Show logo
Explore all episodes

Pie In The Sky

  |  Compiler Team  
Artificial intelligence
Open source

Compiler • • Pie In The Sky | Compiler

Pie In The Sky | Compiler

About the episode

There is a lot of excitement around AI models, but can it meet the expectations set by blockbuster movies? What’s the current inflection point between what’s feasible and what’s not?

The Compiler team talks to both big dreamers and heavy adopters wading into the space, hearing their thoughts on how AI can help scale daunting work, fill in the gaps, and make the fantastic into reality. 

Compiler team Red Hat original show

Subscribe

Subscribe here:

Listen on Apple Podcasts Listen on Spotify Subscribe via RSS Feed

Transcript

Whether it's the idea of a robot doing laundry, or it's the prospect of a fully functional AI assistant that caters to every whim and answers every question. Artificial intelligence is something that has tickled the curiosity of the tech minded for decades, and for them, it's a pretty exciting time. Innovation in the AI space is moving fast. Professionals, companies, and even those who don't follow tech trends that closely. They are all ready to set sail in this new ocean of possibility. But what's the current inflection point between what's feasible and what's not? Which capabilities are real and which are just wishful thinking? This is Compiler an original podcast from Red Hat. I'm Kim Huang. I'm Angela Andrews. And I'm Johan Philippine. We're exploring the aspirations around artificial intelligence. This episode, you're going to hear stories ranging from the fantastic to the more practical. Let's get started. I mean, honestly, it started out with cosplay. That's Kris Kersey. He's a technologist, maker and an avid fan of cosplay, a hobby where people dress up as characters from their favorite pieces of fiction. DragonCon here in the Atlanta area is my home convention, which is kind of the Mecca of cosplay. And that's where it began, as is really wanting to kind of take cosplay to the next level. Kris is in the middle of a very interesting project using AI. You see, when he goes to these conventions and he sees cosplayers, he notices the costuming isn't true to the fiction, and that's understandable. But it's not just that they're also cumbersome. I see a lot of problems when people wear these suits. Again, they are just costumes. Most of them are rigid plastic, but it's not a very comfortable experience. And in fact, you'll notice most people that do the Iron Man cosplay do have the motorized helmets. And that's because walking around with the face plate down is very impractical. You have little slits for eyes, and then you take those small slits and you put lights in front of them, which makes it very difficult to see out of them. I saw that there was kind of a need there. What if we could take some of the tech from the movies, which is, you know, developing a heads up display inside the helmet where you're not trying to see out of the helmet? Now you're aided in that with cameras and displays. So there is that practical aspect of it. Okay. I'm just going to do a check on learning here. Johan, Angela, do you know who Iron Man is? Do you know who Iron Man is? I know who Iron Man is! Thank you. Very much. Johan, we're going to get sued. Stop it. I'm all about Marvel. I'm all about Iron Man. So, yes. Once he said that about the helmet, I was like, that's right. How do you effectively cosplay Iron Man with the actual eyes in the helmet. This is an interesting problem. Yeah. So, yes. Current company being the exception, just in case you don't know, Iron Man is a well known comic book superhero. He's an inventor. He's a technologist in his own right who makes powerful suits to fight off everything from intergalactic invasions to supernatural villains and the like. What I think is fascinating here and Kris talks about it a little bit, is that, you know, the creators of the Iron Man comic never imagined people would be walking around in Iron Man suits, and certainly now there's like Worbla, there's LED lights to put in front of your eyes to get that kind of glowy, like, little ice lit thing that Kris was talking about. And those are like the practical effects, right? But they're not comfortable and they're not true to the tech, right? Oh, no, I mean, those are... they look pretty kind of rigid, plasticky. You can kind of see people not being able to move around very much. You know, there's not a lot of flexibility. Whereas you see Iron Man in the movies and he's just like he's doing backflips. He's like flying through the air. He's doing all sorts of stuff. And that... I don't think we're quite there yet. But it's, you know, I've seen some of these costumes, not up, not up close, but in pictures and in movies and stuff. And they're looking pretty. They're looking pretty sophisticated these days. What's important here, I feel like, is, you know, the creators of the comic never imagined that the tech that they were writing, that they were drawing would be something that would happen in their lifetime. Nothing in modern day would reflect something that a Tony Stark would wear right? But what's interesting here is how that technology is becoming more and more possible in modern day. It was the idea that a lot of the tech that we particularly that we saw in the early, Iron Man, the early Marvel movies that were out there, a lot of that tech was becoming practical. A lot of it was feasible. That was things that we could buy now or, you know, that at least the building blocks were there for us to actually make something that was more real. I love that aspect of taking the vision and making it real. That's what got me excited was what if, what if this is more than cosplay? What if this is something that we could start building some new tech out of? Think about it. AI assistants like Jarvis from Iron Man. Well, we have Alexa and Siri. Helmets and visors with a heads up display? We have virtual headsets that do the same thing. Wrist communicators. We have Apple Watches. These are basically kimoyo beads as far as I'm concerned. You know the beads on the bracelets they wore in Black Panther? Yes, yes, yes. Oh, gosh. You know, it's interesting because it's all becoming more feasible and less science fiction and more daily life. I love that. I love it. Bring it on. I would love some of that tech in my life. I'm trying to have Jarvis schedule my day out and plan my, plan my vacations. Kim, remind me a little bit. It's been a long time since I've seen this Iron Man movies. What is it that Jarvis is able to do for Tony Stark? Literally everything. Jarvis is integrated into not just the Iron Man suit, but also into Stark Tower. Where Tony Stark works, it's integrated into his vehicles. He drives his car, he is a wrist smartphone. Yeah, exactly. Like wrist bracelets. So you can talk to Jarvis and ask him questions about everything from, you know, the, the temperature outside to, you know, his, his, his daily, you know, his dinner with Pepper Potts, his romantic interest. Literally everything it does, like, it's a fully functional AI that, you know, is sapient, sentient and can tell him everything. And this, this is what Kris is trying to build with, with the suit? He's trying to build a suit that is functional and also have a little bit of Jarvis in there? Well, not to the extent of Jarvis where he's telling you, you know, he's doing quantum physics in your, in your years. But, definitely along the lines of a, a fully functional Iron Man suit that has an embedded or integrated AI system. So with that in mind, Kris started the O.A.S.I.S. Project. O.A.S.I.S. stands for Open Armor Systems Integrated Suite. Don't be scared by the armor part. It's just there's no weapons involved. It's just a reference to the suit. The project has started out with just the helmet and the heads up display. It's powered by cameras. So instead of having those little slits with the LED lights in there where you can barely see or breathe. You have a heads up display that shows footage that's being shown in real time, that's being taken by a camera that's outside of the helmet. Right. Also, you have a voice command. So you have, an AI that you can ask kind of like basic questions like what time of the day it is or the temperature outside, things like that. Mhm. Obviously it's not Jarvis level yet, but what Kris has done is still pretty impressive. The dream of what we realized in the, you know, in the films. It... I don't think that that was even something that we were considering at the time, you know, that, that something like that would be possible in, particularly in the, in the short term. And it's been absolutely amazing to see how fast this technology has appeared and then how rapidly it's grown. From, you know, really those early days of, you know, ChatGPT, OpenAI being the only game in town at the very beginning, or at least the only game most people knew about. So what's next for Kris? First he has started taking his work on tour. So he's taking the helmet to different events, so that others can see with their own eyes how science fiction and cosplay can become real. My favorite thing is to go to cons and to take off the helmet, to show it to people, even to let people to try it on themselves, recording their reactions. And, and it... that's just it's extremely rewarding to be able to see the difference between experiencing something and just seeing it in a video. It's been a lot of fun. Wow. Yeah. I, I love it. I as someone who loves science fiction and superheroes and certain fictional characters that have just a place in my heart. Tony Stark, AKA Iron Man is one of them to see, to allow people to try on the helmet because we know the character, we kind of understand what's going on inside. We've seen flashes of it in film that puts you right there. You know who doesn't, who doesn't want to do that? That sounds impressive. And the fact that he's taking his show on the road is it's probably going to inspire other cosplayers and, and technical people who maybe didn't think that this was possible, but it's opening their eyes. Yeah. What are the implications of the use of, like, artificial intelligence to make science fiction a reality? Right. That's like there's like a whole other podcast right there. That's like a whole other show. Don't steal my ideas. Don't do it. And I love how just fun this idea. Right. It's purely a use of AI for entertainment purposes. And it's also like pushing the boundaries of what tech and AI can do. But it's, you know, you're doing it so that you can go to a con and, you know, show off your suit and just have a lot of fun with people. Right? It's... I think it's just fantastic. It is. But I think Kris has done here is phenomenal. It's taking a real life problem. I know it's kind of a first world problem, but still, and addressing it using technology and also, you know, maybe even getting people who weren't even interested in AI to be more invested in it and to kind of explore what's possible. It's the art of the possible. Yeah. Okay. We've given you a tech story around fiction. After the break, we have two stories around the more practical and maybe even essential implementations of AI technology. Stay with us. So far we've only discussed aspirations around AI in terms of pop culture and glitzy movies. There's more? Yes, there's more. As a matter of fact, Johan and I spoke with someone whose use case is a lot more down to earth. Hi. My name is Ty McDuffie. I'm the founder of Griott and Grits, which is a, a project to record oral Black family histories. Now, I'm familiar with the term Griot, but for those of you who are not... The Griot is the historian in the West African Village, and they're charged with, you know, carrying the histories of the villages, not just their own villages, but surrounding villages and sharing them orally. It would be just any stories of importance about the village. It would also be things like family lines, family histories, different, maybe even battles that happened, in the villages. So anything that needed to be preserved, would be what the griot would share. Let's hear from Ty. What inspired him to start Griot and Grits. For my mother's last birthday, we did, this thing that I thought was really neat. We, all the sons, danced with her, and each one would go in and tap, and then another one would come. And I'm gonna tell you, there was not a dry eye in the house when we did it. But what was really interesting, and I noticed after the fact that everybody except the sons were recording, right, on their phones and things, and as I thought about it, I'm like, man, everybody has a totally separate video. Each person has a different perspective of an overall story. How neat would it be if those things could be combined somehow, metaphorically, combined to give a complete picture of what happened from all these different perspectives. And that's how the project was born. At first, Ty started with just recording audio stories, but since then the scope has expanded. What we want to create is a platform by which people can use something, maybe even as simple as their cell phone, to just record the history. And then we do all the work from that point, and we just deliver back a fully finished, fleshed out Ken Burns documentary. So, all you have to really be able to do is turn on your cell phone and hit record, and that's pretty much it. Okay, so we have Ty here. He, you know, he had this great experience with his mom, and realized how, Black history is like, and family histories and Black communities was, you know, kind of going kind of unrecorded, undocumented. I can attest that is a problem. And he decided to make this very kind of manual process to address it. So I'm guessing AI is going to come in at this point, right? It is. So one of Ty's friends is a friend of the podcast, Sherard Griffin . He's a senior director of software engineering at Red Hat. He heard about what Ty was doing, and his response kind of blew the lid off of things. He said, imagine, you know, if we were to... if you would have put on a virtual reality headset and as the story's being told. AI could actually, like, scrub the web in for, public access information and things like that and bring in, let's say you say your grandfather fought in the, Vietnam War. And as you're talking about that, AI would bring in newspaper articles so you could look to the left and see a newspaper article or sound effects of war, or just pictures and things to supplement the story. We could make this a completely immersive experience for people. This sounds a lot like The Fifth Element. When Leloo was learning about humanity and society, and it's what popped into my head, because it was all these things that what whatever she was watching thought humanity was. And in the end, we found out it was quite terrifying. But this is an interesting spin on it where you're just saying things and images and sounds are being presented to you to kind of fully flesh out and make the story more real, more, more liveable, more, more audible, more visual. That's a pretty good application. Good thinking Sherard. What do you think, Johan? I want to see the finished product. Right? I mean, this sounds amazing. This sounds like you'd be able to basically make your own museum experience out of your family history. And, that just. I don't know how... I'd be absolutely moved to see something like that. Yeah, I would as well. And, Griot and Grits is a project you can find on GitHub by the way. We're going to check in with Ty and Sherard again this season. But for now, the most salient point is this: what Ty is describing could be a great use case for AI because it's work that's hard to scale. A large language model could be, you know, a game changer and all of that. It's interesting because I spoke with another person who is looking at cultural impact of AI in a different way. I'm Oindrilla Chatterjee, I am leading AI and digitalization at Adage Automation. Oindrilla was a guest on Compiler before. Remember our episode on mentorship? I do. A lot has changed for her recently. She's working on a new field. It's been, it's been very exciting to be honest. Especially because it's a whole new domain, which is industrial automation. It's a lot of learning about new technologies, about chemistry, about manufacturing, and how to put my background knowledge into use, in this whole new manufacturing space. I wanted to speak to Oindrilla because I want to know what she was experiencing in her work now as a knowledgeable data scientist venturing into spaces outside of the enterprise. We are seeing a big shift from using just edge and OT solutions to more integrated solutions, more IT sort of solutions, which means integrating machine learning solutions. So edge, kind of get that. IT definitely know what that is. OT have either of you heard of it? Yes. New to me. Okay. OT stands for Operational Technology. Angela, do you want to kind of explain what it is in your own words? I can take it if you don't want it. You can take it. I've heard of it. I've actually I've heard the phrase recently, so, yes. Tell us what you know. Okay. Well, the easiest way to think about it is when juxtaposed with IT. IT stands for information technology. Right. So data and applications. Right. Well operational technology is the hardware and software used in industrial processes. So manufacturing, power plants, petrochemical plants, you know, energy, those types of environments. That's more operational technology. Got it. So Oindrilla is excited about the use of AI around the tasks involved with operational automation, predictive modeling to anticipate heavy energy demands, emissions monitoring for more efficient usage of resources, and the handling of manual tasks around typical repetitive processes. For example, AI machine learning, even the new advancements in generative AI are very useful in this sector because they can be... these techniques can be used to either automate the processes, make these processes more efficient, efficient in terms of early detection of failures to even, making the human processes, the operations of humans and the factory side of things more efficient. And then there's tooling like generative AI, natural language processing that also has started to have a huge impact in this sector, especially because these are now helping speed up a lot of manual tasks. I feel like this is a common thread when people talk about AI, like replacing manual tasks to make more time for innovation, but what does that really look like? That's interesting because I don't know if we know the full scope of what it buys us. I know that it has to do with lessening, you know, humans need to do certain things and allowing technology to take it take over. But we all know that this cannot fully replace us. But it's one of those things that you hear and it's like, oh, I'm going to automate or AI myself out of a job. And at the end of the day, humans are still required in a lot of these instances. Yes. And we can well, well, we will talk about the specifics of that in a little bit. But, Johan, I wanted to know, like what you think I'm, you know, I'm a humble podcast producer. I don't I don't expect, you know, AI to change my life. I have been using it a little bit lately, but, you're not really, you're not really using a lot of it right now. So what do you think of this? Well, I'm trying to think of it in specifically this scenario with operational technology, and she's talking about machine learning. And I can see how those processes can be integrated in that way to, you know, reduce the manual, repetitive tasks that humans have to do. And kind of offload that on to machines. But she also mentioned that, generative AI and natural language processing are also having some effects here. And I'm just wondering how these two kind of come together. How is natural language processing, how is large language models able to help with the kind of hardware side of things? My take on it, based on my conversation with Oindrilla, would be, the focus right now for her, work is to, get people to understand the, again, the capabilities of natural language processing, and generative AI as pertains to generation of like tickets and information. So like think for example, if you have some kind of like outage, in, in a plant or in a manufacturing facility, and that could automatically generate a ticket with as much detail as possible so that the person, the very human person who's on the other side of the has to address it, has as much information as they need in order to act accordingly. So it's kind of like, you know, they already have kind of, you know, some automation there. But AI is able to add a little bit more context as to what the problems are, what the errors are in a lot of the software and hardware, so that a human person can act upon it in a way that is more efficient and more optimal. Does that make sense? Yeah. So it can help really generate tickets much more quickly than someone would have to go and manually figure out what the problem is, and then take the time to write out the ticket, make sure it's detailed enough for the person on the other end to figure out what the problem is and how to fix it, right? Yes. That's just one use case. Yeah that's fantastic. Thank you I appreciate that. No worries. Oindrilla despite her expertise is running into some challenges in her work. Mainly the long standing aspects of OT culture. OT and IT they have very different policies. They have different concerns. They have different speeds. So when it comes to IT, for instance, we like to be connected to the cloud. We like to be connected to the internet. We like to take, and make advantage of different high-level applications, which is not the case in manufacturing, because you still want to be in, in a closed environment, you still don't want to be connected. And then there's obviously security concerns. You want to be on premise, air-gapped situations, which also makes mitigating vulnerabilities and cyber security an even bigger risk. So I think it's definitely more of a cultural challenge than also it's a challenge of time. So remember that Command Line Heroes episode on robotics where we talked about who updates the software on a submarine? Oh sure do. It's a very similar situation in operational technology where you have to kind of figure out the specifics to go on site and update a device, because it's not connected, it's air-gapped, it's not connected to anything. And identify what, you know, what systems can, you know, be updated and how, and the stakes remember, these systems are not in an office somewhere. Sometimes they're in very harsh environments. Places like power plants, manufacturing centers, places that handle chemicals. These are environments that are dangerous for human life. And the tech that's needed there is crucial not just for a company's success, but for safety and government regulation. For example, if I'm trying to use a big analytical system to monitor ambient air quality, to detect the amount of carbon dioxide, these systems can sort of mimic those systems very accurately. Just using a lot of statistical models and a lot of AI techniques underneath, and a lot of process and AI expertise. And these models have started to replace some of the more simpler emission monitoring systems and also support some of the systems when there is downtime, for instance, when some of these existing systems are under maintenance of repairs, these can be used as very efficient backups for regulation purposes. That's pretty interesting. That AI can be used to fill gaps. Yeah. In certain systems because of downtime or maintenance. Or maybe the system's busy with tasks. I never thought of it in that particular use case, but what's not possible at this point? Someone saw that this was a very valid use case and figured out a way to help it fill a gap where in their, in their environments. So that's pretty interesting for me. Another thing that's pretty interesting, is that Sneakernet has never gone away. It's alive and well, no matter, no matter the industry. Right. You're always going to have some industry that's going to want an air-gapped. And need it because they, you know, they want to do whatever they can to protect their machines. And that's one of the most tried and true methods out there. Yes, indeed. Yes. But I mean, if we're talking about AI, I'm, I'm just thinking strictly in the way of AI handling repetitive tasks with very little or no errors, because we're all human. We get tired, we, we get hungry. And these systems need to be on around the clock in perpetuity, and they need to be monitored around the clock. Right. There's a lot riding on this. So, you know, having AI, like you said, fill those gaps in monitoring and in optimizing, I think that that's really crucial. I feel like that that work is really worth investing some time in. And I, I really, you know, I'm really curious to see what happens next. And Oindrilla's new field for now, she doesn't have it all figured out yet. I don't really think anyone does. But what we can glean from her story is a sense of a more grounded ideal; AI being brought in to supplement vital technology within a field that many could agree it's sorely needed. All right. Takeaways. Angela. Johan. That's three stories. I'm done. I'm tapped out. What do you think? Oh my goodness, there's so much to unpack here. Iron Man, Iron Man, Iron Man, that's the main thing I want to say. But also, with Oindrilla's story about using AI to kind of supplement and fill gaps. That's very intriguing. That is a very interesting use case. But, as we've seen in Iron Man, Jarvis can go sideways. And there is always those concerns about what would happen if, so if we're using this type of sentient technology in certain applications, what's the worst case scenario and how do we mitigate it? But all very good stories. I can't wait to hear more about, Griot and Grits in maybe upcoming episodes. You're going to hear from all of our guests from this episode again later this season, including real life Tony Stark, Kris Kersey. Don't worry, he'll be back. What about you, Johan? What you think? What I really love about this episode is, and again, you we pointed out a little bit earlier that I'm a little bit more skeptical, a little bit more hesitant to bring in AI into my life at this point. But what I love is, is kind of the breadth of stories that we have, right? I mean, we have one story about using it for kind of entertainment and leisure. We have one story talking about using AI to preserve history and provide education and kind of really fill in those stories with a lot more details. And we have one more about, you know, kind of the commercial and industrial use of AI, which maybe a lot of people have heard of, but again, I'd love to hear more about just the whole gamut of capabilities that AI is entering into all the different fields that it makes possible. Yeah, when it comes to AI, there are a lot of possibilities. Yes, it can do some. Not all of the things we see in movies, some of those future visions, they aren't future anymore. They're here now. On a broader scale, though, AI can take on tasks that are cumbersome, difficult to scale, or even dangerous for humans to perform, and that's worth dreaming about. The human element is still important, both to ideate around the future capability of AI and also to understand its uses in the here and now. Well, I am enjoying these episodes so far, and I hope you are to tell us what you thought about this episode. The other episodes in this series, hit us up on our socials at Red Hat, always using the Hashtag Compiler podcast. We'd love to hear about. What are some of your applications for AI, maybe in your social life, maybe in your professional life? Or are you just using it just for day to day? We would love to hear it more about AI. That's literally what this series is all about, and we'd love to hear from you about it. Are you using generative AI for grits recipes? She went there... Let us know. She went there. I wonder if we're to ask ChatGPT for a grit's recipe whether or not it would include sugar? Say less! Say less Johan! Oh my gosh. And that does it for this episode of Compiler. This episode was written by Kim Huang. Victoria Lawton is the real life my girl Friday. If you know, you know. Thank you to our guests. Ty McDuffie, Kris Kersey, and Oindrilla Chatterjee. Compiler is produced by the team at Red Hat with technical support from Dialect. Our theme song was composed by Mary Ancheta. If you liked today's episode, please follow the show. Write us. Leave a review. Share it with someone you know. It really helps us out. Thank you so much for listening. Until next time, see you later. Bye.

More about artificial intelligence

Usually referring to processes and algorithms able to use data to simulate human intelligence, artificial intelligence can provide insights based on acquired knowledge.

Blog Post

Why agents are the new kingmakers

July 9, 2025  |  Scott McCarty, Stevan Le Meur

...

Application development and delivery
Artificial intelligence
Data science
Hybrid cloud

Blog Post

Replicating success, not reinventing the wheel: AI in the public sector

May 30, 2025  |  Jonny Williams

...

Artificial intelligence
Thumbnail image of Chris Wright and Carlos Costa

Original Shows

Technically Speaking | Inside distributed inference with llm-d

Artificial intelligence
Thumbnail image of Chris Wright and Nick Hill

Original Shows

Technically Speaking | Building more efficient AI with vLLM

Artificial intelligence
Compiler, an original podcast from Red Hat

Original Shows

Context And The True "Cost" Of AI | Compiler

Artificial intelligence
Professional development

About the show

Compiler

Do you want to stay on top of tech, but find you’re short on time? Compiler presents perspectives, topics, and insights from the industry—free from jargon and judgment. We want to discover where technology is headed beyond the headlines, and create a place for new IT professionals to learn, grow, and thrive. If you are enjoying the show, let us know, and use #CompilerPodcast to share our episodes.