Transcript 0:00 [upbeat music] I'm here with Leonard Lee from Next曲, and we're here to talk about all things mobile and wireless. How you doing, Leonard? Oh, doing well. Little bit- Yeah... 0:12 tired because June was an absolutely insane month. Yeah. A lot of things happening. A lot of things that we're probably gonna be talking about. Let's talk just a little bit about the World Cup, though. 0:24 Just, uh, who are you supporting? Well, you know, always the US. Yeah. Whether or not we're gonna win, I guess that's really still a big question mark. I think it is. [laughs] I think it is. Are they still in? 0:40 I've been... You know, to tell you the truth, I've been following F1. Right. The British Grand Prix, so. Right. Well, 0:48 I watched the match of Mexico last night, and that was pretty impressive, so I did not watch the match in Mexico, but I watched the match- Yeah... as it happened, so... Yeah. Well, there's... 1:00 It looks like there's some really great moments, and funny thing is I'm exposed more to the technologies that are enabling the media- [laughs] Yeah, yeah... than the matches themselves. [laughs] Oh, yeah, I know. 1:11 If I wanted to think about that, I could be like, "Well, they put the 5G in there," and I know that. Yeah. But it doesn't really matter to the actual running- Yeah... of the match. There we go. Yeah. 1:22 Yeah, I'm still trying- Yeah... to figure out offsides, so. Oh, well, good luck with that 'cause I don't think anyone ever does. [laughs] So yeah, let's talk. Let's have a conversation. Yeah, yeah. 1:35 Well, I wanted to kick it off with the whole rumors and the, I think it was the Wall Street Journal report on SpaceX showing off a prototype of the SpaceX phone, and- Hmm. Yeah... how likely do you even think that is? 1:55 Well, you know, I think it's probably likely. I mean, um, SpaceX could do it. Right. Um, whether it makes sense or actually creates a market for itself, that's- Right... that's still out. 2:10 Recently, I engaged in a series of posts, I think it was by Dean Bubbly, or where, you know, I... 2:18 There were a number of folks exchanging reactions and responses to the announcement and, you know, I guess there's a camp that thinks that this is gonna be revolutionary because, uh, you can connect anywhere, right? 2:31 And that seems to be- Right, right... the value proposition. And anytime, anywhere, right? Right. Uh, so you're never gonna be- Well-... in a dead zone. 2:40 Yeah, I, I'm not in the camp that thinks you can connect anywhere 'cause if you're indoors, then you need a terrestrial 5G network to connect or a Wi-Fi network, so. 2:51 Yeah, but I think the deeper challenge or reality is that most folks, their anywhere is not everywhere, and that was one of... 3:03 That I mentioned, is like, look, you know, it's great to think that everyone finds value in being connected anywhere. Right. 3:10 But most folks are just spending 99% of their time locally within the range of a terrestrial mobile network sort of coverage area, right? So how valuable is it really to have connectivity 3:29 anywhere unless you're one of those folks who is prone to becoming bear bait by being in the middle of the Yukon somewhere? [laughs] Well, it's valuable if you're a mountain guy, but- Yeah, exactly. 3:43 I mean, you know, I think, uh, you really need to think through some of these scenarios and the implications on market and market potential, and then map that back to how differentiated is the offering, right? 3:57 And so, yeah, there's still a lot of question marks. Like, one of the ones that's, uh, a big one is spectrum. What spectrum are they gonna use? How are they gonna get other handset support? Yeah. 4:09 And, you know, is it gonna be 3GPP based? So. Yeah. I don't see how it couldn't really be 3GPP based. 4:20 Otherwise, they're just locking their own users into a SpaceX handset that can't really communicate with much else. Yeah. Or, you know, I, I mean, it could be dual mode. There's a lot of what- Yeah... 4:36 at, but the real question is how valuable is everywhere, anywhere to people- Right. Right... that don't really care about anywhere and everywhere. But to your point, indoor is, you know- Yeah... 4:51 and I think terrestrial networks get a bad rap these days because of the, the hype around Leo constellations and communication. Yeah. Yeah, indeed. Right? Yeah. 5:02 And there's a lot of hype out there right now, people getting excited about stuff and maybe not having, you know, factoring in all the things that are needed in order for a market to actually form around some of these hype narratives. 5:17 Yes. Yeah. So I guess we'll see. I mean, indeed. Yeah, just think about how long it took for the, for the 5G market to even grow in the first place, but- Yeah... 5:28 uh, that already had a basis for 4G and 3G before that, so yeah. Right. Yeah. Let's move on a bit, though, to what are you seeing around kind of network 5:45 or telecom network AI right now? 'Cause I'm s- it all seems to still be in the pre-commercial stage, shall we say, right now. That's how it seems to me anyway, but maybe, I don't know. 6:00 So what do you, what do you mean by that? Do, are you talking about network automation, like the autonomous network stuff? Or are you talking about AI workloads on the network or in, you know, uh, being, uh- 6:13 Connected or conveyed through the network? I- Both? I think if, if we're t- both, but I mean, I think if we're talking about AI workloads being connected through the network seems to be kind of not a pipe dream- Yeah... 6:27 but it seems to be a way away. Yeah, I mean, we do hear a lot about kinetic tokens and theories about, and maybe even some observations about how AI is having impact on network traffic. Right. 6:40 I, I think the question for me always, uh, when I hear these pronouncements is, uh, well, okay, what do you mean by traffic? Um- Mm-hmm... you know, are you talking about volumes, the interactions or calls being made? 6:56 What I mean by that are, quote-unquote, "API calls". Yeah. [laughs] I'm always so reticent about using that acronym, but, uh, exchanges between agents, I mean, there's all these talk tracks out there at the moment. And 7:10 then when we look at network automation, yeah, I think you're making a good observation. A, a lot of the stuff is still about where we were three years ago- Right... 7:21 quite honest, with the- I mean, I think, I think it's proof of concept right now rather than anything else, right? Yeah, there's some progress but, you know, I... 7:30 What I've been saying for quite some time is that what people are doing is they're experimenting and they're figuring out what's actually true and false- Right... about technology and what they've heard. 7:41 So, uh, and I've been saying for at least two years that we're in a phase of th- we still have yet to really fully transition into this phase of unlearning what we thought was true about AI. 7:59 And we see this constantly changing in terms of the talk tracks that we see from the big tech comp- you know, like the AI labs and stuff. I mean, who talks about AGI anymore? No one. Right. 8:10 All of a sudden, no one cares, right? Mm-hmm. So, but there are all these bets that are being made based on two-year-old, three-year-old assumptions, right? And these bets take a long time to actually make. 8:25 And so everybody is, uh, kind of messing around and slowly learning. Some people are, are-- aren't learning at all. They're in a denial phase, right? They're going- Mm-hmm. Mm-hmm... 8:37 "No, this is revolutionary and it has to work out," right? So- Right. It's not just a river in Egypt. Yeah, so they're constantly pivoting, but then the thing is that you have to think in terms of first principle. 8:48 What is-- What are the foundational limitations of this technology where you can hypothesize or create constructs to the nth degree, but you will never solve? 9:00 And one of the classic ones obviously is hallucinations, right? The other is- Yeah... just simple unreliability of AI, the lack of accuracy. These things are things that you just simply cannot, m- you know, uh, dissolve. 9:18 Um, they're just part of what you're dealing with in terms of a technology, which is large- Right... language models at the end of the day. The hallucinations is a real challenge still, so. Still. Yeah. 9:31 And there are people who are saying they got rid of them somehow. Yeah, I know. How did you do that? [laughs] No, they didn't. And we still have people that are doing it and say-- claiming that claim. 9:41 And, you know- I know. I know... these are the people who actually don't know anything about AI. Right. Yeah. Indeed. You mentioned APIs just a second ago there. 9:53 I always- Oh, by the way, before we move on, before we move on, uplink- I-... is going to be a big deal [laughs] one way or the other, so. Yeah. There you go. So I always check in with you on 5G APIs. So- Yeah... 10:08 what stage are we at right now? Or, or should I be concerned or should I be joyful about this? I actually gave a talk about this at DTW Ignite 2026, and- Right... 10:22 my first slide was, "Stop talking about APIs," 'cause- Right... you know, yeah, this is not what you... This is not the innovation. It's also not what you monetize. 10:34 It, it's really these products, these network products, services. Uh, and they can be both informational and what I call transactional, meaning they tap into, let's say, the hardcore network capabilities. 10:47 These are like actual experiential things that are delivered through like what Ericsson called differentiated connectivity. Um- Yeah... 10:57 they really tap into managing that in, uh, that session and complete life cycle of, let's say, a differentiated network s- service or experience. Th- those are gonna take a while. It's tough because those are... 11:10 Especially if you're looking at global network APIs or so these, quote-unquote, "aggregated" type, um, services that the likes of Aduna, Global, and others are looking to support and enable. 11:27 Those are gonna take a while just because you have a constellation of networks of these participating operators that need to be sort of, uh, interworked and harmonized to a certain level to deliver that product. 11:42 And the problem right now is people haven't really figured out what those products are, right? You have the APIs, but you don't know what to do with them. Right. 11:52 And so this layer on top of the network APIs and the GSMA's, you know, Camara service APIs, that is kinda like the missing bit where really we need to have a movement of innovation that articulates what those solutions are. 12:09 But they're gonna, uh, they're g- uh, this is like, w- in my observation- A pretty green field where there's only a few parties who really know, okay, what do we do with these network APIs and- Right... 12:22 innovations can we do on top of them? Yeah. I mean, what, Nokia, Ericsson, and maybe a couple more? Right. Yeah. Uh, Bridge Alliance is the other one that's pretty... Yeah. 12:36 And they're different flavors, right, of um- Yeah, indeed... I, I hate to call 'em aggregators. I... Exchange plus marketplace, I guess. Right. Sort of how you look at 'em. Mm-hmm. But don't be too frustrated. 12:51 And network APIs aren't going, uh, you know, they're not disappearing. They're, they're gonna be- Oh, no, I know... It's just like your earlier comment about 5G. 12:59 You just need to have everything maturing and advancing and, uh- Yep... having scale, and it takes m- longer than people typically anticipate. Indeed. 13:09 Well, everything takes longer than anyone anticipates, I find, but there we go. Um, well, how have you been going on... 13:18 I, you know, me and you share an interest outside of, uh, telecom and technology and kind of like the guitar playing field as well. How have you been going on, uh, with that recently? 13:31 Oh, uh, I haven't done any live performances lately, but I did put together- Yeah... I, a, um, u- a national anthem tribute for the- Wow... 250th birthday of the United States, so yeah. 13:47 I mean, bringing it all back home really. [laughs] It, it was a single take. I will tell you right now, the national anthem is not that [laughs] easy to play. No. I mean, it sounds simple, but I challenge- Right... 13:58 any guitar player out there to do a single take after, like, 20 minutes of practice. [laughs] And I, I really, you know, don't know how, didn't know how to play it, but, uh, I play by ear. 14:09 But I kinda snubbed one, you know, like, uh, sort of messed up one note. That's not bad out of, like- No, it's not bad. I mean, just turn the fuzz pedal on and get a bit of Hendrix on it. It's okay. It's fine. 14:21 There you go. There you go. There you go. Yeah. Mm-hmm. So that, that's the only sort of public performance [laughs] that I've done lately. 14:27 But I'm definitely looking forward to putting London more prominently in my travel roster, my events roster- Oh, wow. Yeah... next year. 14:36 I'll be out in London again and onstage and hang out with my friends, you know, Dean Bubbly as well as Andrew Collins said. Uh, but please- Right, right... join. Join us. Yes, I will. I'll let you know. If I can, I will. 14:51 Yeah. And you know what? We talk a lot about technology as well as the telco industry at large, so, uh, it's always a, a great conversation and obviously some good entertainment. A good hat, I guess. [laughs] Yes. Yes. 15:08 So Leonard, good conversation, but do you have any other ideas about what we talked about and technology and how the news has been going forward over the last week also? Well, let me think about that. There's so much. 15:24 I'm just trying to figure out... Yeah. Mm. One of the thing... You know, there's headlines and then there's things that are actually happening, thematic shifts that I'm seeing. Right, right. 15:32 I know everyone's crazy about AI RAN, or at least was for a while. Yeah. One of the things that I'm noticing coming out o- of DTW Ignite this year in Copenhagen, there's a disaggregation of that conversation. 15:47 I don't mean in the Open RAN sense. 15:51 More of a, hey look, um, AI will have its role in evolving the RAN, but it's not gonna be the way that I think AI RAN has been portrayed, uh, to this point. 16:08 A- actually, that portrayal is still ongoing. 16:11 I think there's a growing recognition that it's gonna be s- the AI on RAN stuff is gonna be separate, and it's going to live in an entirely different way in terms of placement and the types- Right... of loads. 16:25 And, um, again, going back to my earlier comment, people are just finding out about AI, and they're discovering what's untrue about their former, uh, assumption sets, and that's something that I see that's starting to accelerate in terms of recognition. 16:42 And so it's not a news item. It's just, it's something that I, uh, I'm picking up i- in terms of a, a broader recognition at the technical level, right? 16:52 So these are engineers talking about what's feasible, and it, this is not about these press releases from vendors trying to pontificate what the future of the RAN's gonna look like. 17:03 These are engineers that are going, "You know what? That doesn't make sense, but this makes sense." You know what I'm saying? Right. So. Mm-hmm. Anyways, I thought that was really interesting. 17:12 [laughs] Thanks so much for your time. Absolutely. Thank you. Thanks for having me on. I'm Dan Jones, editor of, uh, Next TNT Networks, talking with Leonard Lee from Nextcloud. We'll see you next week.
