Welcome to Energy Bytes with John Kalfayan and Bobby Neelon. Your essential guide to the intersection of data and energy. This podcast dives deep into the world of energy, shedding light on how data, AI, and technology are revolutionizing this sector. Each episode equips listeners with insights into the most efficient tools and resources, paving the way for a data-driven future in energy. From technical nuances to broader industry trends, Energy Bytes offers an unparalleled perspective on the evolution of the energy industry. Join us as we decode the algorithms of energy, one byte at a time.
Welcome back, everybody, to another episode of of Energy Bytes. John Calfan here.
With my wonderful co-host, Bobby Newland. What's good? You're doing good today, Bobby? Awesome. Today we've got John Archer, the Global Edge Hardware Partner Manager at at Red Hat. Appreciate you coming, man. Thanks for having me. A little closer. A little closer. Yeah. Thanks for having me. Can you hear me hear me there? No, you're It's It's honestly mostly from the the other microphones picking you up, not the the main one, but it it feels weird speaking directly into this giant thing. Yeah. [laughter] Yeah, it it is weird, but closer is better. I mean, I I I'm I'm been aware it. Yeah. Thanks for coming today, man. Appreciate it. Yeah, thanks for having me. It's nice meeting you guys and know we've had one conversation before and really like talking to you guys. Yeah. Yeah, absolutely. And really enjoy the podcast that you guys have. you. Appreciate that. It's a It's needed in this space. Thanks. That It's always good to feel Yeah. Your hypothesis Twice we've got some good feedback. is validated. So Yes. If you like the podcast, perfect timing, go like and subscribe, give us some feedback, give us five stars if you're listening. Um [snorts] we'd appreciate it. John, tell us tell us about yourself. How'd you get into to Energy Tech? Where you from? Kind of how'd you how'd you get to Red Hat? Well, I haven't gone very far. So excited cuz we're here at the near the Canon. And so our office is We have a Intelligent Edge Solution Center here at the Canon can join with Chevron. And um But I went to high school right down the road here, Springwoods. So doesn't feel like I went very where anywhere [laughter] at all. Um but you know, I you know, you grew up in Houston, you do energy, right? Yeah. Or medical, maybe a little bit, but maybe a little material science if you're if you're cool. But um but uh you know, I've been um working in um energy since I was 19 in some ways. Um if I go back to my Stewart Title days when I was a kid, I used I was a file clerk. First time I actually coded anything was at Title, where we were using VisiCalc and I was going to say, what'd you code in? Yeah, VisiCalc DBase were the first things where I got paid to code. Yeah. But the cartridges on the Atari as a kid on the 400 was the first thing I coded. That's awesome. Like a basic cartridge and um It's like I mean I don't It was a How do you even do that? I mean I like I have no clue. Like what what did that even look like to code against to code cartridge reader Yeah, so the the actual computer had a cartridge insert for like games, but then they started building like basic Fortran, Okay. Pascal cartridges um you could plug in there and you'd write to your little floppy or actually cartridge uh cassette too, right? And uh you had to get all the magazines that There was like a couple different Atari magazines back there with code. Uh one of the projects I first did was to build a weather uh device that you could hook up to the Atari computer. So a buddy of mine This is like in junior high. And uh we built a thing to pull down weather data and present it, and we thought we'd do this for like fishing guides. The problem is you need to buy an Atari computer, so it [laughter] was a bad plan, so So but um but that was our big idea when we were in seventh grade was That's awesome. we could pull down the weather data over the air. I mean that's pretty innovative, I assume for I mean running it on an Atari so giving you a time frame of when that was me That's how I learned to solder too. So the first one I soldered a little device you plug into the serial com port. That's my uh man, that's honestly my favorite thing about about this podcast is just ever There There isn't a s a common path for pretty much anyone that we've had on here, right? Like everyone got exposed to tech in their own different way. And a lot of it is under older through direct tech, right? It's like, well, I was interested in MySpace, so I started coding MySpace pages, and then I got into code and learned HTML, and then it waterfalls or like my story is at the same time I was working at a data company, I was also in a fantasy football league with all my friends back home, and I wanted bragging rights, so I started those problems too. so I started, you know, messing around with the API. At one point I was commissioner of like eight leagues. [clears throat] [laughter] So I'm down to two. Yeah, how do you factor in uh injuries? that's destroying me right now. Yeah. [laughter] Yeah, so you had Chubb is what I'm hearing? Is that? My wall, my wife had him, but um. Oh, there were a bunch this last week. Yeah, there was basically every receiver I had [laughter] something like I dropped a Cooper Cup and yeah. Yeah, Cup was hard off the bat. You didn't do your homework, man. Yeah, I really didn't. Yeah, to to his point like I showed up like guess I'm just flying blind. to FF today, downloaded the top 100 whatever so. Yeah, for for a data person I have not yet done what he did so. I I really like the football guys one if you want to use that one. Yeah, they have good good data. Okay. They've curated their data. Right. Right now. fantasy football is tricky because the draft stuff is all very it's not there's pretty much no data behind it other than just the player's historic data, but that player might be on a completely different team with a different offense and a different head coach and a different you know. where data science gaps live. And so it's it's events are hard to represent with data science properly. Um speaking of that, shout out to Amazon. I watched their uh I don't even remember what the the name of the broadcast is, but it's their secondary broadcast that they stream on Prime with the uh their computer vision and analytics running on top of it. It's pretty slick. It highlights all the players, it tells you their speed, shows you the routes they run, it predicts blitzes and stuff like that. It's pretty slick. Let's wait till their contracts are tied to the analytics review. So I guess all you joking it out there, you just got docked pay. No, I I can't wait till that if I'm hoping please make that API available because it would make my analytics really interesting. Um anyway, coming back. So kind of how did you how did you get to to where you're at at Red Hat today? What? Yeah, so you know I I I started Houston Java User Group here a long time ago. Well, I'll go way back. I I was How PC guy for old school Houston guys, Houston Area League PC Users. I was a Delphi user for a long time. Coded probably probably coded more applications in Delphi, Borland Delphi platform. Um group to SQL Windows, Power Builder, all that kind of stuff. And then when Java came out, I became a huge Java nerd. Um and started the H jug here in town, which is still going and being being run by my buddy Jim. He's doing a good job with it. Um but I haven't really coded Java in quite some time. We have Quarkus at Red Hat, which is a great platform for optimizing Java on containers. And but overall, what we've been doing is mostly Python, Julia, R type of content when it comes to coding and now a little bit of Rust. Uh you know, Rust and Web Assembly is kind of like the hot newness, right? The new thing. And um you know, in running on microcontrollers is something we've been looking at. Um there's a a buddy of mine here at Red Hat over in Germany who's built a project called Drogue IoT. And it's an an attempt to, you know, containerize workloads on the microcontrollers. So That's awesome. insane. Yeah. The the stuff that that's coming down the pipe on the edge side and PLC industrial automation stuff is really really Yep. exciting in my opinion. Yeah, Nubix IO is a leader there too. So Jason Shepherd, formerly of Dell, is running that thing and they've got a good a good head start on this too. Um if we could just get the software companies to containerize all of their stuff, we'd be doing better. there's what's the Malcolm Gladwell quote? The future is here, it's not you know, widely distributed. [laughter] Right? Evenly distributed. So sure. Um but yeah, that's the case for a lot of things. There's you know, the container stuff's been around a while. If we go back to like Sun, right? And process jails and all that stuff. Or even LPARs on mainframes. The idea has been there for a long time. Right. But how you break up the resources of a computer containers is now the state of the art for how to do that. Um I can get the best utilization of my hardware by containerizing that stuff. So this is a perfect opportunity for you to explain to people kind of the old way with virtual machines and kind of how Well, even that there's a step between of like virtual machines, Docker, and then where you all got to. I mean Yeah. so Start like high level. I don't know how deep to go here. We can start high level and then we'll we'll prod you for questions to get into it. So so I used VMware like 99, right? To to to [clears throat] like Phoenix BIOS simulation, how they kind of started. And I was running like a KIX region on a an image, right? Great use case, loved VMware back then. But when you build VMware at scale or you know create all these images, it's a copy of the OS in every image. Yeah. So it's using the resources of an OS full stack every time you spin it up. Containers is more like an OO thinking here where I can reuse the base layers and get better utilization cuz I'm not every image is really just the deltas of what you need in that container. Significantly more efficient from a resources perspective, right? And so you know being able to break up the process and isolate it process process process isolation is the core of it. But like what Docker did which is amazing was make an API to make it really easy to use, right? So the registry and the the layers and stuff like that. Um we've come up with a um some improvements on that called Podman. So podman.io is the website. And that comes with Red Hat Enterprise Linux and comes with Red Hat Device Edge, but that's now the most secure way. There's some vulnerabilities in Docker which we can ad alleviate. And uh a lot of the other Linux distributions have picked up Podman and and they're shipping that now. Cool. Um you know Docker kind of with Swarm kind of made a misstep back in the day a little bit and it broke everyone's Kubernetes and that kind of forced this change. Yeah. For like so OCI format became kind of like a way to keep the keep everyone's nose clean. Right. And uh make sure we don't hurt each other's projects and efforts. But um yeah Podman's a way to run containers now. Uh there's a utility to simplify it for new developers called Podman Desktop which is you know active development, heavy development right now. And uh but yeah Podman ecosystem and our you know container registries out there can pull from Docker Hub, pull from Quay. And you can get software up and running very quickly with these with these things. Yeah. And then you know, of course, we use OpenShift to scale it out and do container orchestration. And that's mostly what I've been doing for for since I joined Red Hat was positioning OpenShift at, you know, large super majors here in town and in globally as well. Yeah. So also tell Okay, so now that we kind of understand virtual machines and containers, tell people about what the next step is, which is OpenShift with Kubernetes. Yep. Um just high level. It doesn't have to be super We talked a little about Kubernetes in our one of our second or third episode, but like you talk about high level what that is. I mean the orchestration of the containers and Yeah, so like like I said, I I was at OSCON in 2015 when I learned really what Kubernetes was. Mhm. And Kubernetes early days was, you know, Google's way to like run their run all their systems more efficiently, right? And DevOps it and you know, get get the best utilization of the hardware. That's a critical thing for a search company at that time. Especially a cloud compute company. [laughter] Yeah, cloud compute company. We want want to utilize our resources at the most economic way we can. So getting from VMs to containers in some cases like a 15x 30x better utilization of the same gear. So I could run more with less, right? So OpenShift is a way to do that. We really we had OpenShift existed before Kubernetes. Okay. So OpenShift the way it is with the version three of OpenShift was the first Kube version. Okay. And so, you know, going back, Google and Red Hat got together to really invest heavily in the CNCF and and build the Kube project. It's I mean it's one of the largest open source projects in the world. Okay. Um it's like it's usually in the top three if you look at commits and all that and in the involvement. But Kubernetes has taken over, right? Kubernetes is is it's the established way to architect things and Especially cloud. Cloud Yeah, all the you know, all the cloud guys have their K8s. Yep. Uh OpenShift runs on all the clouds and clouds you haven't heard of too. Yeah. And now it's slowly evolving or being pushed out to edge, right? Like that's kind of the the future. I mean the first my first involvement with Edge was with a company that was pitching modular servers that ran Kubernetes. Yep. But to for the IT manager, the CIO, CTO guys out there listening, what are the benefits? Like why do I need to move from even just containers to Kubernetes or tell me some of the, you know, the the day-to-day benefits of of what that brings to me. Yeah, so like I always tell people and a lot of people were, you know, with the K3s project, that's where I think a lot of people started trying to do this on the Edge or at least doing a constrained device, right? And I know your old company was doing stuff like that and uh but the uh you know, part of the the the goals here is to simplify operations, get operational efficiency down, and improve security. That's really the the two things. And some of this if if it's a like a far Edge device, to reduce the windshield time. Right. Right. So I want to be able to patch successfully. A lot of these OT environments they're, you know, locked down and they, you know, like set it, forget it. Mhm. But there's vulnerabilities associated with that if you look back what's happened at like Saudi Aramco with some of their workloads and there's been a lot of hacks here lately. The idea that the software is not maintained better is what we're trying to fight against, right? We want to be able to you you want to leverage all these new things, you have to have a way to keep it like a living breathing thing. Right. And so, you know, like Edge Ops as a as a pattern taking the, you know, the DevOps idea and be able to push your container to where you need it. Um you need sometimes you need to be able to patch the OS without having to go on site Right. or a guy running around the USB stick, [laughter] right? So we're trying to help that guy not have to run around. So uh part of how we deploy Red Hat Device Edge, which is a way is an alternative way of deploying RHEL, you get the best zero day CVE response and like we're a root on CVE. So we're if something's happening, we're going to be the first responder, right? In that sense. Um so um but what we can do is allow you to patch it automatically. And we have an immutable OS now with this using the RPM-OSTree project. And using GreenBoot, we can create uh a sequence of how you boot the system that'll auto patch and then push out new versions of the image. And and it's transactional. So if it's not safe, it rolls back. So Big big issue with edge that I'm glad y'all are solving out of the gate. A lot of people focus on, you know, the on-site hardware and how's it going to integrate and what's it going to run and what software and blah blah blah. But then one of the last things that people think about is your CI/CD pipeline for these edge devices. And if they're far edge on a drilling rig or on a frac site or in a SCADA system where you might not always have connectivity, what happens when you go to push an update and it Yeah. connectivity fails in the middle of the update? Does it brick your server or does it intelligently know that it needs to roll back because the update didn't come through? That's a That's awesome. I didn't realize it had that level of We've been working on this for a little while. This came out of the Fedora IoT project. So shout out to Peter Robinson who's been working nights and weekends and you know trying to get this thing out the door. Um you know we're on Intel and ARM with this stuff now too. And so we're you know working a lot on ARM chipsets. We're also working on the new RISC-V coming out stuff. So we're going to try to have ubiquity you know Okay. Switzerland across all these different chipsets with this stuff. But the bulk of our workloads are still Intel with this. And so doing things like TPM 2 to commission the box and know that the that box that boots up is our box. Right. And doing that commissioning step is part of this. Um be able to do all encryption at rest and you know do the data assistance. Also incredibly important for edge. network bound edge you know disk encryption is part of this using the Lux project, Clevis and Tang all combined. Um and so we've got this pattern for how to do this right and make this thing a lot more like a a Linux appliance, right? Right. Versus an OS that an IT guy has to administrate. Yeah. Um so you can come build these appliances. Mhm. You know so immutable OS with safe transactional safe patching, and then also with to do a secure bomb and build it like know everything that's running on that box and make sure it's your stuff. If it's signed containers or signed content, you can validate that and if it doesn't, you do the check against the TPM. If it fails, kick it out. Right. So you can commission and decommission automatically without having someone have to babysit it. And so That's awesome. That's a That's a big change in how we can do a, you know, production edge deployment for critical infrastructure Yeah. and not, you know, [clears throat] everyone's doing monitoring type use cases because they're scared. Yeah. But we now have a way to do like closed loop processing in a with this. I think that's We were just having a our last guest, Jeff, we were just having a conversation about, you know, oil field generally speaking goes in kind of evolutionary steps, right? And so it's, you know, it went from writing things down on a notepad or putting them a physical notepad or putting them in a spreadsheet to now we're monitoring things with IoT devices and we're streaming the data. And then the next evolution of that is true closed loop automation where you deploy ML models on the edge and they're You've got federated learning where they're learning locally and pushing the the the learnings up to the mother model in the cloud and that model gets accelerated, you know, the learnings of it get accelerated because of that and all this fun stuff. And so it's really nice to see like we were talking, you know, it's very easy to kind of see how you get to that point, but there are all these little nuanced things that you know, we have to figure out and we have to test and recognize and have they they You got to do the hard things first, right? You can't If you realize you're going to do all this work and you won't pass the digital security team's blessing, you waste a lot of time and money. For sure. So you have to solve the you know, you know, you know, one of my customers has this process for how they check all their You know, we're not going to go impact things that make us money unless we know it's safe, right? So, risk first thinking. And so, that's what we've been aligned to is make sure you do it right versus do it fast. Yeah, I think I saw a a meme or video on Twitter the other day and it was like cyber security, you know, guy watching all the other IT people do stuff and it was like a guy on a worksite and everyone like doing like stupid [ __ ] on the [laughter] on the worksite and he's like just like looking around like what the hell's going on, but I mean I mean and it I get it because like I'm more developer first so like I just want to make sure it work and do it, but then there's a whole other side of it like you have to do it securely and have to do it right like There there's different ilities that an OT guy thinks about than an IT guy thinks about. And the IT guy will like and then they fight, right? These guys are you know, coming from very different angles and some and I'll say this, there's a lot of OT guys that are brilliant, right? But they don't have the same background as an IT guy and they can come in and do things quicker so management goes well that's better, it's faster, it's more agile. Well, the risk associated with doing it agile sometimes is not identified unless you're salty enough to know what the risk is. And so It's funny, it gives me more appreciation for my OT guy cuz his stuff is locked down and he will not budge like no, you're not punching a hole in my network for this like [laughter] But the things they want to do this, they just they have different priorities and so the IT OT conversion everyone talks about is really about educating each other and enabling each other. And so they OT guys can do amazing things. It's just that their tools and their controls are so locked down and so seen like the big brother stuff on that stuff it's a difficult environment to work in. And so the idea here by using open interoperable process controls that XM Mobile's been very public about with the OPAFF and all that uh we're working on helping them, right? Trying to get into a place where you can treat you can treat your OT assets like IT assets Sure. and be able to do it like and getting the human out of the loop which it helps with the de-risking of the deployment. Yeah. So, uh that's that's the bigger story we're working on. I mean I feel like it's maybe a good kicking off point just to maybe even step back and talk about Red Hat. Like, and because I feel like maybe I'm seeing a theme for the different Red Hat that kind of branches, but like just, you know, where Red Hat started as like with the whole Linux distribution, how it compares, you know, on the operating system side, but it seems like I mean, there's a certain mission that you guys serve, you know, that's pretty consistent across the board of like providing like a really robust and, you know, enterprise enterprise grade. Yep. Yeah, we're we're you know, Red Hat has a huge presence in DoD, public sector, federal type places. Like it's actually a mandate over there to use open source, but you do it properly, securely, right? And so, you know, we have great you know, I used to do DoD at Oracle. And you know, that was fun time and all that, but um once uh Red Hat really started sneaking in, it it made sense to, you know, join the winning team. Sure. So, um but overall the you know, Red Hat started over you know, it's never it's almost 30 years old company, right? And um you know, it started as uh you know, the guy named Bob Young who was the founder. And basically, you know, this was a building a Linux distribution off of, you know, Torvalds' work. Yeah. And so, um you know, fast forward a few years, Red Hat became the enterprise thing and really we just started turning off Solaris everywhere. Okay. So, turning off you know, the Unix wars were were were dying and we cleaned up. Sure. And this is how Red Hat really grew. And um and really became a trusted partner for a lot of uh of uh you know, people that built all these applications on Unix, they needed a place to land Yeah. with the Unix distros dying out. So, I look at the clouds today is kind of like the new Unix wars in some ways. Yeah. And proprietary lock in strategy, particularly around the data. And um but overall it's like it's more about the design of what you're building. So, my fight with you know, when I come into a shop, they'll be like, oh, we're mono cloud or an Azure shop, right? Yeah. So, but low and behold, two years later they're they're not mono cloud, right? Cuz they're they don't have they lose the ability to leverage and negotiate if you build to a single cloud, right? And the the the data strategy with some of these clouds is predatory, particularly if you need to get out of there. Yeah. I was going to say it's it's not on the ingestion side as much as it is on the Yeah, the egress you you whoever you're negotiating guys are, you know, Equinix can help solve some problems here, Yeah. but still, depending which cloud you're with, the your egress cost is Yeah, look how cheap S3 is. Oh yeah, just just load it in, right? Yeah. So it's you know, the snowball is a one-way trip, right? The snowmobile, you know, that's a different And when they're bringing an 18-wheeler to your facility to move data, watch out. Yeah. Uh but um No, I mean that's that's We've had this conversation around APIs and stuff and it's like people get pissed off when there's not an API or they have to pay for the API. Like we were talking about it because X, Twitter, Yeah. you know, started locking their stuff down, but it's one of those things where it's like it costs money every time you hit the damn thing. So yes, it's nice to have and it's almost come to be an expectation, but you know, you also have to be responsible with it and I can't tell you how many people we've I've talked to, we've had on the show where it like, yeah, we had an API and then we would have someone that just had like a loop running all of the time, hitting it every however many seconds, and it's just costing us money every time. So we have to put limitations on it and things like that. And it's it's the same thing with a lot of the cloud stuff, right? Yeah, you you lose control over your cost. Yeah. You're to behold you're you're beholden to someone else ways they want to run your business. Now data ownership is one of those things it's the data ownership is a big part of the business here. But yeah, being able to understand how those products are managed on the cloud and you depend upon it, they can make it's like, well, we're going to turn this one off and we're going to go to this one. Good luck migrating, right? [snorts] Um and so but overall I think it's understanding you want part of what OpenShift provides is the ability to run a private cloud anywhere and it could be on the public cloud or the private cloud. And it can address sovereignty concerns. It allows you to choose the data gravity you want, right? And but the idea is I can uh I have I can design things for like it becomes my asset. Right. So if this yeah. Where we really struggle and trying to get people to understand is what what is cloud native mean? Right. Cloud native is more like 12 factors, right? It's it's how you build things, not what services you turn on that someone else owns. Yeah. [laughter] Right. So so your designs should override using those you know leverage those services but building abstraction layers and mediation layers to protect yourself. Yeah. So if your asset is your asset it can deploy anywhere asset build it so for an ISV you limit your customer reach by building something that only runs in certain regions. Yep. Yeah, what we talked about with like um Star here I mean to be able to work in Yep. like say Saudi or some of these other countries like you it has to be able to run Yep. on their soil and like on their servers like and they're not like particularly anything seismic related that's never going to cloud. Yeah. The the first of all the economics don't work for storage. Sure. It's also you you know you pay more to rent things generally. So over long term 20 cloud should be leveraged for on demand things to burst for temporary things. Yeah. Long term stuff Especially when you know what the demand is. Yeah, you need predictability in this. People are you know using tools like cloud ability and stuff like that to try and do forecasting but it's hard to dot all the eyes and cross all the tees when you have changing demands. Not always. and I think one of the things that I mean I saw this first hand when I was selling edge is a lot especially in the energy space, you know a lot of the primarily Azure but it's all I mean it's all the cloud that's their strategy across the board is they'll come in and they'll give these major companies huge amounts of credits Mhm. to get them locked into the platform, right? And so it's like well hey you know the edge allows you to push a lot more of that load down. off of you. Right, but why would I need edge if I am not going to touch you know if I'm only 10% through my credits it's like well are you going to be like that in a year or two like how it's only a matter of time to when you start realizing that there's a cost associated to this and the more you build on it, the higher that cost is going to get. [snorts] Yeah, the the the stickiness of these things is you know the credits is definitely one of their sales strategies. And we deal with that a lot, but you have to have good leadership inside to realize you have to have longer term thinking cuz there is no free lunch with this stuff. and that's I mean uh that's what that's why all the big tech companies exist because they think in the the infinite game, not in the short-term game. Um Well it's also because of so much job jumping. Yep. It's not my problem when I get out of here. Yeah. [laughter] Right? No, yeah. So ultimately the top leadership needs to be able to I mean like and The accountability willing to pay for Red Hat or not. Yeah. Well it it I'm not cuz I mean Linux is free, right? Linux is totally free. Everyone should use free Linux. So but um but But again the whole no such thing as free lunch. I mean like you mentioned earlier to us I mean but you maybe want to talk about where you can like actually save money on the audit side you know like stuff like that. I mean Yeah like uh yeah that was um you know we we we work in utilities as well and we've come out with ways to solve some tough problems around NERC CIP compliance, right? So you can build your applications for a utilities company. Now most utilities companies buy their software. Sure. There's a few who don't and that's a growing trend and we're looking at the projects in the LF Energy and there's some great things around ADMS and DER uh EV charging station projects in there. Um you know but you know there's a lot of pressure to modernize grids Sure. with um um you know bespoke solutions like we're working with GE for example. They're trying to modernize 40 years of tech debt, right? So um but um you know there's all kinds of strategies for these ISVs to leverage these things properly to build a better business for them, right? And having that going back to that cloud-native thing, that needs to be keep their eyes on the prize with that versus something I can turn on very quickly and get value out of but then ultimately it's not the best thing for my business long So it's you have to do the work sometimes, but really you just have to focus on enablement and keep your focus understand how to measure the the outcomes you care about against those those metrics. And so um you know a lot of firms out there are looking for quick wins and quick quick you know quick consumable projects. But ultimately I think it's at a foundational level you have to have good architecture uh guys that understand the cost of their choices. And a lot of guys are [clears throat] it it's hard. Yeah. Like it it it is hard particularly with the pressures a lot of guys are under quarterly or annually. That's what I mean like but people won't bat an eye to spend money on Microsoft or Windows. I mean like so I mean it maybe kind of make some comparisons and contrast between like say Windows, Windows Server and Red Hat and then you've got Linux. I mean I mean could be as simple as someone like that they usually do stuff on Ubuntu or you know their favorite open source Linux. I mean like is it pretty easy to move over to Red Hat? Yeah yeah. So like moving workloads on Red Hat it's you know it's not a big deal. Like actually like if you're a .NET developer .NET runs better on Red Hat. Okay. Like the .NET Core stuff? Yeah .NET Core stuff. But you know going you know the windowing stuff ceiling that doesn't run there. SQL Server runs better on Red Hat. Like by far. Like it's a and it 's weird. It's like you know and Microsoft's cool with this cuz really they care more more about Azure than their Windows revenue these days. Sure. And so Microsoft and Red Hat it's like cats and dogs sleeping together now. It's it you know we we truly are partners. Azure um you know runs Ansible or runs OpenShift. So ARO is Azure Red Hat OpenShift. Um and Ansible and RHEL all run there just fine. And so um but yeah we're we collaborate more now. And it's good for the industry. It's good it's good if we're not all siloed and trying to fight over everything. Collaboration I feel like regardless of what it is generally always benefits the masses. And I hope the oil field continues to get progressive in that. Oil and gas companies, in particular guys who write software in this in this industry, are trying to work better together. Yeah. It's hard. No, I mean a lot of it is, you know, not to demean anybody that is working on this, but it's almost like the bottom-up approach where it's like we're forcing it. We're trying to force it into the companies, the industries and stuff and at some point once you get up the ladder, someone is, you know, putting a a stop to it or throwing up, you know, whatever X, Y and Z risks and stuff like that, but it's one of those things that it's like Colin used used this analogy the other day. It's like when probably 10 15 years ago when we really started getting into the swing of things with the unconventional shale fracks and stuff, how much better would we have been had we shared all of the data and the learnings from all of that stuff? How many reservoirs would we not have completely destroyed the value of because we were putting all these wells too tightly, over fracking them and causing all this communication issue? Like communication [clears throat] is still a problem and we've been doing it we've been talking about it for 6 7 8 years now. Well, and we've said before too like you have really smart people all solving the same problem at different places. Yeah. There's there's there's there's better reducing your engineering costs by sharing the right thing, particularly if you're going to you know there's negative impact from a like a water flood right? And or I'm you know all this horizontal stuff is all crazy, but um but yeah, it's overall we could collaborate in a different way and get better operational efficiencies for everybody and safety and you know there's other things about how you you know socialize some of the equipment and the the best practices. I always you know I this this is a little I'm going to do something here. So fracking came out of an open source mentality, right? Mitchell's started opening up and talking to people, right? Like we wouldn't be in this this this whole thing without a different attitude towards and and open. Yeah. Yeah. So openness and meritocracy wins. Experimentation, right? The ability to just say, hey, let's try something out. bring your competitors in and talking and working out together, the hive mind wins, right? McDonald do really well. Yeah. Yeah. Rising tide lifts all boats, right? Like that's So that mentality like like some of these consortia things are an attempt to do that, but it's still too much monetization first, right? Versus getting it right. loves consortia. [laughter] Yeah. We're we need like a consortium rant counter on on episodes cuz say you know, I participate in these things and they're slow moving is just cuz we're you know, we're hampered with the legal groups mostly in these companies is the main problem. So we're See it's always legal or IT. Sometimes procurement, but going slowing down a lot cuz you have too many engineers in the room too. Yeah. Well Yeah. We also are very good at overcomplicating, overdesigning and also is our skill set too. Also playing who's the smartest guy in the room just to say that I'm the smartest guy in the room. But Alpha geeks are are are [laughter] a problem. even worse when you combine a alpha geek and an alpha engineer into one person. But it's no, I mean it's a it's a
it's crazy to me that the IT the tech industry has all of the has had all these consortia that have spun out so much great things that literally impact everyone's day or life on a daily basis, Bluetooth, Wi-Fi, USB, like those three right out the gate, right? Like everything it is better because of those things. Yep. We don't really have that equivalent in oil and gas. say is not I was going to say in like the EDR is really the first and only thing, but even then we bring that up and it's just like is it just Stockholm syndrome? Is that the one thing we can point out like that it's not even people don't even like it that much. It's just like [laughter] Okay. So It but it's actually a standard that we can use. once they get off of soap, it stops sucking as much. So WebSocket, you know, ETP is a good thing. But yet why call it ETP? Why not just call it web socket? Yeah. Like it's like we have to new verbs and make it complicated. Or we have to get And that's the other thing too. Like we've Bobby and I have had this conversation a number of times is it's like who cares what the data format the final data format is as long as the data is structured in this Like the data structure is way more important. Yeah. You know, what who cares if it's XML, JSON, like whatever text. It doesn't matter because the data science The person on the end of it can figure out how to map it. It just needs to be a structured, repeatable get the same thing every time. I can do something with it. Right. it's getting like so like Okay, so they're here We're getting to some like right now like OPC UA versus Sparkplug B, right? I don't know if y'all are following that kind of stuff. OPC UA has been around forever, right? It's got over engineering things in it. But it's what a lot of these devices use, right? And the OPA is tied to you know, OPC UA formats and the DM the new DM thing is an attempt to clean this up. But yeah, it's still it's it's how do you do stuff in flight and not break things. Yep. Yeah. And so additive engineering processes where I can make changes safely that's what you need to That's what going back to cloud native I want to be able to add new services without impacting current things that we get it right and then be able to roll it out cleanly. That's what this whole edge ops things can do. But the Sparkplug B guys like they're born midstream right now. So like Cirrus Link Arlen you know, if you know these guys uh have done a great job using that format and it's now donated to the Eclipse IoT. And so I think it it's a much more simple, much more clean architecture. And so it's easier to Like you don't have the cognitive load of OPC Foundation all of its concerns. And Sparkplug B is a you know, a a canonical way of getting the data shape that everyone can agree to and can do like the UNS stuff. Right. So you know, I think we're we're looking We're looking to support both. Well, we are Through partners we support both today. And we have our own MQTT that no one uses, but that's Uh, but, you know, most people use Mosquito or a version of Mosquito for that component. But, um, yeah, getting the mediation layer cleaned up and making it easy to do integration. You know, like I would always say it's like a million dollars per integration, you know, endpoint, right? We can get that down and we've got a partner we're we're working with called Litmus IO now who does a good job. We have Ma Con Data Broker supported on Red Hat now, so they they can run as a bare metal OPC UA broker. They have a containerized UBI container running on us and now they've built an OpenShift operator too. So I have this full spectrum of how I want to run my OPC UA process with Red Hat. Uh, so we've been, you know, targeting all these new OT vendors that, you know, if you're going to containerize stuff, we're going to be there to help you. Yeah, sure. If you want to modernize and, you know, clean clean up your tech debt here and make it easier to consume and spin up, make it all declarative deployments and automated deployments, you know, we'll be there to support you. Yeah. I think one other big thing in this space talking to all you operators out there, Bobby, you included, is nothing will change until the operators require it, right? And so having operators, and this is finally starting to happen, right? Like out of the gate it was, okay, well, we're going to need you to be able to stream some on the ops side at least. We're going to need you to stream your data, right? Like that's slowly been working its way into contracts and procurement and stuff like that. But the next step is, okay, not only do we need the data, but we need you to talk to each other and the data needs to communicate with the other dozen service providers that are also on the location who are providing us data as well. And by the way, reminder, this is still all our data. I know you generate it, but it is our data and therefore it needs to be as open and you know, interactive across different platforms, companies, data structures and stuff as we can get it. And I think that is one of the main reasons that a lot of our consortiums don't end up that well is because there are plenty of operators in it, but there's also a lot of service companies in it that have their own agendas and, you know, everyone has to make money. And so there's all these conflicting kind of uh incentives almost that never end up working out very well. Well, it's part of this going back to realization all software companies all companies are software companies, right? yeah. And no one in gas still thinks they're a drill bit company in some cases. Uh drill bits are very important and love drill bits, but [snorts] ultimately it there's not a unique differentiation there. Uh organizations that now I'll say this and not trying to get in trouble, but if you're if you how much you spend on SAP I think determines how much you think of software. So [laughter] um so That's a great That's true. so if you don't value software you don't invest properly. Yeah. You know, I hear I you know all these organizations that run on saying they're agilists now they fundamentally don't get it still. It's like they're not really building things for them. They're still leveraging other people's stuff, so you're beholden to them. Yep. But that's the market, right? And it's the reality of it. Um some organizations are differentiating. There's some people are owning their own assets and they they're gaining the they're going to be leaner, meaner, tougher. They're going to be able to deal with when it goes back down to 40 or under. Yeah. Well, I mean it's just one of those things too like put yourself in the client's shoes. If you're an operator and you can choose option A which doesn't have an integration with whatever on-site vendor that you're using or option B which does have an integration with on-site vendor and they're the same price or close to it why wouldn't you go with the one that has the integration just to make your life easier, right? Like Yeah. the the just the mindset behind a lot of the the reasons why people aren't you know collaborating and opening APIs up and stuff are just kind of archaic and old in all honesty. Yeah. Um and as as this starts to become the expectation in the industry then more and more people will do it. They're going to be forced to do it. Yeah. And so we're we're seeing it. It's it's slow as with most innovation in the oil field, but it is it is happening. Speaking of kind of partnerships and stuff, I know y'all had y'all had a couple press releases within the last week or so. Mhm. Um I wanted to hit on the the Intel ECI and then also the the Guys one. I'm really interested in that one because my background obviously, but Yep. um wanted to ask you about this. Yeah, so this was Intel Innovation was this week, earlier this week, and we had a few I don't know when this will be released, but yeah. a couple weeks to release. just, you know, whatever today is, September or something. But yeah, Intel Innovation, we had a couple things aligned there. You know, Intel's a big partner of us. The spacer in the Canon was started by us and Intel together um to support the OSD community, right? And but we launched Guys AI is a partner of ours. They're being accelerated by Intel program. And they've designed an Edge Ops platform that allows you to deploy to Red Hat Device Edge a containerized workload without having to have real skills there. So it's a easy button, a GUI as it were. It's what it's like a low code for is it explicit Basically, I can commission I can set up locations and set add devices to it and then commission deploy workloads to it without opening up a CLI at all. And are they they're kind of framing that focused on some of these edge use cases around computer vision and their background is they were like computer vision shop. And so they would build a custom CD image for you. So they have in our if you go to Red Hat catalog for containers, they've got like an agent gender thing and they've got apparel logo thing. Yeah. They do car logos. So there's some basic things just to show off their capabilities. But then they can come in and build like a flare monitoring solution or safety PPE thing or whatever, right? And so they were working they were deploying on pies, Raspberry Pies and using Balena OS, and they couldn't get any the security stuff properly. So now they're using Ansible and deploying to RedHat Device Edge. They're using Quay as their container registry. But they've built this nice tool on top of it. That just handles all that. all the security out for you so you don't have to be you don't have to be an expert there. So it really lowers the bar to using RedHat infrastructure. But you know, you could be like a Windows guy or VMware skilled guy that just likes to click on GUIs and you can do that now with our with this this partner platform. great. So it simplifies you know edge deployments at scale. No, I mean I think that's that's a big I've talk I talk about a lot when we talk about edge but you know the ideas are there but there's all these kind of nuances that apply to edge especially especially in oil and gas where you've got you know limited connectivity, you've got ruggedized weather connectivity, you've got people on site, you've got different vendors, different data sources, all this stuff. Um but the ability to manage and deploy to these disparate locations and stuff like we talked about earlier is it's one of the last things that people like really think about, right? It's like oh well I've got it stood up and you know in our lab on this test thing or over at the Chevron you know in the Canon at the Chevron lab and it's running here but okay well how do I manage that when I have a thousand of them out in the field and they all have various levels of connectivity and different networks and all this stuff. Um so it's it's really cool that that you guys have that and I'm personally most of the applications we were looking at when I was at Hivecell um in the oil and gas space, a good number of them were really around computer vision whether like you mentioned flaring, I mean that's a huge one, right? Being able to Yeah. um some but we guess like some drivers on location like when we talk about stuff like water haulers like you pick up like their sticker to see what company it was and then see what time they were in and out and you know like your service contract compilations and your I've been saying this for damn near a decade now but like your entire frac, all the logistics and stuff around your materials, your chemicals, your sand, all the stuff that is constantly moving on a frac site that is still very archaically done through mob tickets and physical pieces of paper. It's like, hey, the trailer leaves the sand mine, it gets weighed on a scale, it's got a, you know, uh QR code on the door and a RFID tag and when it leaves the site, it now knows that this vehicle has left the site and it weighs this much, has this much sand in it. It's going to this other location. If you've got a camera or a reader on that site, you now know when it arrives, you know how much sand is there. You can then force that directly into whatever database and then draw off of it. Like you could automate your entire logistics for all of that stuff just with some basic computer vision stuff. Yeah. The the the use cases for CV is amazing. And and predictive maintenance to end the telemetry off of your pumps and your compressors. Even like some HS and E stuff too. Yeah. No, I mean people on like that's another big one, right? It's like if you know, you're on a frac job or a drilling site and there is a catastrophic event and the well blows out or there's an explosion or there's a fire or whatever, there's still not a great way other than get to the muster station and go through the list of people on the JSA to know how many people were like what happens if the whole thing blows up and no one's out there to notify people, to know who was there or who was be on fire people don't know. So a simple camera and these things are get the cost of this has really come down. Yeah, it's it's it's getting very reasonable to that scale. hands. I mean like [laughter] Yeah. Well, I mean these and there's like in with Lidar and all these emission you know, things IR things and um Well, and and especially now where we talk about we got a new ESG specialist I mean the IRA stuff I mean so if you show how you're mitigating and then also you know, I'm sure it's it's not there yet, you know The rules are coming. acceptable but like the rules are coming but then what Oh, sorry. Sorry about that. Um rules are coming. Yeah. But I mean they're going to run the road but like what they're going to allow you to do like cuz right now people still have to go out and do an inspection but I think this coming soon where like they're going to allow technology to be count as an inspection. got customers doing it already. Some people were proactive in this space around the emissions stuff. Yep. So, you know, you know, like I said, we're we're you know, we're we're doing projects in LF climate where we're like we build stuff more for FIS customers to like score their you know, ETF portfolios for carbon risk, right? That's part of the you know, Goldman Chase kind of guys. But that same data we've built a you know, this OS climate project which can be used by energy companies to share information and data about their ESG standpoints. So, you know, these things are scanning all the 10 K's looking for carbon problems and building portfolios off of that. Right. Uh we've got Airbus now contributing their models and their you know, ESG stances into the into it. And so yeah, that's a project where we would love to get more oil and gas guys like the open footprint project line there a little bit. But but yeah, we've there's there's this data science stuff for computer vision. There's so there's so much value to be had that's low hanging fruit right now. It's the problem is scaling it out and managing it without you know Yeah, they don't even no one knows the problem because they haven't gotten to that point yet, right? Like and that's some folks are doing it, but you know, they run into breaking the device problem that could be patched or they don't know how to do the disconnected stuff so lights out type operations. So, how to do repo syncs and stuff like that we can help with. You bring up data science. I mean like why should someone as a data scientist or data analytics what should I care about Red Hats? It's just an operating system, right? I mean not even like you know so what do you guys offer that is important in this space? we've got a few things here. So, we we've been running like on the Kubernetes. I mean you can run you don't need Kubernetes to run data science. So, the inferencing on a Red Hat device edge is a perfect secure platform for doing it. Now we can pack all those immutable things we talked about. Yep. But we've also put in certain packages working with Intel for their like OpenVINO is in there. One API support's in there. So, I can use the GP GPU to accelerate things. And we do work with NVIDIA too. So we've got a GPU operator that runs on OpenShift. So all the NVIDIA goodness work works too. Um and we're working [clears throat] on like Orne IGX boxes. Gotcha. So for, you know, more edge style the things for, you know, I need a GPU in the field. And you said you all also work on ARM as well as Intel, right? So Yeah. We're we're we're what we love Intel. We love NVIDIA as well. And uh you know, we love TI. We love Qualcomm. We're, you know, we're big happy family here. So but yeah, we want to be able to like enable Red Hat on all these devices. Agnostic. And like I said, we're working in the RISC-V community as well. So and that's a that's an interesting one. I think that will um not challenge ARM but force ARM to clean up some things. Yeah. So we're targeting, you know, SystemReady Okay. type stuff uh with our with our work right now. It's all in the upstream but we've got it booting. Like I've got all the got the Advantech box working. Okay. Yeah. So and and but um Any case, there's there's um you know, basically we need to be able to map the whole breadth of the compute for edge. So edge is, you know, these Atom class Celeron box or core boxes in a lot of cases. But the ARM stuff is more for these low power use cases. And so um you know, being able to, you know, we're never going to we're we're not going to do pies but you know, hardened pies would maybe cuz Fedora 38 runs on pie. Yeah. But we not going to support it. Um but we're going to be able to do these um you know, things that are more ruggedized, famous boxes. And build to like just like these phones are converging all types of apps, the IT the idea is we can converge a bunch of apps through container technology. Right. And so I don't need to I don't I don't need to add another box for every use case. Yeah. So Yeah. put enough beefiness out there on an edge edge node and run multiple workloads. Yeah. So I wanted to ask uh before we get to the kind of speed round tell tell everybody a little bit more about what we what you what you are doing with the the lab over here and the Chevron guys and that whole setup. Because I don't one of the reasons DW exists is because the energy industry doesn't talk enough about the cool stuff that they're doing. And I had the opportunity to go with you on a tour the couple weeks ago and I was shocked at how much cool stuff you'll had and just the setup in general that it and you're much more knowledgeable about it. So I'll let you speak on it but just talk about kind of you know who it's designed for, what they need to do to be able to come in and test and do all that stuff. Well, you can google to find out we have a a landing page on our website. You just search Intelligent Edge Solution Center. Okay. And you can find me that way. But I'm Archer at Red Hat if you want to find out more about it. Um but overall so credit to Chevron. They invested here way before we did, right? They started building their maker space there. And so they're there testing 5G stuff out, Starlink, you know type things. But then all these sensors that was going to say they have that whole sensor which I had no idea, right? LoRaWAN and Wi-Fi sensors for all kinds of upstream and downstream use cases. And so um you know Chris Mallett and Paul Wakowski over there run that space. Uh I work with a lady named Janette Martinez. Um so they're willing to bring I'll let anyone come in there and see what they're working on. Okay. And so you want to set up a tour, you know ping me and I can get Janette will help us. And um you can see all the fun toys over there but they're built an AR VR XR lab over there doing they have a robotics lab. going to say they've got the uh aquatics center. They're adding that building that used to do the Van Gogh thing is uh now becoming aquatics lab for them. And uh but yeah Chevron's investing to help them innovate maker. Uh our connection with them is through the OSTA U. And so we're OSTA U members. Um I'm the focal point in the OSTA U. And my focus has been OSTA U Edge. And so and this is really focused on production APIs which are now finally starting to show up in the OSTA U community. You know it's been focused on the seismic and well log. Yeah, finally. And uh there's there's a bunch of contributors working on it. And it's exciting because so now you know getting into the production domain is really now how we can have a a wider a a bigger impact for us to do For sure. for people that are trying to do like remote valve control Yep. Sure. understand, you know, tying this to trading might start showing up. Yeah. So we can actually start impacting the business cuz everything's been accounting, ESG, right? Like production, you know, all the the business side of production is now being API driven, which is a a a you know, for folks using Aries or you know, these older production Avocet type type stuff now we're finally starting to modernize that ESG community. So it's a very it we've been kind of waiting for it. We did have a remote valve thing that Intel and Exxon built that we had running there. So Modbus control, you know, to do valve control but um and then tied that to inductive automation SCADA. Um so that's in there running there today. Yeah. Um but we're bringing all these other you know, OT workloads that are you know, they're mostly energy focused, but it's you know, retail and some other verticals we're we're addressing too. So but um yeah, it's anyone's welcome to come in, do the tour. We're there to set up to do innovation. We can do workshops, do POCs, pilots. Um we have hardware available for use. So we have hardware from OnLogic, Advantech, and Intel. So if you're devices and some of the Novo gear too. So if you're developing software for edge or for oil field, this is the perfect playground for you to come and experiment, show it off, you know, learn uh get direct feedback from an major oil and gas operator. Um like I said, it's to me it's just it's one of those really one of those things just like a lot of stuff that we do. It's really cool. It's really fascinating. It's opens a lot of opportunity, but hardly ever gets talked about. So I wanted to bring it up just because our doors have been open for a little bit and we're still you know, trying to figure out how to leverage it the right way. And um we've done quite a few workshops with customers. We've had a few open events through Consortia. And uh but yeah, we're we're if y'all want to come check it out and, you know, work with us, we we've got a we've got a stocked candy store over here. So. Yeah, no, it's Well, I mean, it's one of the biggest things, right? When you're with a lot of early energy startups is it's like, hey, I've got this idea because I came from whatever background and I know there's this problem in this space, but I don't know how to I don't have a place to go test my sensor or this software in a pseudo production environment or in a lab environment even. Just being able to have a spot to go test it because it's like you okay, well you developed this software that runs it's this machine learning model. Well, you need inputs from sensors to run the model, but you don't have any sensors and startup and you have no money, you can't go buy them. Yeah. And so we're Simulators only go so far. Exactly. Yeah. So it's yeah, no, I I love that we have these things and we're getting more and more of them. And I I like I said, I just want to give as much exposure to that as possible because it's a Yeah. great resource. Yeah, I love to see y'all guys out there more playing around. So it's Yeah, the I think the community needs like, you know, there's you know, the Maker Space Chevrons built, you know, all these, you know, like Axon's got its Digital Garage and, you know, BP's got something like this, but they're the first ones doing it in kind of an open community way. Right. So that openness again is critical to to knowledge share and to have the information spread more quickly. Yeah. So. But yeah, kudos to Chevron for doing this and and we built actually we built a what I call visual inspection demo with the It's really just detecting if the beer label's applied properly. So so they sponsored us for that spot and we used that at our user conference as a way to show off a way to kill Mettler Toledo devices. I don't know, but in any case, but it's I really just want to get all the equipment to build my own brewery. That's what I'm trying [laughter] to do. So. You're halfway there. You got the bottling stuff, right? So so. But you know, but but I've got to line us out on all the vendors now. Yeah. But it's a it's a good space for like you know, the MTC at Microsoft has a good space too for doing community events, but the I the Canon does as well. Yeah. So, if folks should be aware of the Canon. And Canon has done a great job. Uh they're spread out across the city now pretty well. Yeah. And so really like working with those guys, too. That's cool. We got 5 minutes left. So, we're going to jump into the kind of speed round for you. Okay. Um any last minute things we didn't touch on you think people need to know about? do you I didn't think I really gave you the proper So, I talked about guys I don't really talk about the Intel OPA thing or the So, the Intel ECI announcement that came out um you know I wrote a blog like 3 years ago about hey we're going to start working on this. And you know the you know we were learning about OPA and you know Exxon will say they've been working on it for 13 years. Uh but they you know the OPA started like what 5 6 years ago. And um but really what we were trying to do is start doing open test bed. So, over here we're going to build a test bed and allow people to come in and check it out. And so Exxon's been doing this with Yokogawa for a few years. Um and um and Yokogawa's doing a great job and has built a lot of uh great IP for them. But we want to try to open it up. Right. And so um and figure out how we can go faster and make a better product. Right. And so um so that's something what we have announced here. So, the Intel ECI framework is an is a reference architecture to build an OPA OPA style test bed. Okay. So, we've been working with Schneider and working with Yokogawa. And um you know we're we're we're open to working with anyone who wants to you know come play with us. Cool. But you know trying to build open interoperable process controls is the stated goal here. Help you know do best of breed on a converged device. Build a you know have Rockwell, Honeywell, Schneider, Emerson. just make it agnostic. Yeah. It's basically it's it process control has been kind of proprietary locked in silos. Right. There's a lot of value in opening up things, right? Not like the mediation layer is one thing, but the ability to just you know like I said earlier, treat the OT assets like IT assets. Plug and play. That's the goal. Yeah. And make it easy. Make this standards. Like there's a big I remember I used to go to the the SLC meetings. I don't know if you remember those things. Like it was back before like Energistics and Mimosa and all these guys. And the push was to make energy more like Southwest Airlines, right? Like one jet, one motor, you know, Yeah, easy to work on. Like know how to do the maintenance Yeah, and you can fix it like the It's the same idea. You get much more operational efficiency to your business and better margins Right. if you have a predictable known path and you don't increase the cognitive load, you reduce the cognitive load. So and you have better documented processes and Yeah. just it's more predictable. Sure. So and that's ultimately that's that's what we're trying to get to with the OT side. So more downstream use cases in refineries and chemical plants. But the idea is you allow chemical engineers to experiment. They can do data science versus just hit a button. Okay. So being able to do safe experimentation in this environment is the idea. I could put my Python code on the actuators as close as I can Yeah. and impact the business. Like improve the yield or reduce the cost of the process. Um there's a very exciting thing that just came out. Uh Microsoft has been working on protein folding models. So it's kind of like a stable diffusion idea. Uh-huh. Well, that can be used for catalyst uh chemical creation as well. So who's doing that in I guess, right? So I'm excited about the protein folding because of health reasons, but seeing it applied in different ways to do material science on you know, what have been well established DuPont OEE cycles for years uh and to you know, start improving what's been a very kind of you know, you know, mature process to look for new ways of doing things and create new materials. Yeah. So so that kind of stuff is you know, I want to on that. Yeah, that's really awesome. Yeah, so it's called Evo-Def in case y'all so it's open source model for Microsoft. So Yeah. props to Microsoft here. So Yeah, I mean they've done a great job of making themselves yeah, they were the anti-open source and then like all of a sudden like overnight like became a better Red Hatter. That's what the way I look at it. yeah. I mean like hell they bought GitHub and I mean they were even when they bought it I think they were like the largest contributor on GitHub. I mean like Well yeah, a lot of people were like why are they buying GitHub? This doesn't make sense. And then I remember I don't you might have told me that but I remember hearing that and being like oh well okay. Yeah. Yeah, I I I you know when I was Java guy I remember the days of gray APIs, right? Where they had their own APIs they didn't document and they published APIs that were different slower to hurt their competition. Right? I'd like to think they aren't going to do that again but but overall I think Microsoft's become a great advocate for open source. Yeah. Um and uh really if you're if we're all just focused on the customer and the outcomes they care about that's it all gets solved. Yeah, 100% So and um you know there's all kinds of new things coming out with all this generative stuff. Yeah. Uh you know Red Hat one thing I did want to mention you know we've got generative tools in Ansible now and OpenShift. So we have something called Ansible Lightspeed which will generate playbooks for you. That's really cool. So you can you know you don't have to you can reduce the cost of your Ansible dev with this tool. And it it's built off the you know we have a huge community of Ansible playbooks called Galaxy. Okay. So you anything you want to automate there's LAN party whole setups in in in there and but there's like I want to configure all the Wi-Fi in my or my SD-WAN. Right. Or I want to deploy you know my whole you know suite of applications. Yeah, that's really cool. So all that's been trained on that those Galaxy playbooks and it's all but what we're doing though it's explainable AI. Right. Mhm. showing the sources Very important. So the ethics is is a lot easier to to disseminate Yeah. if you're showing the source. And that's one big problem with all this LLM stuff. big time. People are kind of hiding well what's really bad right now is all the the data usage policy policies being updated behind the scenes. Yeah. A little sketchiness there. But you clicked accept. Yeah, you did. You told you totally just gave us everything. definitely read all of the terms of of service. t's and c's, yeah. and conditions. Yeah. [laughter] Well, thank God we have all the European commissioners reading reading it, not us, so. Um yeah, so I mean I guess we're probably going to jump into that having probably that was probably the open source [clears throat] stuff was probably a good segue into one of the speed around questions. So what's your favorite open source library or repo? Well, for energy stuff [clears throat] the Equinor has done a great little seismic utility. But I wrote a little blog on it, but it's um Now for oil and gas stuff there's yeah like Welly, you know, use that and that was a good one, but the the StatOil um um seismic one I really like. Yeah, that was really good. What about non-oil and gas just in general? Oh good Lord. Um there well there's so much. I know. So I mean I've been playing a lot with um uh Pi-hole. So I don't know if y'all familiar with Pi-hole. So it's a little utility for running out of uh like um a proxy at your house. And so you and you do the at home stuff Home Assistant. Sure. So that stuff's fun to play with and uh and I I do play with Blina too, Home Assistant to that, so it's okay. So and uh but that that's become a little mixture of commercial versus open source. Um but I do a lot of weather stuff too, so I have like three weather stations in my backyard. So I'm like yeah, my wife always jokes she should have been a weatherman or something like that, but um You can never be wrong if you're a weatherman. It doesn't it doesn't matter. Or a reservoir engineer. Yeah, and last night I was playing with uh um um a thing to export out the AirTags off your phone and share it cuz that's all locked down to an individual. I have a little project that does that. That's pretty cool. So this Edge Loaner program we have has you know AirTags in them. Okay. And there's commercial versions of this, but the AirTags are a lot cheaper. Yeah. So [laughter] so I can share share out my share out the AirTag data with that project. Um but it I really I'm geeking out about EvoDiv right now. So the Evodiv thing I mentioned about the you know stable diffusion for you know protein folding. So so it's a um you know I'm dealing with a a a a rare non-Hodgkin lymphoma. So it's a little self-preservation Sure. No, that's So but but like you talk about like when you do things that matter to you or like you know like it's where your um you know your passion can actually drive you know I I'm very interested in how Johnson & Johnson runs their GPU operators. Yeah. So so and if they're using Evodiv yet or have It's still not peer reviewed but [clears throat] uh the the idea of it is you don't have to sequence the amino acids the same way. Um but you really it's like the goal is the shape of protein so it binds to this protein. Okay. And uh so I'm on a I'm a M. D. Anderson lab rat right now Yeah. where they're trying out a they're targeting three proteins on my I have mantle cell lymphoma it's called. And uh those proteins are being bound by this this new drug. Okay. And this is all based upon the COVID mRNA mechanisms Hm. for how drug discovery can be done quickly. But this is a big little leap little leap frog step in the computing you have to do to come up with these drugs. That's very smart. So drug discovery stuff's kind of cool to me. Yeah, it's really neat. Uh I think the people like to kind of [ __ ] on AI a lot for you know it's just like the language models right it's just regurgitate it's it's a search engine right it's just regurgitating what it it found statistically to be the best answer or whatever but where you really see some incredible AI use cases especially in the edge space is around like the medical field right like one of the the uh examples I used at I saw was uh you know they did a study I don't I think it was John Hopkins or one of the big uh universities did a study of right well at what percentage rate could doctors by themselves diagnose breast cancer? Yeah. What percentage rate could an AI computer vision diagnose it? And they were basically comparable. Um I I want to say the doctors were slightly higher than the AI was. But then they did a third test where it was AI or doctor assisted AI. So both of them in tandem and it went from like 80 ish percent I believe Yep. up to over 90 percent which yeah it's literally tens of millions of people a year get early diagnosis or it gets caught or it's treatable versus state you know full on. Yep. And so it's it's I think again everybody gets caught up on the automation side but the first part of the automation step is this human controlled or human assisted AI. Assisted AIs are are I think everyone saw was so scared is this going to replace my job kind of thing. And really it's another tool and Well if it does you're not very good at your job. better like so my mom was a radiologist. Like she used to you know run you know very very manual slow process. But yeah we've run that we've run that model right and there's a Stanford one for skin cancer too. But the problem is the FDA is blocking it at being used right. So it's like we have these tools but the business models Mhm. and the legal stuff is preventing a lot of the usage Yeah. to impact better outcomes for patients and so I don't want to get into [laughter] all that mess so. I'm kidding. Um yeah what is your um favorite book or most interesting book you've read recently or and uh favorite kind of follow on on LinkedIn, Twitter, social media. yeah. Um so favorite book all time I want to go Snow Crash Neal Stevenson. But Cryptonomicon too. I'm a big Neal Stevenson fan. So I haven't like the fall was the last I haven't read his very last book but the are y'all familiar with Neal? No. No. Okay. You need to read Neal Stevenson. Okay. The whole idea of like VR it came out of Snow Crash right. So like virtual worlds all that stuff and um and um but Cryptonomicon is like three stories it's in there's some reality in it. So, it's like, the Enigma engine from World War II being solved. Interesting. And you know, there's Turing, Alan Turing's in the book. Okay. But it's it's also crypto. So, and then he wrote this one series and needed a freaking editor. It was like each book was like 1,000 words or 1,000 pages and the the Baroque Cycle. So, if you got through the Baroque Cycle, you're you're a real Neil Stephenson fan. Yeah. [laughter] But it was really talking about the beginning of banking and currency and stuff like that. So, Neil Stephenson is my favorite author. But I'm reading right now is Team Topologies is really You've probably heard me say cognitive load 500 times here. Probably because I'm reading this book. But it's really talking about how you design teams to get the right balance on cognitive load Okay. and how you design for flow, product flow should go. And so, you know that's interesting. I bet that's fascinating, actually. It's great. We have a We have a group at Red Hat. We have this thing called our Innovation Labs. And they've built a whole library of open practice library it's called. So, for folks that know about the open practice library, it's really the mechanisms for how an effective team should run, how you build software. And Team Topologies is a book recommended by that group. So, shout out to Matt Tanaka for that. So, Cool. So, but favorite LinkedIn follow. Or Twitter, just in general, whether it's energy, tech, or math. I Well, I'm I follow Wade Phillips, you know. So, [laughter] son of Bum, right? So, so Yeah, actually his his sister was my social studies teacher. Oh, yeah? Yeah. So, like I remember like at the time he was coaching the Bills and Doug Flutie was there. So, she had her box of Flutie Flakes and all I get. Oh, yeah? Yeah, I ran it So, I went to a wedding in Youngstown, Ohio just a few weeks ago right when the Hall of Fame game was. Okay, yeah. And Wade Phillips was right on my plane. That's funny. So, I got to sit there and talk to him and his wife for a while. It was really awesome cuz I've you know, I'm Earl Campbell, you know. I'm old. So, but the but you know, Wade Phillips is he was there for he had I think three or players he had coached Okay. were going in. They like you know, where from Cowboys and Mhm. a couple of came out other ones, but like I know he's coaching the Gamblers. So I don't know if y'all remember Houston Gamblers when like when Jim Kelly yeah. Yeah, Jim Kelly Gambler days, you know, and so but Yeah, Wade Phillips is, you know, he's coached everywhere. Every type of every type of environment. Yeah. I was so happy I was happy even though it was Peyton Manning, but when Broncos won with him and so But yeah, follow Wade Phillips and here's a bunch of geeky stuff I follow, but I'd rather talk about Wade Phillips. Fair enough. For sure. One more or Uh What's what's quickly what is something what's the most exciting thing you're seeing or you're excited about in the text energy tech edge space? It doesn't have to be edge specifically, but energy tech or edge. Well, so that's there's a few things that are interesting. I'm I'm getting more onto the utilities grid side. You know, like nothing against oil and gas stuff like that, but our grids are old. Yeah. They're they need like we can't do the renewables without improving the grids. There's a project in Aleph Energy called Everest, Mhm. which is an EV charging station that's fully open source. that I'm trying to get us to pull in. There's one called C Path over there. Couple like ADMS DR stuff and But yeah, I think investment in the grid in It's really a big focus on transmission that like We need a 3X improvement in this. Yeah. We need a 200X improvement in CCUS. Like we're just like I'm you know, thinking about my kids a little bit, right? And particularly after this summer, like kind of like Yeah. This is floating on the edge there of Yeah, we we are we are getting dangerously warmish here. Yeah. And you know, I know it's it'll be in the fact that it snowed five times in the last what? Yeah. Snowing in Houston is also not good. Yeah. Right. So, like I I've had to rip out all my, you know, plants on my Like I I don't know what where to go. Do I buy for hot? Do I buy for cold? No, yeah. [laughter] I mean, yeah yeah, they they've been run by both. I mean, like Like, you know, Moon Valley is taking all my money. So, it's like it's like, you know, trying to keep, you know, green things green and it's it's harder and also a rock garden. Yeah. So, yeah, just I'm just going to start looking like, you know, [snorts] Phoenix here. So. Yeah. Desert plants are the We had a desert sage bush at my last house and I could I if I wanted to try to kill it, I couldn't. And I'm convinced Anytime I I'm going to have to replace stuff at our house now in Katy, it's going to be with desert sage because it's super robust and it's really When it rains, it blooms really pretty purple flowers and when it doesn't, you can cut it back and it doesn't need a ton of water and it it's just a nice healthy thing. But it's such a huge pain in the ass because you're right. It's like, well, my grass is dead now because it hasn't had any water in 6 months. My grass was dead in the winter because it snowed. Like [laughter] Yeah. It's been it's the climate change stuff is, you know, like like we we participate in the COP, right? With that whole thing and there's a guy named Bill Ryder at Red Hat who's trying to drive like climate change projects at scale like through the UN. So, he's doing like he's flew over to Tanzania to like help Like they don't have an alerting system for when it floods over there. And so, like very basic things out in Africa we take for granted they don't have. And so, um so, we're trying to like do stuff with like bee health. So, there's an organization called Enterprise Neurosystem that Red Hat's in Okay. where we're trying to do federated learning globally in a lot of its weather stuff. You know, so IBM had all these weather projects. yeah. They donated a lot of that There's They made a big in donation to Hugging Face with these weather models. And so, but we're now doing acoustics. Acoustics on the bees. And we're trying to do stuff with like oyster beds and clam beds. Those are early indicators of these climate things. Okay. So, we have to monitor that more effectively to know at the macro level what we're doing, right? And some of that's like, you know, pollution too to waterways and stuff like that. So, you know, I like to fish and you know, CCA could get more involved here if they want to. So, like Texas A&M Corpus Christi and those kind of those schools um um so we've been one thing I've been working on uh we have something called Red Hat Research where we like teach graduate students how to like how to code, how to use open source. Yeah. And uh like we did Baylor students over in Bruno Czechia a few years ago. So, trying to expand that to have oil and gas companies like sponsor students and use that for like uh and then just for talent through Red Hat Research. And so we have open source projects. You do your your senior thesis Right. Yeah. with us. That's cool. And uh you know, you build you learn how to build things. We'll teach you how to code and stuff like that and how to use Git and how to use you know, Tekton containers and all that stuff, but Uh-huh. you want to be a driver guy, be a driver guy, but um you know, doing this with more targeted stuff like in these type of projects uh you you can get a lot of data We do a lot of data science out of Boston University and Boston College. um and there's other like we did some schools in Sweden now, but we haven't really got it going here in Texas yet. So, um OSU has a university outreach program we're trying to light up. There's an open source version of OSU in the GitLab now. Okay. So, it's not tied to a commercial offering and I think it's a way to get the university program going on these production APIs. Yeah. No, that'd be great. So, uh so looking for all the university guys listening here to ping me about that. So. Well, that's a perfect segue. How to Where can they find you? How do they ping you? Yeah. So, archer at Red Hat's my email and uh you can come in Canon anytime and and get me. um I am Twitter, but I don't you know, that's Twitter and LinkedIn or Twitter's archerj and LinkedIn is /archer. And then uh I assume Red you'll have a edge section on the website as well. You know, redhat.com has a um um you know, an edge you know, just search for Red Hat device edge. Search that Intelligent Solution Center. Intelligent
blog too? Um I Or no? I do, but it's on Medium and it's it's old. So but So do I. I haven't updated it in years, but three and I drank COVID and I stopped. Every now and then I get a occasionally just [clears throat] post, you know, piehole stuff or Welly stuff on there and It was amazing though like when like you haven't done I don't know anything for like two or three years and like someone's like Yeah. just reaches out cuz like someone else helped her or hey, can you help me do this? Yeah, you get another follow or a like or comment on it and you're like, man, I haven't thought about that in literally years and there are still people finding that. Photo Matt used to come to the H Chuck meetings. The guy that started WordPress, right? He used to live here. Okay. Yeah, I knew he was from so you know, Matt, you know, he's Bay Area now, but you know, so I you know, the WordPress tools are great, but I you know, we I I blog through Red Hat stuff sometimes, but you know, when I do my own stuff, it's Medium. Yeah. Got it. but we we Red Hat has a podcast too I wanted to mention too. So you know, one called Code Comments. It's a kind of like a partner driven thing. You know, we used to have openshift.tv which it's now Red Hat something something something. So we do Twitch live streaming Oh, nice. for like developer content. It's really smart actually. Yeah, and it's on YouTube too, but but using Twitch as a platform for like we'll do live coding things on there. So if you go to developer.redhat.com, you can spin up an OpenShift cluster um and you know, try it out. And we have a bunch of like tutorials and stuff. So if you if you're a developer want to get hands on, developer.redhat.com is a good place to start. Okay. Yeah. Or learn.openshift.com is another good site too. Yeah. So but um Just shoot you an email. Yeah, yeah. Or shoot me an email and I'll send you [laughter] whatever you want. So So John, I appreciate it, man. And this has been great. Yeah. Look forward to doing this again sometime. Yeah. Well, I look forward to seeing you guys at Fuse. Yeah. And I'm I'm sure that'll be a great event. We had some other some of our colleagues that went to it last year's one. So I had one quick question. Is it indoors all or is it outdoor and indoors? It's mostly indoors. We have a outdoor exhibition area where people have equipment. I think we actually have a we're supposed to be having some kind of drone tech come. So I think they'll be letting people Yeah, we haven't been able to bring many drones to remember going to OTC and you could do helicopter flights back in the day. It's been a while, so. Well, cool, man. I appreciate it and we'll talk soon. Thanks everybody. All right, thank y'all. Thanks for having me.