去了趟美国才发现：同样叫科技强国，中美强的地方完全不一样！#圆桌派 #纪实 #时间 #历史 #文化 #聊天 #新闻 #西藏

同样叫科技强国，中美强项大不相同 节目围绕5G技术底座与数字化浪潮展开深度跨界研讨，拆解了频谱资源分配、微基站密度与NSA/SA组网的技术内涵，并澄清了公众对电磁辐射的误读。嘉宾们剖析了1毫秒极限时延与边缘计算如何赋能自动驾驶、远程手术与工业物联网，进而延伸探讨了算法茧房对人类审美的同质化冲击、通用人工智能（AGI）的奇点隐忧，以及万物互联时代的能源重构。全场最终直面技术狂飙中“缺乏刹车机制”的全球治理难题，反思了人类文明在面对不可逆技术演进时的宿命与媒介变革。 通信标准遵循“大卫杜夫定律”，率先入局者通常锁定50%以上的软硬件市场；5G将空口时延压低至1毫秒，较人体约10毫秒的痛觉神经传导快10倍，主要为机器交互与超低时延场景奠定基石。 (9:20) 产业数字化是5G的核心战役，通过基站下沉的“端-边-云”边缘计算架构，美国大河钢铁等案例依托5万个传感器将白领化率推至80%、人均产能拉升至全球均值3倍，解决了高并发工业告警与区块链交易的串行阻塞瓶颈。 (14:08) 美德政府探索向巴西淡水河谷等工业巨头直接配发专用隔离无线电频谱自建工业专网，以保障核心数据安全；这一变革正在倒逼传统电信运营商从垄断网络通道转向提供专业化运维服务。 (44:28) 技术演进呈现单向不可逆性，人类文明在具备某项技术后往往无法克制使用；万物互联与深度学习催生注意力第四产业与个体审美同质化，亟需在全球治理层面设立安全准则与“刹车制动”机制。 (43:13)

《圆桌派》节目中，学者与嘉宾指出5G不仅是网速升级，更是重塑工业制造、能源交通与社会形态的基础设施；然而，伴随万物互联与算法深化的加速演进，人类社会正面临审美同质化、认知蚕食以及缺乏“制动机制”的全球治理挑战。 技术范式与底层架构转变 (0:00) 通信技术遵循十年一代的演化周期，奇数代（1G、3G、5G）往往带来颠覆性架构跃迁，偶数代则负责优化演进。5G（IMT-2020）的核心价值不在于消费端的抢红包或秒速下载，而在于以微秒级时延（空口时延低至1毫秒）与海量并发能力重塑生产端。针对公众对辐射的担忧，专家指出中国执行的电磁辐射限值远严于美国，且高频微波波长极短、穿透与绕射能力弱，基站采取微功率局部覆盖，不会造成有害生物损伤。在部署路线上，基于4G核心网过渡的非独立组网（NSA）无法发挥完整性能，只有彻底重建专用核心网的独立组网（SA）才能支撑起产业互联网的严苛需求。 产业赋能与移动终端的定义重塑 (10:14) 5G与边缘计算（MEC）的结合打破了传统云计算的中心化传输瓶颈。在自动驾驶中，两公里范围内的道路协同依靠基站级边缘服务器实现1毫秒的极低时延制动；在工业场景下，美国大河特钢（Big River Steel）部署5万个传感器后使人均产出达到全球平均水平的三倍，福建投资300亿元的无人化工园区则借此解决海量传感器并发报警造成的网络拥堵。同时，终端硬件正发生解构：轻量化AR眼镜可将重型计算单元剥离至随身包或边缘端；智能汽车被埃隆·马斯克视作“移动通信终端”，其闲置时更可化身为分布式储能节点，在用电低谷充电并在尖峰时段反哺电网以提升能源效率。 认知茧房与算法对人性的异化 (17:38) 随着万物互联将海量行为数据喂养给算法，社会出现了深层次的异化风险。周轶君指出，“即看即买”与流量中心化正在让个体审美高度趋同，社交平台上强调的“自我”反而在技术模板下变得千篇一律。行为经济学视角表明，数字平台针对人类直觉、自动化的“系统1”设计了多巴胺反馈回路，导致用户沉迷短视频，而理性的“系统2”逐渐退化。未来由大数据构建的“数字孪生”不仅能预先推测个人需求，甚至可能通过人机融合让生物与机器的界限彻底模糊，使个体沦为庞大通信网络上的单一数据节点。 狂飙背后的刹车难题与频谱治理 (43:13) 信息化在中国过去数十年的经济增长中扮演了工业与农业现代化的“指数级放大器”，但技术狂飙也暴露出治理短板。窦文涛与嘉宾强调，人类文明往往缺乏“拥有能力却不用”的自制力，所有人都在研究如何将技术释放出笼，却极少有人探寻刹车与钥匙在谁手中。在产业落地层面，德国、美国及巴西淡水河谷公司等允许工业实体直接获取专用频段自建物理隔离的独立专网以保障数据安全；国内通信运营商则需打破传统的封闭垄断思维，从单一的基础设施管道提供方转向专业的专网代维与运维服务商。 媒介解构与“无直播不传播” (48:45) 在媒介形态上，5G时代将呈现“无直播不传播”的趋势，短视频正全量向实时多视角流媒体演进。以体育赛事为例，传统以“球”为中心的官方单视角转播，正被以“人”为中心的个性化多视点平台替代；现场VIP观众手持终端拍摄的高清推流，甚至能作为独立付费视角向全球十万人分发并变现。这种高度碎片化的信息分发模式带来了技术版的“盲人摸象”——尽管所有人都在围观同一场事件，但每个人通过算法与自主机位所接收到的现实却完全截然不同。 从4G改变生活到5G重塑社会，通信网络正向万物互联乃至6G时代的认知互联演进。然而，当单向度的技术飞轮持续运转，人类能否在全球治理中建立起制度与伦理的“制动机制”，并警惕被算法完全数据化，仍是悬而未决的时代叩问。

[6:22] Wu Xiaobo: 4G changed our lives; 5G will change society. [13:27] What you see is what you get, and experience becomes replicable. [17:38] 当年 3G 出来挨的骂，比现在骂 5G 的还要多得多 [22:42] 昨晚你睡得好不好？ [22:42] 我昨晚睡得挺好 [33:14] Elon Musk: You all end up making electric cars, vehicle networking—it's still a car in the end. What I'm making is a mobile communication terminal; I don't call it a car. [34:36] This isn't a car. This is a computer [43:13] 太少人在研究如何踩刹车 [46:25] 窦文涛: 王家卫所有的电影其实就讲了一件事——你回不去了。 [46:25] 窦文涛: 人类从来没有在拥有一种能力之后，能够克制住自己不去使用它。 [46:25] 窦文涛: 在未来的5G甚至6G时代，站在社会食物链最顶层的超级精英，可能正是那个自己从来不用手机的人。 [48:45] 未来一切皆是个性化直播 [48:45] 无直播不传播

[0:00] This frequency issue isn't just a [0:02] problem for China. For instance, when [0:04] Chinese users travel abroad with their [0:05] 4G or 5G phones, they need [0:07] international roaming, so there must be [0:09] global coordination to ensure they're [0:10] on the same frequencies. [0:12] >> Frequency is a type of resource; there [0:13] are several fundamental resources in [0:15] telecommunications. [0:16] >> You see, what everyone is scrambling [0:17] for is this—what we call frequency [0:19] resources. Abroad, there is no such [0:20] thing as a 5G or 4G license in the way [0:22] we think of it. Overseas, frequencies [0:24] are auctioned off for paid use; if you [0:26] secure the frequency, you have obtained [0:27] the license. In China, because the [0:29] companies are state-controlled, having [0:31] the state pay itself to use state [0:33] resources makes no sense, so we have [0:35] never adopted that model. The second [0:38] resource is numbering resources. You [0:40] cannot have telecommunications without [0:42] numbers. The third is satellite orbits; [0:44] look at how the Americans are racing to [0:46] grab those satellite orbits right now. [0:47] >> There is a limited number of them, so [0:49] these are the three. Regarding what [0:52] Professor Wu mentioned about 2018, that [0:55] was when China began trial networks for [0:57] 5G. But theoretically, "5G" is just a [1:00] nickname; its formal name is part of [1:03] the International Mobile [1:05] Telecommunications-2020 group. [1:09] >> Is that American-style Japanese? [1:12] >> That is Japanese-style English. I [1:15] studied at Kyoto University in Japan, [1:17] so you can tell it's Japanese-style [1:20] English? My sense of the language— [1:22] well, even if you didn't understand it, [1:24] you knew it was Japanese-style English. [1:26] >> It is designated for 2020. Why is that? [1:29] Because mobile communication technology [1:31] follows a ten-year cycle. [1:32] >> Right. [1:33] >> The 90s were for 2G, 2000 for 3G, and [1:36] 2010 for 4G. It was supposed to be 5G [1:39] in 2020, but the world couldn't wait, [1:41] especially with the Americans pushing [1:43] for it. The Americans have repeatedly [1:45] released reports calling it a major [1:47] national security strategy, saying we [1:48] cannot fall behind, and so on. This [1:50] highlights that 5G, in this global race [1:52] , follows a rule in the internet [1:54] industry known as the Davidow Law, [1:57] which states that whoever enters the [1:59] market first captures 50%of it. [2:01] >> This is why the Americans are pushing [2:04] to build this network. They are [2:06] fighting for it because those who enter [2:08] first will lead in equipment [2:09] manufacturing and application, likely [2:12] capturing a larger share. So, we are [2:14] already in it. When I heard you mention [2:17] 2020, I found it very interesting. [2:19] Doesn't 2020 sound like the knocking of [2:22] destiny? Ding-dong, ding-dong, 2020. [2:25] >> Right? In Beethoven's era, the knocking [2:28] of destiny was "da-da-da-dum," but now [2:30] in the digital age, it's 2020. [2:32] >> Right, 1010 is computer language, 2020. [2:35] What I mean is, we’ve crossed the [2:38] threshold, but we still know nothing [2:40] about it. So, I have a few ignorant, [2:43] hearsay-based questions—some I heard [2:46] from Zhang Yang. I’m not joking, [2:48] Zhang Yang even apologized later, right [2:51] ? For something so new, we should at [2:54] least ask more questions. For example, [2:58] while in Europe, I heard that 5G hadn't [3:01] passed the radiation environmental [3:03] standards there. Is there any truth to [3:07] that? [3:08] >> Regarding international radiation [3:10] standards, [3:11] >> Actually, every country has its own [3:14] radiation standards. China has its own, [3:16] which is a fraction of the US standard [3:19] in terms of the maximum allowable [3:21] radiation level. [3:22] >> It is far lower than mobile phone [3:24] radiation and far lower than microwave [3:26] oven radiation. [3:27] >> Mm. [3:28] >> Our country has control over this. The [3:31] point Zhang Yang made at the time was [3:33] that 5G would bring massive radiation [3:35] because its frequency is relatively [3:37] high, and high frequencies can produce [3:40] a special kind of electromagnetic [3:42] radiation effect. Ha, how can ordinary [3:44] people understand this? Simply put, the [3:47] thing that causes harm isn't determined [3:50] solely by energy; it’s the product of [3:52] energy and momentum, just like [3:54] electrical power is current multiplied [3:57] by voltage. It’s the product of the [3:59] two. What does that mean? I’ll give a [4:02] simple example. You’d say a car can [4:05] hit and kill someone, right? [4:06] >> But if it’s moving at a low speed or [4:08] idling, it won’t kill anyone. It has [4:10] energy, but not enough momentum. [4:13] >> So it won’t cause harm. Why 5G then? [4:16] Its energy does increase because the [4:18] frequency is higher, which increases [4:20] its capacity to carry information, but [4:22] its momentum isn't that high. With 5G, [4:25] the higher the frequency, the worse its [4:27] diffraction performance. That’s the [4:30] relationship between wavelength and [4:32] frequency, you see. [4:33] >> Since we keep saying 5G frequencies are [4:36] very high, the wavelength is very short [4:38] . With a short wavelength, it can't [4:40] bypass obstacles. [4:42] >> This explains why 5G is expensive; if [4:43] you build a base station on this side [4:45] of a building, there’s no signal on [4:46] the other side. [4:47] >> In that case, it doesn’t need a very [4:49] high power output either. Understand? [4:51] Because it only needs to cover this [4:53] small area, like this courtyard. Once [4:54] it’s outside the courtyard, there’s [4:56] no signal, right? It, well, once [4:58] something blocks them, they get [4:59] obstructed. Back then, with the 700MHz [5:01] frequency used by radio and television, [5:04] a single station in a city could reach [5:06] everyone. Of course, it had higher [5:08] power, but the frequency was low, [5:10] meaning lower energy. So, there was a [5:12] controlled parameter that wouldn't [5:14] cause significant harm to the human [5:16] body. [5:16] >> Hey, that’s a very accessible [5:18] explanation. [5:19] >> So, regarding these 5G base stations, [5:20] does it mean across China in the future [5:22] , we'll need a base station for every [5:23] basketball court? Since the density is [5:24] so high, especially wherever there’s [5:26] an obstruction? [5:27] >> A base station for every basketball [5:28] court? [5:29] >> That wouldn't be enough, would it? [5:31] >> The reason is that there are too many [5:33] obstructions, so it needs to be [5:35] frequent. And, unlike 4G, which was [5:37] built on the foundation of 3 G—you [5:39] know, the modes were just modified back [5:41] then—I heard that for 5G, everything [5:43] has to be wiped out and rebuilt from [5:45] scratch. Is that true? Professor Wu, [5:47] this is also a matter of social concern [5:49] , as it relates to future usage costs. [5:51] >> The 5G networks currently being [5:53] deployed by operators are called [5:55] Non-Standalone (NSA) networks. Uh-huh. [5:57] >> Non-Standalone networking is based on [6:00] existing 4G base stations. [6:02] >> It utilizes some infrastructure that 4G [6:04] can use and adds 5G capabilities on top [6:07] . However, it can't achieve the true [6:09] key features of 5G; for example, it [6:11] can't reach a one-millisecond latency, [6:13] nor can it support one million [6:15] connections per square kilometer. So, [6:17] to truly operate 5G, we need Standalone [6:20] networking, which we call SA. [6:22] >> Xiaobo, tell us your thoughts on the [6:24] arrival of the 5G era. What will it [6:27] bring to our lives? [6:28] >> I think that phrase from the operators [6:30] is quite fitting: 4G changed our lives [6:33] —our information, services, and [6:35] mobile payments. 5G will change society [6:38] . What does "changing society" mean? [6:40] >> It will significantly increase speeds, [6:42] and take connectivity, for instance— [6:44] like autonomous driving. You can't [6:46] achieve that in a 4G environment; you [6:48] can only do it in a 5G environment. The [6:51] vehicle itself needs to be tested, and [6:53] you have to drive it. Given the [6:55] complexity of road conditions, an [6:57] autonomous car needs to monitor the [6:59] situation within 150 to 200 meters at [7:02] all times to be able to react. That [7:04] requires two conditions: first, it must [7:06] have real-time computing speed. [7:07] >> It can't be calculated without greater [7:10] computing power, but 5G can. The second [7:12] issue is that within these 150 meters, [7:14] you must have base stations and the [7:15] right environment. Once you're past [7:17] that, you lose the signal. [7:19] >> Exactly, so it's actually about [7:20] transforming society as a whole, like [7:22] with remote medical care. Phone [7:24] companies are still focused heavily on [7:26] broad entertainment, but I think 5G [7:28] will change much more than that. [7:29] Entertainment and video are just one [7:31] category; there are mobile device [7:33] manufacturers, automobiles, and [7:34] factories—the big data processing [7:36] capability within an entire factory [7:38] environment. That is a very sharp point [7:40] , hitting the root of what 5G is about: [7:42] transforming society. [7:43] >> The Internet of Vehicles, the Internet [7:45] of Things, the Industrial Internet. In [7:46] '94, Kevin Kelly wrote "Out of Control. [7:49] " I read his book back then, and he [7:51] said that in the future, every—he was [7:53] being very extreme—every thread and [7:54] piece of clothing would have a sensor. [7:56] But how could that be possible? Sensors [7:59] cost 6 or 10 dollars, but in the future [8:01] 5G environment, that will be the [8:02] reality. But what Professor Wu just [8:04] said is very important and everyone [8:06] should understand: mobile technology [8:08] research shows that every generation is [8:11] disruptive. [8:11] >> From the first to the third to the [8:13] fifth generation; that's why 5G is [8:14] meant to change society, right there. [8:16] The first generation brought the "brick [8:18] phone," so you didn't have to drag a [8:19] cord around to make calls; you could [8:20] call from anywhere, creating a [8:21] brand-new usage scenario. The very [8:22] first ones were just called wireless [8:24] phones. [8:24] >> Cordless phones. [8:25] >> They were really big. [8:27] >> We used them back then. When I was 29 [8:29] and working at Guangdong Radio, I was a [8:32] big guy. [8:33] >> I have a "fat gene," you know? If I [8:35] didn't starve myself, I’d be huge. [8:37] Back then, I would only eat one meal [8:39] every five days. No, I didn't eat for [8:41] days. At that time, I weighed 168 [8:44] pounds. [8:45] >> Then, a college classmate from Hubei [8:47] used to hang out with me. When he [8:50] walked down the street with me using [8:52] that brick phone, he didn't want to be [8:54] seen with me; he said everyone thought [8:56] he was my henchman. [8:57] >> Yijun, tell us about your understanding [8:59] of 5G. For me, 5G boils down to two [9:02] keywords. As everyone mentioned, [9:04] there's the "Internet of Everything," [9:06] and the other keyword is "no latency." [9:08] That’s what interests me the most: [9:10] the fact that there is no time lag. For [9:12] example, when I go online and order [9:14] something now, I'm already getting [9:16] anxious and impatient. I wish it would [9:18] arrive the next day—or even better, [9:20] if I place the order today, I want it [9:22] delivered today. But with 5G, things [9:24] will be even faster. I feel our [9:26] patience will dwindle further. If I [9:28] want to meet or talk to you, there [9:30] shouldn't be any sense of distance [9:32] between us. I find it a bit hard to [9:34] imagine what interpersonal [9:35] relationships would look like without [9:37] any time delay. [9:38] >> That fits, because we’ve done [9:40] audience surveys regarding 5G. Many [9:43] viewers ask a question—and I think [9:45] this is a challenge for both liberal [9:47] arts and science students—which is: [9:49] aren't we fast enough already? Do we [9:52] really need to be that fast? [9:54] >> Well, let me give you a scenario, the [9:57] one Yijun mentioned. [9:58] >> One of the key features of 5G is what [10:00] we call high reliability and low [10:02] latency, capable of achieving one [10:04] millisecond of latency. What’s the [10:05] concept of one millisecond? If you [10:07] prick your hand with a needle, the time [10:09] it takes to feel the pain is about 10 [10:10] milliseconds; this is even faster than [10:12] that. [10:12] >> Goodness. [10:13] >> Ten times faster. [10:14] >> Oh my god. So, what is it used for? [10:17] Theoretically, it’s still used for [10:19] networking, such as in the control of [10:20] autonomous vehicles. We often hear [10:22] experts say, "Oh, self-driving cars are [10:24] a utopia. By the time the signal [10:26] reaches the cloud, calculates an [10:27] algorithm, and comes back to tell you [10:29] to brake or accelerate, you’d have [10:31] already hit someone seconds ago; [10:32] that’s wrong." But with 5G, they will [10:35] deploy distributed cloud computing [10:37] centers under the base stations, which [10:39] we call edge servers, or mobile edge [10:42] servers. It will issue commands from [10:44] the closest possible point. If a base [10:46] station covers a two-kilometer stretch [10:49] of road, it can see the status of all [10:51] the cars within those two kilometers. [10:53] If a car breaks down, it can emit laser [10:56] waves to warn other cars to yield, and [10:59] so on. In this case, the latency [11:00] between the vehicle terminal and the [11:02] base station, the air interface latency [11:04] , can be as low as one millisecond. If [11:06] it were longer, it wouldn't work; it [11:07] has to be at the edge, right? [11:08] Device-edge-cloud. So, where else did [11:11] you say it would be used? Let me give [11:13] an example, uh. [11:14] >> A friend of mine went to the Consumer [11:15] Electronics Show in Las Vegas at the [11:17] beginning of this year, and he saw [11:18] something. He said he found that [11:19] they’ve decoupled the smart glasses [11:21] ...... [11:22] >> from the headset. What does that mean? [11:24] Today, whenever we use VR/AR, we have [11:26] to wear a headset. You can't walk down [11:28] the street wearing a headset; people [11:30] would think you're an alien, right? [11:31] Right? So what is that headset for? Its [11:34] main function is performing 5,000 [11:36] floating-point operations per second. [11:37] It has to calculate, and even then, the [11:39] speed is quite slow. We get dizzy if [11:42] it's below 50 milliseconds. So, for an [11:44] immersive experience, if I select an [11:46] app to view the Antarctic while wearing [11:48] the headset, I’m no longer here; I'm [11:50] in the Antarctic, looking at the sky, [11:52] the penguins, and the seals. [11:54] >> But it can lag at moments like that, [11:57] you know? For example, if we look up at [11:59] these lights, it shouldn't lag for more [12:02] than 15 seconds—no, not even 15 [12:05] milliseconds. [12:06] >> Now, Americans have come up with a [12:07] solution: "I’ll take the computing [12:09] power from the headset, turn it into a [12:11] small laptop, put it in a lady's makeup [12:13] bag, and you can carry it around or [12:15] leave it at home, as long as they [12:17] communicate with each other." If [12:19] there's 5G coverage, it’s one [12:20] millisecond, so there won’t be any [12:22] latency or lag. So it's like [12:24] downloading, downloading to the [12:25] terminal, right? So the calculation is, [12:27] no, it's on the cloud. Right, you're [12:29] absolutely correct. That is the concept [12:31] of edge computing, the place closest to [12:33] the edge. Now, they can just carry this [12:34] computer, wear glasses cooler than [12:36] Google's, and head out onto the street. [12:38] You go out, see a girl wearing a red [12:39] dress, and think, "Wow, that looks good [12:41] ," and your eyes— [12:42] >> You’ll immediately know her phone [12:44] number and home address, right? [12:45] >> If your eyes stay on her for more than [12:47] three seconds, the eye-tracker on the [12:49] glasses will detect that you're [12:51] interested in her outfit. [12:52] >> It will immediately generate a 3D image [12:55] of you wearing that same outfit, [12:56] floating right before your eyes. You [12:59] look at it and think, "I'd look pretty [13:00] good in that." You blink, and it [13:02] immediately searches to find where it's [13:03] sold and how much it costs, and you say [13:05] , "I can buy this." You blink again, [13:06] and it scans your iris for [13:07] authentication, and by the time you get [13:09] home, the item is already there. [13:10] >> It sounds like a dream. [13:12] >> It won't be about money in the future. [13:14] >> It won't be a "shopping spree" tribe [13:15] anymore, but a "gouged-eye" tribe—you [13:16] won't be able to just blink whenever [13:17] you want. Do you remember that fairy [13:19] tale from when we were kids, about the [13:20] gourd or something? You’d give it a [13:22] shake, and when you got home, your [13:23] hallway would be filled with everything [13:24] you ever wanted. [13:25] >> It’s a lot like that. [13:27] >> The most important application for this [13:29] one-millisecond latency is actually in [13:31] the control of intelligent robots and [13:32] remote vehicles. [13:33] >> And remote diagnostics, right? For [13:36] remote surgery, you can’t have a [13:37] multi-millisecond delay; otherwise, [13:38] you’d operate the scalpel here, but [13:39] it wouldn’t move over there, and all [13:40] sorts of things could happen. Meaning [13:42] you could control a surgical scalpel [13:44] from thousands of miles away, while a [13:46] robotic arm performs the procedure on [13:48] the other end. [13:49] >> Just before National Day this year, at [13:51] a heavy machinery exhibition in Beijing [13:53] , Liugong remotely operated an [13:55] excavator located in Guilin—or [13:57] somewhere in Guangxi—and it was [13:58] operating and working with incredible [14:01] precision. [14:01] >> It means things are replicable; [14:03] essentially, what you see is what you [14:05] get, and experience becomes replicable. [14:07] >> Right. [14:08] >> Also, I think 5G will be a huge [14:10] contribution to blockchain. If you look [14:13] at something like Bitcoin transactions, [14:16] it only handles about 2,000 or so [14:17] transactions per second right now. But [14:20] with this, it could potentially handle [14:21] hundreds of thousands or even millions. [14:23] >> Professor Wu is amazing. I can't [14:25] believe he managed to do this. Let me [14:28] tell you. [14:29] >> 5G and blockchain will empower each [14:32] other. First of all, before 5G, we had [14:34] been working on the Internet of Things [14:35] for years without reaching 5G [14:37] capabilities, but we still needed it [14:38] for things like manhole cover [14:40] monitoring—otherwise, heavy rain [14:41] could wash a cover away, and a little [14:43] girl might walk into it and have an [14:45] accident. We did use some bandwidth for [14:47] that monitoring, but there was no [14:48] profit model. It ran for 13 minutes, [14:50] earning the nickname "Thirteen Wolf." [14:53] >> Right? You know about that, too. [14:55] >> But with the time we have today, we [14:57] don't need to ask it to run for 13 [14:59] minutes; it definitely can't. Because [15:01] the speed of a fiber optic cable and [15:02] the speed of a road don't depend on how [15:04] wide they are, but on how many cars are [15:06] on the road at the same time. [15:08] >> This is a statistical multiplexing [15:10] system—a congestion control problem [15:12] that needs to be managed using [15:13] stochastic processes. Chang'an Avenue [15:15] in Beijing, in front of Tiananmen [15:16] Square, has six lanes in one direction, [15:18] yet it will definitely get jammed. At [15:20] night, there isn't a single car; it’s [15:21] a non-uniform distribution. My doctoral [15:23] thesis was actually on this. [15:25] >> So, " [15:26] >> Thirteen Wolf,""Thirteen Wolf"—anyway [15:28] , I'm just stating a principle here. [15:30] Take industrial digitalization for [15:32] example, like steel enterprises. In [15:34] Arkansas, USA, Big River Steel [15:35] connected its enterprise using over [15:37] 50,000 sensors and high-definition [15:39] cameras; this is a typical application [15:41] of 5G. They said, "Originally, 80%of my [15:43] staff were blue-collar, working in [15:45] front of scorching furnaces, sweating [15:47] and wearing goggles. Now, 80%are [15:49] white-collar, clicking mice in front of [15:52] large screens and drinking coffee, just [15:54] like employees at a stock exchange." [15:56] Their per capita production capacity [15:58] for steel is three times the global [16:00] average precisely because of this. [16:02] Someone told me it's five times that of [16:04] China's, so they have improved [16:05] efficiency, right? Just think about [16:07] this streamlined industrial production; [16:09] there's also a chemical company [16:10] planning to invest nearly 30 billion in [16:12] Fuzhou for a chemical plant. They said, [16:14] "I have no workshops, only pipelines [16:15] and furnaces; I hope to build an [16:17] unmanned factory." Because there are [16:19] chemical corrosives and such involved, [16:20] they want to create an unmanned factory [16:22] . They want to build it near a port so [16:24] that when an oil tanker arrives, they [16:26] can hook up the pipes to take in oil on [16:27] one side and produce materials on the [16:29] other. You can imagine this process; we [16:31] estimate it needs at least 50,000 [16:32] sensors to handle these tasks and [16:34] control unmanned management. Just [16:36] imagine, if a malfunction occurs in one [16:37] spot, it doesn't just alarm in one [16:39] place; it will trigger a chain reaction [16:41] , creating massive concurrent [16:42] communication. And 5G has the [16:43] capability to allow 10,000 vehicles to [16:45] arrive on the second ring road [16:46] simultaneously without causing a [16:47] traffic jam—that is this capability. [16:49] Just now, Professor Wu spoke about why [16:51] it would drive the development of [16:52] blockchain. Because the biggest pain [16:54] point of blockchain is that bookkeeping [16:55] efficiency is extremely low. Your [16:57] previous communication systems were [16:58] sequential; they recorded things one by [17:00] one. By the time you recorded a [17:01] transaction, minutes had passed, but 5G [17:03] can handle that. [17:04] >> The professor has a fairly famous [17:06] viewpoint, which I found to be quite [17:08] abstract, almost like a parlor trick. [17:10] He said that while laypeople like us [17:12] haven't figured out 1G, 2G, etc., he [17:15] claims that 1, 3, and 5—from 1G to 5 [17:18] G—the odd numbers are all [17:19] revolutionary changes. 2, 4, and 6 are [17:22] more about optimization; for instance, [17:26] take 3G, [17:26] >> It's regarding this revolution, [17:29] >> Yes, for example, the first generation, [17:30] although it looked as cool as Teacher [17:32] Dou's, carrying those huge bricks, [17:33] you'd often get cross-talk where others [17:35] 'voices would cut in. By the second [17:37] generation, they became small and [17:38] flexible terminals using digital [17:39] encoding to improve quality, but they [17:41] still only solved the issue of voice [17:42] calls between people. The third [17:44] generation created a new scenario: [17:45] saying that with this phone, you [17:47] wouldn't just make calls, you'd also go [17:48] online. I remember there were many [17:50] critics. I tell you, I'm too used to [17:52] hearing this, because back in the 3G [17:53] days, there were even more people [17:55] criticizing it than those criticizing [17:56] 5G now. [17:57] >> I once heard from someone, an insider [17:59] in the blockchain industry, who told me [18:02] he felt blockchain wasn't created for [18:04] humans, but for AI. Because, in reality [18:06] , humans might not be the best carriers [18:09] of information; our ability to process [18:11] information is actually too weak and [18:13] too slow. So, things like blockchain [18:16] plus 5G are prepared for AI. Now, [18:18] speaking of the Internet of Everything, [18:19] like the "what you see is what you get" [18:21] you mentioned—for example, if you're [18:22] wearing a red dress and walk down the [18:24] street, whoever sees it buys it. I'm [18:25] very worried that after walking around [18:27] once, everyone on the street will be [18:28] wearing a red dress; it will just [18:29] become a situation where everyone is [18:31] the same. For instance, when I watch [18:33] someone like Li Jiaqi selling things, [18:35] the traffic is incredibly high. [18:36] Sometimes I feel that he is just a [18:38] carrier of traffic. We are getting [18:40] faster and faster, and once we all [18:42] become carriers of traffic, it's very [18:43] likely we will all become increasingly [18:45] alike. This has actually already [18:47] happened. For example, ten years ago— [18:50] maybe not quite ten years—I remember [18:52] the cover of Time magazine featured the [18:54] "Me Generation." They thought this [18:56] generation only cared about themselves. [18:58] The cover showed a girl lying on the [19:00] ground holding a phone, pointing it at [19:02] herself, saying "this is me," implying [19:04] that this generation only cares about " [19:06] me." [19:07] >> Right, but if you look at those "me" s [19:08] today, they all look very similar. [19:10] >> Yes. Yes. Many of the young people I [19:14] see who come to complain to me can't [19:15] find their direction in life. I've [19:17] thought about it, and it might be [19:19] because they haven't found their true " [19:21] self," even though they are busy [19:22] photoshopping pictures of themselves [19:24] every day. But every "me" is so similar [19:26] , so where is the unique quality of [19:28] that "me"? It's not there. It's the age [19:31] of technology, so you see Professor Lu [19:32] is talking about tech, but we liberal [19:34] arts students still have our liberal [19:36] arts concerns. Right, what if we all [19:37] start thinking more and more alike? [19:39] Whenever anyone thinks about this issue [19:41] he's talking about, I slowly begin to [19:43] understand why we laymen don't quite [19:45] get 5G yet; as far as we're concerned [19:47] right now, it's just for grabbing red [19:49] envelopes or downloading... [19:50] >> Small red envelopes don't need 5G, no. [19:53] >> They say there are already people using [19:55] 5G to grab red envelopes, and it's [19:57] obviously much faster than 4G. I've [19:59] heard them say that. [20:00] >> That's because of the internet speed [20:02] issue. [20:02] >> Right, internet speed, and they also [20:04] say it's for downloading [20:06] high-definition movies... [20:07] >> It just downloads faster, that's it. [20:10] >> That's the consumer side, but I now [20:12] slowly understand that its true [20:14] revolutionary significance is in [20:16] industrial production. [20:19] >> The changes brought about by that, I [20:21] really have this feeling that we will [20:23] be shrouded by something. You just said [20:26] something I don't agree with. For [20:28] example, if a girl wears a red dress— [20:29] say she's the most beautiful woman in [20:31] the world—do you think girls walking [20:33] down the streets of Beijing or Shanghai [20:34] will all be wearing red dresses? [20:36] >> I believe so, because if it exists, [20:38] there will definitely be convergence. [20:39] It's true that there are many girls' [20:41] faces where you can't tell them apart, [20:42] South Korean faces, the way Chinese [20:44] noses look, but don't you think people [20:45] —the girls themselves—have no sense [20:47] of self? Definitely. I recently went to [20:50] a small...let's not underestimate human [20:53] intelligence and people's [20:55] self-awareness. [20:56] >> This will happen, because I saw it in a [20:58] town, a place that to this day doesn't [21:01] even have a supermarket, yet the girls [21:03] in this town dress exactly like [21:05] everyone you see on the internet in big [21:08] cities. Oh, don't you remember the 80s, [21:10] didn't we do the same in the 80s and [21:12] 90s? Didn't Teacher Dou's so-called [21:14] girlfriends back then also pick things [21:17] from French films? What girl today [21:19] would still choose that? [21:20] >> We mainly wore blue and green Mao suits [21:23] back then, right? [21:24] >> Exactly. [21:25] >> And they are thinking more and more [21:27] alike. [21:27] >> Actually, what I'm more worried about [21:29] is that—I think—it's that part [21:31] about artificial intelligence; Yijun is [21:33] indeed right about that. Even if it [21:35] gets stronger and stronger, like [21:37] Kurzweil said—I saw his point about [21:39] 2045, but he changed it last year, he [21:41] said 2029, so that's only 10 years away [21:43] . If computer intelligence surpasses [21:45] the capability of the human brain, can [21:47] this only happen in a 5G or 6G [21:49] environment? At that point, I'm [21:51] actually quite worried: does the [21:53] machine really have a form of [21:54] consciousness? Many scientists now [21:56] believe that machines have aesthetics, [21:58] which they acquire through learning. If [22:01] you haven't trained them, at this [22:03] current stage—what we call the AI [22:05] stage—there's actually no need for [22:08] concern. Just as we developed cars [22:10] hoping they would be faster than humans [22:12] , and cranes hoping they would be [22:15] stronger, it is perfectly normal that [22:17] an AI we developed can defeat a world [22:19] champion at Go. [22:20] >> But I recall that even Putin expressed [22:22] this concern, asking whether this might [22:24] eventually lead to the development of a [22:26] type of warrior. Ah, as for that, a [22:28] real warrior...yes, they treat it like [22:32] nuclear weapons. Major powers should [22:35] have agreements, viewing it as a [22:37] nuclear weapon, and perhaps agreeing [22:39] that we will not be the first to use it [22:42] . [22:42] >> There are two issues involved here. One [22:45] relates to Artificial General [22:47] Intelligence. Artificial General [22:48] Intelligence means that the AI we talk [22:50] about, even if it beats a world [22:51] champion, can only play Go. [22:52] >> Right. [22:53] >> It can't drive a car or play golf at [22:55] the same time. Therefore, the day it [22:57] can become a warrior or operate an [22:58] autonomous car, it becomes Artificial [23:00] General Intelligence—a collection of [23:02] multiple capabilities. [23:03] >> Do you think we will see this in our [23:05] lifetime? [23:05] >> We will definitely see it. [23:07] >> Oh, so from your perspective... [23:09] >> Do you think there will be robots [23:10] hosting talk shows? [23:11] >> Uh, that possibility exists. Nowadays, [23:14] machines can write scripts and poems, [23:16] but you have to teach the machine all [23:19] those behavioral characteristics. The [23:21] machine can even make some inferences, [23:24] but let me give an example. If you [23:27] teach a group of people and a machine [23:29] some facts, and then you ask them who [23:32] the US President is, they will both [23:34] quickly answer Trump, right? But if you [23:37] ask, "How did you sleep last night?" [23:39] the machine definitely won't be able to [23:40] answer. That's the point; the machine [23:42] must be taught by humans, and that's [23:44] deep learning. [23:44] >> Ah, but the important role of 5G is [23:46] that it realizes the interconnection of [23:48] everything. Actually, the biggest [23:50] impact of connection is that 5G is a [23:52] very fundamental communication and [23:54] network technology. The role of the [23:56] network is to enable us to obtain [23:58] valuable data relations through [24:00] connections. Honestly, 5G might not [24:01] necessarily make money—I've always [24:03] held this view—but the processing, [24:04] manipulation, and application of the [24:06] data behind 5G is what can make it [24:07] profitable. Like the experiment we just [24:09] mentioned, it requires a massive amount [24:11] of learning, and 5G provides the [24:12] nourishment for this learning through [24:14] these connections. So we believe it [24:16] might happen, as humanity is also [24:18] moving in that direction. Could there [24:20] be an AI that, with the help of 5G, [24:23] learns to become a monster? [24:25] >> That is what Hawking predicted, his [24:27] biggest fear of something akin to [24:29] autonomous will, perhaps... [24:30] >> Perhaps this thing could even exist as [24:32] a data stream on the internet, wreaking [24:34] havoc everywhere. It could, then... [24:37] >> Because just as Professor Lu said, for [24:38] example, "Are you asleep yet?" I think [24:40] when people hear this question, they [24:42] wonder, what does it mean to sleep? [24:43] Like lying in bed, not doing anything, [24:45] eyes closed, for those six hours—that [24:46] is called sleeping. But if someone [24:48] defines it, what would it mean for a [24:50] robot to sleep? For instance, if its [24:51] data intake falls below 100 terabytes [24:53] during a certain period, that’s sleep [24:55] , and it might tell you, "I slept quite [24:57] well." [24:57] >> Right, and I have another question, [24:59] that thing called cognitive awakening, [25:01] right? [25:02] >> And also, your point has stimulated my [25:04] brain to think on many levels. Many of [25:07] our viewers have asked about "Black [25:09] Mirror," that famous British sci-fi [25:11] anthology series that imagines many [25:13] futuristic scenarios. Could you, [25:15] Professor Lu, talk about whether the [25:17] arrival of the 5G era will make many of [25:19] the scenes in "Black Mirror" a reality? [25:21] I feel that in "Black Mirror," for [25:24] example, I saw one episode, it implies [25:26] that we could, say, wear full-body [25:29] sensor suits. We could gain, for [25:31] instance, just like you said with VR [25:34] technology, I could be in Hawaii right [25:36] now. [25:37] >> Yes. [25:38] >> I could smell the ocean waves, I could [25:41] even feel the sea breeze, and one [25:43] person could be in New York, another in [25:45] Tokyo, another in Beijing, all having a [25:47] bonfire party together on a beach in [25:49] Hawaii. Right. Even though we are each [25:53] in our own homes, our "lives" or [25:56] whatever you call it, would be in that [25:58] virtual space, and we would truly be [26:01] interacting with a sense of touch. [26:03] >> Each one of us becomes a node in the [26:05] entire network. Actually, I feel it [26:07] turns into something neither human nor [26:09] non-human. Originally, humans were [26:11] humans; without physical contact, we [26:12] weren't together. In the future, we [26:14] will be connected to the entire network [26:16] . [26:16] >> It's called something like, [26:18] machine-biological integration. [26:20] Biological engineering. [26:22] >> You know, in the situations described [26:23] in "Black Mirror," like this kind of [26:25] virtual reality—if the realism is [26:27] high enough, actually, the reason we [26:29] don't want to go in right now is just [26:30] because it isn't realistic enough yet. [26:32] >> If it really gets strong enough, it [26:34] will be like what was in Black Mirror. [26:36] For instance, two buddies, me and Xiao [26:39] Bo, we put on those electrodes, and [26:41] boom, we enter a game. But in the game, [26:43] he’s a beautiful woman, right, and [26:45] I’m a handsome guy. Slowly, in this [26:48] virtual space, we develop feelings for [26:50] each other. You know? When we return to [26:53] the real world, we’re still just two [26:55] guys, right? But it feels a bit [26:57] unfulfilling, because what we were [26:59] playing in there was like, [27:00] >> it was like a dating game. What I mean [27:03] is, this kind of virtual sensation is [27:05] like augmented reality. If it gets [27:07] strong enough, [27:08] >> have we ever considered the changes to [27:10] human personality? [27:11] >> Mm. [27:11] >> Of course it will. What you just [27:14] mentioned is exactly what our [27:15] university’s Professor Zhang Ping, [27:17] who was just named an academician this [27:18] year, could address. He specializes in [27:20] mobile communication technology, and he [27:22] has already held two forums this year [27:24] discussing 6G. He believes that after [27:26] mobile communication technology solved [27:28] person-to-person calling and [27:29] person-to-computer internet access, the [27:31] next step with 5G is to connect the [27:33] Internet of Everything, the Internet of [27:34] Vehicles, and the Industrial Internet. [27:36] Where do we go after that? He thinks 6G [27:39] is about connecting the human soul. [27:41] >> Well, I think—because of the arrival [27:43] of 5G...The question isn't whether [27:46] people have souls. [27:47] >> Right, that’s exactly what he meant [27:49] by "soul." Let me give you an example. [27:51] Not long ago, it might have been [27:53] Zhejiang Mobile that provided police [27:54] with a piece of high-tech headgear; [27:56] there are images of it online that you [27:58] can find. A police officer boards a bus [28:00] to check IDs. After getting on, he [28:02] stares at someone and says, "Please [28:03] show me your identification." The [28:05] person says, "My name is Zhang San." [28:07] The officer says, "No, that's wrong, [28:08] your name is Li Si." At this point, [28:10] because of facial recognition, you [28:11] could even walk into a bank and, before [28:13] you've said a word, they’d say, "Sir, [28:14] you're here—one of your financial [28:16] products expires today." They have big [28:18] data behind them, you know? Then they [28:20] say, "Please sit and have some tea. Are [28:21] you going to renew it?""Or how would [28:23] you like to use that money?""We have [28:25] new products here." They already know [28:27] exactly what you’re thinking. All [28:29] these years, our clicks, browsing, and [28:31] consumption have essentially helped [28:32] internet companies create what we call [28:34] a "network clone" or "digital twin." [28:36] Their initial intent in profiling us [28:38] was to save you the search cost amidst [28:40] massive amounts of data, pushing [28:42] content to you so you can more easily [28:44] find what you're interested in; for [28:46] example, if I once liked something, [28:48] they'll push similar things to me. [28:50] >> This is what you mean by something [28:51] similar to a "quasi-soul." The soul, it [28:54] knows what you're thinking. But if we [28:56] take it a step further, if it knows [28:58] everything I'm thinking, through the [29:00] various interactions and collisions [29:01] between these "souls" you mentioned— [29:03] like how our discussion generates many [29:05] new questions—it will form a learning [29:07] process for the soul, a process of [29:09] reasoning. [29:09] >> At that point, for instance, the four [29:11] of us could just sit here, wear a [29:12] headset, and we wouldn't even need to [29:14] speak. [29:14] >> With a hum, we enter a virtual space— [29:16] this is what Musk wants to do, and [29:18] he’s already doing it—then, [29:20] >> the audience is also connected to this [29:22] virtual space of thought. No, let me [29:24] tell you, I feel there are two things [29:25] happening now. One is what they call " [29:27] calm technology" for the future. You [29:29] see, Bill Gates said years ago that [29:31] everyone should have a computer on [29:32] their desk, but now that technology is [29:34] "noisy," something you can see and face [29:36] head-on. In the future, it will be " [29:38] calm technology"—you won't even know [29:39] where it is, it will already be [29:40] integrated into you. [29:41] >> Actually, in the early days of humanity [29:43] , if we were farmers, we worked 8 to 12 [29:45] hours a day, and after cooking and [29:47] eating, we went straight to sleep. Our [29:49] time was fully occupied by labor, and [29:51] then that eventually became an 8-hour [29:53] workday. Some have woven a cocoon [29:55] around us, wrapping us up so we can no [29:57] longer see the outside world; we’ve [29:59] become frogs in a well. We only ever [30:01] see what we already focus on, the [30:02] viewpoints we believe to be correct, [30:04] which are constantly being reinforced [30:05] for us. This is a problem brought on by [30:07] the internet, a major issue—can we [30:09] save ourselves? I... [30:10] >> So, 99%of people cannot be saved. [30:13] >> You are being pessimistic. [30:14] >> However, on the internet...The internet [30:16] is also beginning to open up those [30:18] applications that lean toward [30:19] personalization. Isn't that how TikTok [30:21] was suddenly able to stand out from [30:23] among other social apps? In the past, [30:25] we had Weibo. Weibo was essentially [30:27] just a concept of an online diary. Its [30:29] information capacity was limited. Since [30:32] I don't know you and you don't know me, [30:34] why would I follow you? I might act out [30:36] , curse, or pick fights just to grab [30:38] attention. But why did TikTok later [30:41] become so popular? Notice how it's very [30:43] grassroots. You don't need to be a big [30:45] influencer, because why? I'm not famous [30:47] , and I don't have that much creativity [30:49] , but I caught a car crashing into a [30:51] utility pole on the street, and it got [30:53] over 10 million views. This is where [30:55] cyberspace, my personal application, [30:57] and real life finally merged. Once it [31:00] merged with reality, we saw that [31:01] everyone focuses on different things, [31:03] leading to personalization. Exactly. [31:06] But if this piece of information—like [31:08] the crash you mentioned—goes viral, [31:10] we might spend all eight hours of our [31:12] day talking about it. [31:14] >> It shapes my thoughts, but from another [31:16] perspective... [31:18] >> I think there's this guy, Kahneman, the [31:20] Nobel laureate in Economics, who works [31:22] in behavioral psychology. What he said [31:24] really resonates with me: the human [31:26] brain has two systems, System 1 and [31:27] System 2. System 1 is intuitive, like [31:29] when I call you Wentao. I don't need to [31:32] think to know your name is Wentao. It's [31:34] about feelings, hunger, liking a girl [31:36] when you see one—it's pure intuition. [31:38] System 2 is rational. [31:39] >> Realizing that woman is my mother, so I [31:41] can't talk nonsense—that's [31:42] rationality. [31:43] >> Or, if I haven't paid for something, I [31:45] can't eat it. Even if I really want a [31:47] cigarette, I have to hold back. [31:49] Kahneman says that humans are actually [31:51] animals driven by System 1; in most [31:53] states, we function via System 1. [31:55] Furthermore, all technological [31:57] revolutions in material civilization [31:59] are aimed at enhancing our System 1. [32:01] That's why 99%of people are just...the [32:03] future? What's truly worrying is that [32:05] after robots emerge, we become lazier [32:07] and increasingly entrust our will and [32:09] choices to System 1. Harari, who wrote [32:11] "Sapiens," came to China, and I was [32:13] quite shocked during his lecture. He [32:14] studies the future of humanity, yet he [32:16] doesn't use a smartphone. [32:17] >> Of course he doesn't, he never uses [32:19] those things. [32:20] >> Right, he doesn't use a smartphone, and [32:21] he meditates for two hours every day. [32:23] It's because he is someone with an [32:24] exceptionally strong System 2. [32:26] >> Oh, so he refuses to be controlled by [32:27] this thing. He doesn't. I think those [32:29] students who spend their days scrolling [32:30] through Kuaishou and TikTok are [32:32] actually people whose System 1 is [32:33] becoming increasingly powerful. Their [32:34] System 1 is getting stronger and [32:35] stronger. [32:36] >> Constantly reinforcing their System 1. [32:38] >> Aren't they like lab mice? [32:39] >> Look at those lab mice, aren't they— [32:40] >> Listen to our show and you'll [32:42] eventually realize. [32:43] >> They are like lab mice; as long as you [32:44] press a button, you get that dopamine, [32:46] that instant pleasure. Look at those [32:48] mice, they keep pressing it until they [32:50] potentially kill themselves from the [32:51] high. Sometimes you just keep giving it [32:53] to them. But look, why do some people [32:55] say, "Oh, I scrolled Kuaishou for two [32:57] hours, TikTok for two hours"? Those two [33:00] hours are when they are 100%controlled [33:02] by their System 1. So why can some [33:03] people stop after 15 or 20 minutes? [33:05] Because their System 2 capability is [33:07] relatively stronger. I believe we still [33:09] need to exercise our System 2. But [33:11] Xiaobo, do you think humans can be [33:14] calculated? [33:14] >> I think there will definitely come a [33:16] day when machines understand us better [33:18] than we do ourselves. I believe that [33:19] day is already approaching. 5G and the [33:22] Internet of Everything are indeed [33:23] intended to generate some—what I mean [33:26] is that 5G connectivity brings about [33:28] the utilization of intelligent data. [33:29] However, most technical experts [33:31] currently—the communications circle [33:33] and the AI circle are isolated from [33:35] each other. They are two different [33:37] circles, and they are more concerned [33:39] with where this can be used. Used for [33:41] machine intelligence, but what problems [33:43] machine intelligence solves hasn't [33:45] reached us humans yet. Let me give you [33:47] an example: when Musk visited China, [33:48] his conversation with the leaders [33:50] wasn't on the same wavelength. Later, [33:52] it turned out that even Jack Ma wasn't [33:53] on the same wavelength either. [33:55] >> Why is that? He said, "You all end up [33:57] making electric cars, vehicle [33:58] networking, it's still a car in the end [34:00] . What I'm making is a mobile [34:01] communication terminal; I don't call it [34:02] a car." [34:03] >> Hmm. [34:03] >> Do you understand what that means? [34:05] >> Oh, what you're saying is so [34:06] interesting. [34:07] >> What is a mobile communication terminal [34:08] ? It’s a mobile device that serves as [34:10] the interface between humans and [34:12] network applications. I have all kinds [34:13] of applications that can come in [34:15] through this terminal. A couple of days [34:17] ago, Samsung was sued for making a [34:18] smart fridge that could go online. You [34:20] put your hand here, and it turns [34:22] transparent. [34:23] >> They made a model where you don't even [34:24] have to open it; you can just see [34:26] what's inside. Look at eggs, once they [34:28] become internet-connected, everything [34:30] can go online. Parents buy them, and [34:31] then they complain to me, saying "I [34:33] took away my child's phone and computer [34:34] , but I saw them online again." It [34:36] turns out they were using the [34:37] refrigerator to get online. So the [34:39] problem arises, right? But the Internet [34:41] of Everything will bring new social [34:42] issues. Let's loop back to Musk's [34:44] vision. If we develop the Internet of [34:46] Vehicles, hoping to make cars smarter, [34:49] our goal isn't just self-driving or [34:50] anything like that. Let me give you an [34:53] example. Why is the government pushing [34:55] electric vehicles so relentlessly? [34:56] First, for energy security. In case [34:59] something goes wrong, what if others [35:01] cut off your oil supply? China still [35:02] has other power generation facilities, [35:04] so you'll still have electricity, and [35:06] your car won't be stranded. More [35:08] importantly, the average car is only [35:10] driven one-twelfth of the day. That's [35:12] an average of two hours per car; it’s [35:15] idle for over 20 hours. You don't need [35:17] that much time to charge. If you have [35:19] an electric vehicle, and it becomes [35:21] smart—let me tell everyone—if it's [35:22] connected, what Musk and the others are [35:24] thinking is that you can sell [35:26] electricity. Because I can know your [35:27] travel radius doesn't exceed 50 [35:29] kilometers, I'll charge you just enough [35:31] to cover that. Then, when electricity [35:32] is expensive, I'll stop it, and when [35:34] it’s cheapest, I’ll charge it for [35:36] you. When social demand for electricity [35:38] suddenly spikes, as long as there's a [35:39] network and an intelligent solution, I [35:41] can ask if you want to sell your power. [35:43] >> You could sell your electricity for [35:44] double the price, you know. [35:46] >> That's what they're doing. Connectivity [35:48] is meant for many intelligent uses. I [35:50] get it now; it can help you live more [35:52] correctly. [35:52] >> We make mistakes in our daily lives all [35:54] the time now. [35:55] >> In the future, it will help you be more [35:57] precise. I went to Taipei a while ago. [35:59] A friend of mine bought a Tesla. I’m [36:01] not advertising for them—they don't [36:03] need us to—but he said something. [36:04] >> He said, "This isn't a car.""This is a [36:08] computer." [36:08] >> Exactly, a computer you can live in. [36:11] >> This completely changed my perspective; [36:13] it is a computer. People used to say a [36:15] car is a sofa on four wheels; in the [36:17] future, a car should be a computer on [36:19] four wheels. [36:19] >> It doesn't even have an engine, no [36:21] engine or transmission. [36:22] >> Sitting inside is pure entertainment; [36:24] you don't even need to look at the road [36:25] , actually. Hey, Xiao Bo, since you [36:27] study economic history, can you predict [36:30] how the 5G era will impact human [36:32] economic models and business operations [36:35] in the future? Can you imagine it? [36:37] >> I believe that many jobs involving [36:40] standardized tasks will disappear. Take [36:42] the automotive industry, for example: [36:44] truck drivers, about 2.5 million jobs [36:46] in the U.S. could be gone. [36:48] >> What do you think they could do instead [36:50] ? Build charging stations, perhaps? [36:52] We'll need so many of them. That point [36:54] touches on a debate that has long [36:56] occupied the fields of economics and [36:57] philosophy, and I have my own [36:59] perspective on it. [37:00] >> When the second wave of humanity, the [37:02] Industrial Revolution, arrived, it [37:04] actually caused panic in Europe because [37:06] the emergence of machines meant one [37:08] machine could do the work of ten [37:10] craftsmen. [37:10] >> That’s why the first thing they did [37:12] in England was smash those machines. [37:14] Most people were laid off. Today, [37:16] people say AI might be the first [37:18] technology in human history to destroy [37:20] more jobs than it creates. Doesn't that [37:23] cause a problem? If this is [37:25] economically irrational—if everyone [37:27] is unemployed—who will consume all [37:29] the goods produced by such efficient [37:31] production capacity? That fear spread [37:33] throughout Europe at the time, leading [37:35] to mass economic depressions and even [37:37] World War I and World War II. But later [37:38] , people realized that because one [37:40] person could do the work of ten, that [37:41] one person could just oversee the [37:42] machine. By then, their wages had [37:44] increased tenfold and their vacation [37:45] time grew, so what were they going to [37:47] do? They wanted leisure and [37:49] entertainment, which gave rise to the [37:51] tertiary sector. That’s why I often [37:53] tell this story. I say the primary [37:56] sector is raising cattle and sheep; the [37:58] secondary sector is slaughtering them [38:00] —Professor Wu is surely familiar with [38:02] this common economic joke—then the [38:05] tertiary sector is eating the beef and [38:07] mutton. What is the quaternary sector? [38:10] Boasting and making a spectacle of [38:12] yourself. [38:12] >> We are absolutely in the sunrise era of [38:14] the quaternary sector, a concept [38:15] debated for years that is now finally [38:17] being realized. Take livestreaming, for [38:19] instance; in the end, they earn more [38:21] money than those doing the hard [38:22] physical labor. A kid once took his [38:25] grandfather's pension and tipped away [38:27] over 200,000 yuan in a year. While we [38:29] don't advocate for that kind of [38:31] nonsense, we do support the creative [38:34] industries, cultural industries based [38:36] on the web—much like what we're doing [38:38] right now, in a way. [38:40] >> And there's one more thing: starting [38:42] last year, a unicorn industry emerged [38:44] that is incredibly profitable—a [38:46] meditation app that helps you clear [38:48] your mind. Everyone, think about it, [38:50] when you receive so much information, [38:52] you’re generally anxious, and it [38:54] turns out the most profitable thing is [38:55] meditation. [38:56] >> Look, look, we aren’t just lecturing [38:58] or spending time giving classes; what [38:59] is this way of making money now? You go [39:01] up to the mountains, pay 20,000 yuan [39:03] each, and they take away your phones [39:04] for silence. [39:05] >> Silence, yeah. [39:06] >> Silence, three days of silence. [39:09] >> The teacher hasn't even said anything [39:10] to you yet, you paid 20,000 yuan, it's [39:11] great, even better than my lectures, [39:13] even better than Teacher Lin's. [39:14] Furthermore, I think it will bring [39:15] about a change in the world, a change [39:17] in worldviews. What kind of change in [39:19] worldviews? Well, I keep hearing you [39:21] all talk about the "Internet of [39:23] Everything," the Internet of Everything [39:24] . You know, the first time I heard that [39:27] term, I thought of that movie called [39:29] Avatar. Is it that movie? Do you [39:31] remember Avatar? That planet in Avatar [39:34] is very interesting. It’s actually [39:37] those trees. And those indigenous [39:40] people on the planet, don't you [39:42] remember them twisting a braid from [39:44] themselves onto the tree, and they’d [39:46] connect to the planet's consciousness? [39:50] In the end, you see, the indigenous [39:52] people living on this planet, their [39:54] entire cycle of life and death seems to [39:56] be linked to the branches of the trees [39:59] on this planet, all through their [40:01] braids—is that how I remember it?— [40:03] they connected instantly. Of course, in [40:06] their philosophy, it’s about [40:08] connecting to the existence of all [40:11] things, connecting to this planet, [40:13] meaning the planet is a networked [40:15] community where humans and other [40:17] species have an interface to connect to [40:20] this common planet. Moreover, the [40:23] Internet of Everything creates a [40:25] management issue; just think, such a [40:27] massive system that knows everything. [40:30] It’s helping us make the most correct [40:32] arrangements, guiding us along. Right? [40:35] What to eat if you're too fat, [40:37] arranging a nutritional meal for you, [40:39] what clothes you should wear, what you [40:42] like—bam, bam, bam, it knows all the [40:44] global resource information. And then [40:46] think about it, this includes where our [40:48] production goes. [40:49] >> I want to ask, where is the center of [40:51] this network? What are the values of [40:53] this network right now? [40:54] >> Is this network decentralized? [40:56] >> Yes, the network is currently moving in [40:57] another direction: decentralization, [40:59] simply because it's too big. Yes, [41:00] blockchain, it’s the fundamental [41:02] doctrine of blockchain. [41:03] >> Because it’s too big? It’s actually [41:05] unsafe; once you capture this [41:07] centralized node, the entire system is [41:09] at risk. The emergence of the Internet [41:11] was meant to be decentralized. [41:12] Traditional telephone networks used a [41:13] hierarchical, radial switching system [41:15] —hub offices, regional offices, and [41:16] local offices—much like our [41:17] administrative structure: the Party [41:18] Central Committee, provincial [41:19] governments, municipal governments, and [41:20] county governments. The Americans asked [41:21] : Can we build a network where the [41:23] nodes are independent of each other, [41:25] where no one is a superior or [41:26] subordinate, and everyone is equal? It [41:28] is called a P2P architecture; " [41:30] peer-to-peer" means equality. Thus, in [41:32] 1969, they used packet-switching [41:34] technology to create the first military [41:36] data command network, ARPANET. Its [41:38] popularization among the public brought [41:40] about today’s internet. That is why, [41:42] in his remarks at Wuzhen a few days ago [41:44] , President Xi mentioned that this year [41:46] marks the 50th anniversary of the human [41:48] internet, referring to that 1969 U.S. [41:50] military data network. The original [41:52] concept was to take those vertical, [41:54] centralized systems and decentralize [41:57] them. Indeed, that was the initial [41:59] intent of the internet at the time. [42:00] Today, blockchain technology has [42:02] emerged, aiming to achieve [42:04] decentralization not just physically, [42:06] but within the software architecture [42:08] itself. [42:09] >> So, [42:10] >> I would like to ask: when everything is [42:11] interconnected, will humanity still be [42:13] the center? Is it possible for "things" [42:15] to surpass humans? [42:16] >> Addressing this issue—the one [42:18] affecting humanity—ultimately [42:19] involves artificial intelligence. While [42:21] some argue that AI might replace [42:23] humanity, others believe that AI should [42:26] be a slave to humanity, not its master. [42:28] >> Exactly. We must design from this [42:30] perspective; human intelligence lies in [42:32] the fact that everything we develop is [42:34] made to be controllable and manageable. [42:36] Let me give you an example. I [42:38] participated in writing China's first [42:40] internet report. In 1993, when the [42:42] Clinton administration proposed the " [42:44] Information Superhighway" initiative, [42:47] >> the formal name was the National [42:48] Information Infrastructure plan, and [42:50] our country launched a project to study [42:52] what it was, why it mattered, and how [42:54] to implement it. I believe many people [42:56] analyzing China's rise—attributing it [42:57] to centralized leadership, the ability [42:59] to pool resources for major projects, [43:01] or our demographic dividend—have [43:02] missed the most important point. The [43:04] most critical point I made in that [43:05] report, which I wrote at the time, was [43:07] a quote from Gu Chaoqi, secretary to [43:09] Zhu Rongji: "Informatization is the [43:10] multiplier for our industrial and [43:12] agricultural modernization; it [43:13] accelerates development." We cited the [43:15] rise of the UK, which occurred because [43:16] they caught the wave of industrial [43:18] modernization. Regarding China's rise [43:19] over the past 40 years, Americans have [43:21] always wondered how China could become [43:22] the world's second-largest economy in [43:24] such a short time. The truth is, we [43:25] caught the express train of information [43:28] technology application. Today, [43:29] blockchain and digital currency bring [43:32] many highly efficient social [43:34] advancements, but they also have many [43:36] problems. Too few of us are studying [43:39] how to solve the "brake" issue, right? [43:41] >> The problem of the "cage"—if you [43:43] solve the cage, it’s released, but [43:45] can it actually be used? [43:46] >> China's current fear regarding [43:47] blockchain lies in where the key is. [43:49] Exactly, where is the key, where is the [43:51] cage, how do you... [43:52] >> But I really don't think this is just [43:54] China's problem. For example, like the [43:56] biological gene editing technology I [43:58] talked about in the last episode, the [44:00] reason it can't be unleashed is because [44:02] the whole world is asking if this thing [44:05] can be reversed. If you modify human [44:07] genes and put them into the body, can [44:10] they be brought back? It’s just as [44:12] you said, can this thing hit the brakes [44:15] , can it be erased? And you mentioned [44:17] this Internet of Everything. [44:19] >> Does it have brakes? [44:20] >> This eventually becomes a global [44:22] governance issue, meaning the world [44:23] might have to sit down and ask, "Where [44:24] are the brakes?" [44:25] >> Is it a matter of human destiny? [44:27] Because... [44:28] >> The Internet of Everything brings with [44:30] it another major issue: the protection [44:33] of security and privacy. Exactly. [44:35] >> Data and network security; so now [44:37] abroad, there is a popular practice [44:39] where governments even allocate [44:40] spectrum frequencies to certain [44:41] enterprises. For instance, Brazil's [44:43] Vale—they aren't short on money. They [44:45] say, "If the carriers can't look after [44:47] my needs, I'll pay for the network [44:49] myself," but you don't have the [44:50] frequency. As I just mentioned, numbers [44:52] , frequency, and orbital slots are [44:53] essential communication resources. So, [44:55] you are given a frequency. The U.S. has [44:58] also given frequencies to some [44:59] companies, with the provision that you [45:00] can only use this frequency for your [45:02] own private network. Mhm. [45:04] >> But it cannot be used for public [45:06] services. Once they have this, the [45:07] companies will fully invest in building [45:09] the network. This is also a point I [45:11] really want to convey. I have always [45:13] believed that the opening up of China's [45:14] telecom industry should reach this [45:16] stage and should reflect on this issue. [45:18] For 5G and the Internet of Everything [45:19] to embrace the real economy, many [45:20] companies have the capability, but if [45:22] you don't give them frequencies, they [45:23] have to borrow from carriers. That’s [45:25] somewhat like reselling network access, [45:26] which is problematic from a regulatory [45:28] standpoint, you know? [45:29] >> Because the frequency was allocated to [45:30] you—a resource allocated to you for [45:32] free—and you go and lease it out; [45:33] isn't that just being a middleman? Do [45:34] they control the right to the access? [45:36] >> Exactly, I’ve always thought this is [45:38] great, but...What surprises me is that [45:40] many of our carrier executives...Most [45:41] people don't really support this [45:43] approach, and it's not just in China. [45:44] Germany has been buzzing about this [45:46] lately; the German government is [45:47] planning to grant spectrum to [45:48] enterprises, and companies are saying [45:49] they’ll build their own private [45:51] networks. There’s also an important [45:52] reason for this: I want to build a [45:53] closed network, an "Internet," an [45:55] isolated network that doesn’t connect [45:56] to the outside, because I have data [45:58] issues. That’s the problem, but [45:59] it’s something that can be done. In a [46:01] 5G environment, if they want to do it, [46:02] you have to allocate them the frequency [46:04] . Now, the cloud providers are getting [46:05] anxious, saying, "You were supposed to [46:07] be my customer, but now that you're [46:08] building your own, who am I going to [46:10] charge?" Then they realize, since the [46:11] companies are investing in the build [46:12] but lack the maintenance and [46:13] operational capabilities, perhaps [46:14] they’ll outsource it to you, right? [46:16] You then become the service provider, [46:17] you don't need to invest, and you can [46:19] still make money, can't you? Many [46:20] people haven't thought this through; [46:21] it’s still that old mentality of [46:23] keeping benefits within one's own house [46:24] . So, Xiaobo, you seem to be an [46:26] optimist when it comes to future [46:27] technology, right? [46:28] >> Even Hawking was deeply worried before [46:30] he passed away, but he might have been [46:32] right that, in the future, we might [46:33] ultimately need regulation. [46:35] >> Yes. [46:36] >> Like rules—for example, currencies [46:37] backed by gold or fiat currency, or [46:38] many of the monetary systems in [46:40] economics back then. Exactly. [46:42] >> Of course, it's just folk rumors, and I [46:43] don't know if it's true, but people say [46:45] the internet speed is slowing down, and [46:46] it's because many resources are being [46:48] diverted to build 5G. [46:48] >> Huh. Is that so? [46:51] >> Is it possible? Why? Doesn't [46:53] non-standalone networking share some [46:55] infrastructure with 4G? So if you [46:57] divert some capabilities to 5G, [46:59] naturally, some 4G performance will [47:00] slow down; that is indeed possible. [47:02] >> You know, so what does this mean? I [47:05] feel that when we talk about technology [47:07] , like the arrival of the 5G era, it [47:10] eventually comes down to human nature. [47:12] There’s a Wong Kar-wai film that says [47:15] all his movies are basically about one [47:17] thing: you can't go back, you know? I [47:19] mean, you can discuss it, but I feel [47:21] like the path of human technological [47:24] development is just like time—can [47:26] time ever go back? Time, no. The [47:29] arrival of the technological era means [47:31] that humans have never been able to [47:33] possess a capability and just choose [47:34] not to use it. Never mind whether this [47:37] capability is good or bad; once humans [47:39] have it, they have to use it. This [47:41] usage will create new problems, but [47:43] even with new problems, we should still [47:45] remain positive. It creates new [47:46] problems, have you noticed? Human [47:48] strategy remains relying on new [47:50] technology to solve these new problems, [47:52] while simultaneously creating even [47:53] newer ones. In short, there are always [47:55] more solutions than problems; that's [47:57] called being an optimist. I looked back [47:59] at Alvin Toffler's "The Third Wave," [48:01] written in 1980, just a couple of days [48:02] ago. He had some things to say; what [48:04] were they? He said that the computer is [48:06] no longer just hardware; it is a form [48:08] of a new civilization. Therefore, we [48:09] have been swept into an era where [48:11] economic and social orders are being [48:13] completely disrupted. He remembered [48:15] saying that we don't know what will [48:16] happen in the future, but we are [48:18] enthusiastically participating in the [48:19] construction of this new civilization. [48:21] We’re almost all in, and though we [48:23] don't know what the future holds, [48:24] everyone is genuinely happy to talk [48:26] about it. Think about someone who [48:28] doesn't use a mobile phone—everyone [48:29] around them holds their phones to help [48:31] them, and their information is manually [48:33] provided to them by others. So I say, [48:34] in the end, in this 5G or 6G era, the [48:37] person at the very top might well be [48:39] the one who doesn't use a phone. [48:40] >> Right, the future elite won't be using [48:42] their own phones. [48:43] >> Exactly. [48:45] >> When I teach my 5G and 6G courses [48:47] online, I often tell this exact story. [48:50] Uh, early this year, the Minister of [48:51] Industry and Information Technology [48:52] said that everything would be [48:53] personalized live streaming in the [48:54] future. What is personalized live [48:56] streaming? It means that before, when [48:58] you watched a soccer game, the main [48:59] camera always followed the ball. [49:01] >> Mm. [49:02] >> In the future, it won't matter whether [49:03] I watch the ball or not; if I want to [49:05] watch Ronaldo or watch Messi, I can. [49:07] Does this exist? This summer, at a [49:09] communications exhibition in Shanghai. [49:11] Liaoning TV collaborated with the [49:13] Japanese on a demonstration. Ah, [49:15] because the Japanese are planning to [49:17] build a personalized live streaming [49:18] platform for next year's Tokyo Olympics [49:20] , a broadcast platform that I think [49:22] will bring massive changes to new media [49:24] . Ah, this change—we tried out this [49:26] broadcast platform. There are so many [49:28] games going on; you click to say you [49:30] want to watch this one, watch soccer, [49:32] and you see Portugal versus Argentina [49:33] is currently being broadcast. You tap [49:35] it, and it brings up 23 different [49:37] angles: 11 players on one side, plus [49:38] the goalkeeper, plus the angle [49:40] following the ball—23 scenes total. [49:41] Which one do you pick? I said I’d [49:44] pick Ronaldo, and I tapped Ronaldo; [49:46] dozens of scenes appeared. Some images [49:48] were quite dim, lagging, and stuttering [49:51] , while some were super clear and right [49:53] in the center. Later I asked him why it [49:56] was so good. He said we simulated this [49:58] person who bought a 5,000-yuan ticket, [50:00] sat in the best seat with a high-priced [50:03] ticket, and used an 8,000-yuan 5G phone [50:05] to live-stream in 4K. [50:06] >> It was filmed by an audience member. [50:08] >> This is what we call "no live-stream, [50:10] no dissemination." Platforms like [50:12] Douyin will become like this in the 5G [50:15] era; everything will be live. So, if I [50:17] am holding this phone and [50:18] live-streaming, I thought, this is a [50:20] great scenario. I’ll just post a QR [50:22] code, and people can pay one yuan to [50:24] watch. If 100,000 people around the [50:26] world watch this, wouldn't he earn [50:28] 100,000 yuan? Goodness, isn't this like [50:30] the blind men touching an elephant? We [50:32] watched the same game, yet we saw [50:33] something completely different.