Latest / UK Column Radio / Has China Already Won the Tech Race? | Silk and Steel

Transcript

Has China Already Won the Tech Race? | Silk and Steel · UK Column Radio

• 0:01
  Speaker 1: Okay, we are live. Welcome to another episode of the
• 0:08
  Speaker 1: Silk and Steel Podcast. I'm your host, Carl Za. Today
• 0:13
  Speaker 1: we have our returning guest, a tech expert on everything
• 0:19
  Speaker 1: China tech and AI, T p Huang. TP Huang has
• 0:23
  Speaker 1: been one of the audiences favorite and since it's been
• 0:28
  Speaker 1: a while, I feel like every six month there's like
• 0:31
  Speaker 1: some tremendous progress that China has made. This is why
• 0:34
  Speaker 1: I have TP back. TP do wish to remain somewhat anonymous,
• 0:40
  Speaker 1: so this is why his camera is off. So I'm
• 0:43
  Speaker 1: just gonna switch the camera to me, but you guys
• 0:47
  Speaker 1: will hear TP's voice. Okay, welcome. First of all, welcome
• 0:52
  Speaker 1: to the show. TP. How are you, sir?
• 0:54
  Speaker 2: Good? Good, good, Carl, Like, I'm really happy to be back.
• 0:57
  Speaker 2: And this was like for the first time for me
• 0:59
  Speaker 2: to do a live stream, so this is exciting.
• 1:02
  Speaker 1: Yeah, this is uh. This we will be able to
• 1:05
  Speaker 1: take some questions from the audience, so at the end,
• 1:09
  Speaker 1: but first I want to get an update from you
• 1:14
  Speaker 1: on China's tech development. Like I said, every six month,
• 1:18
  Speaker 1: I don't know that last time we talked, I think
• 1:20
  Speaker 1: it was like four months ago, and that's in terms
• 1:25
  Speaker 1: of AI development. That's that's almost like like a generation.
• 1:31
  Speaker 1: What do you have for us?
• 1:33
  Speaker 2: Well, you know there's both the software aren't hardware side
• 1:36
  Speaker 2: of things, I think, and I think both of them
• 1:38
  Speaker 2: are doing really well. Uh, the hardware kind of made
• 1:41
  Speaker 2: a lot of noise recently with Huawei and their semiconductor
• 1:45
  Speaker 2: h news about reframing the idea of advanced chip production
• 1:52
  Speaker 2: from north law to health scaling, which was really a
• 1:56
  Speaker 2: very fascinating way of thinking about things. And and they
• 2:01
  Speaker 2: have like this huge amount of innovations in all the
• 2:06
  Speaker 2: semiconductor stuff that they're working on, and I think that
• 2:10
  Speaker 2: people are really not understanding the kind of impressive research
• 2:18
  Speaker 2: and into just all kinds of trip production because it's
• 2:25
  Speaker 2: very much foreign concept to the Koreans, the Taiwanese and
• 2:30
  Speaker 2: Americans because they don't have the expertise to do it.
• 2:33
  Speaker 2: And so that's kind of interesting. And also in the
• 2:37
  Speaker 2: AI world, I think we're seeing in China now like
• 2:41
  Speaker 2: two Trash. We have the big hyperscalers like Bike Dance
• 2:46
  Speaker 2: and Ali Baba who are doing their own stack of
• 2:51
  Speaker 2: ais from the hardware all the way.
• 2:55
  Speaker 3: To the models and the apps.
• 2:58
  Speaker 2: So for example, Ali Baba recently released, there are new
• 3:04
  Speaker 2: Qui models, but not only the Qui model itself. It
• 3:07
  Speaker 2: has also has this entire AI platform and it also
• 3:12
  Speaker 2: showcased all the chips that it designed for the entire
• 3:16
  Speaker 2: AI stack and bite.
• 3:18
  Speaker 3: Dancers doing the same thing.
• 3:20
  Speaker 2: And they have like the most popular AI at in China.
• 3:24
  Speaker 3: It's called Dobao.
• 3:26
  Speaker 2: Anyone's ever used it. It's really good. If you're in China,
• 3:29
  Speaker 2: you should use Dobao.
• 3:30
  Speaker 3: Uh.
• 3:30
  Speaker 2: And then then there are actually the AI startups as
• 3:34
  Speaker 2: I call them, you know, like the the the Deep
• 3:37
  Speaker 2: Sea Kimmy Kimmy, Slash, Moonshots, z Ai and Mini Max.
• 3:44
  Speaker 2: There's a bunch of them, and it's it's very interesting
• 3:46
  Speaker 2: because uh, you know, we talked about involution with China
• 3:51
  Speaker 2: across different texts. Now we have like an evolution months
• 3:55
  Speaker 2: of the year and the AI companies And in some
• 3:58
  Speaker 2: way I wonder if the uh, you know, the compute
• 4:01
  Speaker 2: resources are too widely divided amongst the AI startups because
• 4:06
  Speaker 2: individual ones of these AI startups probably don't have as
• 4:10
  Speaker 2: much compute as they need. And but you know, it
• 4:15
  Speaker 2: is driving a lot of hardworking high startups in China
• 4:18
  Speaker 2: and they're doing some really great research.
• 4:21
  Speaker 1: So I do want to address this question, like the
• 4:24
  Speaker 1: common reframe right now in the America is that, by
• 4:29
  Speaker 1: the way, for the people who are just joining, I'm
• 4:31
  Speaker 1: not I'm only showing myself, not TP, because TP one
• 4:35
  Speaker 1: would like to remain unseen. So you just get to
• 4:39
  Speaker 1: look at my beautiful mud while I ask questions to TP,
• 4:44
  Speaker 1: and TP will respond with his voice. And so one
• 4:48
  Speaker 1: question I have for you TP, is you know a
• 4:50
  Speaker 1: common refram America is that, Okay, China might come out
• 4:54
  Speaker 1: with all these innovative algorithms trying to squeeze more juice
• 4:59
  Speaker 1: out of their compute, but the United States, they have
• 5:04
  Speaker 1: the scale, they have the scale of the compute because
• 5:08
  Speaker 1: there's no restriction on Nvidia AI chips to American AI
• 5:12
  Speaker 1: companies obviously, and the United States will just be able
• 5:17
  Speaker 1: to brud power uh scale by building the scale to
• 5:22
  Speaker 1: overwhelm the Chinese. So the Chinese will lose out whatever
• 5:27
  Speaker 1: advantage they might have in terms of innovative algorithm by
• 5:32
  Speaker 1: just you know, the brood power of compute that's available
• 5:36
  Speaker 1: to the American AI companies. And they and I even
• 5:40
  Speaker 1: hear people say compute is all that matters right now,
• 5:43
  Speaker 1: because we have already reached a point where it's all
• 5:47
  Speaker 1: about scale. What do you say to.
• 5:49
  Speaker 2: That, Oh, well, I mean, this is a very hard
• 5:56
  Speaker 2: question because the people that are talking of the Chinese
• 6:01
  Speaker 2: AI labs are typically talking to Moonshot and zen Ai,
• 6:07
  Speaker 2: who happen to be the uh the AI startups in
• 6:11
  Speaker 2: China that are most lacking off compute resources, but they're
• 6:15
  Speaker 2: also lack money also. So it is very interesting because,
• 6:20
  Speaker 2: as I said before, in China you have kind of
• 6:23
  Speaker 2: like a two chair system. You have the the BYT
• 6:26
  Speaker 2: Dance and Alibap out of the world who you know.
• 6:30
  Speaker 2: They they have these data centers that they lease from
• 6:33
  Speaker 2: subsidiaries in Southeast Asia right mostly in like Malaysia, Thailand
• 6:39
  Speaker 2: and whatever, that have their huge clusters that they.
• 6:42
  Speaker 3: Used to train, and then they have the domestic.
• 6:46
  Speaker 2: Clusters more for inference and just running their apps and
• 6:49
  Speaker 2: things like that. So I don't think the C teams
• 6:53
  Speaker 2: like or the Quinn teams are lacking compute resources to
• 6:56
  Speaker 2: the same degree. But if the question is between someone
• 7:01
  Speaker 2: like deep Sea can mooshat versus Anthropic, then it's obviously
• 7:05
  Speaker 2: the Chinese labs do lack do have much less compute resource.
• 7:12
  Speaker 2: There's this continued belief inside Silicon Valley that once they
• 7:18
  Speaker 2: get to a certain level and then they can just
• 7:21
  Speaker 2: keep iterating on this stuff and get exponentially smarter. I
• 7:30
  Speaker 2: you know, I generally I like to see things play out.
• 7:34
  Speaker 2: I don't like to live on hypotheticals. And the what
• 7:39
  Speaker 2: we've seen is that the the AI portion of it,
• 7:43
  Speaker 2: if you have used the AI recently, you notice that
• 7:45
  Speaker 2: the the agantic stuff, the coding has gotten so much
• 7:48
  Speaker 2: better recently, and that has that's gonna put a lot
• 7:52
  Speaker 2: of software developers out of work, and it's mostly because
• 7:56
  Speaker 2: coding you can actually verify the results if you if
• 8:00
  Speaker 2: you do through a program, you can you can keep
• 8:03
  Speaker 2: generally result there's infinite data. So in terms of on
• 8:08
  Speaker 2: stuff like that that you can verify online, that is
• 8:12
  Speaker 2: getting a lot better stuff you can verify digitally, right,
• 8:16
  Speaker 2: But other stuff it's really depending on the quality of
• 8:20
  Speaker 2: the UH input data and that is something.
• 8:27
  Speaker 3: That is not so discussed.
• 8:29
  Speaker 2: But that's also you know, if you listen to people
• 8:32
  Speaker 2: like Nason Lambert talk about this. The Chinese ecosystem also
• 8:37
  Speaker 2: doesn't seem to have as storyle of like you know, ah,
• 8:45
  Speaker 2: I guess token input as the Western ones, but West
• 8:48
  Speaker 2: has like a pretty developed industry. A lot of these
• 8:51
  Speaker 2: Chinese labs are dependent uh uh you know, themselves to
• 8:56
  Speaker 2: figure it out.
• 8:58
  Speaker 3: So that's actually part of the reason why deep Seek.
• 9:02
  Speaker 2: Is making its own like models so cheap so that
• 9:05
  Speaker 2: people can use it and feed as much token as possible,
• 9:09
  Speaker 2: and then people can then generate uh data, then it
• 9:13
  Speaker 2: can you know, feedback.
• 9:14
  Speaker 3: In esnthetic tokens like these.
• 9:16
  Speaker 2: These are things that Anthropic did to really improve its models,
• 9:20
  Speaker 2: for example, and what Opening what opening ai is doing also,
• 9:24
  Speaker 2: So in terms of like the progress of the Chinese labs,
• 9:28
  Speaker 2: I haven't seen anything to indicate that the Chinese labs
• 9:31
  Speaker 2: are actually falling behind. Depending on the measure you go,
• 9:37
  Speaker 2: you can see in some areas the Chinese labs for coding,
• 9:40
  Speaker 2: for example, are really shrinking the gap with you know,
• 9:44
  Speaker 2: Anthropic uh. In other areas maybe it's uh, it's staying
• 9:49
  Speaker 2: about the same. So generally I would say like that
• 9:53
  Speaker 2: when people say four to six months, I think that's
• 9:55
  Speaker 2: pretty reasonable, and it's been that way for about a
• 9:58
  Speaker 2: year now.
• 9:59
  Speaker 3: And I think if you look at how things are.
• 10:04
  Speaker 2: In Silicon Valley and in China, people are all working
• 10:07
  Speaker 2: very hard night and day on this stuff. So when
• 10:10
  Speaker 2: people asking why the Chinese labs are not closing the gap.
• 10:14
  Speaker 2: I tell them that everyone's working really hard at this stuff,
• 10:17
  Speaker 2: and uh, you know, it's very hard to extend your
• 10:21
  Speaker 2: lead or catch up when everyone's working so hard at it,
• 10:24
  Speaker 2: and there's a lot of smart people around. So from
• 10:27
  Speaker 2: from that aspect, if you only concentrate on the l
• 10:32
  Speaker 2: l M side of things, the you know, the the
• 10:36
  Speaker 2: American labs still have about the same advantage it had
• 10:39
  Speaker 2: six months ago.
• 10:40
  Speaker 3: But there are other areas like you know, world models.
• 10:44
  Speaker 2: Stuff that you use for your robotics and video generation
• 10:48
  Speaker 2: models like with China. Uh you know what Bike Dance
• 10:52
  Speaker 2: came up with a few months ago, with Sea Dance
• 10:54
  Speaker 2: two point zero.
• 10:55
  Speaker 3: Those are the things where China has to.
• 10:57
  Speaker 2: Actually we're already leading or have left ahead of the
• 11:01
  Speaker 2: American competitors.
• 11:04
  Speaker 3: Yeah.
• 11:04
  Speaker 2: So, so I'm not convinced of this entire theory that
• 11:07
  Speaker 2: you know, just there's actually gonna be any growth in
• 11:11
  Speaker 2: the gap or shrinking the gap. But I think a
• 11:15
  Speaker 2: lot of this is hypothetical, and it seems to be
• 11:17
  Speaker 2: very much of a Silicon Valley culture to just you know,
• 11:24
  Speaker 2: sip on their own like Silicon Valley superiority thing and
• 11:28
  Speaker 2: just assume things will work out well for them.
• 11:32
  Speaker 1: Yeah. And another thing to keep in mind this is
• 11:35
  Speaker 1: the AI races is really the Chinese in China versus
• 11:40
  Speaker 1: the Chinese in America, because those are the top AI
• 11:45
  Speaker 1: talent in America. There are Chinese and they're not even
• 11:48
  Speaker 1: Chinese America. They're like Chinese straight from the Chinese universities.
• 11:53
  Speaker 1: So the you know there I saw some Uh, there's
• 11:58
  Speaker 1: some alarm bell recently going off among uh you know
• 12:03
  Speaker 1: the beltwegh that the apparently the Chinese AI are grabbing
• 12:09
  Speaker 1: a lot of market share even in America, and there's
• 12:12
  Speaker 1: discussion of banning US companies from using Chinese AI. Uh
• 12:18
  Speaker 1: do you think there will be more restrictions you know,
• 12:21
  Speaker 1: like uh, you know in the United States, maybe legislative wise,
• 12:27
  Speaker 1: are are or otherwise restriction of Chinese AI applications.
• 12:34
  Speaker 3: I mean, what are they gonna do.
• 12:36
  Speaker 2: They're gonna stop VPNs, They're gonna stop like as s h.
• 12:40
  Speaker 2: They're gonna stop access to get up Like I don't know,
• 12:43
  Speaker 2: I don't know what these proposers.
• 12:44
  Speaker 3: Are are are talking about. Right.
• 12:47
  Speaker 2: So one of the reasons why the Chinese models are
• 12:50
  Speaker 2: getting gaining market share on stuff like open router. If
• 12:53
  Speaker 2: you watch it, you can see it is because they're
• 12:56
  Speaker 2: very cheap to use. Like I use Kiddy two point six,
• 12:59
  Speaker 2: and I think telling people about this on X for
• 13:01
  Speaker 2: a long time, I pay fifteen.
• 13:03
  Speaker 3: Dollars a months. Right now.
• 13:05
  Speaker 2: It does everything that I needed to do. There's nothing
• 13:09
  Speaker 2: I've been using for a long time, for like a
• 13:10
  Speaker 2: couple of months now, and there hasn't been a task
• 13:14
  Speaker 2: my Kimmy code can't do that I can personally do,
• 13:18
  Speaker 2: Like it's better than me I Python coding.
• 13:20
  Speaker 3: Okay, So why do people need.
• 13:23
  Speaker 2: To spend this much money on clock code? I don't understand,
• 13:25
  Speaker 2: Like do people just have so much money that they
• 13:27
  Speaker 2: want to burn? I'm not really sure, Like it must
• 13:30
  Speaker 2: be like a case where people like this is in
• 13:33
  Speaker 2: the insane months of China watchers like that. I don't want
• 13:36
  Speaker 2: to name names, but like I'm like when they complain
• 13:39
  Speaker 2: about code being too expensive and how they use twenty
• 13:44
  Speaker 2: five dollars or just with one prompt, I said, well,
• 13:47
  Speaker 2: why don't you just use Kimmy cod It does everything?
• 13:50
  Speaker 3: You know.
• 13:52
  Speaker 2: I actually have a daily conversation with with Kimmy just
• 13:56
  Speaker 2: so that you can answer me technical questions I need
• 13:59
  Speaker 2: to do for work and for personal research.
• 14:04
  Speaker 1: Yeah, so I have a So let's let's let's talk
• 14:07
  Speaker 1: about hardware side of things, because there's there seems to
• 14:12
  Speaker 1: be an effort to double down chip restriction to China,
• 14:18
  Speaker 1: which is it's a little bit of conflicting because they
• 14:20
  Speaker 1: do want to sell like the water down uh uh
• 14:24
  Speaker 1: Nvidia chips to China, which China is not accepting. And
• 14:27
  Speaker 1: then there's, uh, there's talk about now they're going to
• 14:32
  Speaker 1: widen the restriction on AI chips available to even like Chinese,
• 14:40
  Speaker 1: you know, like like outside even even like not necessarily
• 14:43
  Speaker 1: in China. But any I don't know how that Again,
• 14:46
  Speaker 1: I don't know how they're gonna implement that, but but
• 14:50
  Speaker 1: I have seen that, you know, the people be floating
• 14:53
  Speaker 1: floating around kind of widened the the restriction chip access.
• 15:00
  Speaker 1: What's your take on that?
• 15:03
  Speaker 2: Well, this is also probably why Ali Baba and Tension
• 15:06
  Speaker 2: and also Byte Dance are all developing their own A
• 15:09
  Speaker 2: six or chips. In this case of Ali Baba, it
• 15:14
  Speaker 2: came out with a number recently. Basically it developed their
• 15:19
  Speaker 2: chip called for its personal use recently called m A ninety,
• 15:24
  Speaker 2: which it you know, produced about one hundred thousand in
• 15:27
  Speaker 2: the space of three months, so and also it buys
• 15:31
  Speaker 2: from other people too for its own purposes. And I
• 15:36
  Speaker 2: think that byte Dance is probably on the same track
• 15:39
  Speaker 2: where it's developing designing chips that are probably produced by Samsung.
• 15:45
  Speaker 2: Both of them use Samsung, I think because that way
• 15:49
  Speaker 2: also helps them to get the high bandwidth memories from
• 15:53
  Speaker 2: from Samsung. So the stuff that they're doing is not
• 15:57
  Speaker 2: as support control yet and.
• 16:01
  Speaker 3: It because.
• 16:04
  Speaker 2: They just fall under the expert control whatever. But in general,
• 16:09
  Speaker 2: by procuring more of it and using really fast network
• 16:13
  Speaker 2: better interconnectivity, they can build their own like clusters which
• 16:21
  Speaker 2: are sufficient for their own purposes. Like these guys are
• 16:27
  Speaker 2: sourcing them chips themselves. People think that China is just
• 16:32
  Speaker 2: about Huawei, but all these big techs in China, just
• 16:35
  Speaker 2: as Alphabet is doing with TPUs and Amazon is doing
• 16:39
  Speaker 2: with Trainium, they're all developing their own tips. Not everyone
• 16:44
  Speaker 2: needs and video chips for everything. In fact, as the
• 16:48
  Speaker 2: AI keep getting better, you can write using AI to
• 16:51
  Speaker 2: write better firmware for your chips to make it, you know,
• 16:55
  Speaker 2: be more useful in training and infests. And I think
• 17:01
  Speaker 2: probably over the next year or so, we're going to
• 17:04
  Speaker 2: see a huge step up work in availability of more
• 17:09
  Speaker 2: better ascent.
• 17:12
  Speaker 3: Compute clusters in China.
• 17:15
  Speaker 2: And this is because Huawei has finally got to the
• 17:18
  Speaker 2: point where the software part of it is you know.
• 17:22
  Speaker 3: More up to par.
• 17:24
  Speaker 2: It's been working very intimately with deep Seek and there
• 17:29
  Speaker 2: was a lot of report on like all the work
• 17:31
  Speaker 2: they did with deep Seak to improve the inference and
• 17:34
  Speaker 2: you know, post training on the deep Seek before there's
• 17:39
  Speaker 2: a lot of work done on the hardware side of things,
• 17:42
  Speaker 2: improving the chips. So what they're doing essentially is they're
• 17:47
  Speaker 2: saying each individual chips because the fact that our processes
• 17:52
  Speaker 2: have lower yields, so we need to make the chips
• 17:54
  Speaker 2: a little smaller, a little individual chips are a little
• 17:58
  Speaker 2: smaller and less powerful. But since we're so good at
• 18:01
  Speaker 2: you know, telecom technology and networking technology, we're going to
• 18:06
  Speaker 2: use the most advanced networking tech available to link together
• 18:11
  Speaker 2: our supercomputers and so the so the AI super Trip
• 18:18
  Speaker 2: as they would call it, are really gonna come out
• 18:22
  Speaker 2: in later this year and also uh, you know later
• 18:26
  Speaker 2: in twenty twenty seven. And you can see with their
• 18:29
  Speaker 2: own you know, advertising brochure that they that they believe
• 18:34
  Speaker 2: that the by twenty thirty they don't have these like
• 18:37
  Speaker 2: humongous super tips. Uh, that's probably gonna be sufficient for
• 18:42
  Speaker 2: any kind of compute you need.
• 18:44
  Speaker 3: But obviously the question is.
• 18:47
  Speaker 2: That now the question for many people online is is
• 18:50
  Speaker 2: why we're going to get there before? Like the gat before,
• 18:54
  Speaker 2: Like the American AI companies have like a huge lead
• 18:59
  Speaker 2: over China, these AI labs due to the computer advantage.
• 19:02
  Speaker 2: I tend to doubt it. I tend to think that,
• 19:05
  Speaker 2: you know, the semiconductor export controls are already proving to
• 19:14
  Speaker 2: be a real failure. And uh, at this point, I
• 19:19
  Speaker 2: think people a lot of people are just talking themselves
• 19:22
  Speaker 2: into thinking that, uh, that this has been this great
• 19:26
  Speaker 2: success in limiting Chinese AI development when there's actually no
• 19:31
  Speaker 2: sign of that.
• 19:33
  Speaker 1: So when before well, when deep Seek v four was
• 19:37
  Speaker 1: coming out, I hear some internet chatters saying, oh, the
• 19:41
  Speaker 1: reason deep Sea v four was delayed or takes so
• 19:45
  Speaker 1: long to come out is because they were forced by
• 19:48
  Speaker 1: Chinese government to use the Huawei Ascent chips. Uh that
• 19:52
  Speaker 1: that's that's what caused it delay? Now is that is
• 19:55
  Speaker 1: that just bs? I mean, like did that really impact
• 20:00
  Speaker 1: or do Deep Seat development or do you think that's
• 20:03
  Speaker 1: just you know, just just just be stock.
• 20:07
  Speaker 2: I mean, I don't know why Deep Seas needs to
• 20:09
  Speaker 2: be forced into it, since, like they they lack resource
• 20:15
  Speaker 2: and they also lack compute. So if they can get
• 20:19
  Speaker 2: if they can work with Huawei and get access, they
• 20:22
  Speaker 2: can work with that you know AI ship provider intimately
• 20:26
  Speaker 2: and you know, get the compute they need. That seems
• 20:31
  Speaker 2: to be a huge plus to me. So that I
• 20:33
  Speaker 2: do think there was maybe some of that took some
• 20:36
  Speaker 2: of bandwidth away, But the bigger issue is like from
• 20:40
  Speaker 2: V three to V four they went from around six
• 20:42
  Speaker 2: hundred to seven hundred billion parameters to one point six
• 20:45
  Speaker 2: trillion parameters and then they needed to do our new architectures.
• 20:49
  Speaker 3: There was a lot of research involved, right, So the
• 20:54
  Speaker 3: model itself, if you look at the large version of.
• 20:57
  Speaker 2: Their model, it it looks still very under train at
• 21:01
  Speaker 2: the moment. Like it looks, I think it can get
• 21:03
  Speaker 2: a lot better. The I tend to think that their
• 21:08
  Speaker 2: models will be based on this before and it will
• 21:11
  Speaker 2: continue to improve quite a bit over the rest of
• 21:15
  Speaker 2: this year, and all their models released this year will
• 21:18
  Speaker 2: probably be around one point six to one point seven
• 21:20
  Speaker 2: trillion parameters.
• 21:22
  Speaker 1: And yeah, well Huawei, I think was Huawei a top
• 21:29
  Speaker 1: Honk Chull just recently, I think it was in debate.
• 21:34
  Speaker 1: Uh that just congratulates because basically just thanked US chip
• 21:40
  Speaker 1: sanction on Huawei's achievement. Does this mean Huawei has reached
• 21:47
  Speaker 1: escape velocity? You know, you think they are well on
• 21:51
  Speaker 1: the way to achieve a self sufficiency in this my
• 21:56
  Speaker 1: conductor supply chain.
• 21:58
  Speaker 2: Well, you know, while Way had always said twenty twenty
• 22:01
  Speaker 2: six was going to be the year for something, right,
• 22:05
  Speaker 2: so they had a plan and they had to work
• 22:10
  Speaker 2: through these plants with engineering excellence at each steps. And
• 22:16
  Speaker 2: now that they got their own semiconductor equipment supply chain
• 22:21
  Speaker 2: and also you know modern I guess just just modern
• 22:27
  Speaker 2: machinery and metrology supply chain worked out, and they have
• 22:32
  Speaker 2: their chip design figure it out. It's there's a clear
• 22:36
  Speaker 2: path to word them having competitive chip manufacturing, chip design,
• 22:42
  Speaker 2: chip equipment manufacturing, and ramping up production. It might not
• 22:49
  Speaker 2: move as fast as some people would hope, because you know,
• 22:53
  Speaker 2: it takes time to build the factories to produce these things.
• 22:57
  Speaker 3: So production this.
• 22:59
  Speaker 2: Year might be you know up to I don't know,
• 23:03
  Speaker 2: forty to fifty k way first for months. Maybe next
• 23:07
  Speaker 2: year will be like one hundred and the year after
• 23:09
  Speaker 2: that would one hundred and fifty. So like when people
• 23:11
  Speaker 2: talk about like why it's China not fully selfish sufficient,
• 23:14
  Speaker 2: it's because it takes time and money to build these things,
• 23:17
  Speaker 2: right like, so so that's why it sees itself And
• 23:24
  Speaker 2: I agree with you when you say that escape velocity
• 23:26
  Speaker 2: because it sees the past. So that's why it's celebrating
• 23:30
  Speaker 2: right now. Right But you know, if it hadn't reached
• 23:34
  Speaker 2: this point, it wouldn't be doing so much public unveiling
• 23:38
  Speaker 2: of these things. You know, it kept its mouth shut
• 23:41
  Speaker 2: for many years because it was working on all these technologies.
• 23:46
  Speaker 2: So you know, for a lot of these Western semi
• 23:49
  Speaker 2: experts or mostly most of them are on the internet
• 23:52
  Speaker 2: that just dismissed it, or even Jensen flung himself dismissing
• 23:56
  Speaker 2: it as like, oh, ten years behind TSMC, they are not.
• 24:02
  Speaker 2: Either they're just trying to be protect their own product
• 24:08
  Speaker 2: and not want to deal with reality, or they haven't
• 24:13
  Speaker 2: looked deeply into what's actually happening. Because what Huawei is
• 24:18
  Speaker 2: doing is like a tighter integration in terms of chip
• 24:22
  Speaker 2: production than anything that TSMC has done. Like TSMC cannot
• 24:28
  Speaker 2: compete against Huawei in terms of full stack engineering. TSMC
• 24:33
  Speaker 2: is a great fab right, but it's not. It doesn't
• 24:36
  Speaker 2: make his own equipment. It's not great at firmware, it's
• 24:41
  Speaker 2: not great a software. It doesn't do the things that
• 24:45
  Speaker 2: Huawei does. It's not you know, excellent a telecom. So
• 24:50
  Speaker 2: you cannot possibly do this thing that Huawei do. So
• 24:52
  Speaker 2: you cannot escape expect TSMC to be better a Huawei
• 24:57
  Speaker 2: at the stuff because it's just not.
• 24:59
  Speaker 3: A reason no ord request.
• 25:02
  Speaker 2: And it's that's gonna be very hard for a lot
• 25:04
  Speaker 2: of Taiwanese and Koreans to take when I tell them
• 25:07
  Speaker 2: this stuff, because they worship their their the homegrown champions
• 25:11
  Speaker 2: that you know that like.
• 25:12
  Speaker 3: Did really well in this AI chip thing, right.
• 25:19
  Speaker 1: Well, but does Quawai still rely on the u vs
• 25:23
  Speaker 1: from ASML.
• 25:26
  Speaker 3: Well, I guess that's kind of the thing. You don't.
• 25:28
  Speaker 2: What they're saying is you don't need the uvs to
• 25:32
  Speaker 2: to make the advancements, like uh. Part of the thing
• 25:36
  Speaker 2: is because they're saying they're using this called chip voting. Now,
• 25:39
  Speaker 2: one of the common thing common problems when it comes
• 25:42
  Speaker 2: to UH the Morse law is as transistor shrink. Right,
• 25:48
  Speaker 2: at a certain point, you get to the point where
• 25:50
  Speaker 2: the transistors are so small for the logic unit that
• 25:57
  Speaker 2: you're dealing with, you're dealing with certain physical boundaries that
• 26:03
  Speaker 2: are harder to reconcile with. So on, one problem is
• 26:06
  Speaker 2: what we call s RAMP shrinking. So when you're if
• 26:10
  Speaker 2: you have your typical CPU, it not only has CPU
• 26:14
  Speaker 2: also has caches cashing off memory caches, right, and that
• 26:18
  Speaker 2: uses s ramp at a certain point basically around five
• 26:22
  Speaker 2: nanometer tsmc, some s RAM just cell sizes are not
• 26:28
  Speaker 2: shrinking at the moment at all. The analog features in
• 26:32
  Speaker 2: your chips aren't shrinking right now, like your copper wirings
• 26:36
  Speaker 2: are not going to shrink at a certain point, right
• 26:38
  Speaker 2: like you have all your wirings, All those features don't shrink.
• 26:43
  Speaker 2: The only strength things you can shrink is your transistors itself.
• 26:47
  Speaker 2: So at a certain point, at a certain point, even
• 26:52
  Speaker 2: if you shrink your transistors by fifty percent, the rest
• 26:57
  Speaker 2: of your chip cannot shrink by the same move. So
• 27:01
  Speaker 2: then the question that why we're trying to raise was
• 27:04
  Speaker 2: how do you improve continue to improve your chip performance
• 27:09
  Speaker 2: without you know, while dealing with this stuff and its
• 27:14
  Speaker 2: methodology is we're gonna go from two D to three
• 27:18
  Speaker 2: D we're gonna find, you know, the.
• 27:21
  Speaker 3: Most optimal pass.
• 27:23
  Speaker 2: And what it's doing is is what what it calls
• 27:28
  Speaker 2: what they call logic folding, which is different from the
• 27:31
  Speaker 2: current three D stacking because it is putting multiple like
• 27:36
  Speaker 2: two or three layers away first together and because it's
• 27:40
  Speaker 2: going three D, it is able to stack them so
• 27:45
  Speaker 2: that the UH so that going from one core, one
• 27:51
  Speaker 2: CPU core to another CPU core is a much a short,
• 27:55
  Speaker 2: much shorter pass if you stack them on top of
• 27:57
  Speaker 2: each other, right, and the same with know CPU cors
• 28:01
  Speaker 2: talk to to to the to your as ramp cache.
• 28:03
  Speaker 3: Like if you reduce the.
• 28:04
  Speaker 2: Amount of distance between that by putting on top instead
• 28:07
  Speaker 2: of like somewhere else on the two D board, that's
• 28:10
  Speaker 2: less shorter path needs to travel. That generates less heat. Also,
• 28:17
  Speaker 2: it used to elect less electricity and that increases the
• 28:21
  Speaker 2: performance of the chips. So that's how it's able to
• 28:25
  Speaker 2: raise its clock speed from two point six gigaherts. Eventually
• 28:31
  Speaker 2: its goal is to reach five gigaherts in a few
• 28:34
  Speaker 2: years because it can just stack more way fast together.
• 28:38
  Speaker 2: The main challenge involved is, you know, we have to
• 28:41
  Speaker 2: make some certain advancements in UH. The software designed the
• 28:45
  Speaker 2: e ED as how we design chips. We have to
• 28:48
  Speaker 2: deal with concerns of how to if we generate more
• 28:53
  Speaker 2: heat from having extra layers, how do we get rid
• 28:56
  Speaker 2: of the heat. So those are the kind of challenges
• 28:57
  Speaker 2: that while we're dealing with right now. But just by
• 29:01
  Speaker 2: voting it alone and by designing chips that possibly have
• 29:07
  Speaker 2: cores that occupy multiple levels.
• 29:09
  Speaker 3: You know, you have a top level of your.
• 29:12
  Speaker 2: CPU core and a bottom levels if you CEP core,
• 29:15
  Speaker 2: that's that's kind of connected together.
• 29:17
  Speaker 3: That actually already.
• 29:21
  Speaker 2: Improve the performance so much versus just staying with two
• 29:26
  Speaker 2: D and keep shrinking your transistor further and further.
• 29:31
  Speaker 1: That's great to hear. So it looks like there are
• 29:35
  Speaker 1: you know, they have a path they figure out. What
• 29:41
  Speaker 1: about what what other advance bind would you like to
• 29:46
  Speaker 1: talk about recent progress China made on the tech front.
• 29:52
  Speaker 2: Well, in the the the hard thing is, it's very
• 29:57
  Speaker 2: interesting to see them just jumping ahead in so many
• 30:00
  Speaker 2: areas and the just the entire AI supply chain, so
• 30:04
  Speaker 2: it doesn't stay was just you know, the the chips itself.
• 30:12
  Speaker 2: They made a lot of progress in memory, in the
• 30:14
  Speaker 2: memory chips also, they made a lot of progress in
• 30:17
  Speaker 2: the fiber optics also, the transceivers. They also made a
• 30:23
  Speaker 2: lot of progress in PCBs. So in your AI data centers,
• 30:28
  Speaker 2: each part of your AI data centers needs a board.
• 30:31
  Speaker 2: The control stuff if you imagine like you know, inside
• 30:35
  Speaker 2: your phone, inside your any kind of things you have
• 30:38
  Speaker 2: that has electronics in it in there, it needs a board.
• 30:41
  Speaker 2: It needs to deal with cooling, it needs to deal
• 30:43
  Speaker 2: with power draw and things like that. So it's made
• 30:47
  Speaker 2: a lot of progressing advanced PCBs. And also it's made
• 30:54
  Speaker 2: a lot of advancement in you know, the fiber optic
• 30:59
  Speaker 2: networks that uh communication so so so this allows you
• 31:03
  Speaker 2: to have faster communication between chips, between network notes and.
• 31:08
  Speaker 3: Things like that.
• 31:10
  Speaker 2: There's this other thing that's come out recently called m
• 31:13
  Speaker 2: l CC. That's uh, that's the industry that was you know,
• 31:17
  Speaker 2: dominated by the Japanese, especially a company called Marada, and
• 31:23
  Speaker 2: they're making they're they're catching up to the tech with
• 31:26
  Speaker 2: Marada also recently, although I think m l CC is
• 31:29
  Speaker 2: something that's going to take a little longer. But yeah,
• 31:33
  Speaker 2: so there's so, there's so, there's there's In terms of
• 31:36
  Speaker 2: the AI supply chain, it's making progress on every part
• 31:41
  Speaker 2: of the supply chain. So it's very interesting to see
• 31:43
  Speaker 2: what they're doing in you know, in terms of deployment,
• 31:52
  Speaker 2: the embodied AI deployment in China is probably the most impressive.
• 31:57
  Speaker 2: You're seeing like a lot of these rope body stuff
• 32:01
  Speaker 2: coming out of China now that were they're starting to
• 32:03
  Speaker 2: make into everyday life. And the I saw CCTV report
• 32:09
  Speaker 2: recently about AI toys being like a huge growth engine
• 32:13
  Speaker 2: inside China, and that makes a lot of sense to
• 32:16
  Speaker 2: me because you know, they have the full supply chain
• 32:19
  Speaker 2: over there. They're making the best models so for the
• 32:24
  Speaker 2: for these things, so it's kind of hard for anyone
• 32:27
  Speaker 2: else to really compete there.
• 32:31
  Speaker 1: So yeah, I was personally surprised by the kind of
• 32:35
  Speaker 1: the proliferation of these robots. I thought you would take
• 32:39
  Speaker 1: a lot longer, you know, for them to reach kind
• 32:43
  Speaker 1: of the mass level. But what you know, Okay, now,
• 32:47
  Speaker 1: now I have a question. As impressive as these robots are,
• 32:53
  Speaker 1: do they have any real applications other than being some
• 32:57
  Speaker 1: kind of curiosity toy, because what's a boy to having
• 33:02
  Speaker 1: like humanoid robots?
• 33:05
  Speaker 2: Yeah, so I guess that's still trying to find applications
• 33:09
  Speaker 2: at the moment. The initial applications typically inside the factories
• 33:14
  Speaker 2: themselves and also maybe as some dealerships where they they're
• 33:17
  Speaker 2: kind of just greeting you in front of the store
• 33:19
  Speaker 2: as a novelty factor, right. But by D kind of
• 33:24
  Speaker 2: talked about this recently. It's said that if it has
• 33:27
  Speaker 2: more humanoid robot in its stores, it doesn't need to
• 33:30
  Speaker 2: hire as many people, especially in places like Europe where
• 33:33
  Speaker 2: it's really expensive to hire people.
• 33:37
  Speaker 3: And I know, not alarming to a lot of.
• 33:39
  Speaker 2: People because it's taking away human jobs, but that is
• 33:43
  Speaker 2: kind of like one part of the human robots. The
• 33:48
  Speaker 2: other part is actually probably more helpful, which is home
• 33:51
  Speaker 2: service robots. So instead of having just you know, your
• 33:56
  Speaker 2: rombot at home, I robot at home that goes around
• 33:59
  Speaker 2: clean stuff, although in China it's mostly dreaming by now,
• 34:03
  Speaker 2: and I think robo rock is another big one.
• 34:06
  Speaker 3: There's a couple of companies that are really big with
• 34:08
  Speaker 3: that stuff.
• 34:10
  Speaker 2: But instead of having that, you're actually having a humoid
• 34:13
  Speaker 2: robots that walks around your home, fold your clothes and
• 34:17
  Speaker 2: do some basic dishes and things like that. From what
• 34:20
  Speaker 2: I hear, they're very awkward so far. At this point,
• 34:24
  Speaker 2: they're still not very good at this job. So they're
• 34:27
  Speaker 2: still testing this out and people are still just getting
• 34:30
  Speaker 2: them more for an opity factor than anything else. But
• 34:32
  Speaker 2: you could see like a longer term having humanoid robots
• 34:35
  Speaker 2: at homes to take care of elderlies and also to
• 34:39
  Speaker 2: cling stuff for them that you know that that actually
• 34:44
  Speaker 2: is a big help if you have a society where
• 34:47
  Speaker 2: young people don't or you don't have as many people
• 34:50
  Speaker 2: from the countryside coming to the city to you know,
• 34:54
  Speaker 2: serving those roles or young people who are willing to
• 34:57
  Speaker 2: do like caretaker account of roles.
• 34:59
  Speaker 1: Right or you or you don't want to open immigration
• 35:03
  Speaker 1: like some countries you know they just import, if they
• 35:07
  Speaker 1: just import domestic helpful like Southeast Asia, right, well, you
• 35:12
  Speaker 1: know it creates a problem of its own, but yeah,
• 35:15
  Speaker 1: go ahead.
• 35:15
  Speaker 2: But you know, like China is such a big country
• 35:17
  Speaker 2: that even if it imports like some people from Vietnam
• 35:22
  Speaker 2: and blows right, like, that's still it's like a peanuts
• 35:25
  Speaker 2: compared to like the Chinese population. So you like you
• 35:30
  Speaker 2: can do that in Singapore, but Singapore has six million people.
• 35:35
  Speaker 1: Yeah, exactly. Now, I do want to switch a track
• 35:40
  Speaker 1: a little bit to talk about the Chinese space program
• 35:44
  Speaker 1: unless you have something else you want to cover.
• 35:46
  Speaker 3: No, no, no, space program is way is way more.
• 35:49
  Speaker 1: Interesting, Okay, perfect, So right now we just had the
• 35:55
  Speaker 1: we just have the Blue Origin rocket, New Glen rocket
• 35:58
  Speaker 1: Blue up and this this, I think this is a
• 36:03
  Speaker 1: perfect opportunity to talk about the space race between US
• 36:08
  Speaker 1: and China. You know, both have now have a Moon program,
• 36:14
  Speaker 1: a moon landing program, and one area though China is
• 36:21
  Speaker 1: still lacking behind in the United States in space is
• 36:26
  Speaker 1: recoverable rockets. Right like, this is something that both space
• 36:32
  Speaker 1: x has done very well, and uh, you know Blue
• 36:35
  Speaker 1: Origin is starting to do that. They were able to
• 36:38
  Speaker 1: recover their rocket even though they weren't able to put
• 36:42
  Speaker 1: their load into into orbit. But still you know, China
• 36:48
  Speaker 1: hasn't is not quite there yet. And I just saw
• 36:51
  Speaker 1: a YouTube video actually talking about this. He said space
• 36:55
  Speaker 1: x is have s built up such a lead and
• 36:59
  Speaker 1: you'll be difficult for China to catch up because the
• 37:02
  Speaker 1: reason they gave is, you know, Elon Musk's plan is
• 37:06
  Speaker 1: too because the lower orbit is limited, the space is limited.
• 37:12
  Speaker 1: What Elon Musk is planned is just putting up put
• 37:17
  Speaker 1: up as much satellites as possible to occupy all the space.
• 37:22
  Speaker 1: So by the time China, you know, I don't know
• 37:25
  Speaker 1: how how long it will take. China to catch up.
• 37:28
  Speaker 1: You know, five ten years, there might not be enough
• 37:31
  Speaker 1: space left for you know, the Chinese equivalent to compete
• 37:37
  Speaker 1: with UH with space space X. What do you say
• 37:43
  Speaker 1: to that.
• 37:45
  Speaker 2: Yeah, So that last part has always seen a hypothesis
• 37:49
  Speaker 2: for a while that we are limited in the you know,
• 37:55
  Speaker 2: lower Earth orbit we call LEO by about us under
• 38:00
  Speaker 2: thousand or I don't know, people say that we're we
• 38:03
  Speaker 2: can have most one hundred thousand satellites up there. You know,
• 38:07
  Speaker 2: we'll see, we'll see, like that's still hypothesis because we've
• 38:11
  Speaker 2: seen both China and Chinese satellite providers and Eland Musk
• 38:17
  Speaker 2: himself apply for like two hundred thousand or a million
• 38:21
  Speaker 2: satellite constellations up there. You know, I tend to think
• 38:27
  Speaker 2: that China did a privatize certain part of the space
• 38:33
  Speaker 2: race space industry a little too late, and it hasn't
• 38:37
  Speaker 2: moved fast enough in some of these areas. So we
• 38:40
  Speaker 2: actually saw them attempt to reusable rocky landings in the
• 38:45
  Speaker 2: past six months and they both failed.
• 38:47
  Speaker 3: So this is this just shows like space industry.
• 38:50
  Speaker 2: Especially if you're trying to do something new, it's still
• 38:53
  Speaker 2: very complicated, right, and there's a there's multiple components here.
• 38:57
  Speaker 2: There is the entire supply chain, So you need to
• 39:00
  Speaker 2: produce the rockets, you need to produce the the rock
• 39:05
  Speaker 2: the engines, and also you need the launch sites, you
• 39:09
  Speaker 2: need the satellite superfactories also, so there's a lot of
• 39:14
  Speaker 2: components involved in there, and they're building up all of that,
• 39:17
  Speaker 2: you know, like they they're building this new commercial space
• 39:21
  Speaker 2: launch center now in hein end dollars and it's you know,
• 39:26
  Speaker 2: it's still ramping up to its launch capacity, right and
• 39:29
  Speaker 2: they have these ambitious plans where they're looking to launch
• 39:34
  Speaker 2: and create their own constellations.
• 39:36
  Speaker 3: In the LEO satellites right now, they're.
• 39:40
  Speaker 2: Probably have about three hundred satellites in there between two constellations,
• 39:44
  Speaker 2: whereas you know space X has maybe a few thousands.
• 39:50
  Speaker 2: I don't know, it's under ten salands between five and
• 39:52
  Speaker 2: ten thousand satellites in there. But you know, over the
• 39:56
  Speaker 2: next couple of years, the Chinese constellation are going to
• 39:58
  Speaker 2: increase get a lot lotter in size. I tend to
• 40:03
  Speaker 2: think that this is actually one area where the Chinese
• 40:08
  Speaker 2: military is really kind of admiring space X at the
• 40:11
  Speaker 2: moment because they do value the constellations. Constellation in space
• 40:19
  Speaker 2: is a saying it's I just don't see any circumstance
• 40:25
  Speaker 2: where they you know, they will stop launching.
• 40:29
  Speaker 3: Stuff into the space.
• 40:30
  Speaker 2: Right, Like there there is plans to have at least
• 40:37
  Speaker 2: two or three ten thousand satellite constellation amongst Chinese companies
• 40:43
  Speaker 2: in space. Right, they have a path to get there,
• 40:48
  Speaker 2: so it may be a little delayed, but it's actually.
• 40:57
  Speaker 3: They have.
• 40:59
  Speaker 2: They have supply chain that's getting built up right now
• 41:02
  Speaker 2: to get there. So I think maybe most these things
• 41:06
  Speaker 2: get delayed one or two years.
• 41:07
  Speaker 3: But that's about it.
• 41:09
  Speaker 2: And space X itself, you know, has also challenges because
• 41:15
  Speaker 2: Elon likes to boast a lot.
• 41:18
  Speaker 3: He assumes things. He's very I don't know, he's he's.
• 41:22
  Speaker 2: He's very white in that like he's he's always thinks
• 41:24
  Speaker 2: that things would just go his way, you.
• 41:26
  Speaker 3: Know, Like I don't know, the guy is just very outwardly.
• 41:30
  Speaker 2: Confident about everything that makes no sense to me. Like
• 41:33
  Speaker 2: for example, like recently he talks about how he wants
• 41:36
  Speaker 2: to make his own space data centers and how he's
• 41:40
  Speaker 2: going to produce these great solar factories. But then he
• 41:43
  Speaker 2: complains he went to China with Trump because China would
• 41:46
  Speaker 2: not export these like solar panel equipments to him. And
• 41:51
  Speaker 2: my question is like, why would China, why would trying
• 41:54
  Speaker 2: to export all these leading edge solar panel equipments to you?
• 41:59
  Speaker 3: If if if your goal is to.
• 42:01
  Speaker 2: Stop China from having its own life space constellations, right,
• 42:06
  Speaker 2: and the world is not made just to make your
• 42:09
  Speaker 2: life easier to help you become a trillion aer faster, right,
• 42:12
  Speaker 2: Like his entrd thing when he talks about this is
• 42:15
  Speaker 2: just to get people to give him more money. Basically,
• 42:18
  Speaker 2: if you look at these space X space X like
• 42:24
  Speaker 2: i PO filings, this is not a company that should
• 42:27
  Speaker 2: be valuated where he wants it to be value that right,
• 42:30
  Speaker 2: So he has to keep talking it up all these
• 42:32
  Speaker 2: all these things. What I do see was what the
• 42:36
  Speaker 2: Chinese producers have done recently, which is quite interesting, is
• 42:40
  Speaker 2: that they're they're testing out these really advanced sore cell
• 42:45
  Speaker 2: technologies in space with their space station launches, and they
• 42:51
  Speaker 2: can also test out the latest battery technology up there.
• 42:54
  Speaker 2: So that actually will help a lot with not just
• 42:59
  Speaker 2: you know, satellites, but also lunar.
• 43:03
  Speaker 3: Lunar colony in the future.
• 43:05
  Speaker 2: So I think you know, China probably will get it's
• 43:11
  Speaker 2: moon landing by twenty thirty, that it plan to and
• 43:14
  Speaker 2: that it will be able to prevent, you know, build
• 43:17
  Speaker 2: like a lunar colony over the space of maybe ten
• 43:20
  Speaker 2: years over there. And I'm not convinced that America's Artemis
• 43:25
  Speaker 2: program will be around in ten fifteen years.
• 43:28
  Speaker 1: So but but I'm going to push back on you TV.
• 43:33
  Speaker 4: Even with the Blue Origin rocket blowing up, the newly
• 43:38
  Speaker 4: revised Artemist plan for moon landing human moon landing, it's
• 43:43
  Speaker 4: still ahead of China's plan, which is I think artemists
• 43:48
  Speaker 4: supplant was what like twenty twenty eight or something like
• 43:51
  Speaker 4: China's is twenty thirty, right.
• 43:56
  Speaker 2: Yeah, so I think I think there's a good chance
• 43:59
  Speaker 2: for that US does get to move ahead of China.
• 44:02
  Speaker 2: But you know, like I don't that the Chinese perspective,
• 44:07
  Speaker 2: from what I can see, is that they never expected
• 44:11
  Speaker 2: to be in a space race with America in this
• 44:14
  Speaker 2: This is more like a manufacturer to buy NAZA and
• 44:19
  Speaker 2: people in the US elite circles to get funding for NASA, right, So,
• 44:25
  Speaker 2: you know, that's kind of what it is. I tend
• 44:27
  Speaker 2: to think it's the new guy, JAREDI is, the committed
• 44:30
  Speaker 2: is a lot more realistic. With the NASA's launch schedule
• 44:33
  Speaker 2: than the previous NASA heads.
• 44:37
  Speaker 3: So maybe this.
• 44:39
  Speaker 2: Gives you as a better chance of getting to move
• 44:41
  Speaker 2: by twenty twenty eight, but.
• 44:45
  Speaker 3: It is also just as likely that.
• 44:49
  Speaker 2: The delays that we've seen so far continue all the
• 44:52
  Speaker 2: way to twenty thirties. So I think it's probably like
• 44:58
  Speaker 2: fifty to fifty right now who gets their first. But
• 45:01
  Speaker 2: I don't think it really matters that much because a
• 45:06
  Speaker 2: longer term, it's more about like what you can do
• 45:10
  Speaker 2: once you get there, and the technologies that China has
• 45:14
  Speaker 2: developed that are usable in space, like automous vehicles, you know,
• 45:20
  Speaker 2: the electric vehicles and solar panel technology that they have
• 45:23
  Speaker 2: there is going to allow it to do a lot
• 45:26
  Speaker 2: more once you land these things in space.
• 45:30
  Speaker 1: That's fair, and I think I agree with you. I
• 45:33
  Speaker 1: think the whole idea of a race to the Moon,
• 45:36
  Speaker 1: I think it's more kind of hyped up in American
• 45:41
  Speaker 1: media and American consciousness because now, like the United States
• 45:45
  Speaker 1: have just had the China Derangements syndrome, everything has to
• 45:48
  Speaker 1: peace the competition with China. Whereas the Chinese people, they
• 45:53
  Speaker 1: won't just be happy if they get to the book.
• 45:55
  Speaker 1: I mean, like their goal was never to compete with
• 45:59
  Speaker 1: the United Space to see who get to the Moon first.
• 46:02
  Speaker 1: You know, they will be happy, just they will be
• 46:04
  Speaker 1: put a Chinese astronaut, Chinese tycoon not into the moon
• 46:09
  Speaker 1: and start to build the Chinese Moon space Chinese moon station.
• 46:13
  Speaker 3: Yeah.
• 46:14
  Speaker 2: Yeah, I just want to make sure to explain to
• 46:16
  Speaker 2: people that the moon race itself is not critical to
• 46:21
  Speaker 2: national security or anything like that. Right in terms of
• 46:24
  Speaker 2: space race, the one that actually does care about is
• 46:27
  Speaker 2: its constellations in space.
• 46:30
  Speaker 3: We've seen what.
• 46:30
  Speaker 2: Happens in Iran when you have a satellite in the space, right,
• 46:34
  Speaker 2: we saw what happened in the India Pakistan battle where
• 46:38
  Speaker 2: these countries are relying on Chinese satellites to hit, to
• 46:42
  Speaker 2: see targets, to get electronic intelligence, and to actually target them.
• 46:49
  Speaker 2: Like without the Chinese satellites, the results will be entirely
• 46:53
  Speaker 2: different naturally.
• 46:54
  Speaker 3: So it's it's very it's very.
• 46:57
  Speaker 2: Telling the importance of satellite in national security from that perspective.
• 47:03
  Speaker 2: And then they spent a lot.
• 47:04
  Speaker 3: Of money building BEATO.
• 47:05
  Speaker 2: They tend to spend a lot of money building these
• 47:08
  Speaker 2: high powered, high resolution imagery satellites and also electronic intelligence
• 47:14
  Speaker 2: satellites in space.
• 47:16
  Speaker 1: So how do you see China close the gap, uh
• 47:20
  Speaker 1: with the Space acts because right now there is a
• 47:23
  Speaker 1: there's a space ACX did develop a significant, significant lead
• 47:28
  Speaker 1: in terms of reuseful rockets and you know, cheaper launch
• 47:32
  Speaker 1: it launch vehicles and stuff like that.
• 47:35
  Speaker 2: Yeah, so I think fortunately, you know, there's only so
• 47:38
  Speaker 2: much launch capacity that you have as a country. Right
• 47:42
  Speaker 2: once you launch something, you it doesn't become available to
• 47:46
  Speaker 2: launch again for like a few days that launch area.
• 47:50
  Speaker 2: So it's just a lot of investment in both launch sites,
• 47:55
  Speaker 2: larger rockets, and factories that produced satellites and things like that.
• 48:02
  Speaker 2: And it is entirely possible that you know, space X
• 48:08
  Speaker 2: will always have more satellites in space than the Chinese.
• 48:11
  Speaker 2: Chinese companies like space Sale, right, but that remains to
• 48:15
  Speaker 2: be seen. I think the Chinese ones are just saying,
• 48:18
  Speaker 2: if we get once we get our thirteen thousand or
• 48:21
  Speaker 2: fifty thousand satellite constellation in space, that's enough for us.
• 48:25
  Speaker 2: You know, that's enough for everything that you need. You
• 48:28
  Speaker 2: can have constant flyovers of satellites, right. I fully expect
• 48:33
  Speaker 2: both US and China have a constellation that they need
• 48:37
  Speaker 2: for you know, national security and civilian utilization purposes.
• 48:44
  Speaker 3: The kind of the.
• 48:45
  Speaker 2: Question is like Elon Comes constantly spends all this money
• 48:49
  Speaker 2: that are not his. Really, like, how long can this
• 48:53
  Speaker 2: keep up? You know, how long can Silicon Valley billionaires
• 48:56
  Speaker 2: keep throwing money at him just and just hyping up
• 48:59
  Speaker 2: his docs when you know, at a certain point, you know,
• 49:03
  Speaker 2: the Middle Eastern Southern well stones are stopped. Are not
• 49:06
  Speaker 2: going to be giving money to these guys anymore because
• 49:09
  Speaker 2: you know, you are going to get kicked out of
• 49:11
  Speaker 2: the Middle East? So like, how long can this money
• 49:14
  Speaker 2: keep following in? I don't I don't really know the
• 49:16
  Speaker 2: answer to that.
• 49:18
  Speaker 1: I just want I just want a Chinese competition to
• 49:22
  Speaker 1: uh Starling, you know, like because because starlink do provide
• 49:27
  Speaker 1: a valuable service to areas with you know, like sparse
• 49:33
  Speaker 1: internet coverage and and I you know, competition is good.
• 49:38
  Speaker 1: If China can enter that game, you will help to
• 49:42
  Speaker 1: bring the price down for everybody. So that That's what
• 49:44
  Speaker 1: I'm waiting for.
• 49:46
  Speaker 2: And just keep thinking thinking about this that Starlink doesn't
• 49:50
  Speaker 2: make money, SpaceX doesn't make money, and Starling itself probably
• 49:54
  Speaker 2: the biggest user of the US military and all these
• 49:56
  Speaker 2: other military Like where would Ukrainian military be without starlings?
• 50:00
  Speaker 2: You know, like what happens with SpaceX's Uh income, once
• 50:05
  Speaker 2: did Ukraine work comes to stop?
• 50:08
  Speaker 3: I'd be asking myself that question, you know.
• 50:11
  Speaker 2: So it feels like SpaceX is this giant thing where
• 50:14
  Speaker 2: the US military is just throwing money at it to
• 50:18
  Speaker 2: keep it afloat.
• 50:22
  Speaker 1: But they can't throw a lot of money for a
• 50:26
  Speaker 1: long time because US government can create the money supply
• 50:34
  Speaker 1: for a long time.
• 50:35
  Speaker 2: This is why I don't expect space X to go
• 50:37
  Speaker 2: away anytime soon. Right, This is why, like I think
• 50:42
  Speaker 2: this the Starlings itself is going to be a viable
• 50:46
  Speaker 2: for a long long time. I don't expect it to
• 50:49
  Speaker 2: go anywhere. Uh.
• 50:50
  Speaker 3: And you know, I think that this the entire.
• 50:57
  Speaker 2: Part about offering Internet in remote places. Uh, that's what
• 51:06
  Speaker 2: you know these people talk about through the outside world.
• 51:09
  Speaker 2: But the actual usage for a lot of these constellations
• 51:13
  Speaker 2: are just for military purposes, national security purposes. Yeah. Although
• 51:21
  Speaker 2: the Chinese, if you look at the Huawei model, they
• 51:24
  Speaker 2: do say that they believe six for six G, the
• 51:28
  Speaker 2: the space constellations is going to be a huge part
• 51:31
  Speaker 2: of it.
• 51:32
  Speaker 3: What we don't know it.
• 51:34
  Speaker 2: But the but the thing is that the Space America
• 51:37
  Speaker 2: doesn't have a six G company, right, So it's like,
• 51:42
  Speaker 2: you know, space X doesn't have much that much commercial
• 51:45
  Speaker 2: values different it's starting network, so I think it's revenue
• 51:50
  Speaker 2: stream will be very military based for a long time.
• 51:54
  Speaker 1: I don't even know how much find you coverage is
• 51:57
  Speaker 1: currently in US. You know, I interview guests from over
• 52:01
  Speaker 1: the world, and I can tell about their Internet speed
• 52:04
  Speaker 1: by how fast their video upload to stream yard and
• 52:09
  Speaker 1: I have a couple of guests whose upload speed is
• 52:13
  Speaker 1: slower than mot in Bali, which is which is crazy?
• 52:17
  Speaker 1: Which is really crazy?
• 52:20
  Speaker 2: Yeah, I can't I can't disagree with that one. And
• 52:25
  Speaker 2: you know, I live in the East coast of the
• 52:27
  Speaker 2: US where there's too many people and using it and
• 52:30
  Speaker 2: not enough bandwidth.
• 52:33
  Speaker 1: Yeah. Great, Is there anything else you like to cover?
• 52:40
  Speaker 2: I mean, like, I figure, maybe this's good time to
• 52:42
  Speaker 2: take some questions. Like the space and the AI stuff
• 52:46
  Speaker 2: are very interesting. I'm curious what people what else people
• 52:50
  Speaker 2: care about outside of this?
• 52:54
  Speaker 1: Yes, okay, people, So now for the Q and A,
• 52:58
  Speaker 1: the people in the chats, now is your opportunity to
• 53:02
  Speaker 1: pose a question for the tech expert TP and uh
• 53:07
  Speaker 1: and don't be shy, you know. Uh, then then I'll
• 53:11
  Speaker 1: pull the pull the comments up and then TP will
• 53:14
  Speaker 1: get to answer them. Anybody really no questions, no question
• 53:22
  Speaker 1: for t T Well, I'm surprised, I'm shocked. This is uh,
• 53:27
  Speaker 1: this is this is like a very good opportunity for
• 53:30
  Speaker 1: you to ask real questions.
• 53:33
  Speaker 3: It looks like a lot.
• 53:34
  Speaker 2: Of people have been asking, have been commenting about a
• 53:37
  Speaker 2: wrong ward with the China. Chinese Satellite Assistance, Yeah, it is.
• 53:44
  Speaker 2: It is definitely very helpful. So I did find that
• 53:48
  Speaker 2: they that company, Bizarre Vision, was able to track be
• 53:54
  Speaker 2: one bombers just using AI plus satellite informations, like he
• 54:00
  Speaker 2: was able.
• 54:01
  Speaker 3: To figure out where it is.
• 54:02
  Speaker 2: So there's there's like in the future we're going to
• 54:06
  Speaker 2: get a situation where satellites plus AI is going to
• 54:09
  Speaker 2: be a attract stuff like beach bombers and things like that.
• 54:12
  Speaker 1: Well I thought that they just play sanction on misser
• 54:15
  Speaker 1: Vision or or are they talking about placing sanction on
• 54:19
  Speaker 1: misser Vision.
• 54:20
  Speaker 2: Yeah, well thought like you know that's great, right, Like
• 54:25
  Speaker 2: once you sanctioned, get sanctioned on the US government, then
• 54:27
  Speaker 2: you know you made it.
• 54:30
  Speaker 1: That's true. And and and I think recently China is
• 54:35
  Speaker 1: I think it was reported by Global Times that they're
• 54:38
  Speaker 1: they're gonna do a more general pushback on US anctions.
• 54:44
  Speaker 1: You know, it started with when they when US sanctioned
• 54:47
  Speaker 1: Henley Refineries for taking Iranian crude, and the Chinese government
• 54:53
  Speaker 1: order companies in China not to, uh, not to abide
• 54:58
  Speaker 1: by the U s anction. But now it looks like
• 55:01
  Speaker 1: there's been there will be a general directive to Chinese
• 55:05
  Speaker 1: companies and companies operating in China not to abide by
• 55:08
  Speaker 1: US sanctions targeting Chinese companies. So that that.
• 55:12
  Speaker 2: Was just kind of crazy, right, Like it's one thing
• 55:15
  Speaker 2: to like put sanctions on like your little teapot refineries
• 55:21
  Speaker 2: in Shandle, right, Like it's quite something else to uh,
• 55:25
  Speaker 2: to put sanctions on China's third largest private enterprise, Like
• 55:30
  Speaker 2: what are they?
• 55:32
  Speaker 1: Well, I think the reason that didn't put sanctioned on
• 55:35
  Speaker 1: a little teapot because that's kind of fruitless exercise because
• 55:41
  Speaker 1: first of all, most of the teapot operators they only
• 55:45
  Speaker 1: sell domestically to the Chinese market. You know, I don't
• 55:47
  Speaker 1: know how much is it going to be affected by
• 55:50
  Speaker 1: US tratechery departments sanction And two there are just going
• 55:54
  Speaker 1: to be too many of them to to go after
• 55:56
  Speaker 1: one each each individually. So I mean, so that's why
• 55:59
  Speaker 1: they probably had to go after Friendly because it's a
• 56:03
  Speaker 1: it's a bigger fish that can concentrate on their resources
• 56:07
  Speaker 1: just to target Heady.
• 56:09
  Speaker 2: Yeah, yeah, I guess so. But this is one of
• 56:13
  Speaker 2: those things where only itself is just such a huge player,
• 56:18
  Speaker 2: and not only not only petrol chemicals, but also shipbuilding
• 56:23
  Speaker 2: and marine engines and things like that.
• 56:26
  Speaker 1: One of the world it's one of the world's largest shipbuilders.
• 56:30
  Speaker 1: I just I just learned that recently.
• 56:32
  Speaker 3: Yeah.
• 56:32
  Speaker 2: No, actually it has eighty percent of the global backlock
• 56:37
  Speaker 2: for very large crew carrier. It's just like these these
• 56:41
  Speaker 2: are the kind of dumb things that the US government
• 56:43
  Speaker 2: does because it's run by a bunch of Wall Street guys.
• 56:49
  Speaker 1: Yeah, well, I think it's uh, it's it's that, and
• 56:52
  Speaker 1: it's it's just Hubris. I mean, I think it's a
• 56:56
  Speaker 1: the whole Hubris is pretty kind of its kind of
• 57:01
  Speaker 1: permeate the air in the belway, not just so just
• 57:04
  Speaker 1: a Wall Street people and even even some of the
• 57:08
  Speaker 1: tech bros too, A lot of them, I don't know, man,
• 57:14
  Speaker 1: Even though the ones that you think would know, like
• 57:17
  Speaker 1: the ones that had been to China I had business
• 57:20
  Speaker 1: dealing with China steal. They talk like you know I'm
• 57:24
  Speaker 1: talking about you know, like that that guy worked in China, man,
• 57:31
  Speaker 1: but didn't he working Bay Do?
• 57:34
  Speaker 2: Yeah? Is like if you see some of these China
• 57:37
  Speaker 2: watchers who are real hawks online, right, they are the
• 57:40
  Speaker 2: people that went into China thinking that they're superior to
• 57:43
  Speaker 2: to yellow people, and then they that they can rescue
• 57:47
  Speaker 2: China from communism or whatever I meantame Western Germany, and
• 57:52
  Speaker 2: then and then and then because they couldn't do that,
• 57:55
  Speaker 2: they're really pissed off, and then they decide to do whatever.
• 57:59
  Speaker 2: I feel like that's pready common, right, But if you
• 58:02
  Speaker 2: don't go if you don't go there with that kind
• 58:05
  Speaker 2: of attitude, then it's different. Like a lot of people
• 58:07
  Speaker 2: that just went to China recently, like Bill Gurley, he
• 58:10
  Speaker 2: came back like, Wow, this is really impressive stuff that
• 58:12
  Speaker 2: they're doing over it, and he's a lot more talking
• 58:15
  Speaker 2: about how that there should be more cooperation between the
• 58:17
  Speaker 2: two countries. Right, if you are one of those people
• 58:19
  Speaker 2: that went to China twenty five years ago where US
• 58:23
  Speaker 2: was so far ahead of China technology and then you
• 58:25
  Speaker 2: start trying to catch up, you might just be saying
• 58:30
  Speaker 2: these things because you.
• 58:33
  Speaker 3: You can't accept that things have changed. I don't know.
• 58:38
  Speaker 2: Well, one more thing is, like I noticed someone asks
• 58:40
  Speaker 2: about UV. The thing about UV is now that the
• 58:43
  Speaker 2: Huawei has come out with this new TOWE scaling, UV
• 58:47
  Speaker 2: is not necessarily needed anymore because one snick can produce
• 58:53
  Speaker 2: and HUAWE can produice like five nanimeter chips pretty steadily, right,
• 58:59
  Speaker 2: it can over the long term.
• 59:01
  Speaker 3: And the STRAM scaling.
• 59:03
  Speaker 2: And the black silicon issue, like a lot of the
• 59:05
  Speaker 2: analog scaling issues, wiring scaling issues really become an issue
• 59:12
  Speaker 2: after you get past five nanometer to three nanometer, two
• 59:15
  Speaker 2: nanometers even smaller.
• 59:17
  Speaker 3: So.
• 59:19
  Speaker 2: You once you scale smaller the transistors that the three
• 59:27
  Speaker 2: D folding might not work as well for smaller transistors
• 59:30
  Speaker 2: as the five nanometer transistors. So it's what Huawei's saying
• 59:36
  Speaker 2: is that smaller transistors and UV is not the only
• 59:41
  Speaker 2: way to get to higher chip performances.
• 59:43
  Speaker 3: You can get there just with.
• 59:47
  Speaker 2: Other methodologies. Right, So it's not you know, it's not
• 59:50
  Speaker 2: a rush. You're not being rushed into it anymore. And
• 59:54
  Speaker 2: even if you don't have UV in a couple of years,
• 59:55
  Speaker 2: it's not a big deal. I still think they're going
• 59:58
  Speaker 2: to get to UV pretty soon, because once you get
• 1:00:00
  Speaker 2: to EUV, you can actually produce chips more efficiently than
• 1:00:06
  Speaker 2: like you can produce four or five nanometers with three
• 1:00:09
  Speaker 2: nanometer chips much more efficiently than if you're not using
• 1:00:13
  Speaker 2: it right because you don't have to do this multiple
• 1:00:15
  Speaker 2: exposure nonsense. So and you can increase yield. Everything is
• 1:00:20
  Speaker 2: a lot more efficient with UV. So I think that
• 1:00:23
  Speaker 2: it's still important, but it's not a must. And this
• 1:00:28
  Speaker 2: is kind of like mind blowing to us, like a
• 1:00:30
  Speaker 2: couple of us that were really follow the Chinese semi sector,
• 1:00:34
  Speaker 2: is that we never even thought about like this what
• 1:00:38
  Speaker 2: we call three D logic thing before because we saw
• 1:00:41
  Speaker 2: like basically what happened is in the memory chips sector,
• 1:00:47
  Speaker 2: so the nan chips and then where we store stuff
• 1:00:51
  Speaker 2: that's like you know, SSDs and things like that went
• 1:00:53
  Speaker 2: three D pretty early on because you can't drink nanda
• 1:00:58
  Speaker 2: anymore once it gets to like around eighteen nanometers, and
• 1:01:02
  Speaker 2: then the d rams are going to go three D
• 1:01:07
  Speaker 2: pretty soon because once it gets to around ten nanometers
• 1:01:10
  Speaker 2: cell sizes, it can't shrink anymore.
• 1:01:12
  Speaker 3: We all thought the.
• 1:01:15
  Speaker 2: Actual logic transistors can still shrink, so there's no need
• 1:01:19
  Speaker 2: to actually go into the three D of this but
• 1:01:23
  Speaker 2: you know, obviously, while we decided it's that this is
• 1:01:29
  Speaker 2: their competitive advantage and they're already able to do it,
• 1:01:32
  Speaker 2: and it's very impressive because nobody else is doing to
• 1:01:34
  Speaker 2: the same degree. Like the Twawei is talking about hybrid
• 1:01:39
  Speaker 2: bounding of like one point five micrometers and TSMC is
• 1:01:45
  Speaker 2: at six, so like four time is as wider you know,
• 1:01:50
  Speaker 2: gap between the two layer wafers. That's why the huaiwei
• 1:01:53
  Speaker 2: wan works and the TSMC one is doesn't perform the
• 1:01:58
  Speaker 2: same way.
• 1:02:00
  Speaker 1: Okay, since since the chat is not asking you a question,
• 1:02:04
  Speaker 1: I got a question for you. Uh well, what other
• 1:02:07
  Speaker 1: areas of tech should we uh keep our eyes uh
• 1:02:13
  Speaker 1: you know what other area of tech advancement coming out
• 1:02:16
  Speaker 1: of China issue keep our eyes on.
• 1:02:18
  Speaker 2: So if you look at the fifteenth five year Plan,
• 1:02:21
  Speaker 2: they named the six main categories that really they really
• 1:02:24
  Speaker 2: care about China and body the AI right, bring, computer interest,
• 1:02:30
  Speaker 2: interconnect b C I. I think the hydrogen is a
• 1:02:34
  Speaker 2: major one that they're really focused on, and quantum computing.
• 1:02:40
  Speaker 1: And hydrogen as in hydrogen fuel. Yes, so that's interesting
• 1:02:46
  Speaker 1: that hold on say that's interesting because they China is
• 1:02:51
  Speaker 1: still looking into hydrogen while it's pretty much dominating in
• 1:02:56
  Speaker 1: dB right now.
• 1:02:58
  Speaker 2: Yeah, because you need the hydrid to replace crude in
• 1:03:04
  Speaker 2: a lot of these petrochemical sectors. Right, like we've seen
• 1:03:07
  Speaker 2: with the straightfor move shutdown, Like China still has this
• 1:03:11
  Speaker 2: huge import demand on oil, right, So the way to
• 1:03:16
  Speaker 2: get past that is by fully transitioning the transportation sector
• 1:03:20
  Speaker 2: into non oil stuff. And then for things like chemicals,
• 1:03:27
  Speaker 2: you need to transition to using green methanol or green
• 1:03:31
  Speaker 2: ammonia rather.
• 1:03:33
  Speaker 3: Than rather than just using.
• 1:03:38
  Speaker 2: Crude as your feedstock into your chemical factories. And also
• 1:03:43
  Speaker 2: a major one that's coming right now, and this is
• 1:03:45
  Speaker 2: this is quite interesting actually, is that about one to
• 1:03:49
  Speaker 2: two percent of the global shipping right now uses what
• 1:03:51
  Speaker 2: we call alternative fuel, which is like LNG or methanol
• 1:03:55
  Speaker 2: and stuff like that.
• 1:03:56
  Speaker 3: But in terms of backlocks yearly backlock, that's a TOI.
• 1:04:01
  Speaker 2: So fifty percent of the new ships coming in are
• 1:04:03
  Speaker 2: using can use something other than diesel or not. These
• 1:04:06
  Speaker 2: are like a bunker bunker fuel bunker oial and China
• 1:04:10
  Speaker 2: has built a shumongous supply chain around like uh bunkering
• 1:04:19
  Speaker 2: of container ships using like for l G and for
• 1:04:25
  Speaker 2: a methanol fuel. Right, so they've signed these pretty large
• 1:04:31
  Speaker 2: long term contracts with some of the biggest UH ship
• 1:04:36
  Speaker 2: owners out there, like Maersk and CGM. So that's like
• 1:04:41
  Speaker 2: a major one where they're going to become like a
• 1:04:44
  Speaker 2: major energy export in the future with a lot of
• 1:04:48
  Speaker 2: this green methanol and sustainable aviation fuel stuff.
• 1:04:52
  Speaker 3: Yeah, but that's what people.
• 1:04:55
  Speaker 1: But do you think the recent US you know, right now,
• 1:05:00
  Speaker 1: recently there's a US effort trying to limit the Chinese
• 1:05:05
  Speaker 1: ship building capability and trying to give more work towards
• 1:05:09
  Speaker 1: South Korea, do you think that's going to have an impact?
• 1:05:14
  Speaker 2: Well, maybe if they can actually figure out a rare
• 1:05:16
  Speaker 2: earth solution, then that good. But you know, if they
• 1:05:21
  Speaker 2: do this again trying to just reinstitute the rare earth
• 1:05:24
  Speaker 2: export controls right like, this is going to be the
• 1:05:29
  Speaker 2: US is not going to be able to figure out
• 1:05:31
  Speaker 2: like a rare earth solution critical mineral solutions for like
• 1:05:35
  Speaker 2: I don't know, five or ten years at least. So,
• 1:05:38
  Speaker 2: so that's plenty of time for trying to build up
• 1:05:40
  Speaker 2: its entire like sustainable green fuel shipping industry.
• 1:05:48
  Speaker 1: But what why why eugen Someone pointed out how eugene
• 1:05:53
  Speaker 1: is just a storage of energy, you know, it's a
• 1:05:56
  Speaker 1: it's a it's a it's a it's a vehicle for
• 1:05:59
  Speaker 1: carrying and storing the energy. Uh, it's not like oil.
• 1:06:05
  Speaker 1: So why why why hydrogen when you already have you know,
• 1:06:09
  Speaker 1: advanced being the battery technology can still store electricity.
• 1:06:13
  Speaker 3: Yeah, I mean, like part of the problem with batteries.
• 1:06:17
  Speaker 2: Let's just talk about electricity, right, it's really hard to
• 1:06:20
  Speaker 2: actually store uh energy inside batteries for like, you know,
• 1:06:26
  Speaker 2: several months, Like you can do it in today, right,
• 1:06:29
  Speaker 2: but we have to address the situation where you know,
• 1:06:32
  Speaker 2: maybe in the winter time that you don't get as
• 1:06:34
  Speaker 2: much sun and or whatever, or in summertime it gets
• 1:06:38
  Speaker 2: too hot in places like southern China and you need
• 1:06:41
  Speaker 2: constant air conditioning, and then your hydro dams are out
• 1:06:47
  Speaker 2: of water because it's too hot. So then you need
• 1:06:51
  Speaker 2: like some alternative way to uh uh store energy. And
• 1:06:58
  Speaker 2: it seems like storing stuff in hydrogen or methanol is the.
• 1:07:05
  Speaker 3: Best way to uh to uh deal with this problem.
• 1:07:10
  Speaker 2: Otherwise, you know, obviously they can just keep using coal,
• 1:07:12
  Speaker 2: but I think they're trying to go to zero net
• 1:07:17
  Speaker 2: zero at some point.
• 1:07:18
  Speaker 1: So, but Toyota was one of the third pioneer of hydrogen, right,
• 1:07:25
  Speaker 1: what happened to that?
• 1:07:26
  Speaker 2: Well, I think like generally speaking, the hydrogen you know,
• 1:07:29
  Speaker 2: transportation sector is in the auto transportation sector is dead
• 1:07:33
  Speaker 2: because just can't compete against batteries. But hydrogen in long
• 1:07:40
  Speaker 2: term energy storage still makes a lot of sense. And
• 1:07:43
  Speaker 2: hydrogen in the chemical sector and plastic sector makes a
• 1:07:47
  Speaker 2: lot of sense still. So also you need like hydrogen
• 1:07:51
  Speaker 2: for stuff like fertilizers. So you know, we have this
• 1:07:56
  Speaker 2: fertilizer shortage issue globally right now because of Trump, so
• 1:08:01
  Speaker 2: so that that needs to be fixed. But the other
• 1:08:03
  Speaker 2: ones I think people should focus on is quantum computers.
• 1:08:07
  Speaker 2: And you know, they're putting a lot of effort in
• 1:08:10
  Speaker 2: quantum computing right now.
• 1:08:11
  Speaker 3: And.
• 1:08:13
  Speaker 2: It's very apparently been used to do AI drug developments,
• 1:08:21
  Speaker 2: so might become a real thing in the next few years,
• 1:08:25
  Speaker 2: because biotech is one of those areas where China is
• 1:08:27
  Speaker 2: making a lot of money right now.
• 1:08:29
  Speaker 1: So what about fushion.
• 1:08:34
  Speaker 3: That I tend to think is still probably twenty years away?
• 1:08:38
  Speaker 1: Oh wow, Okay, okay.
• 1:08:40
  Speaker 2: Yeah, I think I've looked into it and that the
• 1:08:43
  Speaker 2: commercialization plan for it is very interesting, but very long.
• 1:08:47
  Speaker 3: Still.
• 1:08:49
  Speaker 2: I tend to think, like in terms of nuclear energy,
• 1:08:53
  Speaker 2: some of the other stuff that they're working on might
• 1:08:56
  Speaker 2: get commercial applications sooner, like you know, we might start seeing, oh,
• 1:09:04
  Speaker 2: we might start seeing like USh uh you know, container
• 1:09:07
  Speaker 2: ships running on h small nuclear reactors.
• 1:09:12
  Speaker 3: Those are the kinds of things we might see. Ah,
• 1:09:15
  Speaker 3: but nuclear fusion.
• 1:09:17
  Speaker 2: You know, we already have a nuclear fusion engine older
• 1:09:20
  Speaker 2: it's called the Sun and it's uh, it's producing plenty
• 1:09:25
  Speaker 2: of like energy for us.
• 1:09:28
  Speaker 1: That's fair. But but but but new nuclear power plant.
• 1:09:33
  Speaker 1: You know, the average construction of a nuclear power plant
• 1:09:37
  Speaker 1: itself is like ten years, right, yeah, so I you know,
• 1:09:43
  Speaker 1: like like then first first you have to develop like
• 1:09:47
  Speaker 1: a path to a commercialization, and then you get to
• 1:09:50
  Speaker 1: build it. I think, yeah, I think that that's fair
• 1:09:52
  Speaker 1: to say twenty years. It's a conservative estimate.
• 1:09:58
  Speaker 2: Well, the funniest thing is Americans are so into this
• 1:10:01
  Speaker 2: nuclear fusion stuff when it can't and build the regular reactors, right,
• 1:10:07
  Speaker 2: Like maybe you should get like be able to build
• 1:10:11
  Speaker 2: you know, your AP one thousand first before you think
• 1:10:14
  Speaker 2: about building nuclear fusion reactors and also building reactors in space.
• 1:10:20
  Speaker 2: The funniest thing I read is is like this idea
• 1:10:23
  Speaker 2: you're going to build like these mini reactors in space
• 1:10:26
  Speaker 2: when you can't even build them on the ground, Like,
• 1:10:29
  Speaker 2: what is what is people thinking about stuff like that?
• 1:10:31
  Speaker 3: You know?
• 1:10:32
  Speaker 1: Yeah, well I think Nathan is asking you China using
• 1:10:36
  Speaker 1: high youguene for trucks. I don't think that's that's the
• 1:10:38
  Speaker 1: thing currently. But Chinese is actually developing like electric heavy
• 1:10:45
  Speaker 1: trucks trucks, which is interesting, by.
• 1:10:48
  Speaker 3: The way, only the hydrogen.
• 1:10:50
  Speaker 2: They did try it for a couple of years, but
• 1:10:52
  Speaker 2: commercially it just couldn't compete against the battery power ones.
• 1:10:55
  Speaker 2: So the battery power heavy duty truck market sharees probably
• 1:11:00
  Speaker 2: while we're thirty percent by now. So that's actually part
• 1:11:03
  Speaker 2: of the reason why their oil demand is down so much,
• 1:11:06
  Speaker 2: because the diesel demand is down, and the diesel demand
• 1:11:09
  Speaker 2: is down because the heavy trucks are not running on
• 1:11:14
  Speaker 2: diesel anymore. They're running on either natural gas or electric.
• 1:11:20
  Speaker 1: Yeah, you know, right now, the whole world is drawing
• 1:11:24
  Speaker 1: down on their strategic petroleum reserve, except China. China hasn't
• 1:11:28
  Speaker 1: really drawn down that much. Funny, it's one point four
• 1:11:31
  Speaker 1: billion barrels of oil in their strategic oil reson.
• 1:11:35
  Speaker 3: So I have a theory about that.
• 1:11:36
  Speaker 2: Actually, I think they're they're importing less crude and not
• 1:11:40
  Speaker 2: drawing down the spr as much because a lot of
• 1:11:44
  Speaker 2: the commercial reserves of chemicals and maybe refined oil product
• 1:11:52
  Speaker 2: are getting drawn down, like in a way that's not
• 1:11:55
  Speaker 2: really visible to us, because I think what's happening is
• 1:11:57
  Speaker 2: a lot of supply chains like oh Sing are so
• 1:12:00
  Speaker 2: much more expensive than they were two months ago. I
• 1:12:03
  Speaker 2: want to I still have some stockpiles left. I want
• 1:12:05
  Speaker 2: to use up the stock piles before I buy new
• 1:12:07
  Speaker 2: ones because I don't want to increase prices. Once I
• 1:12:10
  Speaker 2: increase price, people don't want to pay for the money anymore.
• 1:12:14
  Speaker 2: And we we you know, we saw that in the
• 1:12:16
  Speaker 2: entire memory chips thing, where because memory chips went up
• 1:12:20
  Speaker 2: two three folds in a very short time, the Chinese
• 1:12:23
  Speaker 2: OEMs just refused to buy them for a while.
• 1:12:27
  Speaker 3: So until they really run run down their commercial stocks.
• 1:12:31
  Speaker 2: I feel like there's a lot of this going on
• 1:12:34
  Speaker 2: still speaking but speaking, memory can't last forever, right, Like
• 1:12:38
  Speaker 2: this can last maybe a couple more months, but it
• 1:12:39
  Speaker 2: can't last forever.
• 1:12:41
  Speaker 1: Speaking of memory chips, do you think China will be
• 1:12:44
  Speaker 1: able to bring the price down soon, you know, by
• 1:12:48
  Speaker 1: by you know, just mass pretty production, Yes.
• 1:12:54
  Speaker 2: But I think the best chance for it to come
• 1:12:57
  Speaker 2: down is for this AI bubbles to burst, which this
• 1:13:01
  Speaker 2: Gigles's AI bubble. It's like there's no money coming in
• 1:13:04
  Speaker 2: and you have people like data stacks going on their saying, oh,
• 1:13:07
  Speaker 2: each data each Gigle data center requires twenty five billion
• 1:13:12
  Speaker 2: dollars to build and it can generate fifty billion dollars. No,
• 1:13:15
  Speaker 2: it takes fifty billion dollars to build anything generally twenty
• 1:13:18
  Speaker 2: five billion dollars revenue a year. I'm like, who is
• 1:13:22
  Speaker 2: keep on giving you fifty billion dollars to build these
• 1:13:24
  Speaker 2: data centers?
• 1:13:27
  Speaker 1: Well, that's saying.
• 1:13:28
  Speaker 5: That's how they say they're gonna win over China because
• 1:13:32
  Speaker 5: you US is building thousands of data centers and China
• 1:13:36
  Speaker 5: only have like what like under one hundred or something
• 1:13:40
  Speaker 5: like that.
• 1:13:41
  Speaker 2: But the actual like data center electricity demand that they're
• 1:13:45
  Speaker 2: talking about the currently China is projecting over the fifteen
• 1:13:51
  Speaker 2: to five year plan to add about twenty to twenty
• 1:13:54
  Speaker 2: five gigawad of data centers a year.
• 1:13:57
  Speaker 3: Wow. Right, and that is actually more than in the US.
• 1:14:03
  Speaker 2: Now, it is possible that the Chinese ones use more
• 1:14:05
  Speaker 2: because it's running on less efficient GPUs whatever, so that
• 1:14:08
  Speaker 2: that's possible, right, but it's also possible that you know,
• 1:14:12
  Speaker 2: US just can't build AI data centers because it doesn't
• 1:14:17
  Speaker 2: have the electricity to power it and stuff like that. Right,
• 1:14:21
  Speaker 2: So the monetary part of this is going to have
• 1:14:24
  Speaker 2: to come back and hot people in a little bit,
• 1:14:26
  Speaker 2: because how can you keep paying the ridiculous amount of
• 1:14:30
  Speaker 2: stuff for clock cold when there's these fifteen dollars a
• 1:14:33
  Speaker 2: month models that run perfectly, you know, like this this
• 1:14:37
  Speaker 2: is the part that I don't understand these companies. Just
• 1:14:39
  Speaker 2: like the entire US white collar salary is only so
• 1:14:45
  Speaker 2: much for programmers in North America that the total salary
• 1:14:49
  Speaker 2: is two hundred billion dollars a year, Right, how can
• 1:14:52
  Speaker 2: you generate like one hundred billion dollars a year of revenue, right,
• 1:14:59
  Speaker 2: two hundred bi million dollar revenue from displacing white collar
• 1:15:02
  Speaker 2: jobs when the total salary is only so much?
• 1:15:08
  Speaker 1: And what about thorium reactors in China?
• 1:15:13
  Speaker 3: That's also probably twenty years away?
• 1:15:16
  Speaker 1: Oh years, that's fair.
• 1:15:20
  Speaker 3: Yeah, yeah, I wouldn't.
• 1:15:21
  Speaker 2: I wouldn't put giant hope on these things. I think
• 1:15:24
  Speaker 2: the fast neutron reactors is fur their ahead. Some of
• 1:15:28
  Speaker 2: the other thing they did is really cool. It's a
• 1:15:31
  Speaker 2: super critical reactors that they have. Yeah, that that technology
• 1:15:36
  Speaker 2: is amazing and it can be used for other applications,
• 1:15:39
  Speaker 2: liking the nuclear submarines. It works perfectly for that also there,
• 1:15:45
  Speaker 2: once they get that figured out, they can start building
• 1:15:47
  Speaker 2: nuclear power nail ships everywhere. So that's it's, it's it's
• 1:15:51
  Speaker 2: it's it's gonna be massive. Also, the other thing that
• 1:15:53
  Speaker 2: they got really good at recently that people don't realize
• 1:15:56
  Speaker 2: is gastromize.
• 1:15:59
  Speaker 1: Yeah, turbines not like this is something in high demanding
• 1:16:04
  Speaker 1: the United States right now because they're trying to build
• 1:16:07
  Speaker 1: their own uh they're trying to build their own power
• 1:16:10
  Speaker 1: generation to power those data centers, and they realize they
• 1:16:13
  Speaker 1: cannot rely on American electric quids, so they have to
• 1:16:18
  Speaker 1: build their own uh uh uh power generation using gas turbines.
• 1:16:23
  Speaker 1: And then guess, well, guess where did the gas turbine
• 1:16:26
  Speaker 1: come from?
• 1:16:28
  Speaker 2: Well, the thing about the gas turbine is the Chinese
• 1:16:30
  Speaker 2: ones are typically lost. They haven't got to the ones
• 1:16:33
  Speaker 2: where they're doing those large gas turbines yet, but they
• 1:16:37
  Speaker 2: they've gone pretty good on the twenty five megawatt like
• 1:16:41
  Speaker 2: you know, one mago wat to fifty megawat gas turbines.
• 1:16:46
  Speaker 2: So you're you're seeing quite a lot of exports of
• 1:16:48
  Speaker 2: these smaller sized gas turbines. And these gas turbines are
• 1:16:52
  Speaker 2: just really good for industrial usage, offshore drilling, and also
• 1:16:58
  Speaker 2: also they're great for naval ships. So that's why they
• 1:17:01
  Speaker 2: that's why the Chinese, uh, you know why the Chinese
• 1:17:05
  Speaker 2: resource went into these first, because they're like, oh, we
• 1:17:07
  Speaker 2: don't need this for we don't need nat natural gas
• 1:17:10
  Speaker 2: uh uh, you know, thermal plants right now, we got cold,
• 1:17:13
  Speaker 2: but we do need these smaller gas turbines for for
• 1:17:16
  Speaker 2: our naval ships, right, So that's why they did a
• 1:17:20
  Speaker 2: lot of that.
• 1:17:22
  Speaker 1: Speaking of which, what progress are they making over the
• 1:17:28
  Speaker 1: jet engine right now, because that's still one of the
• 1:17:31
  Speaker 1: area China is still catching up, right.
• 1:17:35
  Speaker 2: Yeah, that area, I think it's catching up pretty quickly.
• 1:17:38
  Speaker 2: I mean in terms of the new aero engine, it's
• 1:17:41
  Speaker 2: gonna be it's probably gonna be all caught up in
• 1:17:45
  Speaker 2: ten years. Until then, you want to see, uh, some
• 1:17:51
  Speaker 2: efforts specially on the C nine one nine project where
• 1:17:53
  Speaker 2: they're trying to have the and this and this you
• 1:17:56
  Speaker 2: know c G two thousand project where they're trying to
• 1:17:58
  Speaker 2: get domestic equivalent to Western engines on the domestic airliners.
• 1:18:04
  Speaker 3: That's probably the big hurdle.
• 1:18:06
  Speaker 2: If they can clear that, then that means they have
• 1:18:10
  Speaker 2: fully caught up in materials, design and all the components
• 1:18:15
  Speaker 2: in the turbofans.
• 1:18:17
  Speaker 1: Are you talking about both the military and civilian Uh?
• 1:18:22
  Speaker 3: Yeah, I guess so, like the domestic ones. It's very
• 1:18:25
  Speaker 3: interesting the aero engines.
• 1:18:28
  Speaker 2: They have this new live aero engines now that they're
• 1:18:31
  Speaker 2: testing out that they've fallen recently. It's it's very competitively
• 1:18:35
  Speaker 2: internationally in terms of costs, in terms of uh, you know,
• 1:18:41
  Speaker 2: performance and things like that. And what that is allowing
• 1:18:46
  Speaker 2: them to do is replace a lot of the engines
• 1:18:51
  Speaker 2: used for drones that used to be like turboprop or
• 1:18:54
  Speaker 2: piston engine drones or turbojet drones throughout turbofan drones going forward.
• 1:19:00
  Speaker 2: And they're really helpful for those all man transport that
• 1:19:04
  Speaker 2: they're building in China, where you know, you have these
• 1:19:06
  Speaker 2: like two tons, five town or tent tongue uh transports
• 1:19:11
  Speaker 2: that just do cargoes without humans in them, right and
• 1:19:15
  Speaker 2: there there's a there's a whole industry around that being
• 1:19:17
  Speaker 2: built up right now.
• 1:19:18
  Speaker 3: It's very it's very interesting like also with you know,
• 1:19:21
  Speaker 3: they're also good for helicopters and things like that.
• 1:19:23
  Speaker 2: So they're very much caught up in a lot of
• 1:19:26
  Speaker 2: these areas. And and the difference between the China and
• 1:19:29
  Speaker 2: Chinese and the Western ones is that they're just so cheap,
• 1:19:32
  Speaker 2: Like you know, the reason why you're seeing like all
• 1:19:38
  Speaker 2: these like cheap weaponry in the in the used by
• 1:19:41
  Speaker 2: the Iranians and maybe the Russians are because they can
• 1:19:45
  Speaker 2: get the engines from China. Like you know, if you
• 1:19:48
  Speaker 2: want like ten thousand dollars turbojet engine right to make
• 1:19:52
  Speaker 2: your drones go a little faster, you just go top
• 1:19:54
  Speaker 2: to Chinese. The same with some you know, if you
• 1:19:58
  Speaker 2: want your set to go a little fast, sure you
• 1:20:00
  Speaker 2: just go talk to Chinese and maybe they'll come.
• 1:20:02
  Speaker 3: They'll give you something new.
• 1:20:06
  Speaker 1: Very interesting. All right, is there anything else you would
• 1:20:10
  Speaker 1: like to talk about?
• 1:20:12
  Speaker 3: No, I think I think I'm good on this front.
• 1:20:15
  Speaker 1: Okay, okay, Chad, you have thirty seconds to come up
• 1:20:21
  Speaker 1: with additional questions for for TP and going once, going twice?
• 1:20:31
  Speaker 1: Can okay? Wait? Wait? Is that a question can leverage
• 1:20:37
  Speaker 1: Elle to get up to speed with almost any topic
• 1:20:41
  Speaker 1: of interest? This need to free Okay, that's not a question. Okay,
• 1:20:45
  Speaker 1: never mind. All right, Well, thank you very much TP
• 1:20:49
  Speaker 1: for making your time to speak to us. This has
• 1:20:52
  Speaker 1: been very enlightening and I'm sure I need to I
• 1:20:58
  Speaker 1: need to have you come back in a few wants
• 1:21:00
  Speaker 1: to keep us an update because things are just moving
• 1:21:02
  Speaker 1: so fast, especially in the AI run, but in the
• 1:21:06
  Speaker 1: other area as well, so on behalf of the community.
• 1:21:12
  Speaker 1: Thank you. Oh wait, latest operating system okay, I have
• 1:21:17
  Speaker 1: there's one question for Nadia. Any latest operating system research
• 1:21:23
  Speaker 1: coming out of China.
• 1:21:27
  Speaker 2: The relevant ones are probably just whatever the Huawei is
• 1:21:29
  Speaker 2: doing with Harmony in getting into work across the Internet
• 1:21:34
  Speaker 2: of things. But other than that, it's mainly the big
• 1:21:38
  Speaker 2: drive is probably just to put AI into operating systems
• 1:21:42
  Speaker 2: now there, and I.
• 1:21:44
  Speaker 3: Would say that's going pretty well.
• 1:21:47
  Speaker 2: If you go to China, if you go to see
• 1:21:48
  Speaker 2: the consumer electronic devices, they have AI built into them
• 1:21:53
  Speaker 2: pretty natively and it's mostly just a lot of the
• 1:21:58
  Speaker 2: ohyam's really a working hard to add integrating AI at
• 1:22:02
  Speaker 2: a much lower level into the into the Android operating
• 1:22:05
  Speaker 2: systems so that they can fully take advantage of us
• 1:22:09
  Speaker 2: the AI features.
• 1:22:14
  Speaker 1: Yeah, all right, thanks thanks for that, and thank you
• 1:22:20
  Speaker 1: again TP for making your time to speak to us
• 1:22:23
  Speaker 1: and thanks everyone for your questions and joining us for
• 1:22:28
  Speaker 1: this session. We will have TP back, don't worry, so
• 1:22:32
  Speaker 1: you have more chances to ask him question in the future.
• 1:22:36
  Speaker 1: Thanks TP, I'm gonna we're gonna sign off now. Thanks
• 1:22:40
  Speaker 1: everyone for tuning in until next time. Thank you,