nonsense, cheap band,
00:10:01
Speaker
know, piece of crap and take it home when I when i went back to back to the UK. Obviously, that was a very, very long time ago. But um yeah, it just the whole culture of it, it just sings to me about the vision and and all of that stuff. So um that's why I kind of based all of that. I you know wear cowboy hats casually. I can't wear them when I'm doing like recording and stuff because i can't put my headphones over them. um But did normally, if you see me out in public, I'll be wearing it. If I do talks, I'm always was wearing a cowboy hat.
00:10:29
Speaker
I love it. and That's fantastic. Quite a renaissance, man. All right. So, Quobix and I are talking
Understanding Open API and JSON Schema
00:10:36
Speaker
now. And so you are in like, um you know, there's something that's been around for a while. Open API, we we kind of talked about that and JSON schema specifically.
00:10:46
Speaker
Tell us a little bit about these technologies for those who are uninitiated that are not familiar with what it would even is this open API thing. And why did why does JSON schema have anything to do with this? Yeah, it's a great question. um So open API is is ah is is the standard or has become the the adopted standard of how we define RESTful APIs out there on the internet. There's another couple of standards that existed.
00:11:07
Speaker
um For example, there was one called RAML. ah that was built by a company called MuleSoft. So a few of them kind of emerged, a little bit like VHS and Betamax. And OpenAPI 1, it wasn't always called OpenAPI, it started out as something called Swagger, which was built by a company called SmartBear.
00:11:26
Speaker
and which is this still still exists as a commercial product, but it's its own thing. It's its own standard and SmartPare donated, that as it was in time, that version of the standard to, um I believe it was the Linux Foundation.
00:11:41
Speaker
It is the Linux Foundation. Anyway, they donated it as ah as an open standard and said, here you go. And then what happened is that became OpenAPI and then it started moving forward and there was a bunch of changes made.
00:11:53
Speaker
um So yeah, ah Swagger actually won the war the API spec wars. Not because it was the best, but because the community was the largest. It had the widest adoption just because it was easier for people to build tooling with.
00:12:07
Speaker
um And there was just more people using it, and it just won. So yeah, that's OpenAPI as outstanding. JSON schema is is the definition of how you to define models within requests and responses and parameters and headers and things like that.
00:12:21
Speaker
And that is another standard. um and The difference is is that Swagger has like a ah form of JSON Schema support. It's like its own limited, it's not it's not really JSON Schema, it's kind of version of it.
00:12:34
Speaker
It uses the same namespaces and things like that. But JSON Schema is actually an open stat and OpenAPI 3.1 adopted JSON Schema as a standard, which means you can now use any tooling that supports JSON Schema, which is a lot of it. There's a lot of great tools out there that do support JSON Schema. You can now use them with OpenAPI because it can actually read or should be able to read those schemas.
00:12:57
Speaker
um And that that great and it creates a lot of problems because most tools are founded on the the hacked version of the the model versus versus the actual implementation.
00:13:09
Speaker
So yeah, I'm kind of building tooling specifically in Go. to ah started about three years ago because I was building a lot this tooling internally inside VMware that we were deploying and running, and people loved it.
00:13:23
Speaker
Developers loved the tooling. It's still operating today in places, I believe. but um We couldn't open source it for different reasons. not going to go into them. but and So that's why i started rebuilding all this stuff outside of the company. So it was available for everyone who had these problems start using it. And that problem, you know is there's multiple companies now doing exactly what I thought they would be doing with the library.
00:13:45
Speaker
and the tooling you know to to solve these OpenAPI problems. So that's why I kind of started doing this was A, the tooling was rubbish, the ecosystem was bad, and there was companies that were doing it internally, but keeping it all for themselves.
00:13:58
Speaker
So I thought, okay, let's break the cycle. yeah So OpenAPI is a way for us to kind of document and share this this is what this API does and here's how you how you talk to it, how you interact. Exactly. the gist of it Yeah, okay yeah it it's so all the endpoints and all the inputs, the parameters and how you would you know make a request to it, the kind of response is going to get back out of it, with the the data shape.
00:14:21
Speaker
So the properties and values and what types they're going to be, whether they're numbers or strings and things like that. So gives you this really,
API Management Maturity
00:14:28
Speaker
really strongly defined contract that you know what goes in, you know exactly what should be coming out, which means you should be able to feed that that spec, that contract, OpenAPI spec, into any different type of tooling, and it will be able to spit out different artifacts from that spec. So, for example, documentation.
00:14:46
Speaker
um and you want to be able to document the API. So you just you know fill out the the spec and complete it you know fully, and you've got great docs automatically. The same thing with SDKs. You can auto-generate code.
00:14:58
Speaker
ah So your your clients will be automatically... you know all the All the code you would need to be able to call your API will be generated in JavaScript or Perl or Python or PHP or Java, whatever it is that you're out there using, you'd you'd that you'd have a code generator for it.
00:15:14
Speaker
And this is what we call like an API first design where you design the API um through open API and then everything else kind falls from that center core. So client generation, code generation, test generation, server generation, backend generation, all that stuff comes from the that the spec. So having a single source of truth for your ah for your API.
00:15:36
Speaker
thanks So when you have all these, you know, a lot of the organizations are facing, you know, we're becoming very API first as you were saying, to you know, design first or API first.
00:15:47
Speaker
um and And when you're inside these big enterprises, you kind of get lost in this, this sea of all these open API documents. How how are you finding like, you know, managing this, bringing everything together and making people aware of this sort of thing? how it what What are people doing with that?
00:16:02
Speaker
Yeah, this is actually, this is this is a great question. um So I think there's there's a few states of maturity that businesses have ah in this in this in this journey. ah you You get, let's let's say an enterprise business is 15 years old or so.
00:16:18
Speaker
um They're going to have lots of different products, some acquisitions, some homegrown. Open API specs are going to be scattered, like you said. Some of them are going to be generated from code. Some of them going to be slapped together. Some some of them are going be thoughtful.
00:16:31
Speaker
um First of all, that like like trump before trying to build like a directory of all the specs, is you've got to get everyone using specs. So whilst you might have a selection of you know specs from that are in different states, you've also got bunch of teams that aren't even thinking about this.
00:16:48
Speaker
haven't even done any work on OpenAPI or any kind of API standard, and they don't have any tooling around it. So they're they're kind of starting from scratch. So there's first, there's there's an like awareness is how do you get everyone to see?
00:17:01
Speaker
So the first thing you do pull everything in that you can find and suck it all in into some form of your own kind of directory to begin with. And there's actually products out there, like Postman have a platform that does this, that allows you to do this kind of spec centralization. But be able to pull it in and do an analysis and figure out where are we in our journeys, how well are these specs put together?
00:17:24
Speaker
If you find that most of those specs go in pretty well, they render out relatively well in documentation, and you can get you know a reasonable kind of quality score out of them. That's okay. we're We're in a good position. Now we need a tool that's going to give us like a central directory where we can allow everyone to explore, search, try them out.
00:17:43
Speaker
But what I'm finding is most enterprise businesses are upside down and inside out when it comes to OpenAPI specifically and having specs.
00:17:54
Speaker
There's parts of the business that are really clean because they're building cloud native, they need OpenAPI. For example, they're using some form of Kubernetes API server, they need to feed something in to be able to to expose rounds and endpoints.
00:18:08
Speaker
So they're really doing well. We could do a lot better, but they're at least they they made that part of their test. And there's yeah the legacy stuff, ah which I see every day in in my day job, which just is playing catch up all the time.
00:18:21
Speaker
And new things might get built with ah with a spec, but most of the stuff is completely undocumented and unspecified. So That level of maturity is like we're we're just getting started. We're just trying to things out.
00:18:32
Speaker
Having us like that kind of a platform solution is just a little bit too early for businesses in that state.
Low-Code Development with OpenAPI
00:18:38
Speaker
So really, we're good but in that point is it's about getting the team's hygiene. Like it's a cultural thing before you even get to worrying about and having a single place all the specs go. It's how do we get teams caring about this, worrying about this, implementing this and making it part of their build.
00:18:53
Speaker
You've got to have all of those how-tos, how do you stitch this tooling together, how do you use this type of tooling, what happens when you do this? and i think that's actually the the hardest part, um which i've had which I actually spend a lot of time during the day doing, is educating and then showcasing and then getting examples of, this is how we do it properly.
00:19:12
Speaker
And then you start to see the spark go off and then teams start making use of it. um So yeah, it's mature businesses, use a platform or build your own. It's really hard though. wouldn't recommend it. I really wouldn't recommend it really, really hard.
00:19:26
Speaker
So yeah, what you know just do on this one, buy. But before then, it is a case of actually getting tooling. So stitching together open source tooling to be able to showcase, you know, we've got quality evaluation. We've got change detection in here. We've got some form of governance going on.
00:19:42
Speaker
we can see what's going on. Getting to that point is really hard. Once you're that point, it's just rinse and repeat and until you're ready for a platform. All right. So do you find that once you get to this level of maturity of the OpenAPI specs being clean and and they're practicing good hygiene and all of that, where you can do kind of that low code, no code kind of a thing and and actually start weaving things together without you know much engineering effort?
00:20:09
Speaker
Yeah, I mean, it's it's a bit of a utopia because I have yet to see ah kind of an organization get to that level of maturity, but that is where it should be. So you have the capability for anyone to be able to explore APIs.
00:20:24
Speaker
um and ah discover them, use them and build automation against them. That is where it should be, right? And adding a new API is a design process because all the code is automatically generated for SDKs and clients.
00:20:40
Speaker
And all the code is automatically generated for the backend, which means you're not so you're not checking any of this code in. You design the spec. The spec is then actually used to generate the code and then go into the directory. So, i mean, that's...
00:20:54
Speaker
really, you know, ah where we where we want to get to, where non-technical folks can ah explore, understand, automate, and create their own workflows and publish those workflows, you know, so then you've got the ability to do kind of high-level education for, like, this is a how you how you do this thing, here is my workflow, and you can capture that and then publish it out, and then then anyone else can pick that up, whether it's a customer or it's, ah you know, another engineer.
AI's Role in OpenAPI
00:21:20
Speaker
So we talked about there's a bit of an AI play here. um There's a there's a trend in the industry that a big push for agentic AI, right? This does play into that some, um you know, if you if you can, you know, present these open API specs such that they're easy to weave together by a human being, presumably an AI agent could do the same kind of is that where you're where you were were headed with that?
00:21:43
Speaker
Yeah. um I mean, and so there's there's ah there's a big boom in MCP as a new concept out there in the market right now. And there's other concepts like A2A.
00:21:57
Speaker
So MCP stands for model context protocol. And A2A is agent to agent protocol without I've always got P on the end. not sure. Anyway, um the the Personally, I believe that an open API spec is all that you need really to be able to understand the capabilities of an API.
00:22:16
Speaker
But what we've created or what the market has created is this layer over the top where it's um it's kind of put this RPC style um communication layer. So it's it's basically JSON schema.
00:22:28
Speaker
over of a post, which is JSON RPC. So it's like an ultra lightweight version of REST. There's no real resources or anything like that. There's no real verbs or anything. It's just post a payload with a schema with context of the conversation.
00:22:45
Speaker
And we're calling it a protocol. So there's there's a lot of hype around that at the moment, really, because that's that's all it is. It's it's a very old fashioned It's a step backwards, really, in terms of um how you think about APIs and the capabilities of APIs. so Backwards from REST, anyway.
00:23:00
Speaker
So I think you know we're going to see a lot more rise for the need of open API or or at least any kind of API you're interacting with, whether it's GraphQL or PROTOBUFF. I think we're going to see a lot more demand for it because AIs are going to want to automate lot of this stuff. MCP is abstracting that, but MCP still needs to be able to understand how to talk to those underlying services and how's it how's it going to need to do that when it needs you know an API contract or some form of an API contract.
00:23:30
Speaker
Not all APIs are going to be HTTP based, obviously. But yeah yeah know it's it's not a replacement. There's there's been a lot of kind of media coverage about MCP being the new standard for how we do APIs. And I'm like, no, I don't.
00:23:43
Speaker
No, I think that's just that's clickbait beyond anything else. um It's doing nothing new or nothing different. And I also am a little concerned that because it's backed by a commercial entity,
00:23:59
Speaker
ah They've got their own interests at heart. So there's every other entity that's playing in AI right now, which is everyone. So everyone wants to be the new protocol thing. So we're we're stuck with we've already seen A2A from Google.
00:24:11
Speaker
We've got MCP from Anthropic. um I'm pretty sure we're about to see a flood more. Maybe not. Maybe maybe it stops here. Maybe this is like oh this is it. We've solved it. But there's there's a lot of problems with MCP, security being one of them. Massive, massive problem.
00:24:27
Speaker
I'm sure there'll be solutions coming up for that. But then we'll start to see deviations of mc MCP, we offer different flavors
Introducing OpenAPI Doctor
00:24:34
Speaker
of MCP. then we will be right back to where we are again, where we were with the spec wars. you know yeah Rammel versus Swagger all over again.
00:24:41
Speaker
But now we've got this added layer. So I'm not betting on any of them, to be honest. I'm standing back and waiting for the hype to fall because I like working on the plateau of product productivity, the slope of enlightenment and the plateau of product productivity.
00:24:56
Speaker
That's where I operate. That's the tools that I build operate. When you need something that's going to work, come see me. When you stay with the hype, you need to go over there because I can't help you. I love that us technologists and then just as an industry, we're so funny. You know, like as you said, OpenAPI has been around for a bit. We've had a great way of explaining, here's what this API does. Here's how you talk to it.
00:25:18
Speaker
And then we came up with this, hey, what if we do this new agentic AI thing? And we're going to hand these agents... tools and but man we need to describe these tools and how do you talk to them let's come up with something brand new let's let's just start we have we obviously there's nothing available so we have to just create this from scratch it's just it's crazy that we do this to ourselves all it's over and over again There's a great XKCD strip that does explains this exactly. i think it begins with something like there's 15 standards and we we need to solve this problem, but none of the standards fit.
00:25:54
Speaker
You know what? We'll create one new standard to rule them all. We'll have one overarching standard that fits all the other standards inside of it. And then the next plan is now we have 16 standards. Mm-hmm. and Yeah, exactly right. Exactly what we do. yeah It'd be great if we could leverage these existing technologies. And, you know, it's it's kind of cool, this this notion of agentic where, you know, you you you ask them the you know these these fancy models to plan out, hey, I need i have to solve this complex problem that that could involve multiple steps. And and you say, OK, here are all the tools at your disposal. And the engine just decides, OK, in order to solve this problem, the first thing you do, A, that's a tool for that. I'm going to talk to that and all that.
00:26:33
Speaker
So I think it and absolutely naturally lends itself. So hopefully we'll see a little more like you're saying in a little more standardization and and people settling on things that are that are reasonable and and not jockeying for control so much as let's let's just be sane about it for once. i mean Yeah. Yeah. I don't know where it's going land. um I feel like it's it every day there's a new spike on the AI hype peak. So there's just, it starts again, like there's a new trends on, there's another trend, there's a new thing happening now.
00:27:04
Speaker
multi-agentic now. And there's another one that's like, most people still haven't even added the word agentic to their marketing material. Now they're going to go back and start adding, you know, modifications to that.
00:27:16
Speaker
That's how fast things are moving, um which it it just lends to, it just lends to From my experience and as your you've experienced as well in the past, just stand back and wait for the wait for the balls to land on the floor because at some point they're all going to come down. They can't keep going up and up and up.
00:27:33
Speaker
We are going to have some form of build come due and then we'll figure out how to make use of all this stuff properly. All right. So you I wanted to maybe dig into you. You have this this product that you're working on that I wanted to hear a lot more about. So tell me tell me about your product that you've been working on, your your passion project, so to speak.
00:27:52
Speaker
Yeah. Great. Thank you. So... I have a company and the company is called Princess Beef Heavy Industries and the acronym is PB33F.
00:28:03
Speaker
And for the last few years, what I've been doing is is building the pieces that I need to build the product. I need all these bits and pieces and they're all really useful pete bits and pieces on their own. um They're individual units of value, like they're they're power tools, if you will.
00:28:18
Speaker
So I've been releasing them as open source. ah Whilst you know the goal is is i I need these things to build what I want, the things that that I need aren't out there in the form that I need, so I'll build them and know'll give them away, give these pieces away for everyone else to be able to go and use. so And um you know they they there's different areas of value.
00:28:38
Speaker
um There's a library that's a complete parser and model and diff engine for open API specs, yeah know huge enterprise-grade ones. I'm talking you know hundreds of Thousands and thousands and thousands of schemes, hundreds of thousands of schemes, massive, um across many, many files, you know just gigantic specifications.
00:29:01
Speaker
And then there's another tool called Vacuum, which is a linter. And what it allows you to do is it's actually it's actually inspired by another open source tool called Spectral. And Spectral will allow you to do, they allow you it's got a rules engine. you can use There's core functions that you can put in and say, like for this path, check if this thing exists or it doesn't exist, or if you've used the right word or uppercase or lowercase.
00:29:25
Speaker
So it allows you to define your own rules or use the built-in rules to define, to check qualities, like have you done this, have you not done this? Is this a good high quality spec? Is it valid? Have you and violated the schema, et cetera?
00:29:38
Speaker
And there's another tool called Wiretap. And Wiretap is like a compliance engine. So it operates as a proxy. So you would send all your client traffic, your UI traffic, or your CLI traffic through it to your ah service.
00:29:53
Speaker
And what it's going to do is look at all the requests and responses and align them with the OpenAPI contract and tell you how compliant your back end or your front end is with that contract. And you shouldn't need this tool.
00:30:05
Speaker
It shouldn't need to exist if you've if you've done it API first. But lo and behold, most people haven't. So that's why it exists. Yeah, they people use it. They depend heavily on it. it gets a lot of use.
00:30:18
Speaker
So yeah, those three those three products are the open source. They're libraries, tools, binaries. They get a lot of uses. A few million downloads now. Nice.
00:30:29
Speaker
A couple of thousand stars across the repos in GitHub. um But again, those are things that I need. They're not like, here, here's open source. Well, thank you. you know There's my gift to the universe.
00:30:41
Speaker
Sure, it's free, but I need it. like I use it. So everything everything that I build is is a feature that I require. Anyway, they're all built with a purpose. to create this this end goal, this universal all-in-one. So you don't have a tool for quality. You don't have a tool for change detection. You don't have a tool for compliance.
00:31:00
Speaker
You have a tool that does all of it. um And that's what the doctor is. So that's the product that I'm building. It's it's called the OpenAPI doctor, but it's themed after um It's a sci-fi show from the UK called Doctor Who. You may have heard of it. You may have heard of it.
00:31:17
Speaker
Everything about it is themed around it. So it's just my own kind of association with it. For example, all the Git commits in the in the repo, all the commits are based off of Doctor Who quotes.
00:31:28
Speaker
So anyway, that's the whole point is that you can... it's ah It's an all-in-one for OpenAPI, so Design Studio. So you can design. Every change you make is tracked. You can see exactly what changed in these unique ways that are completely unavailable anywhere else. Like there's graphs of, you know, diffs that just don't exist in any other tool.
00:31:47
Speaker
You can see the model as a graph. You can see all the references and the relationships between schemas. You can do your quality control. view there's There's vacuum rules built into the doctor because it's powered by vacuum.
00:31:59
Speaker
um There's change detection. haven't built wiretap in yet, so the compliance part isn't available. But it's the idea that not only can you have this as a SaaS service, so it runs as an API, you can build your own products on top of it, or you can run it as part of your CI CD, or you can run it on-prem.
00:32:16
Speaker
It's designed to actually run as an on-prem product, so you know you you want to keep control of your own data, no problem. Or you want to run it as a desktop app. You want to run it as part your dev, like local, like you don't need Docker. You just run the binary and so a desktop UI will pop up and you've got everything you need. So it'll fit into any form factor, run in a Docker container, run as a SaaS service, run on-prem as a service, run as a desktop application or run as a CLI tool.
00:32:41
Speaker
But you've got everything you need in one place. and that's the That's the goal. It's an all-in-one suite. Everything you need, a complete and end-to-end solution for the upstream of OpenAPI, which is contract validation, design, quality, and compliance.
Industry Trends and Advice
00:32:57
Speaker
So the names of all the the tools, i kind of I kind of get those. I can see where you got those, you know, the vacuum and whatnot. Yeah. Doctor, all these things make sense. Okay, you got to tell me about Princess Beef. i like Where did this come from?
00:33:09
Speaker
that's how That's kind of out of left field here again. This is another Cowboy Hat in Virginia deal. What's going here? Well, i was I was thinking about, I think I needed to create some form of an entity because there's sponsorship on the GitHub repos. A few companies reached out to me saying, we want to sponsor you. I'm like, I don't have a bank bank account for business. um So I set an LLC just so I could have that going.
00:33:32
Speaker
um And thinking, well, what do I call it? What am going to call all this stuff? So thought, was thinking about different know boring names, like verbs and the kind of stuff that you see most startups pick, something that's inoffensive and corporate friendly, but sounds fizzy.
00:33:52
Speaker
No, this is boring, boring, it's all boring. So my my daughter was two at the time. um And she had a show, a TV show that she watched. It's called Blaze and the Monster Machines. It's an animated show about monster trucks.
00:34:03
Speaker
And she would say, I want to watch Princess Beef. And we're like, what? What show is Princess Beef? And she'd point at Blaze and the Monster Machine and say, Princess Beef.
00:34:16
Speaker
but There's no mentions of Princesses or Beef or Princess Beef anywhere in the show. That's just what my two-year-old called it. So I thought, well, isn't that a great name for a band? That is a great band name. Isn't that? Because she actually used used to play these like toy drums and she called them the honey drums. I thought Princess Beef and the honey drums.
00:34:35
Speaker
That, now that is a name for a band. That is a band name. That's a great name for a band. But, you know, as I said, my, my music days and my professional music days are long behind me. So um ah can't use it for a band.
00:34:48
Speaker
What um else am I going to use it for? why not call a company princess beef i love it and the the heavy industries actually comes from ah back in the 90s there was ah so hack a hacker think tank called loft heavy industries and um they created a bunch of tools like lock crack for windows NT anyway they were they were like a real aspirational hacker group um they were kind of disbanded in 2000 so only around around for a very short time um but I thought what a great you know, heavy industries, you know, for for software, because it's the antithesis of heavy industries, right? There's no heavy machinery. There's nothing. It doesn't exist. You turn the power off. It's gone.
00:35:30
Speaker
So I was like, well, I love it. Why not bring back Hacker Roots? Because it's all about hacker design. Like that everything's cyberpunk and neon pink and blue. And, you know, it looks like a terminal. So it's it's all kind of in that atmosphere. And why not bring it back and pay homage to Loft?
00:35:48
Speaker
All right. So this has been fantastic. We have a couple of ah segments that we do with every guest. ah One of them, the the first one we're going to do is what we call ship it or skip it.
00:36:01
Speaker
Ship or skip. Ship or skip. Everybody, we got to tell us if you ship or skip. So we have a ah few topics, so I'll just kind of ask you, and ship it or skip it, it's basically like you know hot or not kind of a thing, right? Okay. um So I'll come up with some some topic ideas, and we'll just kind of, do do we ship that? That's good stuff, or skip it? No, it's not good.
00:36:22
Speaker
How about vibe coding? There's a lot of stuff going on around vibe coding these days. What do you think about vibe coding? Skip it. Skip it. Throw it in the trash and burn it. Definitely. Yeah. ah Vibe coding coding, you know, it's great if you're an experienced engineer and you know what you're doing and you know how things work and you want to be able to sketch. It's like taking, it's like like Procreate being on the iPad. It's a supercharger for sketching if you're an artist.
00:36:49
Speaker
If you're an amateur, you know, you'll end up making things look like, you know, a crayon picture like a five-year-old would because you don't understand the tool. And that that's really what I see, you know, vibe coding. AI is a power tool for people who understand that particular domain and they can enhance their output in that domain.
00:37:08
Speaker
But if you don't know what you're doing and you start vibe coding and there was an explosion a month or two ago where a couple of influencers were building like games and like flying games and stuff like that. And, you know, it's part of junk.
00:37:20
Speaker
and It's a big, massive, big blob of garbage that requires a community to to debug. um And whilst it's fun, you know, as a this is this is good for a sketch or POC or to get something off the ground, trusting it as a product or asking people to pay money for it is is danger danger, danger, danger, danger. Because like anyone knows, anyone working software for a long time is the the chaos isn't really the upfront build. It's easy to build something new, but maintaining something over the long term and building something to be maintainable with minimal amount of manpower
00:37:56
Speaker
That's hard. And you're not going to get that from Vibe coding at all. you're goingnna You're going to go the opposite way. So it's not so much the upfront value. It's the what happens next. That's the part that, you know, I'm done now. I'll move on to the next cool thing.
00:38:11
Speaker
And that thing that people are now paying for and all that all the complaints, and it ends up you know collapsing. that's That's the result Vibe coding. So skip it. I agree with you. I do like the nuanced answer of like, you know, for for folks who who do have some experience, it can can be a helpful yeah power tool. It can be a way to accelerate your workflows. But being a consultant, right, in in the consulting business, and you see these fads come and go, right? But there there's then some, you have to pay the piper eventually, right? At at some point, it becomes the point where, oh my goodness, we don't have any idea how to maintain this thing. And that's where
00:38:45
Speaker
mean I mean, I've made a career out of that. I've done a lot of cleanups. So I mean, I'm okay with it. It's fine. Anyway. All right. Another one, we kind of touched on a little bit as we were talking, but like model context, pro call, A2A, like this whole notion of like we're trying to come up, land on some standards.
00:39:02
Speaker
What do you think about this MCP A2A thing? and we are you Are you ship it, skip it? what do you think? ah This is a hard one. I'm kind on the fence um because that the the industry has has picked this and is moving with it.
00:39:14
Speaker
um And that's where the action is. And it's it's moving and there's people using it and there's all kinds of new implementations coming out of it, but but it's evolving so rapidly that I feel like adopting anything or head running headfirst into anything like that right now would be a costly mistake mistake in the long run.
00:39:32
Speaker
But skipping it completely I think it's a bad idea because this is, this is something that is emerging and is what people are talking about. Whether theyre not, whether or not it's going to become the new norm.
00:39:43
Speaker
It's the primordial, primordial soup of where this new idea is coming from. so I'd pay attention. I'd pay attention. I wouldn't completely skip it, but I wouldn't i wouldn't dive in. i don't know if that's a great answer, but it's kind of on the on the fence on this one.
00:40:02
Speaker
Look at it. Definitely understand it, but don't don't run at it like it's Adobe Flash. Don't do that. that's a fit That's a fantastic answer. I'm with you. I think this, for me, don't know. I have this uncanny ability. I'm kind of patting myself on the back here. Sorry.
00:40:20
Speaker
Self-serving, but I feel like I have this uncanny ability of like, it was almost like an instinctual, intuitive sense of when there's just too much complexity right now. And I, my spidey sense is like tangling right now. Like it's this, this world that I don't think we've quite landed on where we're going to eventually kind of settle.
00:40:39
Speaker
I think we're at a ah local minima or local maxima, so to speak. Right. I don't think we've kind of gotten to the point where we're like, we've made that breakthrough to it that'll get us to that global maxima yet. um But yeah, I think there's there's a breakthrough that needs to happen still. My spidey sense is like, no, no, no, don't yeah don't go heavy investing in this stuff quite yet. i'm i'm So I'm with you. I agree with you on that.
00:41:01
Speaker
it's And it's tough to sift through all, the the sifting through that but nonsense. There's a lot of you know noise. you know Noise to signal ratio right now is a little too high. Yeah, every every company has a big AI layer in it now. In every diagram, in every piece of marketing, a big AI box, AI thing. There's AI everywhere.
00:41:21
Speaker
And you ask them, how are you actually using it? Oh, well, we've got a chatbot and we kind of wrapped open AI. I said, okay, great, great. It used to be if you put AI in your product, it was everybody. The joke was, oh, you've got a couple if statements in there. and ah it's It's a little more advanced than that now because chatbots are so easy. But yeah, I'm i'm with you. it's It's watered down a little bit right now. there You can't really see the the true innovation in AI right now because everybody's just no, we're doing AI.
00:41:50
Speaker
All right. um the This was an interesting one. and I'd love to hear your take on it. ah Like the the industry right now, it's funny. We have this, don't know, I call it like the the the Kardashian effect, right? there's Everybody wants to be an influencer, right? They want to want to get out there and influence other people.
00:42:09
Speaker
um What's your take on that? how How do you feel about how we are as an industry now? and I mean, to the world, but I mean, our industry in particular is getting hit with this. I love this question. Um, so, uh, there, there is a number of, I mean, you see them like doing tutorials, creating their own, ah courses, ah trying to get, uh, likes and hits, uh, you know, on social media for anything to do with coding or relating related to coding. Cause it's like, there's, there's a couple of really good influence out, good influences out there that are doing great work. People like the Primer Gen, um, You know, a very experienced guy with a lot of value to add, has done really great training courses, builds really interesting things. Content is entertaining.
00:42:53
Speaker
But the reason why he's got all that credibility is because he actually knows what he's talking about. like not Not at the high level, but then not right down to the the low level, super low level. The the the core fundamental the first principles, core fundamentals computer science, he understands it and is able to explain it in a way that is is helpful.
00:43:09
Speaker
That's something to aspire to. that's But that takes... 15, 20 years of experience. And the problem is, is that you know from what I see as young guys ah wanting to do that. like I want to be a streamer. i want to be you know I want to be on Twitch, and I want to be coding, and I want to be giving advice, but I don't quite know what to give because I don't know what I'm doing yet. But I still want to... It's a little bit like um the people that you know they want to take up skiing, so they go out and buy like the most expensive skis and all the gear, and they've got the Instagram pictures with the stuff on. and
00:43:42
Speaker
you know and and But they can't ski. But they got all the stuff they look like they can. you know they they kind of They give the impression that they can, but they can't actually do it yet. um It's like twenty wanting to um skip the line on experience. you know you want in what ah i want i want the credibility, but I don't want to have to know do the hard work to get there.
00:44:02
Speaker
And I think that's really the key is you can't short-circuit it You can't study your way into this you know in a shortened amount of time. It's experience takes time. It's ah the muscle memory, the the neural the neural pathways in your head that need to be formed from repetitive work, from trial trial and failure, learning. you can't do it. And I think that's what I see with a lot of younger people trying to break into technology, but I don't think kind of diving into the Kardashian effect and trying to become like, you know, some superstar streamer, the people that are making money for it and are famous are because they're experienced and they got something to add.
00:44:40
Speaker
know, they've got real experience to share. and they're turning it into a ah format that is enjoyable and consumable because they know do that too, right? These are two skills that you've got learn through experience and by presenting and storytelling and learning how to and i put up put a storyline together and be able to tell it in an interesting way and provide value. So yeah, the the tech influencers, don't do it until until you' like you've actually got experience. You really know what you're talking about. Not just, I did a course and I'm now
00:45:11
Speaker
yeah trying to um prove that I understand what learned by teaching. Putting yourself into a situation as like a mentor at that point is is is is the mistake. yeah There was a Twitter, the feet well, that's not Twitter, X, conversation I saw. Somebody was asking, you know, what,
00:45:30
Speaker
What do you call those people? You you brought it up both of of like, you know, the people who, who you know, when they want to be a skier and they go and buy all the fancy equipment and they and they buy all those things and and immediately that's what they buy. And they're like, what's the word for that? We need a word for that. And there there's people that are you know, my age that remember back in the day, like the skaters.
00:45:48
Speaker
They were like, ah we've had that word forever. It's called poser. Right? You don't need to invent a new word. We've got a word for that. it's called a poser. That's a great word. that's That's a great word to describe. we We need to revive that because it describes it perfectly, doesn't it?
00:46:03
Speaker
um so It does. So you're skip it, kind of. Yeah, skip it until you have experience. Grind. Like work. Do the work. Do the work. And then you know when you've got the credibility to be able to Build the kind of things that now you're inspired by because genuinely build them because you know how to build them because youve you gain the experience and the skills. You've grown that skill set.
00:46:23
Speaker
ah Then teach others. Then then then be it then be a a role model. But, you know, don't don't don't skip the line because you've got nothing to offer
Lightning Round Fun
00:46:33
Speaker
yet you. You just don't know enough.
00:46:34
Speaker
Yeah, I'm with you. I'm going to skip it as well. I say, you know, it's uncanny. whenever we human beings try to take shortcuts to, to the thing, whatever that is, it's, it just, it doesn't seem to work out well. You think of like, you mentioned muscle memory and I was thinking steroids, right at the, you know, when people do that, there's, there's terrible effects on your body, you know, that sort of thing. Like Whenever we try to take shortcuts, it's like nature smacks us around and says, this this is not how the world or the universe works. You've got to go through the path in order to do something meaningful. Exactly. Yeah. You've you've you've got to do you've got to you've got to earn it, right?
00:47:09
Speaker
I agree. Okay. So right we have another segment of our show that we call the lightning round. And this is this is the most important segment. these is These questions, these are the hard hitting questions that we're really driving This is what people tune in for.
00:47:25
Speaker
Honestly, this, this is, this is where we really get down to the heart of the matter and, and really get to know the, the, you know, the nuts and bolts of the world and the universe. ah So are you, are you ready for this?
00:47:36
Speaker
I am ready. Yeah. Hey man. It's time for the round.
00:47:57
Speaker
Cake or pie? Can i ask a question or do i have to answer it? We'll allow it. One question. Okay. Meat meat kept meat pie or like sweet pie? ah We've never had someone from England on the show yet, so I did not know. i like like Like a steak pie? Like a steak pie? Or like ah like you know like a cheese cheesecake?
00:48:15
Speaker
and we'll go with We'll go with sweet pie. Sweet pie. Cake. Cake. Now, if it was the other way around. Meat pie. Meat pie. Okay. yeah There's no such thing as a meat cake?
00:48:27
Speaker
I mean, there is. i don't know how tasty it would be. But yeah, i'm sure you can make it. I actually use meat cake is as example properties in my test code, funnily enough.
00:48:37
Speaker
Definitely meat pie. i don't know about a meat cake. um It's a crime that like just a steak pie is so hard to find the United States because it is incredible.
00:48:48
Speaker
You can get chicken pot pie, yeah but you can't get like a steak and ale pot pie. Steak and ale? That sounds amazing. Yeah. Okay. Have you ever worn socks with sandals?
00:49:01
Speaker
I'm wearing them right now. Okay. That's a guess. All right. Do you like the word dapper? Yes, I do. When I hear dapper, I hear somebody who's dressed really well in a fancy way that's more fancy than traditional like a business attire, but not super fancy like black tie going to a ball.
00:49:22
Speaker
Dapper means a classier suit than business, not quite, you know, dressed up to go out to a formal occasion. It's in the middle. Dapper covers it perfectly. like that.
00:49:33
Speaker
ands a That was a very thoughtful answer, and I enjoyed it. You get a bonus point for that one. Oh, thank you. For sure. Okay, what is your favorite carnival food? Oh, my goodness.
00:49:45
Speaker
um ah You know, it's probably like ah like a Polish sausage, to be honest. Okay. I'm not a fan of, like, that the funnel cakes or anything fried or the...
00:49:59
Speaker
No. The deep fried stuff, no. Yeah. Polar sausage. Polar sausage. That's a good one. Do you Instagram your food? No, I don't use Instagram. Okay. And i if I did, I don't do the food thing.
00:50:10
Speaker
No, definitely not. the The camera doesn't have to eat first. like That's what my kids say. Never. I mean, maybe like 15 years ago, i tried doing it a few times. when I realized, like, what am I doing? and then I stopped forever.
00:50:21
Speaker
Okay. Okay. So since you're, you're from England and you have a good grasp, grasp of the English language, I can do an imitation. um ah Define the word zeitgeist.
00:50:35
Speaker
The moment is understanding the cultural moment. Zeitgeist is is the is the the feel, the vibe of the world as it exists right now. And the Zeitgeist changes with every generation and time.
00:50:46
Speaker
So it's ah capturing the the gestalt of culture as it stands. Man, that's probably like very close to like the textbook definition. That's fantastic. Great.
00:50:57
Speaker
Thank you. What's your favorite martial art? Oh, that's random. Um... Man, i wish i I wish I knew more names of martial arts because I'm failing now and I can't think of anything except for Kung Fu and that's really, really bad.
00:51:14
Speaker
i don't have one, to be honest. I'm just picking one out of the air. That's good answer. Kung Fu is They're movies. the The bad overdubs of English, you know.
00:51:24
Speaker
I mean, it's, yeah, Kung Fu. Why not? Yeah. i mean it's yeah kung fu what
00:51:33
Speaker
okay Here's a random one. If you could push a button and make everyone in the world 7% happier, but it would also place a worldwide ban on all hair styling products, would you push that button?
00:51:45
Speaker
Yes. Yes. Yes, I would. feel like that's fair. Everyone deserves to be happy at the expense of a slightly less fashionable haircut. Hmm. I think it's fair. It's a good trade-off.
00:51:57
Speaker
Do you think anyone considers you a hipster? No. Not at all. Nope. Not even close. As a younger man, maybe 20 years ago, 25 years ago, 100%.
00:52:09
Speaker
twenty five years ago hundred percent Friends used to call me bohemian, but I'm anything but these days. Gotcha. All right. Well, that concludes the lightning round.
00:52:19
Speaker
That was fantastic. You scored better than most, obviously. You got a bonus point, so that it helps your score. It was fantastic. The Polish sausage, was was the that was the number one answer that we were looking for.
00:52:30
Speaker
Really? Yeah. Somebody says, there you go. Somebody says, Polish sausage. That's right. I'll be darned. Okay, so any closing
Final Thoughts and Farewell
00:52:42
Speaker
remarks?
00:52:42
Speaker
ah we're We're going to sign off here in a few. Any closing remarks from you? Closing remarks. um Don't get lost in the hype. ah AI is has been buzzing for two, three years now.
00:52:55
Speaker
It's going to keep buzzing. There's going to be more and more protocols popping up. MCP, A2A, who knows what's going to come next week. um Don't bet the bank on any of this stuff yet because it's rapidly evolving and changing and growing.
00:53:10
Speaker
And I wouldn't bet my business on it just yet. I'd wait for the dust to settle. um that's That would be the my parting advice is look, touch, but don't buy yet.
00:53:21
Speaker
That means if you can do that. Fantastic. And if our folks want to find you, they can go to qoobix.com. Is that correct?
00:53:32
Speaker
quote but That's right. Q-U-O-B-Q-O-B-X.com or PB33F.io. That's princess beef. And you can grab, you can take a look at the documentation for the different open source stuff, or you can try out the doctor, um or my contact details are on there as well Fantastic.
00:53:48
Speaker
All right. Well, thank you very much, Quobix or Dave, whoever's here right now for joining us today. Well, thank you. um Both of us are here at the moment right now. I really appreciate the opportunity to come talk today.
00:54:02
Speaker
Thank you for taking the time to listen to me rant and listen to me ah vent all of my ideas out my head. Really appreciate it. Thank you. All right. Well, this was fantastic. If you'd like to get in touch with us, drop us a line at the forward slash at Caliberty.com.
00:54:18
Speaker
The forward slash podcast is created by Caliberty. Our director is Dylan Quartz, producer Ryan Wilson, with editing by John Corey and Jeremy Brown. Marketing support comes from Taylor Blessing. I'm your host, James Carmen, and thank you for listening.