I’m a Customer Engineer at Google Cloud. So in the end in Google, we call it a Customer Engineer, but it’s basically a Sales Engineer. Yeah. In the Media and Entertainment team for the UK, I’ve been working in Media Entertainment and kinda cloud technologies for, most of this Yeah. Twenty past twenty-five years or so. Yeah. So, yeah. It’s great to be here. Thanks, Chris. Yeah. And you’ve got sort of, like, a background of quite an interesting background of kind of, music production and some of the first streams that we saw online of some of those, events. That’s right. That’s right. You might have seen some videos. Yeah. Early early days of webcasting. Yeah. So I did my first live streaming in, nineteen ninety seven, actually. Yeah. As you say, it was, it was, like, it was, like, nightclubs. Yeah. Basically, live streaming in nightclubs. So I’ve been streaming a long time. Away for a lot of our, I guess, of what we’re doing with, with video today and how we stream it. Yeah. That’s it. I see people from that back then. They said, what are you doing? I said, well, same thing, but, like Yeah. Slightly higher quality. You paid a little bit more for it, hopefully. Yeah. But yeah. Yeah. And and the quality is a bit better than it was, like, literally modem postage stamp. Yeah. Yeah. Fantastic. So, Chris, I mean, what’s the big things for for Google this year at IBC? I mean, we’ve seen a lot about AI with Google, obviously, with Gemini and, Veo. So, I mean, what are the the the points that you’re highlighting at the show? Yeah. So, I mean, it’s it’s a absolutely exciting time in the industry. You know, and IBC is kind of, always exciting, but there is so much change. Yeah. And the and the pace of change and the rate of change is is is unprecedented. I mean, the the the video industry has been in transformation for, you know, a decade or two Yeah. From, you know, streaming on a modem to now it’s like, you know, mainstay is live streaming. But this disruption that’s coming from AI and what we found at Google is, this year, we’ve released, like, you know, it’s Gemini 1.5, then it’s Gemini 2, then it’s Gemini 2.5, 2.5 Flash, DeepThink model, live API. There’s just features coming out all the time. Then we’ve got Imagen 4, which is our text to image, and Veo 3, which is our text to video. They’re on, like, version three and four, and they’ve only been out in, like, twelve months. Yeah. Yeah. Yeah. So the new and they’re getting new features all the time. So many models, so many things, and so much change. And the infrastructure is changing to keep up with it. We’re getting, you know, V5 TPUs, which is our Tensor Processing Unit, all the latest, AMD and Intel families of CPUs. So, you know, the infrastructure to drive all that new technology Yeah. Is it’s unprecedented here. And then, you know and in video, we think and and in broadcast and in media, we think there’ll be a lot of change too. Yeah. Yeah. It’s difficult. It’s interesting. It’s like I try and keep on top of all all everything new that comes out, but I I find it really difficult to kind of keep up with all the new technologies and the new models that are coming out. So, like, have you got any advice on how we how we stay up to date? What’s the best way to do it in a Google world? I mean, how do you stay up to date with it all? So yeah. It it is a challenge because it is changing so much. And occasionally, we change the names of things. Yeah. You know? And then, and and then sometimes we’ve got two or three ways of consuming one thing. Yeah. So, we we have a, a, an AI blog, which will kind of, give all the highlights that have been out that long. That’s that’s quite good. Yeah. The other thing I find to do is just try them out. Yeah. If you just try something out, if you even if you make five calls to Gemini or five you know, make a video in Veo. You know what Veo is. You won’t forget Yeah. If you’ve made a video in it. And then same with Imagen. And then, yeah, don’t worry about the version numbers too much. Just just just try these things out. And I I I’ve gotta say as well, like, this earlier this year Yeah. When I because I only first got access to Veo. Veo one, it wasn’t for you know, it was really hard to get access to because for for whatever reason. And when I first got access to it and I, and I actually instead of typing a text prompt to video, I did image to video, which is another modality. So image on the input, video on the output. And I just had a picture of me. I thought I’m not gonna use anything controversial, a picture of me. Yeah. Okay. I made this, do this, and have these things going on. And I was just fell off my seat. You know? I was like, what? That’s me doing things that I’ve just told it to do. And it’s like Yeah. It’s like and then, you know, set up till two o’clock in the morning making videos, and then the next day, same thing. And it’s not that often that you get a technology that’s that exciting. Yeah. And game changing. Yeah. Certainly. It So do you think that’s sort of where the biggest change is going to come? It’s kind of on the creation side of these tools. So, you know, obviously, creating videos using AI to do that. I mean, we spoke a little bit earlier about using AI for things like, you know, automated ad breaks and things like that and being able to insert SCTE markers at the right time into streams. So, I mean, where do you think the biggest changes have come? Where are you most excited about the tools that you see? So, you know, where is it gonna happen? And this is the thing. It’s gonna happen everywhere. Yeah. It’s some so it’s gonna be some models for linear, some models for for, offline assets. It’s gonna be in search. It’s gonna be in metadata. It’s gonna be in creative content creation. It’s gonna be in storyboarding. It’s gonna be in, everything. So I I don’t think any one thing will be the main area where AI influences it. It’ll be all parts of the value chain in the industry. Yeah. So from production and post production and, you know, detecting, ad breaks, for example, as we said, in a live environment, in an offline environment, enriching metadata, search of people back catalogs of ten thousand hours of content. Don’t know what’s on there. Can’t watch it all. You can then you do something called, you extract something called video embeddings. You store that in something called a RAG database, and then your your model can use the rag database to know everything that’s in your archive. Yeah. And you say, oh, give me the thing I was I saw this person talking to the so and so or, anything that was shot outside in London that day. And you’ll get all the related visually looking, piece of code. That’s just search. That’s just one use case. Yeah. Yeah. The the the Veo one is the so if you don’t know, Veo is the the model that outputs video. So large language models are kinda generally text in, text out, and, the other big thing is code. A lot of code will come out. Yeah. And that’s language models, but, diffusion models create images and video. And then you’ve got the so so we call them different modalities. And then you might have heard the expression multimodal, and multimodal normally means multimodal on the input side. So you first, you type to it Yeah. But you can also talk to it, and so the modalities is, audio. And then and if you put an image in, it becomes video. Yeah. And then they you know, the next generation of model is video to video. So you put a video in and you said, mate, give me more frames or double the frame rate or, super resolution would be a good example. Upscale something from ten eighty p to four k without just, what do you call it? Nearest neighbor, you know, just traditional upscaling, but using AI to really know what the definition would look like if it was using a a what’s called reverse diffusion model. Yeah. So, I mean, you might are you seeing a lot of that in vendors today when you walk the show floor? Are you seeing a lot of kind of interest in AI tech that Google are producing being used by vendors? Is that So a a bit I’m I must admit it’s first day of the show. Yeah. So, I have walked around, and I I I did sort of notice a lot of the real traditional stuff that still needs to happen. Yeah. I did watch the video, that was created with Veo 3 on the IBC accelerators. So great team from, RAI, the Italian broadcaster, Yle, from Finland, and, ITV were on that accelerator. And they and and, of course, Globo in Brazil, they were one of the main champions on that accelerator. They all got together and built a framework for making content. And, we we we gave them access to Veo 3, and some of the extra features with image to video, and they built and they used a lot of tools actually, not just not just Veo. They used, ComfyUI, which we hosted for them. Yeah. And then, and then to Plutzbox, which is another kind of a a framework tool. And they produced a load of content. So I and I watched that, and that that was packed. That was absolutely packed out. It was like you you couldn’t move. Right. It was over in hall fourteen. Yeah. So the, the obviously, Google are sponsoring the accelerators, and, Zixi are part of the master control room and the cloud accelerator. Is there any kind of standout accelerators that you’ve seen that are really pushing tech? Yeah. So, we we we love the accelerators. Right? It’s the innovation end. And, I know we’ve worked together on them in the past, Chris, and I’ve really enjoyed that. I wish I’d been able to spend more time with you on the master control cloud, actually. Yeah. So, you know, we did the low latency HLS stuff. We did the framework for AI, which I just mentioned. And, my manager, Justin, he worked on the AI agents. So, you know, a bunch of AI agents in the production control room. So, you know, helping with the, in the production control room, assisting, directors and producers. Yeah. Okay. Oh, great. So, like, a lot of transformation from, you know, we call it traditional kit and hardware up into up into cloud environments and sort of, you know, obviously, we’re seeing a lot of that happen. And, you know, over the past ten years, I remember people were talking about kind of lift and shift models where they were just trying to replicate exactly what they’ve got, and then we started introducing the concepts of cloud native and and building things specifically for the cloud. So, I mean, you know, I I can see a trend around kind of remote production that we’re doing today. There’s a lot of remote production moved into the cloud, and I think there’s more and more software that’s making that available and making that possible. Yeah. I think one of the major challenges with remote production has always been around sort of audio, and I think we’re starting to to see some answers for that. I mean, if you look, they they RBC Accelerate we’ve done, we’ve got kind of a commentary track being synced in with a video track now, which is quite cool. So, obviously, kind of remote production is one of those points. I feel like we’ve always had kind of contribution distribution and IP based kind of, streams being sent, for a long time now. Is there any other areas that you think are exciting in broadcast that you see moving to kind of cloud infrastructure? Yeah. Yeah. So yeah. There’s there’s so much you can do so much in the cloud now. Remote production or and we’ve got this Google Distributed Cloud as you know. So you can use cloud like deployment fabric at the edge for those low latency applications. Yeah. And then you can have a com a kind of a combination of the two. So that project we did together, connect and produce anywhere, where we used a Google Distributed Cloud cluster for, at at site production. Yeah. I’m using that cluster again this year because so so it because it’s a cloud like deployment fabric, and instead, I’m just using it for contribution upstream. Yeah. So we’re we’re running a, a IBC Google Cloud Hackfest Yeah. Over in hall fourteen. That’ll be running on Saturday and Sunday. And, we’re taking some live sources from, the ride from hall fourteen, and we’ve got some simulated live sources from Formula E, so some in car cameras and some track cameras. Great. We’re providing all of those. So ten feeds plus audio. We’ve got Shure array microphone, six channel mic, and that’s going up into the cloud as well. And then we fan out in the cloud, as you say, just it’s quite easy to scale at that point. Yeah. And then, and then, the teams will take the, take the feeds. They’ll mix the feeds. A lens that’s cutting marker. And here’s the fun bit. You’ll kinda have in the teams, they’ll be doing, they’ll be using Veo three to create an ad. They’ll they’ll create their ad. They’ll get load that up to Ad Manager, then we’ll use our Video Stitcher to stitch in the ad from Ad Manager into the live stream they’ve just produced. We’re gonna do that with thirty six teams. Right? And it’s like Yeah. All in the cloud. Yeah. And that sort of enables you to do kind of, you know, specified ad breaks for, you know, set regions or perhaps the concept of of people watching this on their consumer devices and having a favorite race or a favorite, like, you know, team that they can have specified ads based on on those players or those those those racing drivers, which is quite Exactly. So personalized. And then, you know, the it it just you just need to think a little bit into the future. But, like, even the the, you know, not a per so when you make an ad creative, you know, you’ll you’ll make one version of it, and then it’ll be targeted to the individual people. But you can make different versions of the ad. And then you can even if you really let your mind run wild, it’s like, you know someone’s watching, you know what they’re interested in. You can make an ad in video for them, which they’re gonna watch in the next ad break. Yeah. And that’s quite scary, isn’t it? Yeah. Yeah. But it’s it’s quite incredible. You know? It could be like, I saw that you watched this thing, like, last week. We’ll make a specified ad based on that thing that you watched last week including Include or or include it into the content. Right? So, you know, product placement is that, oh, right. This person like this product, so that’s that’s the product placement spot that they get filled in and and personalized on that. So I do think Veo and AI will, change the way, eventually, how advertising is consumed in linear, and and in all content. Yeah. I mean, we’re seeing a lot of, you know, a large trend to move to kind of SCTE based ad insertion, for kind of regionalized kind of opt outs and, kind of, distribution. So, you know, based on different region, you can have the SCTE market trigger different adverts as well as kind of, dynamic ad breaks. So if you’re doing a sporting event and you wanna know where to take an ad break. So one of the interesting parts of the sports world, I guess, is is that was always traditionally done by people in the remote location. So on the OB van, they’d they’d trigger when the when the advert was to break or they’d they’d they’d, they’d set the markers in the stream. Now when we start to think about moving kind of that production into the cloud, if you’re doing more remote production in cloud, there’s kind of the handoff between who does certain jobs. So it’s the people that are on-site or the people that are looking after the production cloud system. Right. So it it could get a bit blurry sometimes. So I guess the introduction of, like, an AI based model to to trigger those ads is really interesting and kind of takes that headache away for a lot of people because you you’re doing it in an automated fashion. Yeah. Yeah. It’s great to do. Yeah. So we we think there’s there’s a lot of interest in that. And, you know, the so there’s lots of stuff that’s, you know, still gonna work the same as you say, SCTE marker insertion, linear channels. You know, none of that’s gonna change overnight. There’s all gonna be just a little bit of enhancement with the many of these things using some kind of AI model. I did it when I was working on the the Hackfest. It’s like, these do Terraform. Just ask Gemini. I said, I need to do this in Terraform. Give it to me. It’s like, okay. Okay. This this didn’t make this right to more Terraform. Didn’t write any Terraform myself. Right. Terraform. Just copy and paste. Yeah. Yeah. Okay. Cool. Well, I guess the last kind of mandatory question I have to ask, Chris, is, obviously, looking ahead, so the next kind of twelve to twenty four months, kind of where do you see the most exciting sort of trends and techs? Where do you think we’re gonna go? Anything kind of exciting on the horizon? Yeah. I mean, you know, with without wanting to sound like too much too repetitive. Yeah. But, AI is gonna change everything. Yeah. It’s gonna all the different parts of AI will change all the different parts of all the different businesses, including media, video, broadcast consumption. I think there’ll be, you know, be a big impact in social. I think Veo will have, eventually and or and or, you know, video diffusion models will have a big impact in the way content is created for social, and the way, things are targeted and we learn about things and search. Yeah. And, you know, not least software development. You know, let’s say, I became a Terraform expert without writing any. Yeah. I think it it’s allowing teams to be a lot quicker to produce, you know, higher quality software and put extra features in systems now because it’s a lot easier to to create that software, hopefully. And and what I’d say on that as well, you know, not to say that it’s, you know, to, it’s it’s disruptive. But, there’s all, you know, jobs and but it’s like, if you I think someone said, if you if you if you focus on the door that’s closing, you won’t see all the doors that open, alongside you. Yeah. So it it creates loads of opportunity. It does make people more productive. It’s true. Yeah. I think it will create more jobs in in many places and ways. Yeah. But, like, it’s important to focus on the jobs it creates and enhances and the experiences it improves rather than the kind of, the certain certain, opportunities there. I actually think it allows a bit more of a disruptive approach as well because you get you get more, you know, developers to be able to create more software and interest in software. And, you know, it’s not just kind of the big companies that are creating those. You’re allowing a wider scope of people to be able to access that and create things quicker. Yeah. And and I think more people can do it. Yeah. And that’s just the creation. We think about, like, the QA side of it and using QA processes and Yeah. I think I think all the more power. Right? Yeah. It it it becomes in some ways less specialized, and you might get more people who are developers or more people who are developing products. Right? So that’s, that’s a I think it will improve many things. It’s really exciting time. Brilliant. Okay. Well, thanks for your time today, Chris. Been a pleasure talking to you.