实战练英文｜再不停，人类要灭亡了？！黄仁勋痛批 AI 末世论

实战练英文｜再不停，人类要灭亡了？！黄仁勋痛批 AI 末世论，特朗普突然 CALL-IN 乱入，大骂这个阴谋只会让美国输！

发布于 2026-09-18 · 时长 45:35

字幕摘录

时间	英文	中文
0:00	Oh, no.	哦,没有。
0:04	This is not planned, but we know who it is.	这不是计划,但我们知道是谁。
0:06	Oh, no.	哦,没有。
0:08	No.	没有
0:11	Mr. President.	总统先生
0:13	Oh, yes, sir.	哦,是的,长官
0:15	I got to tell you something.	我得告诉你一件事。
0:18	Ladies and gentlemen, GPU Jesus.	女士们,先生们,GPU耶稣。
0:25	They love you.	他们爱你。
0:26	Thank you.	谢谢

展开字幕全文（1072 条）

序号	英文	中文
1	Oh, no.	哦,没有。
2	This is not planned, but we know who it is.	这不是计划,但我们知道是谁。
3	Oh, no.	哦,没有。
4	No.	没有
5	Mr. President.	总统先生
6	Oh, yes, sir.	哦,是的,长官
7	I got to tell you something.	我得告诉你一件事。
8	Ladies and gentlemen, GPU Jesus.	女士们,先生们,GPU耶稣。
9	They love you.	他们爱你。
10	Thank you.	谢谢
11	I love you back.	我爱你
12	Number one podcast in the world.	世界上第一播客。
13	Wow.	珃
14	We like the new jacket.	我们喜欢新外套
15	Well, you know, I just I felt you guys needed some energy.	嗯,你知道,我只是觉得你们需要一些能量。
16	Yes.	对
17	This is I know we're talking about serious stuff here, but we need to talk about it with	и笵и弧璶 и惠璶酵酵
18	energy.	能源。
19	Yes.	对
20	Let's let's start with this essay from this weekend.	我们从这个周末开始说这篇文章吧
21	Which one?	哪一个?
22	Let's start with Dario's essay because Hemingway involved.	先从达里奥的散文开始,因为海明威参与其中.
23	Actually, did anybody run it through pan gram?	实际上,有没有人运行它通过泛克?
24	I don't even know how much of it was AI	我甚至不知道这是AI的多少
25	helped, but that was a pretty incredible thing.	帮助,但这是一个相当不可思议的事情。
26	And then I think what a lot of people were	然后我想到什么是很多人
27	surprised by was the coalescing of the frontier labs around the essay itself.	令人惊讶的是边疆实验室围绕论文本身。
28	就Jensen, 收拾行李
29	what happened, how you read it, how you interpreted it.	发生了什么事,你是如何读到的,你是如何解释的。
30	And then we'll get into some details that were	然后我们再谈一些细节
31	inside of it, but maybe just the high level thoughts to kick it off.	里面,但也许只是高层次的想法来启动它。
32	Well, first of all,	嗯,首先,我们...
33	there were a lot of stuff in there.	里面有很多东西
34	Yeah.	对
35	And and the first there is a there's a part about	而第一个是关于
36	safety, which we have to take very seriously.	安全,我们必须认真对待
37	Safety is paramount, obviously.	显然,安全是最重要的。
38	Safety and	联合国
39	leadership are not false.	领导不是虚假的。
40	They're false choices.	他们是错误的选择。
41	You're you're able to innovate quickly.	你能够快速创新.
42	You're	琌
43	able to execute quickly.	能够快速执行。
44	And and America is able to lead and to do it safely.	而美国能够领导并安全地做到这一点。
45	I think those are	我认为这些
46	those are false choices.	这些都是错误的选择。
47	But safety is obviously important.	但安全显然很重要。
48	There's a matter of of internal	有内事
49	control that I think he was speaking to.	控制,我认为他是说话。
50	Obviously, the Coxson whistleblower is very	显然,考森吹哨人非常
51	serious matter.	严重的事情。
52	Whenever you have a whistleblower, you have to take it very seriously.	每当有吹哨人的时候,都要认真对待.
53	I thought	我以为
54	Coxson had great courage to put out put out what his concerns were.	考克森有极大的勇气将他所关心的问题说出来
55	And even then there	即使那时
56	were some issues that were kind of conflated within that.	有一些问题有点混为一谈
57	I think the the whistleblowing is	我认为吹口哨是
58	fine.	还好
59	I think the the scientific prediction about the future is less fine because it's not grounded	我觉得未来的科学预测并不那么完美,因为它没有被禁足
60	on science, obviously.	科学,显然。
61	And it was expressed by a scientist, but it was obviously not grounded on	科学家表达了这一点,但显然并非基于
62	science.	科学。
63	And so I take I take issue with that.	所以我认为我不同意这一点。
64	But obviously, the whistleblower part of it,	但很明显,吹哨人的一部分,
65	you know, I think there's just a whole bunch of stuff pausing, pacing.	你知道,我想只是一整堆的东西穷困,步调。
66	Those are all the	这些都是
67	things that they could do if they feel that their company is out of control.	如果他们觉得自己的公司失控了他们可以做的
68	If Coxson saw something, you know, obviously, we didn't we don't know what Coxson saw.	如果考克森看到什么,你知道, 显然,我们不知道考克森看到什么。
69	But if he saw that the company was out of control, and maybe it's a transition from	但是如果他看到公司失控了也许这是从
70	research to engineering, as you know, these labs are research are transitioning from research to	研究到工程,你知道,这些实验室的研究正在从研究转向
71	engineering, extraordinary talent, extraordinary engineering.	工程,非凡的人才,非凡的工程.
72	But obviously, engineering is	但很明显,工程是
73	different than research.	不同于研究。
74	Maybe that transition is clumsy.	也许过渡很笨拙
75	And you know, I don't we don't know what	你知道,我不知道,我们不知道
76	what he saw.	他看见的
77	And ultimately, only he knows.	最后,只有他知道这一点。
78	But if there was a matter of lack of control,	但是,如果有一个问题缺乏控制,
79	that's a different topic.	这话题不同
80	How should the government deal with it?	政府应该如何处理?
81	Now all of a sudden,	现在突然
82	regulation, right?	条例,对不对?
83	I mean, it just covers everything in one block.	我的意思是,它只是覆盖所有在一个街区。
84	Can you just help us sort of unpack?	你能帮我们收拾行李吗?
85	We play we tried to play this game actually this week	我们本周玩过这个游戏
86	on the pod.	在舱。
87	And it was difficult, which is, how do you describe like, you know, my mom calls me,	这很困难,这就是,你怎么形容, 比如,你知道,我妈妈打电话给我,
88	she's like, Jamoth, what is this whole civilizational death thing?	她喜欢,贾莫特, 这整个文明死亡的事情是什么?
89	I don't know	我不知道,你觉得呢?
90	how to explain it to her.	如何向她解释。
91	So when you have very smart people like that, quantize it,	所以,当你有非常聪明的人这样,量化它,
92	and quantify it, I think that's probably what's perturbing to some people.	并量化它, 我认为这可能是困扰一些人。
93	They're like,	他们就像,
94	what does that mean 10% of extinction?	这是什么意思 10%的灭绝?
95	Nobody knows how to explain that to the average person	没人知道怎么跟普通人解释
96	how that's even possible.	这怎么可能
97	Well, first of all, we shouldn't.	首先我们不该这么做
98	Because it's made up.	因为它是编造的.
99	First of all,	首先,我们...
100	I think that we shouldn't because it's made up.	我认为我们不应该因为它是编的。
101	And these are these are well educated.	这些是受过良好教育的。
102	They're called researchers.	他们被称为研究者.
103	Obviously, they're working a lab.	显然,他们正在实验室工作。
104	And so the confluence of these	所以,这些交汇
105	words, and then and then and then the prediction is alarming and troubling, and it shouldn't be	预言令人惊恐和烦恼不应该
106	done.	结束了。
107	It's it's it's irresponsible.	这是不负责任的。
108	Now the fact of the matter is, let's go back and look at the	现在事情的真相是让我们回去看看
109	facts.	事实
110	The facts are there was a prediction that in five years time, radiology will be completely	事实已经预测五年后放射学将完全
111	taken over by artificial intelligence, and there'll be no radiologists in the world that has proven	由人工智能接管世界上没有放射科医生
112	to be exactly the opposite.	完全相反。
113	We need more radiologists than ever in the world.	我们需要的放射学家比世界上任何时候都多。
114	However,	不过,
115	AI has taken over radiology completely, which is great is automated scan reading, which is great.	AI已经完全接管了放射学,这是伟大的是自动化扫描读取,这是伟大的.
116	There was a prediction that within six to 12 months, wasn't it just last year within six to	有人预测在6到12个月内, 是不是去年在6到12个月内
117	12 months, 90% of code would already be generated by AI.	12个月,90%的代码已经由AI生成.
118	That has turned out to be wrong.	这证明是错误的。
119	Within	内部
120	six to nine months, that was predicted last year, 50% of entry jobs will be wiped out.	6到9个月,这是去年的预测, 50%的入职工作将被淘汰。
121	That has proven to be wrong.	这已证明是错误的。
122	Let's see what else what else has proven to be wrong.	看看还有什么其他的证明是错的
123	I mean,	我是说...
124	all of these predictions have been wrong.	所有这些预测都是错误的。
125	Right.	对
126	And if you're a GPT two would be too unsafe to	如果你是GPT 2,那会太不安全了
127	release that llama three would be too unsafe to release one.	释放这只蜥蜴3会太不安全无法释放一只
128	Yeah, we've heard the half of	是的,我们听说了一半
129	white collar jobs would be gone next year jobs apocalypse.	白领工作明年就会消失,
130	Yeah, we have to take accountability.	是的,我们必须承担责任。
131	We have to take account for all of the stupid predictions that were made.	我们必须考虑所有愚蠢的预测。
132	Right.	对
133	Somebody has	有人说过
134	somebody has to take.	得有人来接我
135	And, and so we ought to just keep track of all that.	所以,我们应该一直跟踪这一切。
136	And of course,	当然
137	people do and remind us that those those predictions are inconsistent with ultimately	人们这样做并提醒我们,这些预测与最终的
138	America winning the AI race.	美国赢得AI比赛.
139	The short form for that is some people are saying, you know,	简而言之,有些人说,你知道,
140	they say trust the experts and they use the analog of COVID, which again started with people that	他们说相信专家,他们使用COVID的模拟, 它再次从人们开始
141	were researchers educated people that had an asymmetric awareness of the thing that the rest	研究者受过教育的人对其余的事物有不对称的认识
142	of us did not saying things that ultimately turned out, we find out in facts, not to be true.	我们没有说那些最终结果的东西, 我们发现事实,不是事实。
143	And so there's this war that's happening right now between the trust the experts movement and the,	专家运动和... 信任之间的战争正在发生
144	you know, well, let's just look at the actual history of these predictions.	我们看看这些预测的实际历史
145	And let's just	让我们只是
146	think more methodically.	更有条理地思考
147	Where is this coming from?	这是从哪里来的?
148	Because it's coming from inside the places	因为它是从里面来的
149	that's actually making it like what do you think is the psychological makeup?	其实你觉得心理化妆是什么?
150	Or what is the real	或者什么是真实的
151	incentive?	奖励?
152	Maybe it's a business incentive.	也许这是商业刺激
153	Maybe it's a political incentive.	可能是一种政治刺激.
154	Can you just maybe	你可以只是也许
155	guess or how do you how do you think about what's why they're doing this?	你猜怎么看他们为什么这么做?
156	Well, first of all, I go	好吧,首先,我会去
157	to these are some of the most consequential companies in history.	这些公司是历史上最有影响的公司。
158	Extraordinary engineers,	特别工程师,
159	extraordinary researchers, really fantastic work.	非凡的研究人员,真的很棒。
160	On the one hand, I work very closely with them	一方面,我与他们非常密切地合作
161	as companies to companies.	作为公司对公司。
162	On the other hand, we have to have conversations like this in public.	另一方面,我们必须在公众场合进行这样的对话.
163	And it's really unfortunate.	这是非常不幸的。
164	And I think that that these these companies really ought to be	我认为这些公司真的应该
165	built the way that we used to build companies, which is in silence.	建造了我们以前建造公司的方式,这在沉默中。
166	Right.	对
167	You know, and	你知道吗,和
168	so Jensen, you don't allow anybody in your organization to speak for the entire organization,	Jensen 你不允许你的组织里有人代表整个组织发言
169	especially when they're having like a bad weekend or they rage quit, they're not allowed to tweet on	特别是当他们有一个糟糕的周末或者他们愤怒退出, 他们不允许在推特上
170	your behalf and the organization's behalf.	代表你代表组织
171	Because well, that's that's what they decided when they	因为好,这就是他们决定当他们
172	came to work for us.	来为我们工作。
173	And we told them these are this is the way you behave when you work in our	我们告诉他们,这是你的行为当你在我们工作
174	company.	连队。
175	And and if you would like if you like the culture of our company, which, as you know,	如果你喜欢我们公司的文化如你所知
176	the Nvidia culture and the Nvidia employee base, incredibly happy.	恩维迪亚文化和恩维迪亚员工基地,令人难以置信的快乐。
177	Yeah, they like the fact that	是啊,他们喜欢的事实
178	the company is consistent, that we're stable, that our core values are consistent with taking care of	公司是一贯的,我们是稳定的, 我们的核心价值与照顾
179	the families and creating the conditions by which they can do their life's work, that we do meaningful	家庭,并创造他们可以从事其生活工作的条件,我们做有意义的
180	work, we do it, we do it as quietly as we can.	工作,我们做它, 我们做它,尽可能安静。
181	And we contribute to everybody else's success,	我们为其他人的成功作出贡献,
182	which we're very proud of.	我们为此感到骄傲。
183	And so those kind of core values people are attracted to.	所以这些核心价值被人们所吸引。
184	But when you	但当你
185	come and work in our company, there are also some things that we don't appreciate that you do.	来我们公司工作有些事情我们不欣赏你做的
186	Like,	喜欢,
187	for example, we don't welcome political discourse in our insider company.	举例来说,我们不欢迎我们内部公司的政治言论。
188	Take it home.	带回家去
189	You guys	你们
190	talk about politics outside the company.	谈论公司外的政治。
191	We we are the company is an apolitical company, you know,	我们公司是一个非政治性公司,你知道,
192	we're bipartisan, we want America to succeed.	我们是两党, 我们希望美国成功。
193	And, and we want we want to whatever government	我们想要任何政府
194	is in place, we'll do everything in our power to help America succeed.	已经就位,我们将竭尽全力帮助美国成功。
195	And so, so the discourse	所以,这样的话
196	about about about race and religion and politics and all of that stuff, we tell people do it outside	关于种族、宗教、政治和所有的东西,我们告诉人们在外边做
197	the company.	这家公司。
198	It's not not for us.	这不是为我们。
199	In terms of maybe AI regulation, then more narrowly,	说不定是人工智能的管制更狭义一点
200	Satya was here this morning.	萨特雅今早来过
201	And what he said is, you know, before we talk about regulation that	他说的,你知道, 在我们讨论监管之前
202	could really stymie things, why don't we just get some basics, right?	可能真的有问题,我们为什么不只是得到一些基本,对不对?
203	Why don't we get measurement,	我们为什么不得到测量,
204	right?	对吧?
205	Why don't we get standardization, right?	我们为什么不实现标准化呢?
206	Right?	对吧?
207	Where do you land on get engineering,	你在哪里登陆的工程,
208	right?	对吧?
209	Get the engineering right, right?	把工程弄对了,对吧?
210	Translate the research in a more predictable way, so that	以更可预测的方式翻译研究,以便
211	we're not fear mongering, keep it inside until we're ready to expose it.	我们不害怕闷闷不乐直到我们准备好曝光为止
212	What do you think the	你觉得呢?
213	responses, you know, demos had a proposal, which was sort of this more Finra like organization,	反应,你知道, 演示有一个提议, 这有点像这个Finra像组织,
214	it's not clear what Daria wants this transnational mutated thing that has some sort of control,	并不清楚Daria想要什么跨国变异的东西有某种控制,
215	where do you land on this, the sort of perspective of what what do we need right now?	你在哪里降落, 那种视角我们现在需要什么?
216	You know, regulation should solve actual problems.	你知道,监管应该解决实际问题。
217	And so the question is, what actual problems have	所以问题是到底有什么问题
218	we enjoyed?	我们喜欢吗?
219	Right.	对
220	And, and if you look at look at the actual problems, all of the actual problems	而且,如果你看看实际问题,所有的实际问题
221	so far have come from the labs.	目前来自实验室
222	And the reason for that, the reason for that, and just in their	原因原因原因和他们
223	defense, the reason for that is because they have the most compute.	理由是他们计算得最多
224	Right.	对
225	And the reason for that	还有原因
226	is because they're trying to solve the frontier problems.	因为他们想解决边界问题
227	And so in their defense, and so it's	因此,在他们的辩护, 所以它是
228	sensible that that the labs, the frontier labs will be where the most danger come from.	明智的是,实验室,前沿实验室将是最危险的来源。
229	It is unlikely	这不可能
230	that a high school student did something because they just simply won't have enough compute.	一个高中生做了一些事情因为他们只是没有足够计算。
231	Right.	对
232	And so it's unlikely that a startup will be the reason because they won't have enough compute.	因此,启动不太可能是因为他们的计算不足的原因。
233	In fact, you can look across the planet and everybody won't have enough compute with the	事实上,你可以环顾整个地球每个人的计算都不够
234	exception of the frontier labs.	除了前沿实验室
235	And so so now the question is, if you look at what actually happened,	所以现在的问题是,如果你看看到底发生了什么,
236	and they're doing pioneering work, it's really very hard.	他们正在做开拓性的工作, 这真的很困难。
237	They're transitioning from research	他们正在从研究中过渡
238	to engineering.	到工程学。
239	I could imagine and they're obviously building some of the most consequential	我可以想象,他们显然正在建造一些最导致
240	technology and companies in the world.	科技和公司在世界上。
241	They're building their company, they're building their	他们正在建设他们的公司,他们正在建设他们的公司。
242	culture, they're building the technology, they're building engineering, they're building products	文化科技工程产品
243	all at the same time.	都在同一时间。
244	And so I can understand it's a little bit hair on fire.	这样我才能明白头发有点着火了
245	But nonetheless,	不过
246	the four incidents from one lab, the one giant incident from the other lab,	一个实验室的四起事件另一个实验室的一起巨大事件
247	the first thing that you have to do is just root cause the problem from an engineering perspective.	首先,你必须做的只是从工程的角度来引起问题的根源。
248	What happened?	发生了什么事?
249	What could we have done differently?	我们该怎么办?
250	And what are we going to implement	我们要执行什么
251	and institutionalize, whether it's technology or methods or processes, and make sure that we	并制度化,无论是技术还是方法或过程,并确保我们
252	don't let it happen again?	不要再发生这种事了?
253	Now, I would bet you money that in every single one of those cases	现在,我敢打赌,你的钱在所有这些情况下
254	is within their control in the future to prevent it.	将来他们可以控制它
255	Because the alternative, if it's not in their	因为另一种选择,如果不是在他们身上
256	control, and I'm sure that they are, I'm sure those four incidents won't happen again.	我确信这四起事件不会再发生
257	I'm	и琌
258	sure they root caused it and fixed it.	当然是根引起的并修好了
259	I'm sure they have now much better technology for	我相信他们现在有更好的技术
260	sandboxes and runtimes and monitors and continuous monitors.	沙盒和运行时间监视器和连续监视器
261	And so I'm certain they	所以我确定他们
262	have much, much better technology now.	现在有更好的技术了
263	The alternative is also unlikely, which is for them	选择也不太可能,对他们来说
264	to say, look, we had these incidents.	说,你看,我们有这些事件。
265	After we're done analyzing it, we came to the conclusion we	分析完后我们得出结论
266	don't know anything that happened.	不知道发生了什么
267	And we have no idea how to control it.	我们不知道如何控制它。
268	And we're asking society	我们问社会
269	for help.	寻求帮助。
270	Now, if that's the case, then we ought to, you know, a bunch of companies with engineers	现在,如果是这样的话,那么我们应该,你知道, 一堆公司与工程师
271	ought to send engineers in.	应该派工程师来
272	I mean, and we should advise them if we can.	我的意思是,我们应该建议他们如果我们可以。
273	But I doubt it.	但我怀疑这一点。
274	I think	我觉得
275	they have extraordinary people.	他们有非凡的人。
276	They got this handle.	他们得到了这个把手。
277	But we're not operating in a	但我们不是在一个
278	David, last night, you informed me that there is a Chinese lab, the makers of GLM who are going to	David 昨晚你告诉我有一个中国实验室 GLM的制造者
279	put 3 billion towards a recursive self-improvement run.	投入30亿进行循环自我改进
280	So maybe you could tee that up for James.	所以,也许你可以把它给詹姆斯。
281	That's what was announced as Zipoo.com.	即被宣布为Zipoo.com. (中文(简体) ).
282	The founder just raised 5 billion and said that one	创始人刚筹到50亿,说
283	of their priorities is going to be trying to get to recursive, you know, AI that trains the next AI	他们的当务之急是努力让训练下一个AI的AI重现
284	and to try and automate as much of that as possible.	并尝试和自动化尽可能多的这一点。
285	Yeah, I think that I mean,	是的,我想我是说,
286	this is the new sexy phrase.	这是新的性感短语。
287	But as you guys know, RSI is a combination of a system of ideas.	但你们知道 RSI是思想体系的组合
288	It starts everything with in context stuff.	从背景开始
289	It starts with skills.	以技能为始.
290	It starts with reflection.	它从反射开始。
291	It starts with, you know, reinforcement learning and synthetic data generation.	它从强化学习和合成数据生成开始。
292	And these are all	这些都是
293	very sensible ideas that causes AI to get better at solving a problem, you know, over time.	非常明智的想法, 使得AI得到更好的解决一个问题,你知道, 随着时间的推移。
294	And	还有
295	you could also have low rank.	你也可以低级别。
296	All of that stuff doesn't include the weights.	所有的东西都不包括重量。
297	You could actually	你其实可以
298	improve the weights and it's called LoRa.	提高重量,它叫洛拉。
299	LoRa could be improved in synthetic data generation,	LoRa可以在合成数据生成方面得到改进,
300	reinforcement learning, enhance it without training the base model itself.	强化学习,在不训练基地模型本身的情况下增强它.
301	And then over time,	随着时间的推移,
302	you could train the base model again with all of that experience.	你可以再用这些经验训练基地模型
303	And so I think it's a sensible	所以我觉得这很明智
304	thing that you're going to use the technology to enhance productivity of all kinds of tasks,	你要利用技术来提高各种任务的生产率
305	including building AI.	包括建筑AI.
306	I think that's a very logical idea.	我认为这是一个非常合乎逻辑的想法。
307	And I'm certain that everybody	我敢肯定,每个人都
308	is using it in some degree.	在某种程度上是用它。
309	It's just this phrase is now being used to weaponize the technology in	只是这个短语现在被用来将技术武器化
310	some way and maybe to turn the- As if it's going to spiral out of control is the impression they're	某种方式,也许可以让... 好像要失控了他们的印象是
311	trying to give, but you don't believe that's real.	试图付出,但你不相信这是真实的。
312	No, no, of course not.	不,不,当然不是。
313	And the reason for that is	原因是
314	because you could RSI all day long inside your company, but when you release a product, you've	因为你可以整天在公司里做RSI 但当你发布产品时
315	got to evaluate it, don't you?	得评价一下,不是吗?
316	You have to test it again, don't you?	你得再试一次,不是吗?
317	You have to make sure that	你一定要保证
318	there's no regression, right?	没有倒退,对不对?
319	And so the basic process of control, these labs are going to,	所以控制的基本过程, 这些实验室会,
320	as they move from labs to engineering, they will have much, much better control.	当他们从实验室转移到工程, 他们将拥有很多,更好的控制。
321	And when they have	当他们有
322	much better control, and control comes from methods and knowledge and practice and tools	更好的控制,控制来自方法、知识和做法及工具
323	and technology, all of those things that lead to better control, verification, any of else,	以及技术,所有那些导致更好的控制, 核查,任何其他,
324	it's going to enable RSI to be done inside the company and for good products to be released	它将使RSI在公司内部完成并释放好产品
325	outside the company.	在公司之外,
326	Let's talk about open source for a second.	聊开源一时.
327	I mean, this hugging face,	我的意思是,这个拥抱的脸,
328	we were communicating about this and I said it's going to be one of the most consequential	我们正沟通这个,我说这将是最大的后果之一
329	acquisitions.	采购。
330	I don't even want to call it a transaction because I think it's more important	我甚至不想称之为交易,因为我觉得这更重要
331	than that.	更何况.
332	Give us your first principles explanation of open source versus closed	给我们你关于开源与关闭的第一个原则解释
333	source versus open weights and how the ecosystem should fit together over time.	源与开放式权重,以及生态系统应如何随时间而融合。
334	The world needs both closed models and open models.	世界既需要封闭模型,也需要开放模型.
335	I use as much closed models as I can.	我尽量使用封闭的模型
336	This weekend, I used four of them.	这个周末我用了四个
337	And they work terrifically.	他们的工作非常出色。
338	They're frontier.	它们是边疆.
339	They're	他们是
340	great experience.	伟大的经验。
341	They work incredibly well.	他们的工作非常出色。
342	They're getting better all the time.	他们越来越好。
343	And the way I think about closed models is kind of like bottled water.	我对封闭模型的看法就像瓶装水一样。
344	Water is free,	水是免费的
345	you guys.	你们
346	I don't know if I've told you guys, but water is free.	我不知道我有没有告诉你们但水是免费的
347	I don't want to burst everybody's	我不想把每个人的
348	bubble, but water is free.	但水是免费的
349	And this morning, I used a lot of free water taking a shower.	今天早上,我用了很多免费水洗澡。
350	And so you use the right water in the right places.	所以用正确的水在正确的地方。
351	And this is no different than electricity.	这和电力没什么不同
352	This is no different than all kinds of commodities that we use in the world.	这和我们在世界上使用的各种商品没有什么不同。
353	You need both.	两者都需要
354	Now,	现在,
355	in the case of open, the reason why you need it is because it could be for sovereignty reasons,	在开放的情况下,你之所以需要它,是因为可能是出于主权原因,
356	privacy reasons, proprietary technology reasons.	隐私原因、专利技术原因。
357	Look at the facts.	看看事实
358	The facts are in the last six	事实在最后六个
359	months, $400 billion of venture funding went into AI native companies.	几个月后,4000亿美元的风险投资投入了AI本土公司.
360	80% of them use open models.	其中80%使用开放型号.
361	If not for open models, how could they build their dream?	如果没有公开的模型,他们怎么能建立自己的梦想呢?
362	Because their dream could be different.	因为他们的梦想可能不同
363	Obviously, it would be different than the labs, the frontier labs dreams.	显然,它会不同实验室,前沿实验室梦想。
364	And America has so many	美国有这么多
365	different ways to innovate.	不同的创新方式。
366	That's one of our core strengths.	这是我们的核心优势之一。
367	Great ideas just coming out of the	伟大的想法刚刚出来
368	fountains.	喷泉。
369	And so open models enables that.	所以开放的模型可以实现这一点。
370	Open models enables every single, if we want to win the	开放的模型可以让每一个,如果我们想要赢得
371	AI race, it's not about a few technology companies winning the AI race.	AI种族,与少数科技公司赢得AI种族无关.
372	It's about every company in	是关于每家公司的
373	America.	美国。
374	Every company, every industry, every researcher, every teacher, every student,	每个公司,每个行业,每个研究员,每个老师,每个学生,
375	every startup, everybody wins.	每个开始,每个人都会赢。
376	Some of them will use closed models.	其中一些会使用封闭型号.
377	A lot of them will use open	很多人会打开的
378	models.	模特儿们
379	There's 10 million.	有一千万亿.
380	Does it matter if the open models come from China or the US?	开放型号是来自中国还是来自美国?
381	Well, we're doing everything we can to make a contribution in open models.	我们正在竭尽所能在开放模式中做出贡献。
382	However,	不过,
383	the moment you download, for example, probably the vast majority of the world's contribution	例如,你下载的瞬间, 很可能是世界绝大多数的贡献
384	to open source today is coming from China.	今天的开源是来自中国的
385	They just have a lot more engineers.	他们只是有更多的工程师。
386	They produce	他们生产
387	everything in large scale because it's a larger country.	因为它是一个更大的国家。
388	And so they produce science and math	因此,他们产生科学和数学
389	students in volume.	学生数量。
390	That's one of our disadvantages.	这是我们的缺点之一。
391	They're manufacturing them through amazing	他们正在制造出惊人的
392	universities like Tsinghua University in high volume.	清华大学等高校办学量.
393	Well, they contribute to open source	嗯,他们贡献了开源
394	today.	今时.
395	我们下载Linux。
396	我们下载Kubernetes。
397	We download all the software.	我们下载所有软件。
398	A lot of it has been	很多时候
399	touched by Chinese.	被中国人所感动
400	And once you download it, it's yours.	一旦你下载它,它是你的。
401	We fork it.	我们用叉子
402	We improve it.	我们改进它。
403	We make it	我们成功了
404	ours.	我们的
405	And so when you download one of these Chinese models, it just happens to be made by	所以当你下载这些中国模型时它恰好是由
406	some really great researchers in China, but it's now yours.	中国的一些优秀研究者, 但它现在是你的。
407	Whatever you want to do with it.	随其所欲.
408	So what exactly is the race?	比赛到底是什么?
409	The race?	比赛?
410	Yeah.	对
411	I think that's a really good point.	我觉得这是个好主意
412	My point	我的观点
413	is the race is really about who exploits the technology best.	比赛是真正关于谁利用最好的技术。
414	You know, the last industrial	你知道,最后一个工业
415	revolution, all of the inventors were Maxwell, Volta, Ampere.	革命,所有发明家都是马克斯韦尔,沃尔塔,安佩尔.
416	None of them were American.	没有一个是美国人。
417	They were right.	他们是对的。
418	The last industrial revolution came from Europe, but we exploited it.	上一次工业革命来自欧洲,但我们利用了它。
419	We took	我们拿了
420	advantage of it socially better than anybody else in the world.	社会优势比世界上任何人都好
421	And look how it turned out for us.	看看结果如何
422	I want to make sure that this next generation happens just like this.	我想确保这下一代就是这样发生的。
423	Yeah.	对
424	So why are the communists getting their message out so successfully here right now?	那么,为什么共产主义者得到他们的信息如此成功在这里现在?
425	You know, I think first of all, the narrative is much more practical.	你知道,我认为首先, 叙述更实用。
426	The narrative is much more practical.	叙事更为实用.
427	Nobody's in China is saying that there's end of this and	在中国没有人会说事情会结束的
428	of that and, you know, cataclysmic this and, you know, do more than that.	并且,你知道, 催化这个,你知道, 做更多。
429	They're much more pragmatic about it.	他们对此更加务实.
430	They see AI as a technology that's going to advance their	他们认为AI是一种技术将推进他们的
431	economy, advance your society.	经济,推进你的社会。
432	And they don't have these groups who are basically saying it's going	他们没有这些团体基本上说它去
433	to end civilization.	结束文明。
434	And we're making it up.	我们正在编造它。
435	The part that is frustrating is if it was true,	令人沮丧的是,如果这是真的,
436	then we ought to talk about it and go do something about it.	然后我们应该谈谈并做一些事情。
437	Even if it's true,	即使是真的,我还是觉得
438	we ought to spend more time doing something about it than worrying a bunch of people who can't do	我们应该花更多的时间做一些事情比担心一群人谁不能做
439	anything about it.	任何关于它。
440	It's our job to build it right.	我们的工作就是把它建好
441	Has there ever been a point in history	历史上有过一点
442	where so many people have so vehemently said something that is so untrue?	有这么多人如此强烈地说出如此不真实的话吗?
443	And they're measurably, they're actually demonstrably untrue.	并且可以衡量,它们其实是不真实的。
444	And it actually makes	它实际上使
445	sense it's untrue.	感觉这不是真的
446	It's not based on science.	不以科学为本.
447	It's not based on research.	并非基于研究.
448	Everything that's	所有的东西
449	based on science and research proves otherwise.	基于科学和研究的证明并非如此。
450	Is it a fear of the frontier?	这是对边境的恐惧吗?
451	Humans have never	人类从来没有
452	been there.	去过那里。
453	We've never seen it.	我们从来没有见过它。
454	Therefore, we're scared of it.	因此,我们害怕它。
455	And therefore, it's easy	因此,这很容易
456	to tell everyone to be scared of it.	告诉大家害怕它
457	And I think it could be life experience as well, David.	我认为这也可能是生活经验,大卫。
458	So let me give you an example.	让我举个例子。
459	When I first graduated from school, I was an engineer,	当我刚从学校毕业时, 我是个工程师,
460	and I didn't do that much typing.	我没有做那么多打字。
461	And the reason for that is because I was the first generation	这是因为我是第一代
462	before software became popular.	在软件流行之前
463	We had to go build the computers to make software possible.	我们不得不去制造计算机使软件成为可能。
464	Could you imagine in this generation, every single engineer who came into the world of engineering,	你能想象这代人每一个进入工程界的工程师
465	you spend all your time typing.	你花所有的时间打字。
466	Literally, that's what you do.	从字面上来说,这就是你的工作。
467	When you get a job,	当你找到工作,
468	they give you a laptop, they give you a chair, and you start typing.	他们给你一台笔记本电脑他们给你一把椅子然后你开始打字
469	You type all day long.	你打字一整天。
470	You type from the moment you wake up to the... Well, there was engineering before typing.	你打字从你醒来到... 嗯,有工程在打字之前。
471	And so, can you imagine that the world has a mountain of engineering work to do	所以,你能想象这个世界有一山的工程工作要做
472	where most of it is not typing anymore?	大部分都不再打字了吗?
473	Sure.	当然
474	We had busy engineers before typing.	在打字前我们有很多工程师
475	I think we're	我认为我们
476	when I say typing, I mean coding.	当我说打字时,我的意思是编码。
477	So even at NVIDIA, when software engineers talk to me,	因此,即使在NVIDIA, 当软件工程师跟我说话,
478	I tell them, you're just typing.	我告诉他们你只是在打字
479	I've been saying that forever, but obviously for fun.	我一直这么说但显然是为了好玩
480	And I tell	我告诉
481	them my favorite key is backspace.	我最喜欢的钥匙是后空格
482	And the reason for that is because the best software is the	原因是最好的软件是
483	smallest software.	最小软件。
484	So I want you to use backspace as often.	所以我希望你经常使用后座空间。
485	我们来谈谈NVIDIA
486	Let's do a little teardown of NVIDIA.	让我们做一点撕裂NVIDIA。
487	Teardown meaning just explain the pieces because there's	拆掉意味着只是解释片段因为有
488	a lot of strategy at play.	有很多策略在发挥。
489	Let's start at the absolute bottom.	从绝对的底部开始
490	Oh, no.	哦,没有。
491	This is not planned,	这是没有计划的,
492	but we know who it is.	但我们知道是谁
493	Oh, no.	哦,没有。
494	No.	没有
495	Mr. President?	总统先生?
496	Oh, yes, sir.	哦,是的,长官
497	I got to tell you something.	我得告诉你一件事。
498	If it wasn't because of you calling, I'm on stage with the besties.	如果不是因为你打来电话, 我和最美的女生上台了.
499	I'm on stage with the besties.	我在舞台上与最爱。
500	I'm on stage with Sachs.	我在舞台上与萨克斯。
501	Yeah.	对
502	You know, the whole group.	你知道,整个团体。
503	Yeah.	对
504	Jason's here.	杰森来了
505	Chamath's here.	查玛斯来了
506	Dave and David is here.	戴夫和大卫来了
507	Yeah.	对
508	I'm sitting in front of	иГ玡
509	a few thousand people.	数千人。
510	And we're talking, as it turns out, we were talking about you.	我们正在谈论, 事实证明,我们是在谈论你。
511	Good job, sir.	干得漂亮长官
512	Good job.	干得不错
513	The fact that you saw through all of that, I mean, there's a lot of	你所看到的这些, 我的意思是,有很多
514	complexity and the fact that a matter you just saw through all of that.	复杂和事实,一个你刚刚看到的事物通过这一切。
515	And we're all just really	我们都只是真的
516	grateful.	谢谢
517	Tell him I said hi.	告诉他我打过招呼
518	Do you want to say hi to the crowd?	你想跟观众打个招呼吗?
519	Jason would like to put you on	杰森想让你听
520	the mic.	麦克风。
521	Even Jason.	乃至贾森.
522	Put him on speaker mode.	把他放在扬声器上
523	How do we put on POTUS?	我们怎么把总统?
524	Put him on speaker?	让他开扬声器?
525	Speaker.	扬声器
526	Yeah.	对
527	Right into the microphone.	右进入麦克风。
528	Hang on a second.	等一会
529	Hold on, sir.	等一下,长官
530	We're getting a	我们得到一个
531	microphone.	麦克风
532	Mr. President?	总统先生?
533	Yes.	对
534	You're now talking to the planet.	你现在在和地球说话
535	You see, the great thing about life is that Jensen can develop the most complex	生命的伟大之处在于 Jensen能发展出最复杂的
536	computer chip in the world that nobody can copy for 10 years, but he can't figure out how to put	世界上没人能复制10年的电脑芯片但他却想不出怎么放
537	me on speaker.	我在扬声器上。
538	We have to remember this one.	我们必须记住这个。
539	So it's almost this conspiracy.	所以几乎就是这个阴谋
540	And the happiest	还有最快乐的
541	group is China.	集团是中国。
542	And China is very happy.	而中华极乐.
543	And I could even say in the country, a lot of states	我甚至可以说,在这个国家, 很多州
544	are happy that weren't going to get anything because they're being inundated by people that	很高兴没得到任何东西因为他们被别人淹没
545	want to be there.	想要在场
546	But now all of a sudden you see they're building in Finland.	但现在你突然看到他们在芬兰建造。
547	They want to build	他们想要建造
548	one.	一个
549	Google wants to build a big one in Finland, which I'm not happy about because they weren't	Google想在芬兰建造一个大型的, 我对此不满意,因为他们不是
550	able to get permitting.	可以得到许可。
551	And I'm telling you, it's all a hoax.	我告诉你,这都是个骗局。
552	The data centers are great	数据中心很棒
553	and they make people wealthy and they make states wealthy.	他们使人民富裕,他们使国家富裕。
554	And it's the oil of the next 20, 25 years	这是未来20,25年的石油
555	is bigger than the internet and the AI, you know, much more so.	比互联网和人工智能还要大得多
556	And they're just playing right into	他们只是玩
557	the hands of a lot of people that don't want to see it happen.	很多人的手不想看到它发生。
558	And that could be political people	这可能是政治人物
559	that could also be China.	那也可能是中国。
560	And we're not going to let that happen.	我们不会让这种事发生的
561	It's a, it's a hoax.	这是一个,这是一个骗局。
562	And	还有
563	you're right.	你说得对。
564	We're not going to let that happen, sir.	我们不会让这发生,先生。
565	No, we're not going to let it happen.	不,我们不会让它发生的。
566	Robots are not going to be taking over the world.	机器人不会占领世界的
567	And that's not going to happen.	然则不可得也.
568	You know, my uncle	你知道吗,我叔叔
569	was a, the top, probably maybe the best of all time, frankly, but professors at MIT for 41, 42	但麻省理工的教授41 42人
570	years and known as being one of the most brilliant men.	多年,并被称为最辉煌的人之一。
571	And he was, he was there for 41 years as	他在那里呆了41年,因为...
572	the top.	顶部。
573	He was like at the top, top of the ladder, top of, did many things.	他像在顶端,顶端梯子,顶端,做了很多事情。
574	Jensen knows all about	詹森知道这一切
575	it, but did many things.	但它做了很多事情。
576	So I have a little genetic, a little genetic strength if you believe in	所以我有一点遗传力,一点遗传力如果你相信
577	resource theory, but I do.	资源理论, 但我有。
578	I have genetics.	我有遗传学
579	That explains why you know so much about AI.	难怪你这么了解AI
580	Well, I know about AI.	嗯,我知道关于AI。
581	I know.	我知道,我记得
582	I also have common sense about AI.	我也对AI有常识.
583	The robots will not be taking over.	机器人不会接手
584	The AI will not be taking over the rest of the world.	AI不会接管世界其他地方.
585	The whole thing is a hoax.	整件事都是骗局
586	Now with that,	现在有了这个
587	we have to be a little bit careful.	我们一定要小心一点
588	We have to very be, you know, we have to do things and we	我们必须非常,你知道, 我们必须做一些事情,我们
589	have to do them prudently, but that doesn't mean we're going to stop an industry because, you know,	但这并不表示我们要阻止一个行业因为你知道
590	as we work on the next 10 years about how to destroy it.	未来10年我们如何摧毁它
591	So I'm with you all the way.	故与诸人俱.
592	I didn't	我没有
593	even know how you felt about it.	甚至知道你对这件事的感受
594	I assumed you felt the same way as me.	我猜你跟我的感觉一样
595	And we're going to lead.	而将领之.
596	And I have an expression.	我有话要说
597	It's whoever wins, AI wins.	谁赢谁赢,AI赢谁赢
598	That's how big it is.	何等为大.
599	It's bigger than the	它比那个大
600	internet.	互联网。
601	And whoever wins, AI wins.	不管谁赢了 AI赢了
602	And we can't let this kind of stuff happen.	我们不能让这种事发生
603	And that includes,	包括,
604	very much includes data centers.	非常包括数据中心。
605	There are communities that were dying that have data	有些社区正在死亡有数据
606	centers right now, and now they're wealthy communities, really wealthy communities.	现在的中心,现在它们都是富裕的社区, 非常富裕的社区。
607	We're going to make sure that we're going to make sure that everybody, we're going to make sure that	令诸众生.
608	everybody wins in the AI race in America, every industry, every company, every state, every people.	每个人在美国的AI比赛中都获胜, 每一个行业,每一个公司,每一个州,每个民族.
609	Good.	不错
610	Well, I feel strongly about it and I have the position that can do something about it.	好吧,我对此有强烈的感受我的立场可以有所作为。
611	We're	我们是
612	not going to let that stuff happen.	不会让这种事发生的
613	So I have no idea who's at the meeting.	所以我不知道是谁在开会
614	I have no idea who the	我不知道是谁
615	I'm talking to, but I'll see.	我说话,但我会看到。
616	Did you, did you hear that?	你听到了吗?
617	Did you hear that?	你听到了吗?
618	Thousands of people	数千人
619	are clapping for you, sir.	在为你鼓掌,先生。
620	But he's done an amazing job and David has done an amazing job and	但他做了一个惊人的工作大卫做了一个惊人的工作
621	good luck to everybody.	祝大家好运
622	And we're going to stay with the future.	则与未来相续.
623	The country has never done better.	国家从未如此出色。
624	We have $20 trillion of investment coming into the country.	我国有20万亿投资进入我国.
625	And that's as opposed to much less	这比不上少得多
626	than one trillion under sleepy Joe Biden.	超过一万亿困在乔拜登。
627	And that was for four years.	那已经四年了
628	This is in one year.	这是一年之内。
629	So, you know, it's really the country is there's never the country has never seen anything like it	所以,你知道,这是真的国家是从来没有国家从来没有见过这样的东西
630	and we're going to keep it going.	我们会继续前进的
631	And so thank you all very much.	所以,非常感谢大家。
632	Thank you, Mr. President.	谢谢你,总统先生。
633	I'll call you back later.	晚点再打给你
634	Thank you, Mr. President.	谢谢你,总统先生。
635	Thank you.	谢谢
636	I thought that was unique.	我认为这是独特的。
637	I thought it was a bit.	我以为是有点
638	Did you know that was happening?	你知道吗?
639	I thought it was a bit.	我以为是有点
640	Yeah, that was like, no, it's real.	是啊,那是像,不,这是真实的。
641	I thought it was a bit at first.	初以为有一点.
642	And I was like, put up on speakerphone.	我当时想,用扬声器
643	Wow.	珃
644	And he calls you.	他打电话给你
645	Well, how do you, how do you	你怎么样,你怎么样
646	cause you any hour of the night?	让你在晚上有时间吗?
647	Right?	对吧?
648	Well, we were, we were in the, we were in the oval	嗯,我们是在,我们是在, 我们在椭圆形
649	that time when he called you.	当他打电话给你的时候
650	I was sleeping.	我在睡觉。
651	Oh, you're asleep.	哦,你睡着了。
652	And he like said, wake him up.	他喜欢说,叫醒他。
653	I felt so bad because he's like, who's coming to this dinner and we go through the list.	我觉得很糟,因为他喜欢, 谁来参加这个晚餐,我们走过名单。
654	He's	琌
655	like, well, what about Jensen?	詹森怎么办?
656	I said, no, sir, we I he's on vacation because he had to postpone	我说,不,先生,我们是在度假因为他不得不推迟
657	vacation for five years.	休假5年.
658	And he's like, get him on the phone.	他就像,让他打电话。
659	But why do you think he sees	但你认为他为什么看到
660	through the hoax?	- 通过骗局?
661	It's, it's, this is the thing.	这是,这是,这是事情。
662	It was quite an extraordinary thing.	这是相当不寻常的事情。
663	It was,	这是,
664	it's polling minus 80.	这是投票减80。
665	So for anyone else that's sitting in the oval office, you're going to do	所以,对于任何其他坐在椭圆形办公室的人, 你会做
666	what's popular.	何为流行.
667	You're representing the people.	你代表着人民
668	This is what everyone wants.	这是每个人都想要的。
669	They want to shut	他们想关门
670	down the data centers and AI.	在数据中心和AI。
671	It seems to be the popular thing in the moment, but he says it's a	这似乎是现在流行的东西, 但他说这是一个
672	sure.	当然
673	And the reason for that is because a lot of people are falling for it.	之所以如此,是因为很多人都爱上它。
674	And so the fact of	所以事实上
675	matters, it's complicated.	事情很复杂
676	You know, at first, I mean, if you look at the story, if you look at	你知道,一开始,我的意思是, 如果你看的故事,如果你看
677	the stories, it's all anchored on two things.	故事,这一切都建立在两件事上。
678	The first thing that it was anchored on was national	第一件事是国家
679	security.	警卫。
680	And recently that was all blown, blown to bits.	最近一切都被炸成碎片
681	Right.	对
682	And so no, no, that story is no	所以,不,不,那个故事是没有
683	longer anchored on national security.	以国家安全为基础。
684	Now it's anchored on safety.	现在它锚定在安全上。
685	Now, if you want AI to be	现在,如果你想AI是
686	safe, the first thing is we need to make sure that the labs that are building it are in control.	安全,第一件事是我们需要确保建造它的实验室是控制。
687	They're good tests for them.	对他们来说是很好的测试
688	If we would like to have third parties to make sure that a third	如果我们想让第三方确保第三个
689	party evaluator, third party evaluators are available, that's no different than financial	当事人评估师,第三方评估师都有,这与财务没有区别.
690	control.	控制系统
691	You guys know we have auditors and the auditors are quite, quite, they don't have to be	你们知道我们有审计师审计师相当,相当, 他们不必是
692	as expert as we are in our business, but they just have to ask the right questions.	作为我们的专家,我们的工作, 但他们只需要问正确的问题。
693	And I think I	我觉得我
694	heard somebody say that it's good to have independent auditors or evaluators, but they just have to have	听到有人说有独立审计员或评估员是好的,但他们只需要
695	multiple.	多个
696	I agree with that too.	我也同意这一点。
697	Just as there's multiple evaluated and auditors, it makes sure	就像有多个评估和审计员,它确保
698	that one company doesn't become, you know, pilled or somehow influenced for whatever reason.	一个公司不会,你知道的, 被麻醉或不知何故的影响。
699	And so	这样
700	you, you know, there's a lot of different ways that you could solve this.	你,你知道,有很多不同的方式你可以解决这个问题。
701	And so I think the	所以我认为
702	number one thing is let's build the technology safely.	第一,让我们安全地建造这个技术。
703	Let's make sure that the testing of it	让我们确保测试它
704	is safe.	这是安全的。
705	And I recognize completely that what is being built is extraordinary.	我完全认识到,正在建设的东西是非凡的。
706	But these are	但是这些是
707	extraordinary companies and we ought to hold them to, to extraordinary standards.	超凡公司,我们应该让他们达到超凡标准
708	And they want to	他们想要
709	be, and they want to be.	和他们想要。
710	I wanted to go back to open source for a second.	我想回开源一会儿。
711	A year ago, we weren't	一年前,我们不是
712	taking it very seriously.	认真对待
713	It was two years, 18 months behind.	是两年,18个月后。
714	The one thing that, you know, one	只有一件事,你知道,一个
715	of the, as you guys know, one of the challenges when you're on the call with President Trump is	你们知道,当你们和特朗普总统通话时遇到的挑战之一是
716	hard to say something.	很难说些什么。
717	I'm going to get in trouble for that.	我会惹上麻烦的
718	I'm sure he's going to call me up on	我相信他会打电话给我
719	that.	那个
720	But anyhow, what I was going to tell him and all of you is that AI is creating an enormous	但无论如何,我打算告诉他和你们所有人是AI正在创造一个巨大的
721	number of jobs.	工作数量。
722	The thing that he wanted more than anything at the beginning of the administration,	他最想要的东西在行政之初,
723	and my first phone call with him, my first time I met him, is that he wants to create jobs in America.	我和他的第一个电话我第一次见到他是想在美国创造就业机会
724	He wants to re-industrialize the United States.	他想让美国重新工业化.
725	He wants to make sure that the United States has	他想确保美国
726	the energy to support the next industrial revolution.	支持下一个工业革命的能量
727	Without energy, there's no industrial	没有能源就没有工业
728	growth.	增长。
729	And so he wants to make sure that there's energy growth, that there's job growth,	所以他希望确保能源增长,就业增长
730	that we're re-industrializing the supply chain.	我们正把供应链重新工业化
731	Look at everything that we're doing right now.	看看我们现在所做的一切
732	All of it is happening right now as we speak.	所有这一切正在我们发言的时候发生。
733	We're creating more jobs than ever.	我们创造的就业机会比以往任何时候都多。
734	We're creating	我们正在创造
735	software jobs.	软件工作。
736	We were just talking about earlier, $400 billion of venture financing	我们刚刚谈的是4000亿美元的风险融资
737	went into the AI industry just recently.	最近才进入AI行业
738	Six months.	六个月。
739	Well, that's created a ton of jobs.	这创造了很多工作
740	That's created a ton of jobs.	这创造了很多工作
741	It's created, obviously, enormous amount of demand for compute,	很明显,它创造了大量的计算需求,
742	which I'm happy about, which is also creating a lot of demand for data centers.	我很高兴,这也造成了对数据中心的大量需求。
743	And we have	我们有
744	to talk about that.	谈谈这个。
745	I think I was talking to Governor Abbott of Texas, and he wants to appeal	我想我在跟德州州长阿博特说话他想上诉
746	to the industry to make sure that we are empathetic to the small communities as we're building data	给工业,以确保我们爱小型社区,因为我们正在建立数据
747	centers all across America, just to be better listeners.	整个美国的中心, 只是为了成为更好的听众。
748	Let's actually talk about that for a	我们来谈谈吧
749	second.	再来
750	What's incredible about NVIDIA, if you break down the component parts, is you've	NVIDIA的不可思议之处是如果你打破了组件,你是
751	effectively had to become the bank of AI to get the ecosystem going.	为了让生态系统得以发展,实际上必须成为大赦国际的银行。
752	And you've had to do it at	你不得不这么做
753	all the levels.	所有级别。
754	You just did this thing with Cloverleaf where you're doing land power shell.	你刚和Cloverleaf做了这件事你正在做土地动力壳。
755	You did this great thing with BlackRock and Goldman and all these folks to essentially create	你和BlackRock和Goldman一起做了一件伟大的事情和所有这些人一起创作
756	financing capability.	筹资能力。
757	Walk us through your capital allocation strategy.	告诉我们你的资本分配策略
758	What has to happen	发生什么
759	to get a broader ecosystem of folks to be able to come in and underwrite this next phase?	让人们拥有一个更广泛的生态系统来承担下一个阶段?
760	As you guys know, this is a new industrial revolution, and every aspect of it is true.	正如你们所知,这是一场新的工业革命, 它的每一个方面都是真实的。
761	This new industry requires manufacturing.	这一新产业需要制造业。
762	Just as the electricity, internet, and now AI,	就像电,互联网, 现在AI,
763	we power anything.	我们把任何东西都给力
764	We can find anything.	我们可以找到任何东西。
765	Now with AI, we can ask and know anything.	现在用人工智能,我们可以问和知道任何事情。
766	Isn't that right?	是吗?
767	And so that's our future.	这就是我们的未来
768	We tap into the ether, and we can ask it of anything	我们钻进乙醚我们可以问它什么
769	we want, and it could explain it to us.	我们想要, 它可以解释给我们听。
770	Now, in order for that to happen, it's got to produce	现在,为了做到这一点,它必须生产
771	the intelligence.	情报
772	And so that's a production process, which is the reason why this infrastructure	因此,这是一个生产过程, 这就是为什么这个基础设施
773	has to get built.	一定要盖起来
774	But once you get the infrastructure built, the question is,	但一旦你得到了基础设施建设, 问题是,
775	what about all of the other layers across the United States?	那全美国其他的层层呢?
776	This industry isn't just about	这个行业不仅仅是
777	the model.	那个模特儿
778	It's not just about the chips.	不只是筹码
779	It's mostly about the applications on top.	主要是关于上面的应用.
780	It's mostly	主要是这样
781	about the infrastructure layer, the data centers, and all the infrastructure, the construction,	关于基础设施层,数据中心, 以及所有基础设施,建筑,
782	the electricity, the power generation.	电力,发电。
783	All of that is involved.	所有这一切都涉及其中。
784	And so I look across the entire	所以,我看着整个
785	ecosystem and look for bottlenecks.	生态系统和寻找瓶颈。
786	And if there are places where extraordinary companies are	如果有的地方非凡的公司
787	being built.	正在建设中。
788	Constraints.	制约因素。
789	Extraordinary companies being built.	特别公司正在建设中。
790	Maybe it's a	也许它是一个
791	supply chain that has to get scaled up so that when we're ready to deploy compute, that they'll	供应链必须扩大,这样,当我们准备好部署计算,他们会
792	be ready for us, land, power, shell.	准备好我们,土地,权力,炮弹。
793	And so this is no different than looking at the supply chain	所以这跟看供应链没什么区别
794	upstream.	上游地区。
795	I probably think about the long-term supply chain more than most because our company	我可能比多数人更关心长期供应链因为我们的公司
796	is really large.	实在是很大。
797	And in order for us to succeed, a whole bunch of companies has to support me.	为了我们成功一大堆公司必须支持我
798	Wendell Corning has to support me.	温德尔·康宁必须支持我
799	Lu.
800	当然是TSMC
801	And memory companies.	还有记忆公司
802	And so	这样
803	we started working with all of these companies long before the revolution, the growth came,	我们很早就开始和这些公司合作了在革命开始前增长就来了
804	so that the growth could happen.	从而实现增长。
805	Now I'm doing it downstream.	现在我在下游做
806	Cycle tends to be that the earnings over time, over long stretches of time,	周期往往是,收入随着时间推移, 长时间,
807	tends to move up the stack right towards the application layer where you can	倾向于将堆栈向上移动,向上移动到应用程序层
808	over-earn for larger periods of time.	过度学习的时间更长。
809	I mean, you bought Hugging Face.	我的意思是,你买了哈金斯脸。
810	Now you're sort of in the,	现在,你有点在,
811	actively in the serving business.	积极从事服务行业.
812	I mean, it seems pretty natural that products like Open	我的意思是,这是很自然的,像开放这样的产品
813	Router make a lot of sense.	路由器很有意义
814	It seems pretty obvious that there are better versions of ways	很明显,有更好的方法
815	to build things like Bedrock.	像贝德洛克那样的建筑
816	I'm sure you think about it.	我相信你会考虑的
817	What's the natural conclusion?	自然结论是什么?
818	Because	因为
819	it seems like the folks up here have no issue trying to move down and you have the best balance	看来这里的人没有问题想搬下来而你有最好的平衡
820	sheet, these incredible engineers, and you have the proven experience to make it right and engineer	床单,这些不可思议的工程师, 而你有证明的经验使它正确和工程师
821	the product and get it out.	把产品拿出来
822	So how do you think about looking up and saying, I could probably do	所以,你觉得怎么样抬头说,我可能做
823	that?	那个?
824	The reason why NVIDIA runs every single model in the world, it's incredible.	NVIDIA在世界上运行每一个模型的原因,令人难以置信.
825	Last year,	去年,我们...
826	about a year and a half ago, the only thing we ran was OpenAI.	大约一年半前我们唯一跑的是OpenAI
827	And now look at amazing models	现在看看惊人的模型
828	are available.	备有。
829	The Metamuse is available.	梅塔缪斯号可以使用.
830	你让Grok有空
831	Grokbot是不可思议的。
832	We now run Gemini and Anthropic is scaling up on our platform as well.	我们现在运行双子座和安特罗皮克在我们的平台上也正在扩大。
833	Since a year and a half ago,	从一年半前开始
834	you got all these Frontier AI models that are now open that are available.	你得到了所有这些边疆AI模型现在开放,是可用的。
835	So the number of models	所以模型的数量
836	that are growing, there's a whole bunch of companies that I won't mention that are building	有一大堆公司,我不会说正在建设
837	Frontier models as well.	边疆模式也是。
838	And the number of AI labs are growing.	而AI实验室的数量也在增长.
839	The ineffables, the reflections,	无能者,反射者,
840	the list goes on, the physical intelligence, the list goes on.	名单继续身体情报名单继续
841	And so all of these labs are building	所以这些实验室都在建设
842	关于NVIDIA。
843	And the reason for that is because as a company, I'd rather for us to help everybody	原因是作为一个公司,我宁愿帮助每个人
844	succeed instead of taking a slice out.	成功而不是取出一块。
845	And so we would go up as far as we need to,	这样我们就可以越走越远了
846	but as low as possible.	但尽可能低。
847	Our strategy is go up as far as we need to and as low as possible.	我们的战略是尽可能提高,尽可能降低。
848	And the reason for that is because if I do that, if I solved it, if not for NVIDIA creating KooDNN,	原因是,如果我这样做, 如果我解决了它, 如果不是NVIDIA创建KoodNNN,
849	all of the frameworks wouldn't exist.	所有的框架都不存在。
850	If not for us creating Megatron core, then all of the	如果不是我们创造了威震天核心那么所有的
851	scale training wouldn't have happened.	规模训练不会发生。
852	So we go and we invent all the technology necessary as far	所以我们去发明所有必要的技术
853	as we need to.	因为我们需要。
854	And then we let a thousand flowers bloom.	然后我们让千花开花
855	And so that posture allows us to be,	这样姿势就能让我们
856	quite frankly, the only- Let's be honest, I agree with you.	坦白说,唯一的... 老实说我同意你的看法
857	The pushback would be that	后退是那样的
858	it really would be great to have more competition at the hyperscale layer.	在超尺度层进行更多的竞争会很好
859	And I think you've done a great job supporting the Neo clouds.	我认为你已经做了一个伟大的工作支持尼欧云。
860	There are some, and by the way,	有些,顺便说一句,
861	I think you introduced me to NVIDIA is superb.	我认为你把我介绍给NVIDIA很棒
862	Great.	不错
863	Everything.	一切众生.
864	They're amazing,	他们是惊人的,
865	but we need like 50 of these guys.	但我们需要50个这样的人
866	We need a hundred of them.	我们需要一百个
867	We need a thousand of them.	我们需要一千个
868	And it just may take some.	它只是可能需要一些。
869	Yeah.	对
870	You know, it's just, I'm surprisingly uncompetitive.	你知道,这只是, 我惊讶地没有竞争力。
871	Really?	真的吗?
872	Yeah.	对
873	That's not my thing.	这不是我的事。
874	You know, my thing is kind of like, for example,	你知道,我的事情有点像,例如,
875	I'd be more than happy with five hyperscalers.	我更喜欢5个超标器
876	However, the reason I noticed the early customers	然而,我注意到早期顾客的原因
877	of all the Neo clouds, all the, what we call NCPs, all the early customers were the hyperscalers.	所有新云,所有,我们称之为NCP, 所有早期的客户都是超规模的。
878	Exactly.	没错
879	And the reason for that is because the hyperscalers plan once a year,	原因是超标器每年计划一次
880	but the market dynamics is so volatile right now that they're always almost wrong.	但是现在的市场动态非常动荡, 以至于它们总是几乎是错误的。
881	And so with all	┮Τㄆ薄
882	these regional clouds who are agile, they can move fast.	这些区域云是敏捷的, 它们可以快速移动。
883	They know their state or they know their	他们知道自己的状态或他们知道自己的状态
884	country.	国家。
885	They know their region.	他们知道自己的地区。
886	They're securing land, power and shell in a way.	他们从某种意义上保证了土地、电力和炮弹的安全。
887	That's hard for	这很难
888	somebody who sits in Seattle or sits in Palo Alto to be able to see the planet.	有人坐在西雅图或者坐在帕洛阿尔托能看到地球。
889	And so we now have	所以我们现在有了
890	basically a large scale distributed network of companies that are building securing land,	基本上是一个大规模的分布式网络公司正在建设保障土地,
891	power, shell for us.	电力,贝壳给我们。
892	And now countries realize it's strategic.	现在各国意识到这是战略性的。
893	So many countries are saying,	这么多国家说,
894	I'm going to take my power and only give it to my own companies.	我要夺权,只给自己的公司.
895	Well, Nvidia is in that country	恩维迪亚在那个国家
896	as well, and we could help the Neo clouds in that country grow.	也可以帮助这个国家的尼欧云层生长
897	And so whether it's, whether it's	所以,无论它是,是否
898	Firmus in Australia, we just did a whole bunch of stuff in Australia, brought on two more gigabytes,	在澳大利亚的Firmus 我们刚在澳大利亚做了一堆事情带来了两个千兆字节
899	Southeast Asia, of course, IOH and others bring on a few gigabytes.	东南亚,当然,IOH等带来了几千兆字节.
900	And so we're building a gigawatts.	所以我们正在建造千兆瓦。
901	So we're building, you know, we're scaling up, you know, it's pretty clear though.	所以,我们正在建设,你知道, 我们正在扩大,你知道, 这是相当清楚的。
902	I just want to get this one thing in.	我只想把这个东西弄进去
903	It's pretty clear that you're going pretty high up and getting	很明显,你上台了
904	very focused on open source.	非常专注于开源。
905	Obviously you have your nemotrons doing exceptionally well.	很明显,你的坏蛋们做得特别好
906	I use	我用这个
907	them often hugging face poolside and Laguna, very, very solid product that you're now aqua hiring,	他们经常拥抱脸池边和拉古纳, 非常,非常扎实的产品,你现在正在雇用,
908	hiring, whatever it is.	雇佣,无论是什么。
909	And then you have your open source stack for self-driving also very disruptive.	然后,你有了开源堆栈用于自我驾驶也非常干扰。
910	So we are the frontier model in five domains.	所以我们是五个领域的前沿模式。
911	Yeah.	对
912	And so you don't seem to build	所以你似乎没有建设
913	products to get the silver medal.	获得银牌的产品。
914	You seem to go for the gold.	你似乎去的黄金。
915	So are you going for the gold and	所以,你要去的黄金和
916	will you have the best hands down open source model?	你会有最好的手下开源模型?
917	And then part B to that is can open source	B部分可以打开源头
918	catch up to frontier models and are you the person to do it?	追上边疆模特儿,你做得到吗?
919	So the logic, the logic Jason is that	所以逻辑,逻辑杰森是
920	that we will build it because one we can, we have the skills to do it and because our customers need	我们将建造它,因为一个我们能够, 我们有做它的技能和因为我们的客户需要
921	us to do it.	我们这样做。
922	Right.	对
923	So for example, Alpamayo is the world's first thinking self-driving car.	例如,阿尔帕马约是世界上第一个思考自我驾驶的汽车.
924	And by thinking, by reasoning, you don't need as much data as, you know, you don't have to train	而且通过思考,通过推理, 你不需要像,你知道, 你不需要训练
925	on a few billion hours of road data because you could reason about it, break down the problem into,	在几亿小时的公路数据上因为你可以解释这个问题把问题分解为
926	Oh, I've seen this before.	哦,我见过这个。
927	It's not exactly the same, but it's largely the same as that.	虽然不完全相同,但大致相同.
928	Okay.	摆
929	And so, so Alpamayo, why is it necessary?	那么,阿尔帕马约,为什么有必要呢?
930	Well, there's a whole bunch of car companies.	有很多汽车公司
931	Every car in the world is going to be autonomous, but beyond that, every ag tech, every truck,	世界上的每辆车都会是自主的但除此之外每台高科技每辆卡车
932	every van, and most of them aren't big enough in scale to be able to build that whole stack.	每辆面包车, 并且大部分都不够大规模足以建造整个堆。
933	So I'll build an extraordinary stack for them.	所以我会为他们建造一个非常的堆栈。
934	They do last mile adapting for their application.	他们做最后一英里适应他们的应用。
935	Now, everything that moves in the future could be autonomous.	现在,未来一切动静都可以自主.
936	If not for us building some of the,	如果不为我们建造一些,
937	some of the biology models, the world wouldn't have it.	一些生物学模型, 世界不会有它。
938	The, the ESM two protein found language	两种蛋白质找到语言
939	model.	型号。
940	We created that ESM fold, open fold, alpha fold two, all the stuff with, with coup	我们创建了ESM折叠, 打开折叠, α折叠2, 所有的东西与, 政变
941	variant.	变种.
942	All of that stuff, technology wouldn't have existed if we didn't build it.	所有这些东西, 如果我们不建造它,科技就不会存在。
943	One of my favorites, protein, a complexa is, you know, synthesizing next generation proteins	我最喜欢的蛋白质,一个复合体就是合成下一代蛋白质
944	and it's binding.	它的束缚。
945	It's groundbreaking stuff.	这是突破性的东西。
946	We built that.	我们造的
947	And so we'll build that because	所以我们要建立它,因为
948	Lily needs it and, and, you know, Mark needs it and others need it and they don't have the	莉莉需要它,而且,你知道, 马克需要它,其他人需要它他们没有
949	capability to do it or they they're not yet there.	要不然他们还没到那里
950	And so we can make it a real contribution.	这样我们才能做出真正的贡献。
951	So I	所以说
952	do everything out of need.	做一切无所求故.
953	I'm not trying to disrupt.	我不是想打搅你
954	I mean, we don't wake up in the morning,	我的意思是,我们不会早上醒来,
955	try to disrupt anybody, or just wake up in the morning, try to help everybody.	试图破坏任何人, 或者只是早上醒来, 试图帮助每个人。
956	Jensen, what about, what about competitive threats that might be emerging to your core	詹森,那又怎样,你的核心可能出现的竞争威胁呢?
957	business?	生意吗?
958	Can you just comment?	你能发表评论吗?
959	We're just so nice.	我们很好
960	Yes.	对
961	Well, I know this is, well, I actually want to, I want to just get your,	嗯,我知道这是,嗯, 其实我想,我想只是得到你的,
962	let's just call it a take.	就当是一桩吧
963	What's your take on Tarafab, a hundred million square foot facility	你对塔拉法布有什么看法? 一亿平方英尺的设施
964	Elon's announced.	伊隆公布.
965	And if anybody could do it, he can.	如果任何人都能做到这一点,他可以。
966	And the two of us were on a flight together	我们俩一起搭飞机
967	to a country and.	去一个国家,然后。
968	With a person who sometimes calls you on the phone.	和一个有时打电话的人在一起
969	It was a nice plane.	飞机不错
970	And, and, and we had like, you know, and, you know, Elon likes to talk about	而且,而且,我们喜欢,你知道, 而且,你知道,埃隆喜欢谈论
971	these things.	这些东西。
972	And, and so we spent a lot of time talking about it.	所以,我们花了很多时间谈论它。
973	Is that anybody could do it?	有人能做到吗?
974	He could do it because I mean, you could, you design chips, you don't fab them.	他可以这样做,因为我的意思是,你可以, 你设计芯片,你不喜欢它们。
975	Could your chest be fab there?	你胸口会痛吗?
976	Or is it?	是吗?
977	Well, we know, we, we know a lot about	我们知道,我们知道很多
978	process technology because we're pushing the limits of everything.	工艺技术,因为我们在推动一切的极限。
979	And, you know,	还有,你知道,
980	because we scale at such large scale, we have incredible memory technology inside the company	因为我们的规模如此之大我们公司内部有惊人的记忆技术
981	with the world's best 30s company.	与世界最佳30强公司.
982	You know, we got lots of amazing.	你知道,我们有很多惊人的。
983	So your take is you've	所以,你的选择是你已经
984	talked a lot about it.	谈了很多关于它。
985	So we could just, yeah, we could talk about it.	所以我们可以,是的,我们可以谈论它。
986	And, and you can't discourage	而且,你无法抑制
987	Elon from doing it, which is one of his incredible, that's his superpower.	伊隆从做到这一点,这是他不可思议的之一,这就是他的超能力.
988	And once you decide to go	一旦你决定离开
989	do something, it's hard to stop them.	做点什么很难阻止他们
990	And so I can, can you give us your take on where China is	所以,我可以,你可以给我们你承担中国在哪里
991	with advanced lithography systems?	高级的石刻系统?
992	Um, native grown, they're going to get there by 2030 by 2030.	嗯,土生土长, 他们将在2030年之前到达那里。
993	Yeah.	对
994	And 2030 is just around the corner.	2030年就在拐角处。
995	Yeah.	对
996	Also that's sorry, how long will all be dead at	还有,对不起,你们要死多久?
997	that time?	那个时候?
998	So, and does, and for China, does that mean the switch is flipped and then that's all	所以,确实,而且对于中国来说, 这是否意味着开关被翻了,然后就这样
999	going to go into, um, uh, mainland fabs almost immediately?	马上就要进入大陆了?
1000	You know, the way to think about	你知道, 思考的方式
1001	China is really good at high volume production.	中国的产量确实很高。
1002	And this is just matter of time.	而这只是时间问题。
1003	Yeah.	对
1004	And I,	还有我
1005	you know, I, I think in, I, I think in decades as well, you know, I've been around a long time and,	你知道,我,我想, 我,我想在几十年的时间里,你知道,我一直周围的时间和,
1006	you know, for Nvidia, I've got to think about what happens next decade, decade after that.	你知道,对恩维迪亚, 我必须思考接下来的十年,十年之后会发生什么。
1007	So two or three years, it's just a click.	因此,两三年,这只是一个点击。
1008	It's nothing.	没什么
1009	And so as far as they're concerned,	至于他们,
1010	they're already there.	他们已经到了
1011	They're already there.	他们已经在那里。
1012	Yep.	对
1013	That's an Elon and Gwen.	这是伊隆和格温的作品
1014	We've got to run	我们得走了
1015	America.	美国。
1016	We got to run.	我们得走了。
1017	Yeah.	对
1018	Speed run.	速度运行。
1019	So we got a speed run.	所以我们得到了一个速度运行。
1020	Well, slowing down is definitely	慢下来肯定
1021	the wrong strategy.	错误的策略
1022	Well, I mean, it, it feels apparent, I think to most of us in the industry	嗯,我的意思是,它,感觉很明显, 我想我们大多数的行业
1023	that we're kind of in the AGI moment and it's a definition, obviously as smart as any other	我们处于AGI的时刻,它是一个定义,显然和任何其他人一样聪明
1024	human.	人类
1025	I think we're already there.	我想我们已经到了
1026	We're there.	我们到了
1027	Right.	对
1028	And so then superintelligence is the next	接下来是超级情报
1029	waypoint based on what you see based on your customer base, based on your history.	根据你的客户基础,根据你的历史,看路标
1030	And Jason,	还有杰森
1031	I think we're there too.	我想我们也到了
1032	You think we're at superintelligence.	你以为我们在超情报机构吗?
1033	Yeah.	对
1034	When you,	当你
1035	when you, when you take a narrow segment, a narrow segment, I mean, my, my self-driving car,	当你,当你走一个窄段,一个窄段, 我的意思是,我的,我的自驾车,
1036	I don't want you to make me an omelet.	我不想你给我做煎蛋卷
1037	I just want you to drive the car.	我只是想让你开车。
1038	Right.	对
1039	That is super	好厉害
1040	intelligent.	聪明
1041	It's better than a human.	比人类强.
1042	Yeah.	对
1043	Like one 10th of the accident, right?	就像事故的第10个对吧?
1044	Exactly.	没错
1045	Yeah.	对
1046	Synthesizing proteins, you know, doing virtual screening of proteins.	合成蛋白质,你知道,对蛋白质进行虚拟筛选。
1047	We're already	已经是了
1048	there.	这里
1049	Yeah.	对
1050	Are you having fun being on the frontier of humanity?	你在人类的前沿玩得开心吗?
1051	I like it.	我喜欢
1052	Yeah.	对
1053	Ladies and gentlemen, I like it.	女士们先生们我喜欢
1054	I like it.	我喜欢
1055	And then guys, guys, it's, it's,	然后伙计们,伙计们,这是,这是,
1056	it's great there.	这是伟大的。
1057	The future is great and we want to get there.	未来是伟大的,我们想到达那里。
1058	Listen, a lot of us don't have to	听着,我们很多人不用这样
1059	work, but I got to tell you, it's too good not to be so fun.	工作,但我得告诉你, 这是太好不这么好玩。
1060	Right.	对
1061	And so, so I want every we,	所以,所以我要我们每一个人,
1062	I want to be there.	我想去那边
1063	I want all of you guys there with me.	我要你们都跟我走
1064	We're all going to be there.	我们都会去的
1065	We're going to be enormously successful together as a humanity.	作为人类,我们将会取得巨大的成功。
1066	And, and in the meantime, we got	在此期间,我们得到了
1067	to encourage them, urge them on.	鼓励他们,敦促他们继续。
1068	They're doing really, really important work, as you guys know,	他们正在做真正,真的很重要的工作, 你们知道,
1069	and I want them to succeed.	我希望他们成功。
1070	I also would love for us to tone down the, the, the drama and make,	我也很想让我们低调戏剧和制作,
1071	and most importantly, we need all of America to come with us.	最重要的是,我们需要所有的美国与我们一起。
1072	That's how we make it.	我们就是这样做到的

该视频共有字幕 1080 条。解锁更多字幕为会员功能，请移动到价格

下载安装