推荐微信扫码登录以获得更好的体验!另:🎉建议更新到最新版,以体验新功能!🎉 版本更新记录→

百度小米掌门人齐发警告:AI热潮与回报的差距正在拉大|附文稿

字幕摘录

时间英文中文
0:00Imagine breaking your arm, you go to the hospital, you look at an x-ray, right, and the doctor says,想象一下,你的手臂断了, 你去医院, 你看X光,对,医生说,
0:06I have no idea what I am looking at.我不知道自己在看什么
0:08Oh wow.哦,哇呜。
0:09That is a terrible feeling.这是一种可怕的感觉。
0:10Yeah.对
0:11But that is exactly这正是
0:13how two major tech CEOs feel right now.两位首席技术执行官现在的感受
0:16Yes, they do.是的,他们这样做。
0:17They're looking at the massive artificial他们正在寻找巨大的人工
0:19intelligence industry, you know, AI.情报行业,你知道,AI。
展开字幕全文(553 条)
序号英文中文
1Imagine breaking your arm, you go to the hospital, you look at an x-ray, right, and the doctor says,想象一下,你的手臂断了, 你去医院, 你看X光,对,医生说,
2I have no idea what I am looking at.我不知道自己在看什么
3Oh wow.哦,哇呜。
4That is a terrible feeling.这是一种可怕的感觉。
5Yeah.对
6But that is exactly这正是
7how two major tech CEOs feel right now.两位首席技术执行官现在的感受
8Yes, they do.是的,他们这样做。
9They're looking at the massive artificial他们正在寻找巨大的人工
10intelligence industry, you know, AI.情报行业,你知道,AI。
11对,AI。
12And the picture is just completely blurry to them.照片对他们来说完全模糊。
13It is very blurry.它非常模糊。
14The numbers just do not add up at all.数字完全没有增加。
15So we are jumping right into our deep所以我们正跳进我们的深渊
16dive today.潜水今天。
17We have a really fascinating article from the South China Morning Post.南华早报有篇很吸引人的文章
18Yeah, it is a是啊,这是一个
19great source.伟大的来源。
20The focus is artificial intelligence and our mission today is simple.重点是人工智能,我们今天的任务很简单。
21We want to find我们要找到
22out something important for you, the listener.拿出对你很重要的东西,听者。
23Exactly.没错
24We need to know if AI is just a giant我们要知道AI是不是个巨人
25wave of excitement or if it has real value for your future.兴奋的浪潮,或者如果它有 真正的价值你的未来。
26Because right now we have a very因为现在我们有一个非常
27serious warning and it comes from the absolute top of the industry.严重警告,它来自行业的绝对顶端。
28We do.没错
29The leaders of two major二大领导人
30Chinese tech companies are sounding the alarm.中国科技公司发出警报.
31I mean, they are really worried.我的意思是,他们真的很担心。
32We are talking我们在谈
33about Baidu and Xiaomi.关于白度和小美
34Robin Li is the CEO of Baidu and Lei Jun is the CEO of Xiaomi.罗宾·李是白度的首席执行官,雷军是小米的首席执行官.
35Right,对
36they recently spoke at the APEC CEO Summit in Shenzhen.他们最近在深圳亚太经合组织首席执行官峰会上发言。
37That is a huge meeting for business这对生意来说是一个巨大的会议
38leaders, right, in the Asia-Pacific region.亚太地区的领导人,没错。
39Yeah, a very big deal.是啊,一个非常大的问题。
40And they both delivered a really他们都送了一个真正的
41harsh truth there.严酷的真相
42What did they say?他们说什么?
43Well, they said companies are spending way too much money on他们说公司花太多钱了
44AI.大赦国际。
45Like, they are pouring in billions of dollars.他们正在以数十亿美元注资。
46Okay, but what is the problem?好吧,但有什么问题?
47The problem is问题是
48profit.利润。
49They are not making nearly enough profit.他们赚的利润都不够
50It just does not justify those massive costs.它只是不能证明这些巨额费用是合理的。
51Wow.珃
52So people are starting to ask questions.所以人们开始问问题。
53They wonder if the AI hype has gone too far.他们怀疑AI的狂言是否过分了.
54Yeah.对
55And we should probably explain hype really quickly.我们也许该快点解释一下
56Good idea.不错
57Hype just means when people get way too excited about a new thing.假话只是指当人们对新事物太兴奋时
58Right.对
59Everyone talks about it.皆说此经.
60But the real facts do not match the excitement.但真实事实与兴奋不符.
61We hear about AI everywhere.我们到处听到AI的消息。
62But the actual business reality is different.但实际商业现实不同.
63The profits are just not there yet.利润还没到
64No, they are not.不,他们不是。
65And because of this gap,由于这个差距,
66people are getting nervous.人们越来越紧张。
67Very nervous.非常紧张。
68They are worrying about an AI bubble, right?他们担心的是人工智能泡沫 对吧?
69Yes.对
70Let's explain the word bubble for our listeners.向听众解释泡泡一词.
71In business, a bubble is a very在生意上,泡泡是一个非常
72specific thing.具体的事情。
73It is.没错
74It means investors put massive amounts of money into something.这意味着投资者投入了大量资金。
75And the price goes up incredibly fast.价钱涨得很快
76Right.对
77The price shoots up.价格涨了
78But the real但真实的
79world value is actually very low.世界价值实际上很低。
80Exactly.没错
81It looks huge and impressive from the outside.从外部看,它看起来很大,令人印象深刻。
82But eventually the market wakes up.但最终市场醒来.
83They realize他们意识到
84there is no solid foundation.没有坚实的基础。
85And?还有呢?
86And the bubble pops.和泡沫弹出。
87Yeah, the bubble pops.是啊,泡泡弹。
88And the prices crash hard.价格暴跌
89This fear of a crash is really growing right now.这种对坠机的恐惧现在真的越来越严重了。
90The market got very jumpy recently.最近市场变得非常紧张
91Very jumpy.非常紧张
92And part of that fear started after a big event.部分恐惧始于一个重大事件之后。
93Jensen Huang resigned.黄詹森辞职.
94He was the CEO of NVIDIA.他是NVIDIA的首席执行官.
95Yes.对
96And his resignation was a massive shock to the system.而他的辞呈是对系统的巨大冲击.
97因为NVIDIA非常重要
98It is.没错
99I mean, they make the specific computer chips for AI.我的意思是,他们做特定的计算机芯片 AI。
100Right.对
101Those chips power almost all modern AI.这些芯片几乎都是现代的人工智能
102NVIDIA is basically the engine of theNVIDIA基本上就是这台机器的引擎
103whole industry.整个行业。
104They are the engine.他们是引擎。
105So when the leader of that essential company leaves, people panic.所以当那个重要公司的领袖离开时 人们会惊慌失措
106They do.没错
107Yeah.对
108They worry the whole engine might just stop working.他们担心引擎可能停止工作。
109So let me ask you this.我问你
110Are these smart companies just throwing money at a wall and hoping it这些聪明的公司是不是只是把钱扔在墙上 希望它
111sticks?棍子吗?
112Honestly, it really looks like that right now.老实说,它看起来真的像现在。
113Really?真的吗?
114Just guessing?只是猜测?
115Well, the technology itself is amazing.技术本身很惊人
116We can all see that.我们都可以看到这一点。
117But the business plan is completely missing.但商业计划却完全缺失.
118Right.对
119They build these massive smart computer systems.他们建造这些庞大的智能计算机系统.
120Yeah.对
121But they do not know who will actually pay for them.但是,他们不知道谁会支付他们的实际费用。
122Why is it so incredibly hard to make a profit?为什么这么难赚钱?
123Lei Jun from Xiaomi talked about this, right?小美的李俊谈这个,对吧?
124He did.没错
125He said the initial investment is just huge.他说最初的投资是巨大的。
126You spend so much money up front.你花这么多钱前面。
127But the actual everyday uses for AI are very limited right now.但目前AI的实际日常用途非常有限.
128Exactly.没错
129He pointed out a major flaw.他指出了一个重大缺陷。
130He says very few companies can close the commercial loop.他说很少有公司可以关闭商业圈子.
131The commercial loop.商业圈.
132Let's break that down.让我们打破这一点。
133What does that mean?这是什么意思?
134It is a simple idea.这是一个简单的想法。
135A commercial loop just means turning a cool technology into actual profit.商业循环只是意味着将一种酷的技术转化为实际利润.
136So you spend money to build something, then you sell it, and the money loops back to you as profit.所以,你花钱来制造一些东西, 然后卖掉它, 钱循环回给你作为利润。
137Right.对
138But right now, companies spend fortunes building these massive AI tools.但是现在,公司花费财富 建造这些庞大的AI工具。
139And they have not figured out how to get the customer to pay for them.他们还没想好怎么让顾客付钱
140Exactly.没错
141The money goes out, but it does not loop back in.钱会花掉,但不会回流
142The loop stays open.循环保持开放。
143Yeah.对
144And an open loop means the company is just bleeding money.而公开循环意味着公司正在流血的钱。
145They're losing cash every day.他们每天都在损失现金。
146Lei Jun also noted something else.黎军也注意到了别的东西.
147He said there is a lack of truly他说没有真正的
148revolutionary AI products.革命性AI产品.
149Right.对
150I mean, we have AI that can chat with us.我的意思是,我们有AI 可以和我们聊天。
151That is cool.这很酷。
152We have AI that can draw funny pictures for us.我们的人工智能可以为我们绘制有趣的图片。
153Yeah.对
154But are those things completely changing how businesses operate?但是这些事情是否完全改变了企业的运作方式?
155Not really.没有
156They are interesting little toys, but they are not essential tools.它们是有趣的小玩具,但它们不是必不可少的工具。
157Exactly.没错
158Businesses will not pay top dollar for a toy.企业不会为玩具支付最高的美元。
159And Robin Lee from Baidu questioned the massive costs, too.而来自拜都的罗宾·李也质疑巨额成本.
160There is a fierce race right now.现在比赛激烈
161Oh, a massive race.哦,一场大规模的比赛。
162The United States and China are both running hard.美国和中国都在艰难地奔跑.
163They both want to build the biggest, most advanced AI models in the world.他们都希望建立世界上最大,最先进的AI模型.
164Right.对
165But Robin Lee is looking at the crazy costs of this race.但是罗宾 李在看这场比赛的疯狂代价.
166And he is wondering something important.他想知道一些重要的东西。
167He is wondering if it is actually worth it, right?他想知道这是否值得,对不对?
168Building these giant general purpose models.建造这些巨大的通用模型.
169Yes.对
170He questions the price tag.他质疑价格标签。
171It sounds like a funny situation.这听起来像一个有趣的情况。
172It is like building a giant expensive race car,就像造一辆昂贵的赛车
173but nobody knows how to drive it to the store.但没人知道怎么开到店里
174That is a perfect way to look at it.这是一个完美的方法来看待它。
175You have this amazing machine.你有一个惊人的机器。
176It can drive 300 miles per hour.它能每小时开300英里.
177But your customer just needs to go buy milk down the street.但你的顾客只需要去街上买牛奶
178Exactly.没错
179The machine is completely overpowered for that simple job.机器完全超能力完成这项简单的任务。
180You spend a fortune on special fuel for a tiny trip.你花一大笔钱买特别燃料 做个小旅行
181We have this powerful engine, but we do not have the wheels to get anywhere useful.我们拥有这个强大的引擎,但是我们没有车轮去任何有用的地方。
182And that raises a really big question for the industry.这给业界带来了一个大问题。
183What is that?这是什么?
184Well, if these giant AI models do not work well for everyday business right now, when will they?那么,如果这些巨型AI模型 现在对日常业务不起作用, 他们什么时候会?
185Ah.摆
186Robin Lee actually gives us an interesting answer for that.罗宾 李真的给了我们一个有趣的答案
187He looks back at history.他回顾历史.
188Right.对
189He compares AI to the Industrial Revolution.他把AI比作工业革命.
190Specifically, he talks about the discovery of electricity.具体来说,他谈到电力的发现.
191This is such a crucial comparison.这是如此关键的比较。
192Think about when humans first controlled electricity.想想人类何时开始控制电力
193Right.对
194It did not make factories rich immediately, did it?它没有使工厂 立即富有,是吗?
195No, it did not.不,它没有。
196In fact, it took decades for electricity to deliver real economic benefits.事实上,电力需要几十年才能带来真正的经济利益。
197Decades?几十年?
198Why did it take so long?为什么这么长时间?
199I mean, electricity is amazing.我的意思是,电是惊人的。
200It is.没错
201You flip a switch and you have instant power.你打开一个开关,你就有瞬间的力量。
202Right.对
203It seems like it should be an instant success.似乎应该马上取得成功。
204Well, you have to picture how the old factories worked back then.你得想象一下那个时候老工厂是如何工作的
205Okay.摆
206I am picturing it.我在想象着
207They ran on giant steam engines.他们跑在巨大的蒸汽机。
208You had one massive steam engine right in the middle of the room.你有一个巨大的蒸汽机 就在房间中间。
209And all the machines were connected to that one big engine.所有机器都和那台大引擎相连
210Yes.对
211They used this crazy system.他们用这个疯狂的系统。
212Long leather belts and heavy metal poles.长皮带和重金属杆.
213So the power had to be shared from that one spot.因此,权力必须从那个地方分享。
214Exactly.没错
215The layout of the whole building was dictated by that single power source.整个建筑的布局是由单一的电源决定的。
216So when electricity finally arrived, what did the factory owners do?所以当电力终于到来时,工厂的业主做了什么?
217They did the simplest thing possible.他们做了最简单的事情
218They took out the big steam engine.他们拿出了大蒸汽机。
219And they put a giant electric motor in the exact same spot.他们把一个巨大的电动机 放在同一个位置。
220Yep.对
221Right in the middle of the room.正恁么时.
222Did that actually help them make more money?这真的帮了他们赚更多的钱吗?
223No, it didn't.不,没有。
224Not at all?一点也不?
225The dangerous belts were still there.危险的腰带还在那里。
226The bad layout was still there.糟糕的布局还在那里。
227Oh, I see.原来如此
228They completely missed the real magic of electricity.他们完全错过了真正的电魔法.
229Yes, they did.是的,他们做到了。
230So what is the real magic?那么,什么是真正的魔法吗?
231The magic is that you can put a tiny, cheap electric motor on every single machine.魔力在于你可以在每个单机上安装一台小型,廉价的电动机.
232Wow.珃
233So you do not need a central power source anymore.所以你不再需要中央电源了
234Exactly.没错
235You can move the machines anywhere you want in the building.你可以把机器移动到任何你想要的建筑物。
236You can redesign the whole factory floor.你可以重新设计整个工厂的地板。
237Yes.对
238You can create a logical assembly line.可以创建逻辑组装线.
239The product moves smoothly from one worker to the next.产品从一个工人顺利地移动到另一个工人。
240But factories had to figure that out first.但工厂必须先弄清楚这一点。
241They had to completely change how they worked.他们必须彻底改变工作方式。
242Right.对
243They had to change their organizational structure.他们不得不改变组织结构。
244Let's explain organizational structure.我们来解释一下组织结构
245It just means how a company sets up its teams, how workers do their daily jobs.它只是指公司如何组建团队,工人如何做日常工作.
246So they had to change all of that to actually fit the new technology.因此,他们不得不改变这一切 真正适应新技术。
247Yes.对
248And Robin Lee says the exact same thing is true for AI today.还有罗宾 李说今天AI也一样
249Small and medium businesses cannot just buy an AI program and plug it in.中小企业不能仅仅购买AI程序并插上.
250No, they cannot.不,他们不能。
251They cannot just add it to their old rigid way of working.他们不能仅仅把它加在他们老的僵硬的工作方式上。
252Right.对
253If you just give a worker an AI chatbot but keep all the old rules, nothing changes.如果你只是给一个工人一个AI聊天员... ...但保持所有旧规则,没有什么变化。
254The company has to rethink the whole process.公司必须重新思考整个过程.
255They need a new structure, one that aligns with what AI can actually do.他们需要一个新的结构,一个符合AI实际所能做的结构.
256Okay.摆
257Let's unpack this because I have to push back on this idea.我们把这个打包好 因为我得回击这个主意
258Okay.摆
259Go ahead.来吧。
260Today, the modern business world moves much faster than the industrial revolution.今天,现代商业世界的发展速度远快于工业革命.
261Oh, absolutely.当然
262Much faster.再快一点
263Do businesses really have decades to figure this out?企业真的有几十年的时间来解决这个问题吗?
264That is the real problem right there.这就是真正的问题。
265In the past, companies had the luxury of time.过去,公司有奢侈的时间。
266Right.对
267But today, technology moves incredibly fast.但是今天,技术的移动速度非常快。
268If a company waits decades to adapt now, they will simply go out of business.如果一家公司要等几十年才适应 现在,它们就会失去业务。
269The competition will just crush them.竞争只会压垮他们
270Exactly.没错
271Waiting is too dangerous.等待太危险了
272So these CEOs suggest a faster solution.所以这些首席执行官建议更快的解决办法.
273They want to find real value right now, today.他们想要找到真正的价值 现在,今天。
274Yes.对
275They want to stop focusing on the big general models.他们希望不再专注于大通用模型.
276They want to focus on specific practical tasks instead.他们希望侧重于具体的实际任务。
277Right.对
278Robin Lin says the real economic value is in the application layer.罗宾 林说实际经济价值在应用层.
279The application layer.应用层.
280Let's break down that term for our listeners.让我们把这个词分解给我们的听众。
281Okay.摆
282Think of your smartphone.想想你的智能手机。
283The background code is the foundation.背景代码是基础.
284You never really see it.你从来没有真正看到它。
285Right.对
286I do not interact with the operating system code.我不与操作系统代码互动。
287But the application layer is the apps on your screen, the map, the camera.但应用层是屏幕上的应用程序,地图,相机.
288The messaging app.腾讯应用.
289The application layer means the specific tools that people actually use every day.应用层是指人们每天实际使用的具体工具.
290Yes.对
291You cannot sell a customer just the hidden foundation.你不能只卖隐藏的基金会
292You have to sell them a working app.你得卖个工作应用软件
293So Robin Li is saying we spend way too much time and money building the hidden foundation.所以李罗宾说我们花太多时间和钱 来建立隐藏的基础
294Exactly.没错
295We need to build more useful apps.我们需要建立更有用的应用程序。
296And Lei Jun from Xiaomi completely agrees with this.而小米的雷军也完全同意这一点.
297He does.他这样做。
298He says the greatest products will combine AI with hardware.他说最伟大的产品会把AI和硬件结合起来.
299So he wants to put the AI brain inside physical devices.所以他想把人工智能大脑放进物理设备里.
300Things you can touch and use.你可以触摸和使用的东西。
301Yes.对
302He predicts huge value in AI smartphones.他在AI智能手机中预测了巨大的价值.
303He also talks about AI vehicles.他还谈到AI车辆.
304Like smart cars.像聪明的汽车。
305And he specifically mentions humanoid robots for factories.他特别提到了工厂的人造机器人。
306Humanoid robots.人类机器人.
307That means robots that look and move just like humans.这意味着机器人的外观和动作就像人类一样.
308Right.对
309This is actually a very smart strategy.这其实是一个非常聪明的策略。
310Why is that?为什么?
311Well, think about a factory today.好吧,想想今天的工厂。
312Factories are already built for human workers.工厂已经为人类工人建造。
313Oh, I see.原来如此
314They have stairs.他们有楼梯。
315They have human-sized tools.他们有人类规模的工具。
316Exactly.没错
317They have narrow walkways meant for people.他们有狭窄的走道 专为人们。
318So if you build a robot shaped like a human, you do not have to rebuild the factory.所以,如果你建造一个像人类形状的机器人, 你不必重建工厂。
319Right.对
320The robot can just step in.机器人可以进来
321It can use the existing human tools.它可以利用现有的人类工具。
322It brings the AI brain into the physical world very smoothly.它使AI大脑非常顺利地进入物理世界.
323That makes total sense.这是完全合理的。
324Now Baidu has a slightly different vision for finding immediate value.现在拜都对于寻找眼前的价值有稍有不同的看法.
325Robin Li says domain-specific AI agents are the real answer.Robin Li说特定域的AI特工才是真正的答案.
326Domain-specific.具体领域.
327Let's explain that one.我们来解释一下
328That means the AI is trained for just one specific job.这意味着人工智能只接受一项具体工作的培训。
329Right.对
330It is not a general genius.这不是一个一般的天才。
331It does not know how to write poetry.它不知道如何写诗。
332It does not know how to draw pictures.它不知道如何画画。
333It just knows its own small world perfectly.它完全了解自己的小世界
334It is a highly trained specialist.这是一个训练有素的专家。
335And there is a brilliant example from our source article.我们的源头文章中有一个辉煌的例子。
336The Xingdao port, right?兴岛港,对吧?
337Yes.对
338Look at the port of Xingdao in China.望中国兴道港.
339It is a massive shipping port.这是一个庞大的港口。
340Millions of large metal shipping containers move through it every single year.每年有数百万个大型金属运输集装箱经过该地。
341And Baidu uses a domain-specific AI right there at the port.而Baidu在港口使用一个特定域名的AI.
342What does it do?这有什么用?
343It only does one single job.它只做一项工作。
344It looks at all the shipping containers waiting on the dock.它看着码头上所有等待的运输集装箱.
345Then it calculates the absolute most efficient way to load those containers onto the cargo boats.然后它计算出把这些集装箱装上货船的绝对最有效的方法.
346Wow.珃
347Loading a massive cargo ship is a giant complicated puzzle.装上大型货船是一个巨大的复杂谜题.
348It really is.没错
349You have very heavy boxes.你们的箱子很重
350You have very light boxes.你有非常轻的盒子。
351And they're all going to different cities around the world?他们都要去世界各地不同的城市?
352Right.对
353If you load them wrong, the ship becomes unbalanced.如果你装错了, 船变得不平衡。
354It is dangerous.这是危险的。
355Or you might bury the boxes for the first city at the very bottom of the pile.或者把盒子埋在第一座城市的底部
356Exactly.没错
357A human planner takes a very long time to solve that puzzle.一个人类的策划者需要很长的时间来解决这个谜题。
358They have to calculate weights and destinations manually.他们需要人工计算权重和目的地.
359But the AI?但人工智能吗?
360The AI solves it perfectly in seconds.人工智能在几秒钟内就能完美解决
361And that delivers immediate real gains.这能立即带来真正的收益。
362The port saves time.端口省时.
363They save expensive fuel.他们节省了昂贵的燃料。
364They save money.他们节省了钱。
365This happens today.这是今天发生的。
366They do not have to wait decades for a return on their investment.它们不必等待几十年的投资回报。
367No, they don't.不,他们没有。
368So instead of one giant AI brain that knows everything,而不是一个知道一切的巨型AI大脑
369we really need thousands of small AI brains.我们真的需要几千个小的AI大脑。
370Yes.对
371Small brains doing specific jobs, like stacking boxes or driving a delivery truck.小脑做特定的工作,如堆箱或驾驶送货卡车.
372The GI brain is way too expensive to run for a simple task.GI的大脑太贵了 无法运行一个简单的任务。
373But a small, focused brain is cheap and fast.但是,一个小的,集中的大脑是便宜的和快速的.
374And it finally closes that commercial loop we talked about earlier.它终于关闭了我们之前谈到的商业环路。
375It does.没错
376The port pays for the loading service.港口支付装卸服务费。
377The money goes out.钱都花出来了
378And the profit actually comes back to the AI company.而利润实际上回到了AI公司.
379The loop is closed.循环是关闭的。
380Yes.对
381It creates real, undeniable business value.它创造了真实的,不可否认的商业价值.
382It actually solves a specific problem.它实际上解决了一个具体问题。
383Now, you might think these tech companies would stop spending so much money on the现在,你可能会认为这些技术公司 将停止花这么多钱
384giant AI brains then.巨型AI大脑。
385You would think so, yeah.你会这么想的,是的。
386If the small brains are better for profit, they should just switch strategies completely.如果小大脑更有利于盈利,他们就应该完全切换策略.
387But there is a surprising twist at the end of the story.但故事的结尾有一个令人惊讶的曲折。
388The business world is really that simple, is it?商业世界真的那么简单,是吗?
389No, it is not.不,不是这样的。
390These massive companies are trapped in a very tough position right now.这些大型公司现在处于非常艰难的境地。
391Let's look at Baidu's own numbers.试看百度自号.
392Their profits just dropped sharply.他们的利润急剧下降。
393Very sharply.非常尖锐。
394They dropped by 46% for the September quarter.9月季度下降了46%。
395That is almost half their profit.这几乎是他们的一半利润。
396Simply gone.就这么走了
397Right.对
398Their main business is search advertising.他们的主要业务是搜索广告.
399And that revenue is sliding down.收入正在下降。
400People are changing how they search for information online.人们正在改变他们在网上搜索信息的方式。
401And AI was supposed to help them boost those profits.而AI应该帮助他们增加这些利润.
402Baidu had an early AI chatbot, right, called Erniebot?拜都有一个早期的AI聊天机,对不对,叫厄尼博特?
403Yes, Erniebot.对,厄尼博特
404Back in 2023, it was a clear leader in China.2023年回国,是中国明显的领导人.
405But now, that lead is faded.但是现在,铅已经淡出。
406The market boosts fast.市场快速推进.
407Very fast.很快的,你应该知道
408Rivals have completely overtaken them.狂欢已经完全超越了他们。
409Fast-moving startup companies like DeepSeq and Moonshot AI are beating them.DepSeq和月光AI等快速移动的启动公司正在殴打他们.
410Big established rivals like Tencent and Alibaba have also passed their models.像Tencent和Alibaba这样的大竞争对手也通过了他们的模式.
411And we should note something interesting here.我们应该注意到一些有趣的事情
412Alibaba actually owns the South China Morning Post.阿里巴巴实际拥有"南华早报".
413Right.对
414The paper that published our source article today.今日发表来源文章的论文.
415But the market data they report highlights a universal struggle.但他们所报道的市场数据突出了一场普遍的斗争.
416Baidu is not the only one feeling the pressure.贝都不是唯一感受到压力的人
417No, the whole industry is struggling with this.不,整个行业都在挣扎着
418The market is very crowded.市场非常拥挤。
419The competition is intense for Baidu.比赛对白都激烈.
420They are losing ground.他们正在失去立足点。
421And here is the big twist.这是大转弯。
422Even with falling profits.即使利润下降。
423Even with a crowded market.即使市场拥挤。
424Even with their own CEOs warning about an AI bubble.即使他们自己的CEO警告过一个AI泡沫.
425Baidu says there is an urgent shortage of computational power.拜都说计算功率紧急不足.
426Computational power.计算力.
427Let's explain that simply.让我们简单解释一下
428It basically means the sheer computer brain power needed to run AI.基本上意味着运行AI所需的完全计算机脑电.
429So AI requires thousands of special computer chips working together at high speeds.因此AI需要上千个特殊的计算机芯片高速合作.
430Right.对
431It takes massive amounts of electricity and special hardware.它需要大量的电力和特殊硬件.
432You need giant data centers to process all that information.你需要巨大的数据中心 处理所有的信息。
433Yes.对
434And to get more of this power, Baidu is building a massive new computer center next year.为了获得更多的这种能量,白杜明年正在建造一个大型的新计算机中心.
435It is a supernode called Tianqi.这是一个超级节点,叫天奇。
436A supernode is just a massive cluster of computers working together as one giant brain.一个超级节点只是一个庞大的计算机集群,作为一个巨大的大脑一起工作.
437And it will use 1,024 highly advanced computer chips.它将使用1024个高度先进的计算机芯片。
438Right.对
439And these chips are made by their own in-house company.这些芯片是由自己的内部公司制造的.
440A company called Kunlunxin.一家名为昆仑信的公司.
441So they are building their own hardware to run their own AI.所以他们正在建造自己的硬件来运行自己的AI.
442Yes.对
443They are trying to control the whole system.他们试图控制整个系统。
444Wait.单单
445Let's stop and look at this closely.我们停下来仔细看看
446Because this is completely contradictory.因为这是完全矛盾的。
447It really is.没错
448Let's unpack it.我们把它拆开
449They say AI might be a bubble.他们说AI可能是一个泡沫。
450Their profits are falling heavily.他们的利润正在严重下降。
451Their old AI models are losing the race.他们的老AI模型正在输掉比赛.
452Yeah.对
453But they are still building massive, incredibly expensive new computer centers.但他们仍在建造庞大的,令人难以置信昂贵的新计算机中心。
454Yes, they are.是的,他们是。
455I mean, why would they do that?我的意思是,他们为什么要这样做?
456It makes no sense.这根本说不通
457Well, from their perspective, they have absolutely no choice.从他们的角度来看,他们绝对别无选择。
458No choice?没有选择?
459Why not?为什么不呢?
460It is a forced race.这是一场被迫的比赛。
461It is like a trap.这就像一个陷阱。
462You mean if they stop building, they will lose the race entirely?你是说如果他们停止建设 他们就会完全输掉比赛?
463Exactly.没错
464If they stop spending billions on the giant models, the other companies will pull ahead.如果他们停止在巨型模型上花费数十亿美元,其他公司就会拉上前台.
465And the technology is moving too fast.技术进展太快了。
466Right.对
467If Baidu pauses for even one year, they will never catch up.如果拜都暂停了一年 他们就永远追不上了
468So they are completely trapped by the competition.因此他们完全被竞争者所困.
469They are.他们是。
470They must keep spending huge amounts of money just to stay in the game.他们必须继续花大钱 留在游戏中。
471They hope their core business can survive long enough.他们希望他们的核心业务能够持续足够长的时间.
472Right.对
473They are desperately waiting for that application layer to arrive.他们急于等待申请层的到来。
474They need those specific, useful tools to finally start making real money.他们需要这些具体而有用的工具,以便最终开始赚真正的钱。
475Exactly.没错
476They are building the incredibly expensive foundation right now.他们正在建造一个非常昂贵的基础。
477And they hope the house will pay for itself later.他们希望房子以后能自讨苦吃
478Yes.对
479They warn the entire industry about a dangerous bubble.他们警告整个行业 一个危险的泡沫。
480But they must keep blowing air into that exact same bubble.但他们必须不断把空气吹进同样的气泡里
481Wow.珃
482Because if the bubble is real, they must own a piece of it.因为如果气泡是真实的, 他们必须拥有它的一部分。
483Right.对
484If it becomes the future of the global economy,如果它成为全球经济的未来,
485they simply cannot afford to be left behind.他们根本不能被抛在后面。
486It is a massive financial gamble.这是一个巨大的金融赌博。
487A huge gamble.一场巨大的赌博。
488It is like a high-safe game of poker,这就像一个高安全性的扑克游戏,
489and everyone is pushing all their chips into the middle of the table.大家把薯片都推到桌子中间
490Yeah.对
491That is a great way to put it.说来话长
492Well, we've covered a lot of important ground today.嗯,我们已经覆盖了很多重要的地方今天。
493Let's bring all these ideas together for you, our listener.让我们把这些想法集中起来吧 我们的听众
494The AI industry is going through a painful change right now.AI产业正在经历痛苦的变化
495Yes.对
496It is moving away from the excited hype stage.它正在远离激动的狂欢舞台。
497And it is entering the hard work stage.它正在进入艰苦的工作阶段。
498The era of just being amazed by a chatting computer is ending.被聊天电脑惊奇的时代即将结束。
499Businesses do not want magic tricks anymore.企业不再想要魔术
500Right.对
501The real winners in the future will not just build the biggest, smartest models.未来真正的赢家们将不仅建立最大,最聪明的模型.
502No.没有
503The winners will build useful tools for very specific jobs.获奖者将为非常具体的工作建立有用的工具。
504They will build the small brains that stack shipping boxes efficiently.他们将建立能高效堆放运货箱的小大脑。
505They will build the software that smoothly runs a robot on a factory floor.他们会在工厂地板上 建造一个能顺利运行机器人的软件
506Exactly.没错
507They will finally close that open commercial loop.他们最终会关闭开放的商业环路。
508They will turn cool, expensive technology into real, measurable business value.它们将把酷酷、昂贵的技术转化为真正、可衡量的商业价值。
509They will actually solve real problems.他们会解决真正的问题。
510And this matters directly to you.这直接关系到你
511You will see these changes in your own life and work very soon.你会看到这些变化 在自己的生活和工作很快。
512Yes.对
513You absolutely will.你会的
514So the next time you see a new AI tool on the market or in your workplace,所以,下一次你看到一个新的AI工具 在市场上或你的工作场所,
515ask yourself a simple question.问自己一个简单的问题。
516Does this tool just look cool?这个工具看起来很酷吗?
517Or does it actually solve a real problem for me?还是它真的解决了我真正的问题?
518Is it the giant, incredibly expensive race car?是那辆巨大的 昂贵的赛车吗?
519Or is it a simple, effective trip to the store?还是只是去商店的一次简单有效的旅行?
520Focus on the tools that actually do the work.专注于实际工作的工具.
521Exactly.没错
522Now, I want to leave you with a final thought to Molliver.现在,我想留给你一个最后的想法 莫利弗。
523Something to think about after we finish.等我们做完后再想想
524Okay.摆
525Let's hear it.颂曰.
526We talked a lot about the discovery of electricity today.我们今天谈了很多关于电力的发现。
527We did.我们做到了。
528We talked about how old factories had to completely change their organizational structure.我们谈到了老工厂如何彻底改变组织结构。
529Right.对
530They could not just plug the new power into the old system.他们不能把新动力塞进旧系统
531They had to tear down the old belts.他们不得不拆掉旧皮带.
532They had to move all of the machines around.他们必须把所有机器都搬走
533They had to rebuild themselves to use the new electric power properly.他们不得不重建自己,以正确使用新的电力。
534And it was painful and slow.痛苦而缓慢。
535Very slow.非常缓慢。
536Many of those old companies failed.很多老公司都失败了
537So think about the modern business world today.所以想想今天的现代商业世界.
538Okay.摆
539If modern companies must do the exact same thing to use AI properly,如果现代公司必须做同样的事情才能正确使用AI,
540who will win the future?谁能赢得未来?
541That is a great question.问得好
542Will it be the old, massive companies trying desperately to adapt their old ways?会不会是那些老的,大规模的公司 拼命地尝试着调整自己的老方式呢?
543Or will it be brand new companies,或者是全新的公司,
544companies built entirely from scratch today with AI at their core?公司今天完全从零开始建设 以AI为核心?
545Right.对
546New companies will not have any old belts to tear down.新公司将没有任何旧的皮带可拆.
547They start fresh.他们开始新鲜。
548That is the real question for the next decade of business.这是下一个十年生意的真正问题。
549Yes, it is.没错
550Keep a close eye on those startups.盯紧那些创业者
551Thank you for joining us on today's Deep Dive.感谢您参加今天的深潜
552Thanks for listening.谢谢你的倾听。
553We will explore more fascinating topics with you next time.下次我们会和你们一起探讨更迷人的话题
该视频共有字幕 555 条。解锁更多字幕为会员功能,请移动到 价格