2025-07-21 Eric Schmidt  at Moonshot podcast

影片字幕

00:00:03 When do you see what you define as
00:00:03 你什麼時候看到你定義的

00:00:03 When do you see what you define as digital super intelligence?
00:00:03 何時會看到您定義的數位超級智慧？

00:00:04 digital super intelligence?
00:00:04 數位超級智慧？

00:00:04 digital super intelligence? Uh, within 10 years.
00:00:04 數位超級智慧？呃，10年內。

00:00:06 Uh, within 10 years.
00:00:06 呃，10年內。

00:00:06 Uh, within 10 years. The AI's ability to generate its own
00:00:06 呃，10年內。人工智慧能夠產生自己的

00:00:09 The AI's ability to generate its own
00:00:09 人工智慧能夠產生自己的

00:00:09 The AI's ability to generate its own scaffolding is imminent. Pretty much
00:00:09 人工智慧即將擁有自己建造鷹架的能力。

00:00:12 scaffolding is imminent. Pretty much
00:00:12 腳手架即將搭建。差不多

00:00:12 scaffolding is imminent. Pretty much sure that that will be a 2025 thing. We
00:00:12 腳手架即將搭建。幾乎可以肯定，這將是2025年的事。我們

00:00:15 sure that that will be a 2025 thing. We
00:00:15 肯定是2025年的事。我們

00:00:15 sure that that will be a 2025 thing. We certainly don't know what super
00:00:15 肯定是2025年的事。我們當然不知道超級

00:00:16 certainly don't know what super
00:00:16 當然不知道什麼超級

00:00:16 certainly don't know what super intelligence will deliver, but we know
00:00:16 我們當然不知道超級智慧會帶來什麼，但我們知道

00:00:18 intelligence will deliver, but we know
00:00:18 情報會傳遞訊息，但我們知道

00:00:18 intelligence will deliver, but we know it's coming.
00:00:18 情報將會公佈，但我們知道它即將到來。

00:00:19 it's coming.  00:00:19 它來了。

00:00:19 it's coming. And what do people need to know about
00:00:19 它來了。人們需要知道什麼

00:00:21 And what do people need to know about
00:00:21 人們需要了解什麼

00:00:21 And what do people need to know about that?
00:00:21 人們需要了解什麼呢？

00:00:21 that?  00:00:21 那？

00:00:22 that? You're going to have your own polymath.
00:00:22 那個？你將會擁有自己的博學者。

00:00:24 You're going to have your own polymath.
00:00:24 你將會擁有自己的博學者。

00:00:24 You're going to have your own polymath. So, you're going to have the sum of
00:00:24 你將擁有自己的博學之人。所以，你將擁有

00:00:25 So, you're going to have the sum of
00:00:25 所以，你會得到

00:00:25 So, you're going to have the sum of Einstein and Leonardo da Vinci in the
00:00:25 所以，你將得到愛因斯坦和列奧納多達文西的總和

00:00:28 Einstein and Leonardo da Vinci in the
00:00:28 愛因斯坦和李奧納多達文西在

00:00:28 Einstein and Leonardo da Vinci in the equivalent of your pocket. agents are
00:00:28 愛因斯坦和達文西在你的口袋裡。特工是

00:00:30 equivalent of your pocket. agents are
00:00:30 相當於你的口袋。代理是

00:00:30 equivalent of your pocket. agents are going to happen. This math thing is
00:00:30 相當於你的口袋。代理將會發生。這個數學問題是

00:00:31 going to happen. This math thing is
00:00:31 會發生。這個數學問題

00:00:32 going to happen. This math thing is going to happen. The software thing is
00:00:32 會發生。數學的事情會發生。軟體的事情

00:00:33 going to happen. The software thing is
00:00:33 會發生。軟體的事情

00:00:33 going to happen. The software thing is going to happen. Everything I've talked
00:00:33 會發生。軟體的事情會發生。我所說的一切

00:00:34 going to happen. Everything I've talked
00:00:34 會發生的。我所說的一切

00:00:34 going to happen. Everything I've talked about is in the positive domain, but
00:00:34 會發生。我談論的都是正面的方面，但是

00:00:36 about is in the positive domain, but
00:00:36 大約在正域中，但是

00:00:36 about is in the positive domain, but there's a negative domain as well. It's
00:00:36 是關於正域的，但也有一個負域。它是

00:00:38 there's a negative domain as well. It's
00:00:38 還有一個負域。它是

00:00:38 there's a negative domain as well. It's likely, in my opinion, that you're going
00:00:38 還有一個負面領域。在我看來，你很可能會

00:00:40 likely, in my opinion, that you're going
00:00:40 在我看來，你很可能會

00:00:40 likely, in my opinion, that you're going to see.
00:00:40 在我看來，你很有可能會看到這一點。

00:00:48 Now, that's a moonshot, ladies and
00:00:48 女士們，這真是個偉大的夢想

00:00:48 Now, that's a moonshot, ladies and gentlemen.
00:00:48 女士、先生們，這是一個偉大的夢想。

00:00:53 Hey, everybody. Welcome to Moonshots.
00:00:53 大家好，歡迎收看《Moonshots》。

00:00:53 Hey, everybody. Welcome to Moonshots. I'm here live with my Moonshot mate,
00:00:53 大家好！歡迎收看《Moonshots》。我和我的《Moonshot》夥伴一起直播。

00:00:54 I'm here live with my Moonshot mate,
00:00:54 我和我的 Moonshot 夥伴在這裡直播，

00:00:54 I'm here live with my Moonshot mate, Dave London. Uh we're here in our Santa
00:00:54 我和我的 Moonshot 夥伴 Dave London 一起在這裡直播。呃，我們在聖誕老人

00:00:57 Dave London. Uh we're here in our Santa
00:00:57 戴夫倫敦。呃，我們在聖誕老公公這裡

00:00:57 Dave London. Uh we're here in our Santa Monica studios and we have a special
00:00:57 戴夫倫敦。呃，我們在聖莫尼卡工作室，我們有一個特別的

00:00:58 Monica studios and we have a special
00:00:58 莫妮卡工作室，我們有一個特別的

00:00:58 Monica studios and we have a special guest today,
00:00:58 莫妮卡工作室，今天我們有一位特別嘉賓，

00:01:00 guest today,  00:01:00 今天的客人，

00:01:00 guest today, Eric Schmidt, the author of Genesis. We
00:01:00 今天的嘉賓是《創世紀》的作者艾瑞克·施密特。我們

00:01:03 Eric Schmidt, the author of Genesis. We
00:01:03 艾瑞克‧施密特，《創世紀》的作者。我們

00:01:03 Eric Schmidt, the author of Genesis. We talk about China. We're going to talk
00:01:03 艾瑞克‧施密特，《創世紀》的作者。我們討論中國。我們將討論

00:01:05 talk about China. We're going to talk
00:01:05 談中國。我們將討論

00:01:05 talk about China. We're going to talk about, you know, digital super
00:01:05 談中國。我們要討論的是數位超級

00:01:06 about, you know, digital super
00:01:06 關於數字超級

00:01:06 about, you know, digital super intelligence. We'll talk about, you
00:01:06 關於數位超級智慧。我們會討論你

00:01:08 intelligence. We'll talk about, you
00:01:08 情報。我們會討論，你

00:01:08 intelligence. We'll talk about, you know, what people should be thinking
00:01:08 智力。我們會討論人們應該思考什麼

00:01:10 know, what people should be thinking
00:01:10 知道人們應該想什麼

00:01:10 know, what people should be thinking about over the 10 years.
00:01:10 知道，人們在未來 10 年內應該思考什麼。

00:01:11 about over the 10 years.
00:01:11 大約過去10年。

00:01:11 about over the 10 years. And we're talking about the guy who has
00:01:11 大約十年前。我們談論的是

00:01:14 And we're talking about the guy who has
00:01:14 我們正在談論的這個人

00:01:14 And we're talking about the guy who has more access to more more actionable
00:01:14 我們討論的是那些擁有更多機會獲取更多可操作資訊的人

00:01:16 more access to more more actionable
00:01:16 更多存取權限，更多可操作性

00:01:16 more access to more more actionable information than probably anyone else
00:01:16 比其他任何人都能獲得更多可操作的信息

00:01:17 information than probably anyone else
00:01:17 比其他任何人都了解

00:01:17 information than probably anyone else you could think of. So, it should be
00:01:17 比你能想到的任何人都更了解資訊。所以，它應該是

00:01:20 you could think of. So, it should be
00:01:20 你能想到的。所以，應該是

00:01:20 you could think of. So, it should be should be pretty exciting.
00:01:20 你能想到的。所以，這應該會很令人興奮。

00:01:21 should be pretty exciting.
00:01:21 應該會非常令人興奮。

00:01:21 should be pretty exciting. Incredibly brilliant. All right, stand
00:01:21 應該會很精彩。太棒了。好的，站起來

00:01:23 Incredibly brilliant. All right, stand
00:01:23 太精彩了。好，站起來

00:01:23 Incredibly brilliant. All right, stand by for a conversation with the Eric
00:01:23 太棒了。好的，請稍等，聽聽 Eric 的對話。

00:01:25 by for a conversation with the Eric
00:01:25 與 Eric 的對話

00:01:25 by for a conversation with the Eric Schmidt, CEO or past CEO of Google and
00:01:25 與 Google CEO 或前 CEO Eric Schmidt 的對話

00:01:28 Schmidt, CEO or past CEO of Google and
00:01:28 施密特，Google 執行長或前任首席執行官

00:01:28 Schmidt, CEO or past CEO of Google and an extraordinary investor and uh and
00:01:28 施密特，Google CEO 或前 CEO，也是一位傑出的投資者，呃，

00:01:30 an extraordinary investor and uh and
00:01:30 一位非凡的投資者，呃，

00:01:30 an extraordinary investor and uh and thinker in this field of AI.
00:01:30 一位傑出的投資者和人工智慧領域的思想家。

00:01:32 thinker in this field of AI.
00:01:32 人工智慧領域的思想家。

00:01:32 thinker in this field of AI. Let's do it.
00:01:32 人工智慧領域的思想家。我們一起行動吧。

00:01:33 Let's do it.  00:01:33 我們開始吧。

00:01:33 Let's do it. Eric, welcome back to Moonshots.
00:01:33 開始吧！ Eric，歡迎回到《Moonshots》節目。

00:01:35 Eric, welcome back to Moonshots.
00:01:35 Eric，歡迎回到 Moonshots。

00:01:35 Eric, welcome back to Moonshots. It's great to be here with you guys.
00:01:35 Eric，歡迎回到 Moonshots。很高興和大家在一起。

00:01:36 It's great to be here with you guys.
00:01:36 很高興和你們在一起。

00:01:36 It's great to be here with you guys. Thank you. It's been uh it's been a long
00:01:36 很高興和你們在一起。謝謝。好久不見了

00:01:39 Thank you. It's been uh it's been a long
00:01:39 謝謝。好久不見了

00:01:39 Thank you. It's been uh it's been a long road since I first met you at Google. I
00:01:39 謝謝。自從我在谷歌第一次見到你以來，走過了漫長的道路。我

00:01:42 road since I first met you at Google. I
00:01:42 自從我在谷歌第一次見到你以來，我

00:01:42 road since I first met you at Google. I remember uh our first conversations were
00:01:42 自從我在谷歌第一次見到你以來，我記得我們第一次談話是

00:01:44 remember uh our first conversations were
00:01:44 記得我們第一次談話是

00:01:44 remember uh our first conversations were fantastic. Uh it's been a crazy month in
00:01:44 記得我們第一次談話非常愉快。嗯，這是一個瘋狂的月份

00:01:48 fantastic. Uh it's been a crazy month in
00:01:48 太棒了。呃，這是一個瘋狂的月份

00:01:48 fantastic. Uh it's been a crazy month in the world of AI, but I think every month
00:01:48 太棒了。嗯，這是人工智慧世界瘋狂的一個月，但我認為每個月

00:01:50 the world of AI, but I think every month
00:01:50 人工智慧的世界，但我認為每個月

00:01:50 the world of AI, but I think every month from here is going to be a crazy month.
00:01:50 人工智慧的世界，但我認為從現在開始每個月都會是一個瘋狂的月份。

00:01:52 from here is going to be a crazy month.
00:01:52 從現在開始將會是一個瘋狂的月份。

00:01:52 from here is going to be a crazy month. And so I'd love to hit on a number of
00:01:52 從現在開始，這將是一個瘋狂的月份。所以我想談一些

00:01:55 And so I'd love to hit on a number of
00:01:55 所以我想談談

00:01:55 And so I'd love to hit on a number of subjects and get your your take on them.
00:01:55 因此，我很樂意談論一些話題並聽聽你的看法。

00:01:57 subjects and get your your take on them.
00:01:57 主題並了解你對它們的看法。

00:01:57 subjects and get your your take on them. I want to start with probably the most
00:01:57 了解你對這些主題的看法。我想先從最

00:01:59 I want to start with probably the most
00:01:59 我想從最

00:01:59 I want to start with probably the most important point that you've made
00:01:59 我想從你提出的最重要的一點開始

00:02:00 important point that you've made
00:02:00 你提出的重要觀點

00:02:00 important point that you've made recently that got a lot of traction, a
00:02:00 你最近提出的一個重要觀點得到了廣泛的關注，

00:02:02 recently that got a lot of traction, a
00:02:02 最近，這個主題引起了很大的關注，

00:02:02 recently that got a lot of traction, a lot of attention, which is that AI is
00:02:02 最近，人工智慧引起了廣泛關注，

00:02:05 lot of attention, which is that AI is
00:02:05 很多關注，那就是人工智慧

00:02:05 lot of attention, which is that AI is underhyped when the rest of the world is
00:02:05 許多關注，這是因為當世界其他地方都在關注人工智慧時，人工智慧卻沒有得到充分重視。

00:02:07 underhyped when the rest of the world is
00:02:07 當世界其他地方都在

00:02:07 underhyped when the rest of the world is either confused, lost, or think it's,
00:02:07 當世界其他地方感到困惑、迷茫或認為它…時，它卻沒有受到重視。

00:02:09 either confused, lost, or think it's,
00:02:09 要嘛困惑，要嘛迷茫，要嘛認為，

00:02:10 either confused, lost, or think it's, you know, not impacting us.
00:02:10 要麼感到困惑，要麼感到迷茫，要麼認為它不會影響到我們。

00:02:13 you know, not impacting us.
00:02:13 你知道，這不會影響我們。

00:02:13 you know, not impacting us. We'll get into in more detail, but quick
00:02:13 你知道，不會影響我們。我們會更詳細地講，但很快

00:02:16 We'll get into in more detail, but quick
00:02:16 我們會更詳細地講解，但很快

00:02:16 We'll get into in more detail, but quick most important point to make there.
00:02:16 我們將進行更詳細的介紹，但最重要的一點是要在這裡提出。

00:02:19 most important point to make there.
00:02:19 這是最重要的一點。

00:02:19 most important point to make there. AI is a learning machine. Yeah.
00:02:19 最重要的一點是，人工智慧是一台學習機器。是的。

00:02:21 AI is a learning machine. Yeah.
00:02:21 人工智慧是一種學習機器。是的。

00:02:21 AI is a learning machine. Yeah. And in network effect businesses, when
00:02:21 人工智慧是一台學習機器。是的。在網路效應企業中，當

00:02:24 And in network effect businesses, when
00:02:24 在網路效應企業中，當

00:02:24 And in network effect businesses, when the learning machine learns faster,
00:02:24 在網路效應企業中，當機器學習得更快時，

00:02:26 the learning machine learns faster,
00:02:26 學習機器學得更快，

00:02:26 the learning machine learns faster, everything accelerates.
00:02:26 學習機器學習得更快，一切都加速了。

00:02:28 everything accelerates.
00:02:28 一切都在加速。

00:02:28 everything accelerates. It accelerates to its natural limit. The
00:02:28 一切都加速了。它加速到了它的自然極限。

00:02:31 It accelerates to its natural limit. The
00:02:31 它加速到自然極限。

00:02:31 It accelerates to its natural limit. The natural limit is electricity.
00:02:31 它加速到它的自然極限。這個自然極限就是電力。

00:02:35 natural limit is electricity.
00:02:35 自然極限是電力。

00:02:35 natural limit is electricity. Not chips,
00:02:35 自然極限是電力。不是晶片，

00:02:36 Not chips,  00:02:36 不是洋芋片，

00:02:36 Not chips, electricity really. Okay.
00:02:36 不是晶片，是電。好的。

00:02:39 electricity really. Okay.
00:02:39 確實是電。好的。

00:02:39 electricity really. Okay. So that gets me to the next point here,
00:02:39 真的有電。好的。接下來我要講的是，

00:02:41 So that gets me to the next point here,
00:02:41 這引出了我的下一個觀點，

00:02:41 So that gets me to the next point here, which is uh a discussion on AI and
00:02:41 這引出了我的下一個主題，關於人工智慧和

00:02:44 which is uh a discussion on AI and
00:02:44 這是關於人工智慧的討論

00:02:44 which is uh a discussion on AI and energy. So, we saw recently was Meta
00:02:44 這是關於人工智慧和能源的討論。我們最近看到的是 Meta

00:02:47 energy. So, we saw recently was Meta
00:02:47 能量。我們最近看到的是 Meta

00:02:47 energy. So, we saw recently was Meta recently announcing uh that they signed
00:02:47 能量。我們最近看到 Meta 宣布他們簽下了

00:02:50 recently announcing uh that they signed
00:02:50 最近宣布他們簽署了

00:02:50 recently announcing uh that they signed a 20-year nuclear contract with uh with
00:02:50 最近宣布，他們與呃簽署了一份為期 20 年的核子合約

00:02:53 a 20-year nuclear contract with uh with
00:02:53 與呃簽訂了20年的核子合約

00:02:53 a 20-year nuclear contract with uh with Constellation Energy. We've seen Google,
00:02:53 與 Constellation Energy 簽訂了為期 20 年的核子合約。我們已經看到谷歌，

00:02:56 Constellation Energy. We've seen Google,
00:02:56 星座能源。我們已經看到谷歌，

00:02:56 Constellation Energy. We've seen Google, Microsoft, Amazon, everybody buying
00:02:56 星座能源。我們看到谷歌、微軟、亞馬遜等都在收購

00:02:59 Microsoft, Amazon, everybody buying
00:02:59 微軟、亞馬遜等都在購買

00:02:59 Microsoft, Amazon, everybody buying basically nuclear capacity right now.
00:02:59 微軟、亞馬遜，現在每個人都在買核能。

00:03:02 basically nuclear capacity right now.
00:03:02 目前基本上是核能。

00:03:02 basically nuclear capacity right now. That's got to be weird
00:03:02 基本上是現在的核能容量。這肯定很奇怪

00:03:05 That's got to be weird
00:03:05 這一定很奇怪

00:03:05 That's got to be weird uh that private companies are are
00:03:05 這一定很奇怪，私人公司

00:03:08 uh that private companies are are
00:03:08 呃，私人公司是

00:03:08 uh that private companies are are basically taking over into their own
00:03:08 嗯，私人公司基本上接管了

00:03:10 basically taking over into their own
00:03:10 基本上接管了他們自己的

00:03:10 basically taking over into their own hands what was utility function before.
00:03:10 基本上將先前的效用功能掌握在自己手中。

00:03:14 hands what was utility function before.
00:03:14 之前的效用函數是什麼。

00:03:14 hands what was utility function before. Um,
00:03:14 之前的效用函數是什麼？嗯，

00:03:14 Um,  00:03:14 嗯，

00:03:14 Um, well, just to be cynical, I I'm so glad
00:03:14 嗯，好吧，只是憤世嫉俗地說，我很高興

00:03:17 well, just to be cynical, I I'm so glad
00:03:17 好吧，只是憤世嫉俗地說，我很高興

00:03:17 well, just to be cynical, I I'm so glad those companies plan to be around the 20
00:03:17 好吧，只是憤世嫉俗地說，我很高興這些公司計劃在 20

00:03:19 those companies plan to be around the 20
00:03:19 這些公司計劃在 20

00:03:19 those companies plan to be around the 20 years that it's going to take to get the
00:03:19 這些公司計劃在 20 年左右的時間內實現

00:03:21 years that it's going to take to get the
00:03:21 需要花上幾年的時間才能實現

00:03:21 years that it's going to take to get the nuclear power plants built.
00:03:21 建造核電廠需要花費數年時間。

00:03:23 nuclear power plants built.
00:03:23 座核電廠建成。

00:03:23 nuclear power plants built. In my recent testimony, I talked about
00:03:23 核電廠建成。在我最近的證詞中，我談到了

00:03:26 In my recent testimony, I talked about
00:03:26 在我最近的證詞中，我談到

00:03:26 In my recent testimony, I talked about the the current expected need for the AI
00:03:26 在我最近的證詞中，我談到了目前對人工智慧的預期需求

00:03:28 the the current expected need for the AI
00:03:28 目前對人工智慧的預期需求

00:03:28 the the current expected need for the AI revolution in the United States is 92
00:03:28 目前美國對人工智慧革命的預期需求是 92

00:03:31 revolution in the United States is 92
00:03:31 美國革命是92

00:03:31 revolution in the United States is 92 gawatt of more power.
00:03:31 美國的革命是92千兆瓦以上的電力。

00:03:33 gawatt of more power.
00:03:33 兆瓦的更多功率。

00:03:33 gawatt of more power. For reference, one gawatt is one big
00:03:33 兆瓦的更多功率。作為參考，一兆瓦是

00:03:37 For reference, one gawatt is one big
00:03:37 參考一下，一兆瓦是一大

00:03:37 For reference, one gawatt is one big nuclear power station. And there are
00:03:37 參考一下，一千兆瓦相當於一座大型核電廠。

00:03:39 nuclear power station. And there are
00:03:39 核電廠。還有

00:03:39 nuclear power station. And there are none essentially being started now.
00:03:39 核電廠。目前基本上沒有啟動任何核電廠。

00:03:41 none essentially being started now.
00:03:41 目前基本上還沒開始。

00:03:41 none essentially being started now. And there have been two in the last
00:03:41 目前基本上沒有啟動。過去已經有兩個

00:03:42 And there have been two in the last
00:03:42 過去兩年

00:03:42 And there have been two in the last what, 30 years built. There is
00:03:42 過去30年已經建造了兩座。

00:03:44 what, 30 years built. There is
00:03:44 什麼？ 30年建成。

00:03:44 what, 30 years built. There is excitement that there's an SMR, small
00:03:44 什麼？建了 30 年。 SMR，小型的，令人興奮。

00:03:46 excitement that there's an SMR, small
00:03:46 很高興有 SMR，小

00:03:46 excitement that there's an SMR, small modular reactor coming in at 300
00:03:46 令人興奮，SMR，小型模組化反應器即將問世，300

00:03:48 modular reactor coming in at 300
00:03:48 模組化反應器即將投入使用

00:03:48 modular reactor coming in at 300 megawws, but it won't start till 2030.
00:03:48 模組化反應器的發電量為 300 兆瓦，但要到 2030 年才能啟動。

00:03:51 megawws, but it won't start till 2030.
00:03:51 megawws，但要到 2030 年才會開始。

00:03:51 megawws, but it won't start till 2030. As important as nuclear, both fision and
00:03:51 兆瓦，但要到 2030 年才會開始。與核能一樣重要，核分裂和

00:03:54 As important as nuclear, both fision and
00:03:54 與核能一樣重要，裂變和

00:03:54 As important as nuclear, both fision and fusion is, they're not going to arrive
00:03:54 儘管核能、裂變和聚變都很重要，但它們不會到來

00:03:56 fusion is, they're not going to arrive
00:03:56 融合是，他們不會到達

00:03:56 fusion is, they're not going to arrive in time to get us what we need as a
00:03:56 融合是，他們不會及時到達，提供我們所需的

00:03:59 in time to get us what we need as a
00:03:59 及時提供我們所需的東西

00:04:00 in time to get us what we need as a globe to deal with our many problems and
00:04:00 及時為我們帶來地球所需的一切，以解決我們面臨的許多問題，

00:04:01 globe to deal with our many problems and
00:04:01 地球來處理我們許多的問題，

00:04:01 globe to deal with our many problems and the many opportunities that are before
00:04:01 地球來應對我們面臨的許多問題和許多機會

00:04:03 the many opportunities that are before
00:04:03 眼前的眾多機會

00:04:03 the many opportunities that are before us. Do you think uh so if if you look at
00:04:03 我們面前有很多機會。你覺得，呃，如果你看看

00:04:05 us. Do you think uh so if if you look at
00:04:05 我們。你覺得呃，如果你看看

00:04:05 us. Do you think uh so if if you look at the sort of three-year timeline toward
00:04:05 我們。你覺得，呃，如果看三年的時間表，

00:04:07 the sort of three-year timeline toward
00:04:07 三年時間表

00:04:07 the sort of three-year timeline toward AGI, do you think if you started a a
00:04:07 通用人工智慧的三年時間表，你認為如果你開始

00:04:10 AGI, do you think if you started a a
00:04:10 AGI，你認為如果你開始 AA

00:04:10 AGI, do you think if you started a a fusion reactor project today that won't
00:04:10 AGI，你認為如果你今天啟動一個核融合反應器項目，那不會

00:04:12 fusion reactor project today that won't
00:04:12 今天的聚變反應器專案不會

00:04:12 fusion reactor project today that won't come online for five, six, seven years,
00:04:12 目前的聚變反應器工程五、六七年內不會上線，

00:04:15 come online for five, six, seven years,
00:04:15 上線五、六、七年了，

00:04:15 come online for five, six, seven years, is there a probability that the AGI
00:04:15 上線五、六、七年後，AGI 是否有可能

00:04:17 is there a probability that the AGI
00:04:17 AGI 的機率

00:04:17 is there a probability that the AGI comes up with some other breakthrough
00:04:17 AGI 是否有可能取得其他突破

00:04:19 comes up with some other breakthrough
00:04:19 又有了一些突破

00:04:19 comes up with some other breakthrough fusion or otherwise that makes it
00:04:19 提出了一些其他突破性的融合，或其他什麼使得它

00:04:20 fusion or otherwise that makes it
00:04:20 融合，或其他方式

00:04:20 fusion or otherwise that makes it irrelevant before it even gets online?
00:04:20 融合還是其他原因使得它在上線之前就變得無關緊要？

00:04:22 irrelevant before it even gets online?
00:04:22 上線前就無關緊要？

00:04:22 irrelevant before it even gets online? A very good question. We don't know what
00:04:22 甚至在它上線之前就無關緊要了？這個問題問得好。我們不知道

00:04:24 A very good question. We don't know what
00:04:24 這個問題問得好。我們不知道

00:04:24 A very good question. We don't know what artificial general intelligence will
00:04:24 這個問題問得很好。我們不知道通用人工智慧會做什麼。

00:04:27 artificial general intelligence will
00:04:27 通用人工智慧將

00:04:27 artificial general intelligence will deliver. Yeah. And we certainly don't
00:04:27 通用人工智慧將會實現。是的。而我們當然不會

00:04:29 deliver. Yeah. And we certainly don't
00:04:29 交付。是的。我們當然不會

00:04:29 deliver. Yeah. And we certainly don't know what super intelligence will
00:04:29 交付。是的。我們當然不知道超級智能會做什麼

00:04:31 know what super intelligence will
00:04:31 知道超級智慧會做什麼

00:04:31 know what super intelligence will deliver, but we know it's coming.
00:04:31 不知道超級智慧會帶來什麼，但我們知道它即將到來。

00:04:34 deliver, but we know it's coming.
00:04:34 交付，但我們知道它即將到來。

00:04:34 deliver, but we know it's coming. So, first we need to plan for it. And
00:04:34 交付，但我們知道它即將到來。所以，首先我們需要為此做好計劃。

00:04:36 So, first we need to plan for it. And
00:04:36 所以，首先我們需要做好計劃。

00:04:36 So, first we need to plan for it. And there's lots of issues as well as
00:04:36 所以，首先我們要先做好規劃。還有很多問題，

00:04:38 there's lots of issues as well as
00:04:38 還有很多問題

00:04:38 there's lots of issues as well as opportunities for that. But the fact of
00:04:38 這其中有許多問題，也有很多機會。但事實上

00:04:39 opportunities for that. But the fact of
00:04:39 機會。但事實上

00:04:40 opportunities for that. But the fact of the matter is that the computing needs
00:04:40 機會。但事實上，計算需要

00:04:42 the matter is that the computing needs
00:04:42 問題是計算需要

00:04:42 the matter is that the computing needs that we name now are going to come from
00:04:42 問題是，我們現在所說的運算需求將來自

00:04:44 that we name now are going to come from
00:04:44 我們現在命名的將來自

00:04:44 that we name now are going to come from traditional energy suppliers in places
00:04:44 我們現在提到的這些能源將來自一些地方的傳統能源供應商

00:04:47 traditional energy suppliers in places
00:04:47 各地傳統能源供應商

00:04:47 traditional energy suppliers in places like the United States and the Arab
00:04:47 美國和阿拉伯國家等地的傳統能源供應商

00:04:49 like the United States and the Arab
00:04:49 就像美國和阿拉伯

00:04:49 like the United States and the Arab world and Canada and the Western world.
00:04:49 例如美國、阿拉伯世界、加拿大和西方世界。

00:04:51 world and Canada and the Western world.
00:04:51 世界和加拿大以及西方世界。

00:04:51 world and Canada and the Western world. And it's important to note that China
00:04:51 世界、加拿大和西方世界。值得注意的是，中國

00:04:54 And it's important to note that China
00:04:54 值得注意的是，中國

00:04:54 And it's important to note that China has lots of electricity. So if they get
00:04:54 值得注意的是，中國電力充足。所以如果他們

00:04:57 has lots of electricity. So if they get
00:04:57 有很多電。所以如果他們得到

00:04:57 has lots of electricity. So if they get the chips, it's going to be one heck of
00:04:57 有很多電。所以如果他們得到晶片，那將是一次巨大的

00:04:59 the chips, it's going to be one heck of
00:04:59 晶片，這將是一個

00:04:59 the chips, it's going to be one heck of a race.
00:04:59 晶片，這將是一場激烈的比賽。

00:04:59 a race.  00:04:59 一場比賽。

00:04:59 a race. Yeah. They've been scaling it uh you
00:04:59 比賽。是的。他們一直在擴大規模，呃，你

00:05:02 Yeah. They've been scaling it uh you
00:05:02 是的。他們一直在擴大規模，呃，你

00:05:02 Yeah. They've been scaling it uh you know at two or three times. The US has
00:05:02 是的。他們已經擴大了兩到三倍。美國

00:05:04 know at two or three times. The US has
00:05:04 知道兩三次。美國已經

00:05:04 know at two or three times. The US has been flat for how long in terms of
00:05:04 知道兩三次。美國已經平穩多久了

00:05:05 been flat for how long in terms of
00:05:05 持平多久了

00:05:06 been flat for how long in terms of energy production?
00:05:06 就能源生產而言，持平了多久？

00:05:06 energy production?
00:05:06 能源生產？

00:05:06 energy production? Um from my perspective uh infinite. In
00:05:06 能源生產？嗯，從我的角度來看，是無限的。在

00:05:09 Um from my perspective uh infinite. In
00:05:09 嗯，從我的角度來看，無限。在

00:05:09 Um from my perspective uh infinite. In fact,
00:05:09 嗯，從我的角度來看，無限。事實上，

00:05:10 fact,  00:05:10 事實，

00:05:10 fact, electricity demand declined for a while
00:05:10 事實上，電力需求一度下降

00:05:12 electricity demand declined for a while
00:05:12 電力需求一度下降

00:05:12 electricity demand declined for a while as has overall energy needs because of
00:05:12 電力需求暫時下降，整體能源需求也下降，原因是

00:05:14 as has overall energy needs because of
00:05:14 整體能源需求也因

00:05:14 as has overall energy needs because of conservation and other things.
00:05:14 由於節約和其他原因，整體能源需求也是如此。

00:05:16 conservation and other things.
00:05:16 保護和其他事。

00:05:16 conservation and other things. But the data center story is the story
00:05:16 節能減排和其他事情。但資料中心的故事才是關鍵

00:05:19 But the data center story is the story
00:05:19 但資料中心的故事就是故事

00:05:19 But the data center story is the story of the energy people, right? And you sit
00:05:19 但資料中心的故事是能源人的故事，對吧？你坐在

00:05:22 of the energy people, right? And you sit
00:05:22 是能量人，對吧？然後你坐著

00:05:22 of the energy people, right? And you sit there and you go, how could these data
00:05:22 能源界人士，對吧？你坐在那裡，然後想，這些數據怎麼可能

00:05:23 there and you go, how could these data
00:05:23 好了，這些數據怎麼可能

00:05:23 there and you go, how could these data centers use so much power? Well, and
00:05:23 看完這些，你就會想，這些資料中心怎麼會耗費這麼多電力？嗯，

00:05:26 centers use so much power? Well, and
00:05:26 中心耗電那麼多？嗯，

00:05:26 centers use so much power? Well, and especially when you think about how
00:05:26 中心耗電那麼多？嗯，尤其是當你想到

00:05:28 especially when you think about how
00:05:28 尤其是當你想到

00:05:28 especially when you think about how little power our brains do. Well, these
00:05:28 尤其是當你想到我們的大腦有多弱的時候。嗯，這些

00:05:30 little power our brains do. Well, these
00:05:30 我們的大腦幾乎沒有什麼力量。嗯，這些

00:05:30 little power our brains do. Well, these are our best approximation in digital
00:05:30 大腦所能做到的微乎其微。好吧，這是我們在數位領域的最佳近似值

00:05:33 are our best approximation in digital
00:05:33 是我們數字方面的最佳近似值

00:05:33 are our best approximation in digital form of how our brains work. But when
00:05:33 是我們用數字形式對大腦運作方式的最佳近似。但是當

00:05:35 form of how our brains work. But when
00:05:35 大腦運作的方式。但是當

00:05:35 form of how our brains work. But when they start working together, they become
00:05:35 大腦運作的方式。但當它們開始協同工作時，它們就變成了

00:05:37 they start working together, they become
00:05:37 他們開始合作，他們成為

00:05:37 they start working together, they become superbrains. The promise of a superbrain
00:05:37 他們開始合作，成為超級大腦。超級大腦的前景

00:05:40 superbrains. The promise of a superbrain
00:05:40 超級大腦。超級大腦的前景

00:05:40 superbrains. The promise of a superbrain with a 1 gawatt for example data center
00:05:40 超級大腦。超級大腦的前景是，擁有1千兆瓦的資料中心

00:05:43 with a 1 gawatt for example data center
00:05:43 以 1 千兆瓦的資料中心為例

00:05:43 with a 1 gawatt for example data center is so palpable. People are going crazy.
00:05:43 以1千兆瓦的資料中心為例，其影響顯而易見。人們都快瘋了。

00:05:46 is so palpable. People are going crazy.
00:05:46 太明顯了。人們都瘋了。

00:05:46 is so palpable. People are going crazy. And by the way, the economics of these
00:05:46 如此明顯。人們都瘋了。順便說一句，這些經濟

00:05:48 And by the way, the economics of these
00:05:48 順便說一下，這些的經濟

00:05:48 And by the way, the economics of these things are unproven. How much revenue do
00:05:48 順便說一句，這些事情的經濟效益尚未得到證實。

00:05:51 things are unproven. How much revenue do
00:05:51 事尚未得到證實。

00:05:51 things are unproven. How much revenue do you have to have to have 50 billion in
00:05:51 事情還沒有被證實。你需要多少收入才能擁有500億美元的

00:05:53 you have to have to have 50 billion in
00:05:53 你必須擁有 500 億美元

00:05:53 you have to have to have 50 billion in capital? Well, if you depreciate it over
00:05:53 你得有500億的資本？好吧，如果你把它貶值

00:05:55 capital? Well, if you depreciate it over
00:05:55 資本？好吧，如果你把它貶值

00:05:55 capital? Well, if you depreciate it over three years or four years, you need to
00:05:55 資本？好吧，如果你在三年或四年內折舊，你需要

00:05:57 three years or four years, you need to
00:05:57 三年或四年，你需要

00:05:57 three years or four years, you need to have 10 or 15 billion dollars of capital
00:05:57 三年或四年，你需要有100億或150億美元的資本

00:06:00 have 10 or 15 billion dollars of capital
00:06:00 擁有 100 億或 150 億美元的資本

00:06:00 have 10 or 15 billion dollars of capital spend per year just to handle the
00:06:00 每年需要花費 100 到 150 億美元來處理

00:06:03 spend per year just to handle the
00:06:03 每年花費只是為了處理

00:06:03 spend per year just to handle the infrastructure. Those are huge
00:06:03 每年僅用於維護基礎設施就花費龐大。

00:06:05 infrastructure. Those are huge
00:06:05 基礎設施。這些基礎設施非常龐大

00:06:05 infrastructure. Those are huge businesses and huge revenue, which in
00:06:05 基礎設施。這些都是龐大的業務和龐大的收入，

00:06:07 businesses and huge revenue, which in
00:06:07 企業和巨額收入，其中

00:06:07 businesses and huge revenue, which in most places is not there yet.
00:06:07 企業和巨額收入，而大多數地方尚未實現這一目標。

00:06:10 most places is not there yet.
00:06:10 大多數地方尚未實現。

00:06:10 most places is not there yet. I'm curious, there's so much capital
00:06:10 大多數地方還沒有。我很好奇，那裡有這麼多資本

00:06:12 I'm curious, there's so much capital
00:06:12 我很好奇，有這麼多資本

00:06:12 I'm curious, there's so much capital being invested and deployed right now in
00:06:12 我很好奇，現在有這麼多的資本被投資和部署在

00:06:15 being invested and deployed right now in
00:06:15 目前正在投資和部署

00:06:15 being invested and deployed right now in SMRs in in nuclear bringing Three Mile
00:06:15 目前正在投資和部署 SMR，將三英里核電廠

00:06:17 SMRs in in nuclear bringing Three Mile
00:06:17 SMR 核電廠將帶來三英里

00:06:17 SMRs in in nuclear bringing Three Mile Island back online. uh in in fusion
00:06:17 SMRs 核電廠讓三哩島核電廠重新投入運作。呃，核融合

00:06:20 Island back online. uh in in fusion
00:06:20 島嶼重新上線。呃，融合

00:06:20 Island back online. uh in in fusion companies. Why isn't there an equal
00:06:20 島嶼復育在線。呃，在核融合公司。為什麼沒有平等的

00:06:22 companies. Why isn't there an equal
00:06:22 公司。為什麼沒有一個平等的

00:06:22 companies. Why isn't there an equal amount of capital going into making uh
00:06:22 公司。為什麼沒有同等數量的資本投入到呃

00:06:25 amount of capital going into making uh
00:06:25 投入的資金量

00:06:25 amount of capital going into making uh the entire you know chipset and compute
00:06:25 投入製造整個晶片組和計算機的資本

00:06:29 the entire you know chipset and compute
00:06:29 你知道的整個晶片組和計算

00:06:29 the entire you know chipset and compute just a thousand times more energy
00:06:29 你知道整個晶片組和計算只是一千倍的能量

00:06:30 just a thousand times more energy
00:06:30 只是能量增加了一千倍

00:06:30 just a thousand times more energy efficient?
00:06:30 能源效率僅僅提高了一千倍？

00:06:31 efficient?  00:06:31 有效率嗎？

00:06:31 efficient? There is a similar amount in going in
00:06:31 效率？投入的資金也差不多

00:06:33 There is a similar amount in going in
00:06:33 進入時也有類似的數量

00:06:33 There is a similar amount in going in capital. There are many many startups
00:06:33 投入的經費也差不多。有很多新創公司

00:06:35 capital. There are many many startups
00:06:35 資本。有很多新創公司

00:06:35 capital. There are many many startups that are working on non-traditional ways
00:06:35 資本。有許多新創公司正在以非傳統的方式開展業務

00:06:37 that are working on non-traditional ways
00:06:37 正在以非傳統方式工作

00:06:37 that are working on non-traditional ways of doing chips. The transformer
00:06:37 正在研究非傳統的晶片製造方法。變壓器

00:06:39 of doing chips. The transformer
00:06:39 做晶片。變壓器

00:06:39 of doing chips. The transformer architecture which is what is powering
00:06:39 晶片的製造。變壓器架構就是為

00:06:41 architecture which is what is powering
00:06:41 架構是驅動力

00:06:41 architecture which is what is powering things today has new variants. Every
00:06:41 當今驅動事物的架構有了新的變體。每個

00:06:44 things today has new variants. Every
00:06:44 今天的事物都有新的變種。每個

00:06:44 things today has new variants. Every week or so I get a pitch from a new
00:06:44 當今事物都有新的變體。每週我都會收到來自新

00:06:46 week or so I get a pitch from a new
00:06:46 一週左右，我收到了一位新

00:06:46 week or so I get a pitch from a new startup that's going to build inference
00:06:46 一週左右，我收到一家新創公司的推銷，他們打算建立推理

00:06:48 startup that's going to build inference
00:06:48 一家將建構推理的新創公司

00:06:48 startup that's going to build inference time, test time computing which are
00:06:48 新創公司將建構推理時間、測驗時間計算，

00:06:50 time, test time computing which are
00:06:50 時間，測試時間計算

00:06:50 time, test time computing which are simpler and they're optimized for
00:06:50 時間，測試時間計算更簡單，並且針對

00:06:52 simpler and they're optimized for
00:06:52 更簡單，並且針對

00:06:52 simpler and they're optimized for inference. It looks like the hardware
00:06:52 更簡單，而且它們針對推理進行了最佳化。看起來硬體

00:06:56 inference. It looks like the hardware
00:06:56 推理。看起來硬體

00:06:56 inference. It looks like the hardware will arrive just as the software needs
00:06:56 推論。看起來硬體會隨著軟體需求的增加而到來

00:06:59 will arrive just as the software needs
00:06:59 將在軟體需要時到達

00:06:59 will arrive just as the software needs expand.
00:06:59 將會隨著軟體需求的擴展而到來。

00:07:00 expand.  00:07:00 展開。

00:07:00 expand. And by the way, that's always been true.
00:07:00 展開。順便說一句，這一直都是事實。

00:07:02 And by the way, that's always been true.
00:07:02 順便說一句，這一直都是事實。

00:07:02 And by the way, that's always been true. We old-timers had a phrase um grove
00:07:02 順便說一句，這一直都是事實。我們老一輩有句俗語，叫“樹林”

00:07:05 We old-timers had a phrase um grove
00:07:05 我們老一輩有一句話叫“樹林”

00:07:05 We old-timers had a phrase um grove giveth and gates take it away. So Intel
00:07:05 我們老一輩有句俗話：格羅夫給予，蓋茲奪走。所以英特爾

00:07:08 giveth and gates take it away. So Intel
00:07:08 給予，蓋茲拿走。所以英特爾

00:07:08 giveth and gates take it away. So Intel would improve the chipsets right way
00:07:08 給予，蓋茲拿走。所以英特爾會立即改進晶片組

00:07:11 would improve the chipsets right way
00:07:11 可以正確改進晶片組

00:07:11 would improve the chipsets right way back when
00:07:11 會立即改進晶片組

00:07:12 back when  00:07:12 回到

00:07:12 back when and the software people would
00:07:12 當時軟體人員

00:07:13 and the software people would
00:07:13 軟體人員會

00:07:13 and the software people would immediately use it all and suck it all
00:07:13 軟體人員會立即使用並吸收所有

00:07:16 immediately use it all and suck it all
00:07:16 立刻用完並吸完

00:07:16 immediately use it all and suck it all up. I have no reason to believe
00:07:16 立刻用完，然後承受一切。我沒有理由相信

00:07:19 up. I have no reason to believe
00:07:19 起來。我沒有理由相信

00:07:19 up. I have no reason to believe that that's that that law grove and
00:07:19 起來。我沒有理由相信那就是那片法律樹林，

00:07:22 that that's that that law grove and
00:07:22 那就是那片法律樹林

00:07:22 that that's that that law grove and gates law has changed. If you look at
00:07:22 那就是格羅夫和蓋茲定律已經改變了。如果你看看

00:07:24 gates law has changed. If you look at
00:07:24 蓋茲定律已經改變了。如果你看看

00:07:24 gates law has changed. If you look at the gains in like the Blackwell chip or
00:07:24 蓋茲定律已經改變了。如果你看看像布萊克韋爾晶片或

00:07:27 the gains in like the Blackwell chip or
00:07:27 像 Blackwell 晶片或

00:07:27 the gains in like the Blackwell chip or the AS uh the the 350 chip in AMD,
00:07:27 像 Blackwell 晶片或 AMD 的 AS 350 晶片的收益，

00:07:30 the AS uh the the 350 chip in AMD,
00:07:30 AMD 的 AS 350 晶片，

00:07:30 the AS uh the the 350 chip in AMD, these chips are massive supercomputers
00:07:30 AMD 的 AS 350 晶片，這些晶片是大型超級計算機

00:07:33 these chips are massive supercomputers
00:07:33 這些晶片是巨大的超級計算機

00:07:33 these chips are massive supercomputers and yet we need according to the people
00:07:33 這些晶片是巨大的超級計算機，但根據人們的說法，我們需要

00:07:36 and yet we need according to the people
00:07:36 但我們需要根據人民

00:07:36 and yet we need according to the people have hundreds of thousands of these
00:07:36 但根據人們的說法，我們需要數十萬個這樣的

00:07:38 have hundreds of thousands of these
00:07:38 有數十萬個這樣的

00:07:38 have hundreds of thousands of these chips just to make a data center work.
00:07:38 僅維持一個資料中心的運作就需要數十萬個這樣的晶片。

00:07:40 chips just to make a data center work.
00:07:40 晶片只是為了讓資料中心運轉。

00:07:40 chips just to make a data center work. That shows you the scale of what this
00:07:40 晶片只是為了維持資料中心的運作。這顯示了這個

00:07:42 That shows you the scale of what this
00:07:42 這顯示了這件事的規模

00:07:42 That shows you the scale of what this kind of thinking algorithms. Now you sit
00:07:42 這展示了這種思考演算法的規模。現在你坐下

00:07:44 kind of thinking algorithms. Now you sit
00:07:44 這類思考演算法。現在你坐下

00:07:44 kind of thinking algorithms. Now you sit there and you go what could these people
00:07:44 這類思考演算法。現在你坐在那裡，想著這些人能做什麼

00:07:46 there and you go what could these people
00:07:46 在那裡，你去看看這些人

00:07:46 there and you go what could these people possibly be doing with all these chips?
00:07:46 你想這些人用這些晶片能做什麼？

00:07:49 possibly be doing with all these chips?
00:07:49 這些晶片可能用來做什麼？

00:07:49 possibly be doing with all these chips? I'll give you an example. We went from
00:07:49 所有這些晶片可能做什麼？我給你舉個例子。我們從

00:07:51 I'll give you an example. We went from
00:07:51 我舉個例子給你聽。我們從

00:07:51 I'll give you an example. We went from language to language which is what
00:07:51 我舉個例子給你聽。我們從一種語言學到另一種語言，這就是

00:07:53 language to language which is what
00:07:53 語言到語言，也就是

00:07:53 language to language which is what chatbd can be understood at to reasoning
00:07:53 語言與語言之間的轉換，也就是 Chatbd 能理解的推理

00:07:55 chatbd can be understood at to reasoning
00:07:55 chatbd 可以理解為推理

00:07:55 chatbd can be understood at to reasoning and thinking. If you want to look at an
00:07:55 Chatbd 可以理解為推理和思考。如果你想看一個

00:07:57 and thinking. If you want to look at an
00:07:57 並思考。如果你想看一個

00:07:57 and thinking. If you want to look at an open eye example look at open oi03
00:07:57 思考一下。如果你想看一個睜開眼睛的例子，你可以看看 oi03

00:08:01 open eye example look at open oi03
00:08:01 睜開眼睛範例查看睜眼 oi03

00:08:01 open eye example look at open oi03 which go does forward and back
00:08:01 睜開眼睛的例子看看打開 oi03，它執行前進和後退

00:08:02 which go does forward and back
00:08:02 哪個 go 可以前進和後退

00:08:02 which go does forward and back reinforcement learning and planning.
00:08:02 它進行前向和後向強化學習和規劃。

00:08:04 reinforcement learning and planning.
00:08:04 強化學習和規劃。

00:08:04 reinforcement learning and planning. Now the cost of doing the forward and
00:08:04 強化學習和規劃。現在進行前向和

00:08:06 Now the cost of doing the forward and
00:08:06 現在進行正向和

00:08:06 Now the cost of doing the forward and back is many orders of magnitude besides
00:08:06 現在，執行前進和後退的成本是多個數量級的，此外

00:08:09 back is many orders of magnitude besides
00:08:09 回溯是許多數量級，此外

00:08:09 back is many orders of magnitude besides just answering your question for your
00:08:09 除了回答你的問題之外，還有許多數量級的

00:08:11 just answering your question for your
00:08:11 只是回答你的問題

00:08:12 just answering your question for your PhD thesis or your college paper that
00:08:12 只是回答你的博士論文或大學論文的問題

00:08:15 PhD thesis or your college paper that
00:08:15 博士論文或你的大學論文

00:08:15 PhD thesis or your college paper that planning the back and forth is
00:08:15 博士論文或大學論文中提到，反覆規劃是

00:08:17 planning the back and forth is
00:08:17 來回規劃是

00:08:17 planning the back and forth is computationally very very expensive. So
00:08:17 規劃來回的計算成本非常高。所以

00:08:19 computationally very very expensive. So
00:08:19 計算成本非常高。所以

00:08:19 computationally very very expensive. So with the best energy and the best
00:08:19 計算成本非常高。所以用最好的能量和最好的

00:08:21 with the best energy and the best
00:08:21 用最好的能量和最好的

00:08:21 with the best energy and the best technology today we are able to show
00:08:21 憑藉當今最好的能源和最好的技術，我們能夠展示

00:08:23 technology today we are able to show
00:08:23 今天的科技讓我們能夠展示

00:08:23 technology today we are able to show evidence of planning. Many people
00:08:23 如今，我們能夠展示規劃的證據。很多人

00:08:26 evidence of planning. Many people
00:08:26 有計劃的證據。很多人

00:08:26 evidence of planning. Many people believe that if you combine planning and
00:08:26 規劃的證據。很多人認為，如果將規劃和

00:08:28 believe that if you combine planning and
00:08:28 相信如果你結合規劃和

00:08:28 believe that if you combine planning and very deep memories you can build human
00:08:28 相信如果你結合計畫和非常深刻的記憶，你就可以建立人類

00:08:31 very deep memories you can build human
00:08:31 你可以建立非常深刻的記憶

00:08:31 very deep memories you can build human level intelligence. Now of course they
00:08:31 非常深的記憶，你可以建構人類層次的智慧。當然，現在他們

00:08:34 level intelligence. Now of course they
00:08:34 水平的智力。當然，他們

00:08:34 level intelligence. Now of course they will be very expensive to start with but
00:08:34 級智能。當然，它們的起步成本會非常高，但是

00:08:37 will be very expensive to start with but
00:08:37 一開始會非常昂貴，但是

00:08:37 will be very expensive to start with but humans are very very industrious and
00:08:37 一開始會非常昂貴，但人類非常勤勞，

00:08:39 humans are very very industrious and
00:08:39 人類非常勤勞，

00:08:39 humans are very very industrious and furthermore the great future companies
00:08:39 人類非常勤勞，未來偉大的公司

00:08:41 furthermore the great future companies
00:08:41 另外還有偉大的未來公司

00:08:42 furthermore the great future companies will have AI scientists that is
00:08:42 此外，未來的公司將擁有人工智慧科學家

00:08:43 will have AI scientists that is
00:08:43 將擁有人工智慧科學家

00:08:43 will have AI scientists that is non-human scientists AI programmers that
00:08:43 將有人工智慧科學家，也就是非人類科學家，人工智慧程式設計師

00:08:46 non-human scientists AI programmers that
00:08:46 非人類科學家 AI 程式設計師

00:08:46 non-human scientists AI programmers that as opposed to human programmers who will
00:08:46 非人類科學家 AI 程式設計師，與人類程式設計師不同，他們會

00:08:48 as opposed to human programmers who will
00:08:48 而人類程式設計師會

00:08:48 as opposed to human programmers who will accelerate their impact. So, if you
00:08:48 而不是人類程式設計師，他們會加速他們的影響。所以，如果你

00:08:51 accelerate their impact. So, if you
00:08:51 加速其影響。所以，如果你

00:08:51 accelerate their impact. So, if you think about it, going back to you're the
00:08:51 加速它們的影響。所以，如果你仔細想想，回到你是

00:08:52 think about it, going back to you're the
00:08:52 想想吧，回到你是

00:08:52 think about it, going back to you're the author of the abundance thesis, as best
00:08:52 想想看，回到你是豐裕論的作者，最好

00:08:54 author of the abundance thesis, as best
00:08:54 豐裕論的作者，作為最好的

00:08:54 author of the abundance thesis, as best I can tell, Peter, you've talked about
00:08:54 豐富論題的作者，據我所知，彼得，你談到了

00:08:55 I can tell, Peter, you've talked about
00:08:55 彼得，我知道你談到了

00:08:55 I can tell, Peter, you've talked about this for 20 years. You saw it first. It
00:08:55 彼得，我看得出來，你已經談論這個話題20年了。你最先看到了。它

00:08:58 this for 20 years. You saw it first. It
00:08:58 20年來一直如此。你先看到的。它

00:08:58 this for 20 years. You saw it first. It sure looks like if we get enough
00:08:58 20年來一直如此。你先看到了。看起來如果我們得到足夠的

00:09:00 sure looks like if we get enough
00:09:00 看起來如果我們得到足夠的

00:09:00 sure looks like if we get enough electricity, we can generate the power
00:09:00 看起來如果我們有足夠的電力，我們就可以發電

00:09:03 electricity, we can generate the power
00:09:03 電力，我們可以發電

00:09:03 electricity, we can generate the power in in the sense of intellectual power to
00:09:03 電力，我們可以從智力的角度來發電

00:09:05 in in the sense of intellectual power to
00:09:05 就智力而言

00:09:05 in in the sense of intellectual power to generate abundance along the lines that
00:09:05 就智力力量而言，創造豐富的資源，

00:09:07 generate abundance along the lines that
00:09:07 沿著這條線產生豐富的

00:09:07 generate abundance along the lines that you predicted two decades ago.
00:09:07 按照你二十年前預測的方式產生豐富的資源。

00:09:08 you predicted two decades ago.
00:09:08 你二十年前就預測到了。

00:09:08 you predicted two decades ago. Every week, I study the 10 major tech
00:09:08 你二十年前就預測到了。每週我都會研究十大科技

00:09:11 Every week, I study the 10 major tech
00:09:11 每週我都會學習十大科技

00:09:11 Every week, I study the 10 major tech meta trends that will transform
00:09:11 每週，我都會研究 10 大科技元趨勢，它們將改變

00:09:13 meta trends that will transform
00:09:13 元趨勢將改變

00:09:13 meta trends that will transform industries over the decade ahead. I
00:09:13 未來十年將改變各行各業的元趨勢。我

00:09:15 industries over the decade ahead. I
00:09:15 未來十年的產業。我

00:09:15 industries over the decade ahead. I cover trends ranging from humanoid
00:09:15 未來十年的產業趨勢。我涵蓋了從人形機器人到

00:09:17 cover trends ranging from humanoid
00:09:17 涵蓋從人形到

00:09:17 cover trends ranging from humanoid robots, AGI, quantum computing,
00:09:17 涵蓋人形機器人、AGI、量子運算等趨勢，

00:09:19 robots, AGI, quantum computing,
00:09:19 機器人、通用人工智慧、量子計算、

00:09:19 robots, AGI, quantum computing, transport, energy, longevity, and more.
00:09:19 機器人、AGI、量子計算、運輸、能源、長壽等等。

00:09:22 transport, energy, longevity, and more.
00:09:22 運輸、能源、長壽等等。

00:09:22 transport, energy, longevity, and more. No fluff, only the important stuff that
00:09:22 交通、能源、壽命等等。沒有廢話，只有重要的事情

00:09:25 No fluff, only the important stuff that
00:09:25 沒有廢話，只有重要的事情

00:09:25 No fluff, only the important stuff that matters, that impacts our lives and our
00:09:25 沒有廢話，只有重要的事情，影響我們的生活和我們的

00:09:27 matters, that impacts our lives and our
00:09:27 事情，影響我們的生活和我們的

00:09:27 matters, that impacts our lives and our careers. If you want me to share these
00:09:27 很重要，會影響我們的生活和事業。如果你想讓我分享這些

00:09:29 careers. If you want me to share these
00:09:29 職業。如果你想讓我分享這些

00:09:29 careers. If you want me to share these with you, I write a newsletter twice a
00:09:29 職業。如果你想讓我跟你分享這些，我每週都會寫兩次新聞通訊

00:09:31 with you, I write a newsletter twice a
00:09:31 和你一起，我每週寫兩次新聞通訊

00:09:31 with you, I write a newsletter twice a week, sending it out as a short
00:09:31 和你一起，我每週寫兩次時事通訊，以短篇形式發送

00:09:33 week, sending it out as a short
00:09:33 週，以短篇形式發送

00:09:33 week, sending it out as a short two-minute read via email. And if you
00:09:33 週，透過電子郵件發送兩分鐘短文。如果你

00:09:35 two-minute read via email. And if you
00:09:35 透過電子郵件閱讀，需要兩分鐘。如果你

00:09:35 two-minute read via email. And if you want to discover the most important meta
00:09:35 透過電子郵件閱讀，需要兩分鐘。如果你想發現最重要的元

00:09:37 want to discover the most important meta
00:09:37 想要發現最重要的元

00:09:37 want to discover the most important meta trends 10 years before anyone else,
00:09:37 想要比別人提早 10 年發現最重要的元趨勢，

00:09:39 trends 10 years before anyone else,
00:09:39 比其他任何人都領先 10 年，

00:09:39 trends 10 years before anyone else, these reports are for you. Readers
00:09:39 領先其他人十年，這些報告是為你準備的。讀者們

00:09:41 these reports are for you. Readers
00:09:41 這些報告是寫給你們的。讀者們

00:09:41 these reports are for you. Readers include founders and CEOs from the
00:09:41 這些報告是寫給你的。讀者包括來自

00:09:43 include founders and CEOs from the
00:09:43 包括來自

00:09:43 include founders and CEOs from the world's most disruptive companies and
00:09:43 包括世界上最具顛覆性的公司的創始人和首席執行官，以及

00:09:45 world's most disruptive companies and
00:09:45 世界上最具顛覆性的公司和

00:09:46 world's most disruptive companies and entrepreneurs building the world's most
00:09:46 世界上最具顛覆性的公司和企業家正在打造世界上最

00:09:48 entrepreneurs building the world's most
00:09:48 企業家打造世界上最

00:09:48 entrepreneurs building the world's most disruptive companies. It's not for you
00:09:48 企業家打造世界上最具顛覆性的公司。這不適合你

00:09:50 disruptive companies. It's not for you
00:09:50 顛覆性的公司。這不適合你

00:09:50 disruptive companies. It's not for you if you don't want to be informed of
00:09:50 顛覆性的公司。如果你不想被告知

00:09:52 if you don't want to be informed of
00:09:52 如果你不想被告知

00:09:52 if you don't want to be informed of what's coming, why it matters, and how
00:09:52 如果你不想了解即將發生的事情，為什麼這很重要，以及如何

00:09:54 what's coming, why it matters, and how
00:09:54 即將發生什麼，為什麼重要，以及如何

00:09:54 what's coming, why it matters, and how you can benefit from it. To subscribe
00:09:54 即將推出的內容、重要性以及您如何從中受益。訂閱

00:09:56 you can benefit from it. To subscribe
00:09:56 您可以從中受益。訂閱

00:09:56 you can benefit from it. To subscribe for free, go to dmadis.com/tatrends.
00:09:56 您可以從中受益。免費訂閱，請造訪 dmadis.com/tatrends。

00:10:00 for free, go to dmadis.com/tatrends.
00:10:00 免費，請造訪 dmadis.com/tatrends。

00:10:00 for free, go to dmadis.com/tatrends. That's dmandis.com/tatrends
00:10:00 免費，請造訪 dmadis.com/tatrends。網址是 dmandis.com/tatrends

00:10:03 That's dmandis.com/tatrends
00:10:03 這是 dmandis.com/tatrends

00:10:03 That's dmandis.com/tatrends to gain access to trends 10 plus years
00:10:03 造訪 dmandis.com/tatrends 了解 10 多年的趨勢

00:10:06 to gain access to trends 10 plus years
00:10:06 了解 10 多年的趨勢

00:10:06 to gain access to trends 10 plus years before anyone else.
00:10:06 比其他任何人都領先 10 多年，掌握趨勢動態。

00:10:07 before anyone else.
00:10:07 比任何人都先。

00:10:08 before anyone else. Let me throw some numbers at you just to
00:10:08 比任何人都早。我先給你一些數字，只是為了

00:10:09 Let me throw some numbers at you just to
00:10:09 讓我給你一些數字，只是為了

00:10:09 Let me throw some numbers at you just to reinforce what you said. you know, we
00:10:09 讓我給你一些數字來支持你的說法。你知道，我們

00:10:10 reinforce what you said. you know, we
00:10:10 強化你所說的。你知道，我們

00:10:10 reinforce what you said. you know, we have a couple companies in the lab that
00:10:10 強化你所說的。你知道，我們實驗室裡有幾家公司

00:10:12 have a couple companies in the lab that
00:10:12 實驗室裡有幾家公司

00:10:12 have a couple companies in the lab that are doing voice customer service, voice
00:10:12 實驗室裡有幾家公司正在做語音客戶服務，語音

00:10:14 are doing voice customer service, voice
00:10:14 正在做語音客服，語音

00:10:14 are doing voice customer service, voice sales with the new, you know, just as of
00:10:14 正在用新的語音客戶服務、語音銷售，你知道，就像

00:10:16 sales with the new, you know, just as of
00:10:16 新的銷售情況，你知道，就像

00:10:16 sales with the new, you know, just as of the last month.
00:10:16 新的銷售情況，你知道，就上個月的情況而言。

00:10:16 the last month.  上個月 00:10:16。

00:10:16 the last month. Sure.
00:10:16 上個月。當然。

00:10:17 Sure.  00:10:17 當然。

00:10:17 Sure. And the value of these these
00:10:17 當然。這些的價值

00:10:20 And the value of these these
00:10:20 這些的價值

00:10:20 And the value of these these conversations is 10 to $1,000. And the
00:10:20 這些對話的價值是10到1000美元。

00:10:23 conversations is 10 to $1,000. And the
00:10:23 對話是 10 到 1,000 美元。

00:10:23 conversations is 10 to $1,000. And the cost of the compute is, you know, maybe
00:10:23 對話的成本是 10 到 1,000 美元。計算成本大概是

00:10:25 cost of the compute is, you know, maybe
00:10:25 計算成本大概是

00:10:25 cost of the compute is, you know, maybe two three concurrent GPUs is optimal.
00:10:25 運算成本是，也許兩個三個並發 GPU 是最佳的。

00:10:28 two three concurrent GPUs is optimal.
00:10:28 兩個三個並發 GPU 是最佳的。

00:10:28 two three concurrent GPUs is optimal. It's like 10 20 cents. And so they would
00:10:28 兩三個並發 GPU 是最佳選擇。大約是 10 美分 20 美分。所以他們會

00:10:31 It's like 10 20 cents. And so they would
00:10:31 大概是10美分20美分。所以他們會

00:10:31 It's like 10 20 cents. And so they would buy massively more compute to improve
00:10:31 大概是 10 到 20 美分。所以他們會購買大量的運算資源來提高

00:10:34 buy massively more compute to improve
00:10:34 購買大量運算資源來提高

00:10:34 buy massively more compute to improve the the quality of the conversation.
00:10:34 購買大量運算資源來提升對話品質。

00:10:36 the the quality of the conversation.
00:10:36 對話的品質。

00:10:36 the the quality of the conversation. There aren't even close to enough. We we
00:10:36 對話的品質。遠遠不夠。我們

00:10:38 There aren't even close to enough. We we
00:10:38 遠遠不夠。我們

00:10:38 There aren't even close to enough. We we count about 10 million concurrent phone
00:10:38 遠遠不夠。我們統計了一下，大約有 1000 萬支手機同時在線上。

00:10:39 count about 10 million concurrent phone
00:10:39 統計約1000萬支手機同時在線

00:10:39 count about 10 million concurrent phone calls that should move to AI in the next
00:10:39 估計未來 1000 萬通並發電話應該會轉移到人工智慧

00:10:42 calls that should move to AI in the next
00:10:42 下一步應該轉移到人工智慧

00:10:42 calls that should move to AI in the next year or so.
00:10:42 呼籲在明年左右轉向人工智慧。

00:10:44 year or so.  00:10:44 一年左右。

00:10:44 year or so. And and my view of that is that's a good
00:10:44 一年左右。我認為這是一個很好的

00:10:46 And and my view of that is that's a good
00:10:46 我認為這是一件好事

00:10:46 And and my view of that is that's a good tactical solution and a great business.
00:10:46 我認為這是一個很好的戰術解決方案和一項偉大的事業。

00:10:48 tactical solution and a great business.
00:10:48 戰術解決方案和偉大的業務。

00:10:48 tactical solution and a great business. Let's look at other examples of tactical
00:10:48 戰術解決方案和偉大的事業。讓我們看看其他戰術解決方案的例子

00:10:50 Let's look at other examples of tactical
00:10:50 讓我們來看看其他戰術的例子

00:10:50 Let's look at other examples of tactical solutions that are great businesses.
00:10:50 讓我們來看看其他優秀的戰術解決方案的例子。

00:10:52 solutions that are great businesses.
00:10:52 解決方案對於企業來說非常有幫助。

00:10:52 solutions that are great businesses. And I obviously have a conflict of
00:10:52 解決方案對企業來說很重要。我顯然有一個衝突

00:10:53 And I obviously have a conflict of
00:10:53 顯然我有一個衝突

00:10:53 And I obviously have a conflict of interest talking about Google because I
00:10:53 顯然，談論 Google 有利益衝突，因為我

00:10:55 interest talking about Google because I
00:10:55 我對談論谷歌很感興趣，因為我

00:10:55 interest talking about Google because I love it so much. So with that as in
00:10:55 我很喜歡談論谷歌，因為我太喜歡它了。所以

00:10:57 love it so much. So with that as in
00:10:57 太喜歡了。所以

00:10:57 love it so much. So with that as in mind, look at the Google strength in
00:10:57 非常喜歡。所以考慮到這一點，看看谷歌在

00:10:58 mind, look at the Google strength in
00:10:58 看看 Google 的實力

00:10:58 mind, look at the Google strength in GCP. Now Google Google's cloud product
00:10:58 看看 Google 在 GCP 的實力。現在 Google 的雲端產品

00:11:01 GCP. Now Google Google's cloud product
00:11:01 GCP。現在是 Google 的雲端產品

00:11:02 GCP. Now Google Google's cloud product where they have a completely fully
00:11:02 GCP。現在是 Google 的雲端產品，他們擁有完全

00:11:04 where they have a completely fully
00:11:04 他們有一個完全

00:11:04 where they have a completely fully served enterprise offering for
00:11:04 他們擁有完全服務的企業產品

00:11:06 served enterprise offering for
00:11:06 為企業提供

00:11:06 served enterprise offering for essentially automating your company with
00:11:06 為企業提供本質上自動化的公司服務

00:11:08 essentially automating your company with
00:11:08 本質上實現公司自動化

00:11:08 essentially automating your company with AI.
00:11:08 本質上利用人工智慧實現公司自動化。

00:11:09 AI.  00:11:09 人工智慧。

00:11:09 AI. Yeah.  00:11:09 人工智慧。是的。

00:11:09 Yeah.  00:11:09 是的。

00:11:09 Yeah. And the remarkable thing and this is to
00:11:09 是的。最了不起的是

00:11:12 And the remarkable thing and this is to
00:11:12 值得注意的是

00:11:12 And the remarkable thing and this is to me is shocking is you can in an
00:11:12 最令人驚訝的是，對我來說，這是令人震驚的，你可以在

00:11:14 me is shocking is you can in an
00:11:14 令我震驚的是，你可以

00:11:14 me is shocking is you can in an enterprise write the task that you want
00:11:14 令我震驚的是，你可以在企業中寫下你想要的任務

00:11:18 enterprise write the task that you want
00:11:18 企業寫下你想要的任務

00:11:18 enterprise write the task that you want and then using something called the
00:11:18 企業寫下你想要的任務，然後用一個叫做

00:11:19 and then using something called the
00:11:19 然後使用一種叫做

00:11:19 and then using something called the model context protocol you can connect
00:11:19 然後使用模型上下文協議，你可以連接

00:11:21 model context protocol you can connect
00:11:21 您可以連接的模型上下文協議

00:11:21 model context protocol you can connect your databases to that and the large
00:11:21 模型上下文協議，你可以將你的資料庫連接到它和大型

00:11:23 your databases to that and the large
00:11:23 你的資料庫和大型

00:11:23 your databases to that and the large language model can produce the code for
00:11:23 你的資料庫和大型語言模型可以產生程式碼

00:11:26 language model can produce the code for
00:11:26 語言模型可以生成

00:11:26 language model can produce the code for your enterprise. Now, there's 100,000
00:11:26 語言模型可以為你的企業產生程式碼。現在有 10 萬個

00:11:29 your enterprise. Now, there's 100,000
00:11:29 你的企業。現在有 10 萬個

00:11:29 your enterprise. Now, there's 100,000 enterprise software companies,
00:11:29 你的企業。現在有10萬家企業軟體公司，

00:11:31 enterprise software companies,
00:11:31 企業軟體公司，

00:11:31 enterprise software companies, middleware companies that grew up in the
00:11:31 企業軟體公司、中介軟體公司在

00:11:33 middleware companies that grew up in the
00:11:33 成長於

00:11:33 middleware companies that grew up in the last 30 years that I've been working on
00:11:33 在過去 30 年裡，我一直在研究成長的中間件公司

00:11:35 last 30 years that I've been working on
00:11:35 過去 30 年我一直在努力

00:11:35 last 30 years that I've been working on this that are all now in trouble because
00:11:35 過去 30 年來我一直致力於此，但現在都陷入困境，因為

00:11:37 this that are all now in trouble because
00:11:37 現在都陷入困境了，因為

00:11:37 this that are all now in trouble because that that interstitial connection is no
00:11:37 現在都陷入困境了，因為間隙連接不存在

00:11:39 that that interstitial connection is no
00:11:39 那個間隙連接是沒有的

00:11:39 that that interstitial connection is no longer needed
00:11:39 不再需要這種間隙連接

00:11:40 longer needed  00:11:40 不再需要

00:11:40 longer needed with their business
00:11:40 不再需要他們的業務

00:11:41 with their business
00:11:41 他們的生意

00:11:41 with their business and and and of course they'll have to
00:11:41 他們的生意，當然他們必須

00:11:42 and and and of course they'll have to
00:11:42 當然他們必須

00:11:42 and and and of course they'll have to change as well. The good news for them
00:11:42 當然，他們也必須改變。對他們來說，好消息是

00:11:44 change as well. The good news for them
00:11:44 也改變了。對他們來說好消息

00:11:44 change as well. The good news for them is enterprises make these changes very
00:11:44 變化。對他們來說，好消息是企業非常

00:11:46 is enterprises make these changes very
00:11:46 企業做出這些改變非常

00:11:46 is enterprises make these changes very slowly. If you built a brand new
00:11:46 企業做出這些改變非常緩慢。如果你建立了一個全新的

00:11:49 slowly. If you built a brand new
00:11:49 慢慢來。如果你建造了一個全新的

00:11:49 slowly. If you built a brand new enterprise um architecture for ERP and
00:11:49 慢慢來。如果你為 ERP 建構了一個全新的企業架構，

00:11:52 enterprise um architecture for ERP and
00:11:52 企業 ERP 架構和

00:11:52 enterprise um architecture for ERP and MRP, you would be highly tempted to not
00:11:52 企業架構 ERP 和 MRP，你很可能會不

00:11:55 MRP, you would be highly tempted to not
00:11:55 MRP，你很可能會不

00:11:55 MRP, you would be highly tempted to not use any of the ERP or MRP suppliers, but
00:11:55 MRP，你很可能會不使用任何 ERP 或 MRP 供應商，但是

00:11:58 use any of the ERP or MRP suppliers, but
00:11:58 使用任何 ERP 或 MRP 供應商，但

00:11:58 use any of the ERP or MRP suppliers, but instead use open- source libraries,
00:11:58 不要使用任何 ERP 或 MRP 供應商，而是使用開源函式庫，

00:12:01 instead use open- source libraries,
00:12:01 而是使用開源函式庫，

00:12:01 instead use open- source libraries, build essentially use BigQuery or the
00:12:01 而是使用開源函式庫，本質上使用 BigQuery 或

00:12:03 build essentially use BigQuery or the
00:12:03 建置本質上使用 BigQuery 或

00:12:03 build essentially use BigQuery or the equivalent from Amazon, which is Red
00:12:03 建構基本上使用 BigQuery 或亞馬遜的同類產品，即 Red

00:12:05 equivalent from Amazon, which is Red
00:12:05 亞馬遜的等價物，也就是紅色

00:12:05 equivalent from Amazon, which is Red Redshift, and essentially build that
00:12:05 亞馬遜的類似產品，也就是 Red Redshift，本質上就是建構這個

00:12:07 Redshift, and essentially build that
00:12:07 Redshift，本質上建構了

00:12:07 Redshift, and essentially build that architecture and it gives you infinite
00:12:07 Redshift，本質上建構了這個架構，它為你提供了無限的

00:12:08 architecture and it gives you infinite
00:12:08 建築，它給你無限

00:12:08 architecture and it gives you infinite flexibility and the computer system
00:12:08 架構，它給你無限的彈性和電腦系統

00:12:10 flexibility and the computer system
00:12:10 靈活性和電腦系統

00:12:10 flexibility and the computer system writes most of the code. Now,
00:12:10 彈性，電腦系統編寫了大部分程式碼。現在，

00:12:13 writes most of the code. Now,
00:12:13 編寫了大部分程式碼。現在，

00:12:13 writes most of the code. Now, programmers don't go away at the moment.
00:12:13 編寫了大部分程式碼。現在，程式設計師不會消失。

00:12:15 programmers don't go away at the moment.
00:12:15 程式設計師此刻不會消失。

00:12:15 programmers don't go away at the moment. It's pretty clear that junior
00:12:15 程式設計師目前還沒有消失。很明顯，初級

00:12:17 It's pretty clear that junior
00:12:17 很明顯，小

00:12:17 It's pretty clear that junior programmers go away. The sort of
00:12:17 很明顯，初級程式設計師已經消失了。

00:12:18 programmers go away. The sort of
00:12:18 程式設計師走了。

00:12:18 programmers go away. The sort of journeymen, if you will, of the
00:12:18 程式設計師消失了。如果你願意的話，

00:12:20 journeymen, if you will, of the
00:12:20 如果你願意的話，

00:12:20 journeymen, if you will, of the stereotype because these systems aren't
00:12:20 如果你願意的話，這些系統並不是刻板印像中的熟練工

00:12:22 stereotype because these systems aren't
00:12:22 刻板印象，因為這些系統不是

00:12:22 stereotype because these systems aren't good enough yet to automatically write
00:12:22 刻板印象，因為這些系統還不夠好，無法自動寫入

00:12:24 good enough yet to automatically write
00:12:24 還不夠好，無法自動寫入

00:12:24 good enough yet to automatically write all the code. They need very senior
00:12:24 還不夠好，無法自動編寫所有程式碼。他們需要非常高級的

00:12:27 all the code. They need very senior
00:12:27 所有代碼。他們需要非常高級的

00:12:27 all the code. They need very senior computer scientists, computer engineers
00:12:27 所有代碼。他們需要非常資深的電腦科學家和電腦工程師

00:12:28 computer scientists, computer engineers
00:12:28 電腦科學家、電腦工程師

00:12:28 computer scientists, computer engineers who are watching it, that will
00:12:28 正在觀看的電腦科學家和電腦工程師們，

00:12:30 who are watching it, that will
00:12:30 誰在觀看，那將

00:12:30 who are watching it, that will eventually go away.
00:12:30 正在觀看它的人最終都會消失。

00:12:32 eventually go away.
00:12:32 最終消失。

00:12:32 eventually go away. One of the things to say about
00:12:32 最終會消失。關於

00:12:33 One of the things to say about
00:12:33 有一件事要說

00:12:33 One of the things to say about productivity, and I call this the San
00:12:33 關於生產力，我稱之為 San

00:12:35 productivity, and I call this the San
00:12:35 生產力，我稱之為 San

00:12:35 productivity, and I call this the San Francisco consensus because it's it's
00:12:35 生產力，我稱之為舊金山共識，因為它是

00:12:36 Francisco consensus because it's it's
00:12:36 弗朗西斯科共識，因為它是

00:12:36 Francisco consensus because it's it's largely the view of people who operate
00:12:36 弗朗西斯科共識，因為這主要是那些運營

00:12:39 largely the view of people who operate
00:12:39 很大程度上是那些運營

00:12:39 largely the view of people who operate in San Francisco,
00:12:39 很大程度上是舊金山人的觀點，

00:12:40 in San Francisco,
舊金山時間 00:12:40

00:12:40 in San Francisco, goes something like this. uh we're just
00:12:40 在舊金山，情況是這樣的。呃，我們只是

00:12:43 goes something like this. uh we're just
00:12:43 就像這樣。呃，我們只是

00:12:43 goes something like this. uh we're just about to the point where we can do two
00:12:43 就像這樣。呃，我們很快就能做兩件事了

00:12:45 about to the point where we can do two
00:12:45 我們可以做兩個

00:12:45 about to the point where we can do two things that are shocking. The first is
00:12:45 我們可以做兩件令人震驚的事。第一是

00:12:48 things that are shocking. The first is
00:12:48 令人震驚的事。首先是

00:12:48 things that are shocking. The first is we can replace most programming tasks by
00:12:48 令人震驚的事。首先，我們可以用

00:12:50 we can replace most programming tasks by
00:12:50 我們可以用以下方式取代大多數程式設計任務

00:12:50 we can replace most programming tasks by computers and we can replace both most
00:12:50 我們可以用電腦取代大多數程式設計任務，我們可以取代大多數

00:12:54 computers and we can replace both most
00:12:54 我們可以更換大多數電腦

00:12:54 computers and we can replace both most mathemat mathematical tasks by
00:12:54 我們可以用計算機取代大多數數學任務

00:12:56 mathemat mathematical tasks by
00:12:56 數學任務

00:12:56 mathemat mathematical tasks by computers.
00:12:56 計算機完成的數學任務。

00:12:57 computers.  00:12:57 電腦。

00:12:57 computers. Now you sit there and you go why? Well,
00:12:57 電腦。現在你坐在那裡，你會問為什麼？嗯，

00:12:59 Now you sit there and you go why? Well,
00:12:59 現在你坐在那裡，你會問為什麼？嗯，

00:12:59 Now you sit there and you go why? Well, if you think about programming and math,
00:12:59 現在你坐在那裡，你會問為什麼？好吧，如果你考慮一下程式設計和數學，

00:13:02 if you think about programming and math,
00:13:02 如果你考慮程式設計和數學，

00:13:02 if you think about programming and math, they have limited language sets compared
00:13:02 如果你考慮程式設計和數學，它們的語言集相比有限

00:13:05 they have limited language sets compared
00:13:05 相較之下，他們的語言能力有限

00:13:05 they have limited language sets compared to human language. So close they're
00:13:05 與人類語言相比，它們的語言集有限。它們如此接近

00:13:07 to human language. So close they're
00:13:07 人類語言。它們如此接近

00:13:07 to human language. So close they're simpler computationally
00:13:07 類似人類語言。它們在計算上更簡單

00:13:09 simpler computationally
00:13:09 計算更簡單

00:13:09 simpler computationally and they're scale free. You can just do
00:13:09 計算起來更簡單，而且它們是無標度的。你可以這樣做

00:13:12 and they're scale free. You can just do
00:13:12 而且它們是無尺度的。你可以這樣做

00:13:12 and they're scale free. You can just do it and do it and do it with more
00:13:12 而且它們是無尺度的。你可以不斷地重複，用更多的

00:13:13 it and do it and do it with more
00:13:13 做這件事，做得更多

00:13:14 it and do it and do it with more electricity. You don't need data. You
00:13:14 這樣做，用更多的電力。你不需要數據。你

00:13:16 electricity. You don't need data. You
00:13:16 電力。你不需要數據。你

00:13:16 electricity. You don't need data. You don't need real world input. You don't
00:13:16 電力。你不需要數據。你不需要現實世界的輸入。你不需要

00:13:18 don't need real world input. You don't
00:13:18 不需要現實世界的輸入。你不需要

00:13:18 don't need real world input. You don't need telemetry. You don't need sensors.
00:13:18 不需要現實世界的輸入。你不需要遙測。你不需要感測器。

00:13:20 need telemetry. You don't need sensors.
00:13:20 需要遙測。不需要感測器。

00:13:20 need telemetry. You don't need sensors. Yeah.
00:13:20 需要遙測。不需要感測器。是的。

00:13:21 Yeah.  00:13:21 是的。

00:13:21 Yeah. So, it's likely in my opinion that
00:13:21 是的。所以，我認為

00:13:23 So, it's likely in my opinion that
00:13:23 所以，我認為

00:13:23 So, it's likely in my opinion that you're going to see worldclass
00:13:23 所以，我認為你很可能會看到世界級的

00:13:24 you're going to see worldclass
00:13:24 你會看到世界級的

00:13:24 you're going to see worldclass mathematicians emerge in the next one
00:13:24 你會看到世界級的數學家在下一屆比賽中出現

00:13:27 mathematicians emerge in the next one
00:13:27 數學家在下一個

00:13:27 mathematicians emerge in the next one year that are AI based and worldclass
00:13:27 未來一年將出現一批基於人工智慧的世界級數學家

00:13:30 year that are AI based and worldclass
00:13:30 基於人工智慧且世界一流的

00:13:30 year that are AI based and worldclass programmers that going to appear within
00:13:30 年內將會出現以人工智慧為基礎的世界級程式設計師

00:13:32 programmers that going to appear within
00:13:32 程式設計師將會出現

00:13:32 programmers that going to appear within the next one or two years. When those
00:13:32 未來一兩年內會出現的程式設計師。當這些

00:13:35 the next one or two years. When those
00:13:35 未來一兩年。當這些

00:13:35 the next one or two years. When those things are deployed at scale, remember
00:13:35 未來一兩年。當這些技術大規模部署時，記住

00:13:37 things are deployed at scale, remember
00:13:37 事情是大規模部署的，記住

00:13:37 things are deployed at scale, remember math and programming are the basis of
00:13:37 事物是大規模部署的，記住數學和程式設計是

00:13:39 math and programming are the basis of
00:13:39 數學和程式設計是

00:13:39 math and programming are the basis of kind of everything, right? It's an
00:13:39 數學和程式設計是一切的基礎，對吧？這是一個

00:13:41 kind of everything, right? It's an
00:13:41 什麼都有，對吧？這是一個

00:13:41 kind of everything, right? It's an accelerate accelerant for physics,
00:13:41 什麼都行，對吧？它是物理學的加速劑，

00:13:43 accelerate accelerant for physics,
00:13:43 加速物理學的促進劑，

00:13:43 accelerate accelerant for physics, chemistry, biology, material science.
00:13:43 加速物理、化學、生物、材料科學的促進劑。

00:13:46 chemistry, biology, material science.
00:13:46 化學、生物、材料科學。

00:13:46 chemistry, biology, material science. So, going back to things like climate
00:13:46 化學、生物、材料科學。所以，回到氣候等議題

00:13:47 So, going back to things like climate
00:13:47 那麼，回到氣候議題上

00:13:47 So, going back to things like climate change, can you imagine if we and this
00:13:47 那麼，回到氣候變遷的問題上，你能想像如果我們和這個

00:13:49 change, can you imagine if we and this
00:13:49 改變，你能想像如果我們和這個

00:13:49 change, can you imagine if we and this goes back to your original argument,
00:13:49 改變，你能想像如果我們回到你最初的論點，

00:13:51 goes back to your original argument,
00:13:51 回到你最初的論點，

00:13:51 goes back to your original argument, Peter, imagine if we can accelerate the
00:13:51 回到你最初的論點，彼得，想像一下如果我們可以加速

00:13:53 Peter, imagine if we can accelerate the
00:13:53 彼得，想像一下如果我們可以加速

00:13:53 Peter, imagine if we can accelerate the discoveries of the new materials that
00:13:53 彼得，想像一下，如果我們能夠加速新材料的發現，

00:13:55 discoveries of the new materials that
00:13:55 新材料的發現

00:13:55 discoveries of the new materials that allow us to deal with a carbonized
00:13:55 新材料的發現使我們能夠處理碳化

00:13:56 allow us to deal with a carbonized
00:13:56 讓我們處理碳化

00:13:56 allow us to deal with a carbonized world.
00:13:56 讓我們應對碳化世界。

00:13:57 world.  00:13:57 世界。

00:13:57 world. Yeah.  00:13:57 世界。是的。

00:13:58 Yeah.  00:13:58 是的。

00:13:58 Yeah. Right. It's very exciting. Can I love to
00:13:58 是的。沒錯。這很令人興奮。我可以

00:14:01 Right. It's very exciting. Can I love to
00:14:01 對。這很令人興奮。我可以

00:14:01 Right. It's very exciting. Can I love to drill in about
00:14:01 對。這很令人興奮。我可以詳細談談

00:14:03 drill in about  00:14:03 深入探討

00:14:03 drill in about you first?
00:14:03 先深入了解你？

00:14:03 you first?  00:14:03 你先來？

00:14:04 you first? I just want to hit this because it's
00:14:04 你先來？我只是想打這個，因為它

00:14:05 I just want to hit this because it's
00:14:05 我只是想打這個，因為它

00:14:05 I just want to hit this because it's important the potential for there to be
00:14:05 我只是想強調這一點，因為

00:14:09 important the potential for there to be
00:14:09 重要的是存在

00:14:09 important the potential for there to be I don't want to use the word PhD level
00:14:09 重要的是存在潛力，我不想使用「博士程度」這個詞

00:14:12 I don't want to use the word PhD level
00:14:12 我不想用「博士程度」這個詞

00:14:12 I don't want to use the word PhD level you know other than uh thinking in the
00:14:12 我不想用「博士程度」這個詞，你知道，除了思考

00:14:15 you know other than uh thinking in the
00:14:15 你知道除了思考

00:14:15 you know other than uh thinking in the terms of research PhD level AIS and uh
00:14:15 除了從研究博士程度的 AIS 角度思考之外

00:14:19 terms of research PhD level AIS and uh
00:14:19 博士研究水準 AIS 和呃

00:14:19 terms of research PhD level AIS and uh that can basically attack any problem
00:14:19 博士級 AIS 研究，基本上可以解決任何問題

00:14:23 that can basically attack any problem
00:14:23 基本上可以解決任何問題

00:14:23 that can basically attack any problem and solve it uh and solve math if you
00:14:23 基本上可以解決任何問題，呃，還可以解決數學問題，如果你

00:14:26 and solve it uh and solve math if you
00:14:26 並解決它，如果你

00:14:26 and solve it uh and solve math if you would in physics. uh this idea of an AI,
00:14:26 並解決它，就像在物理中解決數學問題一樣。這個人工智慧的想法，

00:14:29 would in physics. uh this idea of an AI,
00:14:29 在物理學中。呃，人工智慧的想法，

00:14:29 would in physics. uh this idea of an AI, you know, intelligence explosion. Um Leo
00:14:29 物理學領域。呃，人工智慧的概念，你知道，智慧爆炸。嗯，Leo

00:14:33 you know, intelligence explosion. Um Leo
00:14:33 你知道，智力爆炸。呃，Leo

00:14:33 you know, intelligence explosion. Um Leo Leopold put that at like 26 27
00:14:33 你知道，智力爆炸。嗯，Leo Leopold 把它算作 26 27

00:14:38 Leopold put that at like 26 27
00:14:38 Leopold 把它放在 26 27

00:14:38 Leopold put that at like 26 27 uh heading towards digital super
00:14:38 Leopold 把它放在 26 27 呃，朝著數字超級

00:14:40 uh heading towards digital super
00:14:40 呃，走向數位超級

00:14:40 uh heading towards digital super intelligence in the next few years. Do
00:14:40 呃，未來幾年將走向數位超級智慧。

00:14:42 intelligence in the next few years. Do
00:14:42 未來幾年內，人工智慧的發展將如何？

00:14:42 intelligence in the next few years. Do you buy that time frame?
00:14:42 未來幾年內，情報工作將會取得長足進展。你認同這個時間框架嗎？

00:14:44 you buy that time frame?
00:14:44 您購買那個時間框架？

00:14:44 you buy that time frame? So again, I consider that to be the San
00:14:44 你買這個時間框架嗎？所以，我再說一遍，我認為這是聖

00:14:46 So again, I consider that to be the San
00:14:46 所以，我再次認為這是聖

00:14:46 So again, I consider that to be the San Francisco consensus. I think the dates
00:14:46 所以，我再次認為這是舊金山共識。我認為日期

00:14:48 Francisco consensus. I think the dates
00:14:48 弗朗西斯科共識。我認為日期

00:14:48 Francisco consensus. I think the dates are probably off by one and a half or
00:14:48 弗朗西斯科達成共識。我認為日期可能相差一年半或

00:14:51 are probably off by one and a half or
00:14:51 可能相差一個半或

00:14:51 are probably off by one and a half or two times,
00:14:51 可能相差一倍半或兩倍，

00:14:52 two times,  00:14:52 兩次，

00:14:52 two times, which is pretty close. So a reasonable
00:14:52 兩次，非常接近。所以一個合理的

00:14:55 which is pretty close. So a reasonable
00:14:55 非常接近。所以一個合理的

00:14:55 which is pretty close. So a reasonable prediction is that we're going to have
00:14:55 非常接近。所以一個合理的預測是，我們會有

00:14:58 prediction is that we're going to have
00:14:58 預測是，我們將會有

00:14:58 prediction is that we're going to have specialized soants in every field within
00:14:58 預測是，我們將在各個領域擁有專門的

00:15:01 specialized soants in every field within
00:15:01 各領域的專業解決方案

00:15:01 specialized soants in every field within five years.
00:15:01 五年內在每個領域都有專門的 soants。

00:15:03 five years.  00:15:03 五年。

00:15:03 five years. That's pretty much in the bag as far as
00:15:03 五年。這幾乎是板上釘釘的事了

00:15:05 That's pretty much in the bag as far as
00:15:05 基本上已經確定了

00:15:05 That's pretty much in the bag as far as I'm concerned.
00:15:05 就我而言，這基本上是板上釘釘的事了。

00:15:06 I'm concerned.  00:15:06 我很擔心。

00:15:06 I'm concerned. Sure.
00:15:06 我很擔心。當然。

00:15:06 Sure.  00:15:06 當然。

00:15:06 Sure. And here's why. You have this amount of
00:15:06 當然。原因如下。你有這麼多

00:15:08 And here's why. You have this amount of
00:15:08 原因如下。你有這麼多

00:15:08 And here's why. You have this amount of humans and then you add a million AI
00:15:08 原因如下。你有這麼多的人類，然後你又增加了一百萬個人工智慧

00:15:11 humans and then you add a million AI
00:15:11 人類，然後增加一百萬個人工智慧

00:15:11 humans and then you add a million AI scientists to do something, your slope
00:15:11 人類，然後你加入一百萬人工智慧科學家來做某事，你的斜率

00:15:13 scientists to do something, your slope
00:15:13 科學家做了一些事情，你的斜率

00:15:13 scientists to do something, your slope goes like this. Your rate of
00:15:13 科學家做了一些事情，你的斜率是這樣的。你的速率

00:15:14 goes like this. Your rate of
00:15:14 就像這樣。你的

00:15:14 goes like this. Your rate of improvement, we should get there.
00:15:14 就像這樣。你的進步速度，我們應該要達到。

00:15:17 improvement, we should get there.
00:15:17 改進，我們應該達成目標。

00:15:18 improvement, we should get there. The real question is once you have all
00:15:18 改進，我們應該能達到目標。真正的問題是，一旦你擁有了所有

00:15:19 The real question is once you have all
00:15:19 真正的問題是，一旦你擁有了所有

00:15:19 The real question is once you have all these sants, do they unify?
00:15:19 真正的問題是，一旦擁有了所有這些聖人，他們會統一嗎？

00:15:23 these sants, do they unify?
00:15:23 這些聖人，他們統一了嗎？

00:15:24 these sants, do they unify? Do they ultimately become a superhum?
00:15:24 這些聖人，他們會統一嗎？他們最終會成為超級人類嗎？

00:15:26 Do they ultimately become a superhum?
00:15:26 他們最終會成為超級英雄嗎？

00:15:26 Do they ultimately become a superhum? The term we're using is super
00:15:26 他們最終會成為超級人物嗎？我們用的是超級

00:15:28 The term we're using is super
00:15:28 我們使用的術語是超級

00:15:28 The term we're using is super intelligence, which implies intelligence
00:15:28 我們使用的術語是超級智能，這意味著智能

00:15:30 intelligence, which implies intelligence
00:15:30 智力，這意味著智力

00:15:30 intelligence, which implies intelligence that beyond the sum of what humans can
00:15:30 智力，這意味著智力超越了人類所能達到的總和

00:15:32 that beyond the sum of what humans can
00:15:32 超越人類所能

00:15:32 that beyond the sum of what humans can do.
00:15:32 超越了人類所能做到的一切。

00:15:33 do.  00:15:33 做。

00:15:33 do. The race to super intelligence, which is
00:15:33 確實如此。超級智能競賽

00:15:36 The race to super intelligence, which is
00:15:36 超級智慧競賽

00:15:36 The race to super intelligence, which is incredibly important because imagine
00:15:36 超級智慧競賽非常重要，因為想像一下

00:15:38 incredibly important because imagine
00:15:38 非常重要，因為想像一下

00:15:38 incredibly important because imagine what a super intelligence could do that
00:15:38 非常重要，因為想像一下超級智慧可以做什麼

00:15:40 what a super intelligence could do that
00:15:40 超級智慧能做到什麼程度

00:15:40 what a super intelligence could do that we ourselves cannot imagine, right?
00:15:40 超級智慧能做什麼是我們自己無法想像的，對吧？

00:15:42 we ourselves cannot imagine, right?
00:15:42 我們自己都無法想像，對吧？

00:15:42 we ourselves cannot imagine, right? There it's so much smarter than we and
00:15:42 我們自己都無法想像，對吧？它比我們聰明得多，

00:15:45 There it's so much smarter than we and
00:15:45 它比我們聰明得多

00:15:45 There it's so much smarter than we and it has huge proliferation issues,
00:15:45 它比我們聰明得多，而且有嚴重的擴散問題，

00:15:47 it has huge proliferation issues,
00:15:47 它有嚴重的擴散問題，

00:15:47 it has huge proliferation issues, competitive issues, China versus the US
00:15:47 它存在巨大的擴散問題、競爭問題，中國與美國

00:15:49 competitive issues, China versus the US
00:15:49 競爭問題，中國與美國

00:15:49 competitive issues, China versus the US issues, electricity issues, so forth. We
00:15:49 競爭問題、中美問題、電力問題等等。我們

00:15:52 issues, electricity issues, so forth. We
00:15:52 電力問題等等。我們

00:15:52 issues, electricity issues, so forth. We don't even have the language for the
00:15:52 電力問題等等。我們甚至沒有語言來描述

00:15:54 don't even have the language for the
00:15:54 甚至沒有語言可以表達

00:15:54 don't even have the language for the deterrence aspects and the proliferation
00:15:54 甚至沒有關於威懾和擴散的語言

00:15:56 deterrence aspects and the proliferation
00:15:56 威懾方面和擴散

00:15:56 deterrence aspects and the proliferation issues of these powerful models
00:15:56 這些強大模型的威懾面向和擴散問題

00:15:58 issues of these powerful models
00:15:58 這些強大模型的問題

00:15:58 issues of these powerful models or the imagination.
00:15:58 這些強大模型的問題或想像。

00:15:59 or the imagination.
00:15:59 或想像。

00:15:59 or the imagination. Totally agree. In fact, it's it's one of
00:15:59 或想像。完全同意。事實上，這是

00:16:00 Totally agree. In fact, it's it's one of
00:16:00 完全同意。事實上，這是

00:16:00 Totally agree. In fact, it's it's one of the great flaws actually in the original
00:16:00 完全同意。事實上，這是原版最大的缺陷之一

00:16:02 the great flaws actually in the original
00:16:02 原文存在重大缺陷

00:16:02 the great flaws actually in the original conception. You remember Singularity
00:16:02 最初的設想其實有很大缺陷。你還記得奇點嗎？

00:16:04 conception. You remember Singularity
00:16:04 概念。你還記得奇點嗎

00:16:04 conception. You remember Singularity University and Ray Curtzwhile's books
00:16:04 概念。你還記得奇點大學和雷柯茨維爾的書嗎

00:16:06 University and Ray Curtzwhile's books
00:16:06 大學與雷‧柯茨維爾的書

00:16:06 University and Ray Curtzwhile's books and everything. And we kind of drew this
00:16:06 大學和雷·柯茨維爾的書等等。我們畫了這樣一幅畫

00:16:08 and everything. And we kind of drew this
00:16:08 等等。我們畫了這樣一幅畫

00:16:08 and everything. And we kind of drew this curve of rat level intelligence, then
00:16:08 等等。我們畫出了老鼠智力水平的曲線，然後

00:16:10 curve of rat level intelligence, then
00:16:10 老鼠智力水平曲線，然後

00:16:10 curve of rat level intelligence, then cat, then monkey, and then it hits human
00:16:10 老鼠智力曲線，然後是貓，然後是猴子，最後是人類

00:16:12 cat, then monkey, and then it hits human
00:16:12 先是貓，然後是猴子，最後是人類

00:16:12 cat, then monkey, and then it hits human and then it goes super intelligent. But
00:16:12 貓，然後是猴子，然後是人類，然後是超級智慧。但是

00:16:14 and then it goes super intelligent. But
00:16:14 然後它就變得超級聰明。但是

00:16:14 and then it goes super intelligent. But it's now really obvious when you talk to
00:16:14 然後它就變得非常聰明。但現在當你和

00:16:16 it's now really obvious when you talk to
00:16:16 現在當你和

00:16:16 it's now really obvious when you talk to one of these multilingual models that's
00:16:16 當你與這些多語言模型交談時，這一點很明顯

00:16:19 one of these multilingual models that's
00:16:19 這些多語言模型之一

00:16:19 one of these multilingual models that's explaining physics to you that it's
00:16:19 這些多語言模型之一可以向你解釋物理學

00:16:21 explaining physics to you that it's
00:16:21 向你解釋物理學

00:16:21 explaining physics to you that it's already hugely super intelligent within
00:16:21 向你解釋物理學，它已經非常聰明了

00:16:23 already hugely super intelligent within
00:16:23 已經非常智能

00:16:23 already hugely super intelligent within its savant category. And so Dennis keeps
00:16:23 已經是天才型的超級智商了。所以丹尼斯繼續

00:16:26 its savant category. And so Dennis keeps
00:16:26 屬於天才類別。所以丹尼斯

00:16:26 its savant category. And so Dennis keeps redefining AGI day as well when it can
00:16:26 是專家類別。所以丹尼斯一直在重新定義通用人工智慧，當它能夠

00:16:29 redefining AGI day as well when it can
00:16:29 重新定義 AGI 日

00:16:29 redefining AGI day as well when it can discover relativity the same way
00:16:29 重新定義 AGI 的日子，當它能以同樣的方式發現相對論時

00:16:30 discover relativity the same way
00:16:30 用同樣的方式發現相對論

00:16:30 discover relativity the same way Einstein did with data that was
00:16:30 用愛因斯坦的方法發現相對論，

00:16:32 Einstein did with data that was
00:16:32 愛因斯坦用數據

00:16:32 Einstein did with data that was available up until that date. That's
00:16:32 愛因斯坦利用當時已有的數據進行了計算。

00:16:34 available up until that date. That's
00:16:34 截至該日期。

00:16:34 available up until that date. That's when we have AGI.
00:16:34 直到那個日期為止都可用。那時我們就有了通用人工智慧 (AGI)。

00:16:35 when we have AGI.
00:16:35 當我們擁有 AGI 時。

00:16:35 when we have AGI. So long before that.
00:16:35 當我們擁有通用人工智慧的時候。這比那要早得多。

00:16:37 So long before that.
00:16:37 那是很久以前了。

00:16:37 So long before that. Yeah. So I think it's worth getting the
00:16:37 很久以前。是的。所以我覺得值得

00:16:38 Yeah. So I think it's worth getting the
00:16:38 是的。所以我認為值得

00:16:38 Yeah. So I think it's worth getting the timeline right.
00:16:38 是的。所以我認為有必要確定時間軸。

00:16:39 timeline right.  00:16:39 時間軸右側。

00:16:39 timeline right. Yeah.
00:16:39 時間線正確。是的。

00:16:40 Yeah.  00:16:40 是的。

00:16:40 Yeah. So the following things are baked in.
00:16:40 是的。所以以下內容已經融入其中。

00:16:42 So the following things are baked in.
00:16:42 因此，以下內容已融入其中。

00:16:42 So the following things are baked in. You're going to have an agentic
00:16:42 因此，以下內容已包含在內。你將擁有一個代理

00:16:44 You're going to have an agentic
00:16:44 你將有一個代理

00:16:44 You're going to have an agentic revolution where agents are connected to
00:16:44 你將經歷一場代理革命，代理將連結到

00:16:46 revolution where agents are connected to
00:16:46 革命中特工與

00:16:46 revolution where agents are connected to solve business processes, government
00:16:46 革命，代理人連結起來解決業務流程，政府

00:16:48 solve business processes, government
00:16:48 解決業務流程、政府

00:16:48 solve business processes, government processes and so forth. They will be
00:16:48 解決商業流程、政府流程等等。他們將

00:16:51 processes and so forth. They will be
00:16:51 流程等等。它們將

00:16:51 processes and so forth. They will be adopted most quickly in companies in
00:16:51 流程等等。這些方法將在以下公司最快被採用：

00:16:54 adopted most quickly in companies in
00:16:54 在公司採用速度最快

00:16:54 adopted most quickly in companies in country companies that have a lot of
00:16:54 在擁有大量

00:16:55 country companies that have a lot of
00:16:55 擁有大量

00:16:55 country companies that have a lot of money and a lot of uh time latency
00:16:55 擁有大量資金和大量時間延遲的國家公司

00:16:58 money and a lot of uh time latency
00:16:58 金錢和大量的時間延遲

00:16:58 money and a lot of uh time latency issues at stake. It will adop be adopted
00:16:58 金錢和大量的時間延遲問題都岌岌可危。它將被採用

00:17:00 issues at stake. It will adop be adopted
00:17:00 問題。它將被採納

00:17:00 issues at stake. It will adop be adopted most slowly in places like government
00:17:00 問題至關重要。在政府等地方，採用速度最慢

00:17:02 most slowly in places like government
00:17:02 最慢的地方是政府

00:17:02 most slowly in places like government which do not have an incentive for
00:17:02 最慢的地方是政府，因為政府沒有動力

00:17:04 which do not have an incentive for
00:17:04 沒有動力

00:17:04 which do not have an incentive for innovation. Um and fundamentally are job
00:17:04 沒有創新動力。嗯，從根本上來說，

00:17:07 innovation. Um and fundamentally are job
00:17:07 創新。嗯，從根本上來說，就是工作

00:17:07 innovation. Um and fundamentally are job programs and redistribution of income
00:17:07 創新。嗯，根本是就業計劃和收入再分配

00:17:09 programs and redistribution of income
00:17:09 計劃和收入再分配

00:17:09 programs and redistribution of income kind of programs.
00:17:09 計劃和收入再分配類計劃。

00:17:10 kind of programs.
00:17:10 種程序。

00:17:10 kind of programs. So call it what you will. The important
00:17:10 這類程式。隨便你怎麼稱呼它。重要的是

00:17:12 So call it what you will. The important
00:17:12 隨便你怎麼稱呼它。重要的是

00:17:12 So call it what you will. The important thing is that there will be a tip of the
00:17:12 隨便你怎麼稱呼它。重要的是，

00:17:13 thing is that there will be a tip of the
00:17:13 事情是會有提示的

00:17:13 thing is that there will be a tip of the spear in places like financial services,
00:17:13 問題是，在金融服務等領域，

00:17:17 spear in places like financial services,
00:17:17 涉足金融服務等領域，

00:17:17 spear in places like financial services, certain kind of bio biomedical things,
00:17:17 涉足金融服務、某些生物醫學領域，

00:17:19 certain kind of bio biomedical things,
00:17:19 某些生物醫學的東西，

00:17:19 certain kind of bio biomedical things, startups and so forth. And that's the
00:17:19 某些生物醫學領域，新創公司等等。這就是

00:17:20 startups and so forth. And that's the
00:17:20 新創公司等等。這就是

00:17:20 startups and so forth. And that's the place to watch. So all of that is going
00:17:20 新創公司等等。這是值得關注的地方。所以所有這些都將

00:17:24 place to watch. So all of that is going
00:17:24 值得關注的地方。所以所有這些都將

00:17:24 place to watch. So all of that is going to happen. The agents are going to
00:17:24 值得關注的地方。所以這一切都會發生。特工們將

00:17:26 to happen. The agents are going to
00:17:26 發生。特工們將

00:17:26 to happen. The agents are going to happen. This math thing is going to
00:17:26 會發生。代理將會發生。這個數學問題將

00:17:27 happen. This math thing is going to
00:17:27 發生。這個數學問題將會

00:17:28 happen. This math thing is going to happen. The software thing is going to
00:17:28 會發生。數學的事情會發生。軟體的事情也會發生。

00:17:29 happen. The software thing is going to
00:17:29 會發生。軟體的事情將會

00:17:29 happen. The software thing is going to happen. We can debate the rate at which
00:17:29 會發生。軟體的事情會發生。我們可以討論一下

00:17:31 happen. We can debate the rate at which
00:17:31 發生。我們可以討論

00:17:31 happen. We can debate the rate at which the biological revolution will occur,
00:17:31 發生。我們可以討論生物革命發生的速度，

00:17:33 the biological revolution will occur,
00:17:33 生物革命將會發生，

00:17:34 the biological revolution will occur, but everyone agrees that it's right
00:17:34 生物革命將會發生，但每個人都同意這是正確的

00:17:36 but everyone agrees that it's right
00:17:36 但每個人都同意這是正確的

00:17:36 but everyone agrees that it's right after that. We're very close to these
00:17:36 但大家都同意它就在那之後。我們非常接近這些

00:17:38 after that. We're very close to these
00:17:38 之後。我們非常接近這些

00:17:38 after that. We're very close to these major biological understandings. Um in
00:17:38 之後。我們非常接近這些主要的生物學理解。嗯

00:17:41 major biological understandings. Um in
00:17:41 主要的生物學理解。嗯

00:17:41 major biological understandings. Um in physics you're limited by data but you
00:17:41 主要的生物學理解。嗯，在物理學中，你會受到數據的限制，但你

00:17:43 physics you're limited by data but you
00:17:43 物理學你受到數據的限制，但是你

00:17:43 physics you're limited by data but you can generate it synthetically. There are
00:17:43 物理學會受到資料的限制，但你可以合成資料。

00:17:45 can generate it synthetically. There are
00:17:45 可以合成。有

00:17:45 can generate it synthetically. There are groups which I'm funding which are
00:17:45 可以合成它。我資助的一些團隊

00:17:47 groups which I'm funding which are
00:17:47 我資助的團體

00:17:47 groups which I'm funding which are generating physics um essentially um
00:17:47 我資助的團體正在研究物理學

00:17:50 generating physics um essentially um
00:17:50 生成物理 嗯 本質上 嗯

00:17:50 generating physics um essentially um models that can approximate algorithms
00:17:50 產生物理模型，本質上是可以近似演算法的模型

00:17:53 models that can approximate algorithms
00:17:53 可以近似演算法的模型

00:17:53 models that can approximate algorithms that cannot be they're incomputable. So
00:17:53 可以近似演算法的模型是不可計算的。所以

00:17:55 that cannot be they're incomputable. So
00:17:55 不可能，它們是不可計算的。所以

00:17:55 that cannot be they're incomputable. So in other words you have a a essentially
00:17:55 這不可能，它們是不可計算的。換句話說，你有一個本質上

00:17:57 in other words you have a a essentially
00:17:57 換句話說你基本上有一個

00:17:57 in other words you have a a essentially a foundation model that can answer the
00:17:57 換句話說，你有一個可以回答這個問題的基礎模型

00:17:59 a foundation model that can answer the
00:17:59 一個可以回答這個問題的基礎模型

00:17:59 a foundation model that can answer the question good enough for the purposes of
00:17:59 一個基礎模型，可以很好地回答這個問題

00:18:01 question good enough for the purposes of
00:18:01 這個問題夠好了

00:18:01 question good enough for the purposes of doing physics without having to spend a
00:18:01 這個問題對於做物理研究來說已經夠好了，而不需要花費

00:18:03 doing physics without having to spend a
00:18:03 無需花費

00:18:03 doing physics without having to spend a million years doing the computation of
00:18:03 做物理研究，而不必花一百萬年的時間進行計算

00:18:05 million years doing the computation of
00:18:05 百萬年進行計算

00:18:05 million years doing the computation of you know quantum chromodnamics and
00:18:05 百萬年進行量子色動力學計算

00:18:07 you know quantum chromodnamics and
00:18:07 你知道量子色動力學和

00:18:07 you know quantum chromodnamics and things like that. Yep.
00:18:07 你知道量子色動力學之類的東西。是的。

00:18:08 things like that. Yep.
00:18:08 諸如此類。是的。

00:18:08 things like that. Yep. Um, all of that's going to happen.
00:18:08 諸如此類的事。是的。嗯，這些都會發生。

00:18:11 Um, all of that's going to happen.
00:18:11 嗯，所有這些都會發生。

00:18:11 Um, all of that's going to happen. The next questions have to do with what
00:18:11 嗯，所有這些都會發生。接下來的問題是

00:18:14 The next questions have to do with what
00:18:14 接下來的問題與

00:18:14 The next questions have to do with what is the point in which this becomes a
00:18:14 接下來的問題是，這在什麼時候會變成一個

00:18:17 is the point in which this becomes a
00:18:17 是這成為一個

00:18:17 is the point in which this becomes a national emergency
00:18:17 是這成為國家緊急狀態的時刻

00:18:19 national emergency
00:18:19 國家緊急狀態

00:18:20 national emergency and it goes something like this.
00:18:20 國家緊急狀態，情況如下。

00:18:22 and it goes something like this.
00:18:22 事情是這樣的。

00:18:22 and it goes something like this. Everything I've talked about is in the
00:18:22 大致如下。我討論的所有內容都在

00:18:24 Everything I've talked about is in the
00:18:24 我所談論的一切都在

00:18:24 Everything I've talked about is in the positive domain, but there's a negative
00:18:24 我談論的都是正面的方面，但也存在著負面的方面

00:18:26 positive domain, but there's a negative
00:18:26 正域，但有一個負域

00:18:26 positive domain, but there's a negative domain as well. The ability for
00:18:26 正域，但也有負域。

00:18:28 domain as well. The ability for
00:18:28 域。

00:18:28 domain as well. The ability for biological attacks, um, uh, obviously
00:18:28 領域也是如此。生物攻擊的能力，嗯，呃，顯然

00:18:31 biological attacks, um, uh, obviously
00:18:31 生物攻擊，嗯，呃，顯然

00:18:31 biological attacks, um, uh, obviously cyber attacks. Imagine a cyber attack
00:18:31 生物攻擊，嗯，呃，顯然是網路攻擊。想像網路攻擊

00:18:34 cyber attacks. Imagine a cyber attack
00:18:34 網路攻擊。想像網路攻擊

00:18:34 cyber attacks. Imagine a cyber attack that we as humans cannot conceive of,
00:18:34 網路攻擊。想像一下我們人類無法想像的網路攻擊，

00:18:37 that we as humans cannot conceive of,
00:18:37 我們人類無法想像，

00:18:37 that we as humans cannot conceive of, which means there's no defense for it
00:18:37 我們人類無法想像，這意味著沒有防禦措施

00:18:38 which means there's no defense for it
00:18:38 這表示它無法防禦

00:18:38 which means there's no defense for it because no one ever thought about it.
00:18:38 這意味著對此沒有防禦措施，因為沒有人想到過它。

00:18:40 because no one ever thought about it.
00:18:40 因為沒有人想過這個問題。

00:18:40 because no one ever thought about it. Right? These are real issues. A
00:18:40 因為沒人想過這個問題。對吧？這些都是現實問題。

00:18:42 Right? These are real issues. A
00:18:42 對吧？這些都是現實問題。

00:18:42 Right? These are real issues. A biological attack, you take a virus, I
00:18:42 對吧？這些都是現實問題。生物攻擊，你感染了病毒，我

00:18:44 biological attack, you take a virus, I
00:18:44 生物攻擊，你感染病毒，我

00:18:44 biological attack, you take a virus, I won't obviously go into the details. You
00:18:44 生物攻擊，你會感染病毒，我當然不會詳述。你

00:18:46 won't obviously go into the details. You
00:18:46 顯然不會講細節。你

00:18:46 won't obviously go into the details. You take a virus that's bad and you make it
00:18:46 顯然不會講細節。你拿一個壞病毒，然後你製造它

00:18:49 take a virus that's bad and you make it
00:18:49 感染一種有害病毒，然後你就可以製造它

00:18:49 take a virus that's bad and you make it undetectable by some changes in its
00:18:49 拿一個有害的病毒，透過改變它的某些部分使它無法被偵測到

00:18:52 undetectable by some changes in its
00:18:52 無法透過其內部的一些變化來檢測

00:18:52 undetectable by some changes in its structure, which again I won't go into
00:18:52 無法透過結構上的一些變化來檢測，我不會再深入討論

00:18:53 structure, which again I won't go into
00:18:53 結構，我不會再深入討論

00:18:53 structure, which again I won't go into the details. We released a whole report
00:18:53 結構，我不再贅述。我們發布了一份完整的報告

00:18:55 the details. We released a whole report
00:18:55 細節。我們發布了一份完整的報告

00:18:55 the details. We released a whole report at the national level on this issue. So
00:18:55 細節。我們在國家層級就這個問題發布了一份完整的報告。所以

00:18:59 at the national level on this issue. So
00:18:59 在國家層面。所以

00:18:59 at the national level on this issue. So at some point the government and not it
00:18:59 在國家層面。所以某種程度上，政府而不是

00:19:02 at some point the government and not it
00:19:02 在某種程度上，政府而不是它

00:19:02 at some point the government and not it doesn't appear to understand this now is
00:19:02 在某種程度上，政府似乎不明白這一點

00:19:03 doesn't appear to understand this now is
00:19:03 似乎現在不明白這一點

00:19:04 doesn't appear to understand this now is going to have to say this is very big
00:19:04 似乎不明白這一點現在不得不說這是非常大的

00:19:07 going to have to say this is very big
00:19:07 不得不說這非常重要

00:19:07 going to have to say this is very big because it affects national security,
00:19:07 必須說這是件大事，因為它影響國家安全，

00:19:09 because it affects national security,
00:19:09 因為它影響國家安全，

00:19:09 because it affects national security, national economic strengths and so
00:19:09 因為它影響國家安全、國家經濟實力等等

00:19:10 national economic strengths and so
00:19:10 國家經濟實力等

00:19:10 national economic strengths and so forth. Now China clearly understands
00:19:10 國家經濟實力等等。現在中國清楚地明白

00:19:13 forth. Now China clearly understands
00:19:13 現在中國清楚明白

00:19:13 forth. Now China clearly understands this and China is putting an enormous
00:19:13。現在中國清楚地認識到了這一點，並且正在投入巨大的

00:19:16 this and China is putting an enormous
00:19:16 中國正在投入巨大的

00:19:16 this and China is putting an enormous amount of money into this. We have
00:19:16 中國正為此投入巨額資金。我們

00:19:18 amount of money into this. We have
00:19:18 投入的經費。我們有

00:19:18 amount of money into this. We have slowed them down by virtue of our chips
00:19:18 我們投入了大量資金。我們利用晶片減緩了他們的速度

00:19:21 slowed them down by virtue of our chips
00:19:21 借助我們的晶片減慢了他們的速度

00:19:21 slowed them down by virtue of our chips controls but they found clever ways
00:19:21 我們利用晶片控制來減慢他們的速度，但他們找到了聰明的方法

00:19:23 controls but they found clever ways
00:19:23 但他們找到了巧妙的方法

00:19:23 controls but they found clever ways around this. There are also
00:19:23 但他們找到了巧妙的方法來解決這個問題。還有

00:19:25 around this. There are also
00:19:25 左右。還有

00:19:25 around this. There are also proliferation issues. Many of the chips
00:19:25 圍繞這一點。還有擴散問題。許多晶片

00:19:27 proliferation issues. Many of the chips
00:19:27 擴散問題。許多晶片

00:19:27 proliferation issues. Many of the chips that they're not supposed to have, they
00:19:27 擴散問題。許多他們不該擁有的晶片，

00:19:29 that they're not supposed to have, they
00:19:29 他們不應該擁有

00:19:29 that they're not supposed to have, they seem to be able to get. And more
00:19:29 他們本不該擁有的東西，卻似乎都能得到。還有更多

00:19:30 seem to be able to get. And more
00:19:30 似乎可以得到。還有更多

00:19:30 seem to be able to get. And more importantly, as I mentioned, the
00:19:30 似乎能夠得到。更重要的是，正如我所提到的，

00:19:32 importantly, as I mentioned, the
00:19:32 重要的是，正如我所提到的，

00:19:32 importantly, as I mentioned, the algorithms are changing. And instead of
00:19:32 重要的是，正如我所提到的，演算法正在改變。

00:19:34 algorithms are changing. And instead of
00:19:34 演算法正在改變。

00:19:34 algorithms are changing. And instead of having these expensive foundation models
00:19:34 演算法正在改變。這些昂貴的基礎模型

00:19:36 having these expensive foundation models
00:19:36 擁有這些昂貴的基礎模型

00:19:36 having these expensive foundation models by themselves, you have continuous
00:19:36 擁有這些昂貴的基礎模型，你就有持續的

00:19:38 by themselves, you have continuous
00:19:38 你自己，你有連續的

00:19:38 by themselves, you have continuous updating, which is called test time
00:19:38 你自己不斷更新，稱為測試時間

00:19:39 updating, which is called test time
00:19:39 更新，稱為測試時間

00:19:39 updating, which is called test time training. That continuous updating
00:19:39 更新，這稱為測試時間訓練。持續更新

00:19:41 training. That continuous updating
00:19:41 訓練。持續更新

00:19:41 training. That continuous updating appears to be capable of being done with
00:19:41 訓練。這種持續更新似乎可以透過

00:19:44 appears to be capable of being done with
00:19:44 似乎可以完成

00:19:44 appears to be capable of being done with lesser power chips. So, so we I there
00:19:44 似乎可以用功率較小的晶片來實現。所以，我們

00:19:47 lesser power chips. So, so we I there
00:19:47 功率較小的晶片。所以，我們在那裡

00:19:47 lesser power chips. So, so we I there are so many questions that I think we
00:19:47 功率較小的晶片。所以，我們有很多問題，我認為我們

00:19:49 are so many questions that I think we
00:19:49 有很多問題，我認為我們

00:19:49 are so many questions that I think we don't know. We don't know the role of
00:19:49 我覺得還有很多問題我們不知道。我們不知道

00:19:51 don't know. We don't know the role of
00:19:51 不知道。我們不知道

00:19:51 don't know. We don't know the role of open source because remember open source
00:19:51 不知道。我們不知道開源的作用，因為記得開源

00:19:53 open source because remember open source
00:19:53 開源，因為記得開源

00:19:53 open source because remember open source means open weights, which means everyone
00:19:53 開源，記住開源意味著開放權重，這意味著每個人

00:19:55 means open weights, which means everyone
00:19:55 表示開放權重，這意味著每個人

00:19:55 means open weights, which means everyone can use it. A fair reading of this is
00:19:55 表示開放權重，這意味著每個人都可以使用它。對此的合理解讀是

00:19:57 can use it. A fair reading of this is
00:19:57 可以使用它。對此的合理解讀是

00:19:57 can use it. A fair reading of this is that every country that's not in the
00:19:57 可以使用它。公平的解讀是，每一個不屬於

00:19:59 that every country that's not in the
00:19:59 所有不在

00:19:59 that every country that's not in the west will end up using open source
00:19:59 所有非西方國家最終都會使用開源

00:20:01 west will end up using open source
00:20:01 西方最終將使用開源

00:20:01 west will end up using open source because they'll perceive it as cheaper
00:20:01 西方最終會使用開源，因為他們認為開源更便宜

00:20:03 because they'll perceive it as cheaper
00:20:03 因為他們會認為它更便宜

00:20:03 because they'll perceive it as cheaper which trans transfers leadership in open
00:20:03 因為他們會認為這比較便宜，跨性別者在公開場合轉移領導權

00:20:05 which trans transfers leadership in open
00:20:05 哪些跨性別者公開轉移領導權

00:20:05 which trans transfers leadership in open source from America to China. That's a
00:20:05 哪些轉變將開源的領導權從美國轉移到了中國。這是一個

00:20:07 source from America to China. That's a
00:20:07 從美國到中國。這是一個

00:20:07 source from America to China. That's a big deal, right? If that occurs.
00:20:07 病毒來源從美國轉移到中國。這可是個大問題，對吧？如果真的發生了。

00:20:09 big deal, right? If that occurs.
00:20:09 沒什麼大不了的，對吧？如果真的發生了。

00:20:09 big deal, right? If that occurs. Um how much longer do the chip bans if
00:20:09 有什麼大不了的？如果發生這種情況。嗯，如果

00:20:12 Um how much longer do the chip bans if
00:20:12 嗯，如果晶片禁令持續多久

00:20:12 Um how much longer do the chip bans if you will hold and how long before China
00:20:12 嗯，如果禁令繼續執行，中國什麼時候才能

00:20:14 you will hold and how long before China
00:20:14 你們會撐多久？中國

00:20:14 you will hold and how long before China can answer?
00:20:14 你們會撐多久，中國才能回答？

00:20:16 can answer?  00:20:16 可以解答嗎？

00:20:16 can answer? What are the effects of the current uh
00:20:16 能回答嗎？目前的影響是什麼？

00:20:18 What are the effects of the current uh
00:20:18 目前的影響是什麼

00:20:18 What are the effects of the current uh government's policies of getting rid of
00:20:18 當前政府的政策擺脫

00:20:20 government's policies of getting rid of
00:20:20 政府的政策擺脫

00:20:20 government's policies of getting rid of foreigners and foreign investment? what
00:20:20 政府驅逐外國人和外國投資的政策是什麼？

00:20:22 foreigners and foreign investment? what
00:20:22 外國人和外國投資？什麼

00:20:22 foreigners and foreign investment? what happens with the Arab U data centers
00:20:22 外國人和外國投資？阿拉伯 U 資料中心的情況如何？

00:20:24 happens with the Arab U data centers
00:20:24 發生在阿拉伯 U 資料中心

00:20:24 happens with the Arab U data centers assuming they work and I'm generally
00:20:24 發生在阿拉伯 U 資料中心，假設他們工作，我一般

00:20:26 assuming they work and I'm generally
00:20:26 假設它們有效，我通常

00:20:26 assuming they work and I'm generally supportive of them um if those things
00:20:26 假設它們有效，我通常會支持它們，嗯，如果這些事情

00:20:29 supportive of them um if those things
00:20:29 支持他們，如果這些事情

00:20:29 supportive of them um if those things are then misused uh to help train train
00:20:29 支持他們，如果這些東西被濫用，呃，以幫助訓練

00:20:32 are then misused uh to help train train
00:20:32 然後被濫用來幫忙訓練

00:20:32 are then misused uh to help train train models. The list just goes on and on. We
00:20:32 然後被濫用來訓練模型。這樣的例子不勝枚舉。我們

00:20:34 models. The list just goes on and on. We
00:20:34 型號。這個名單很長。我們

00:20:34 models. The list just goes on and on. We just don't know. Okay. Can I ask you
00:20:34 型號。這個清單很長。我們不知道。好的。我可以問你

00:20:36 just don't know. Okay. Can I ask you
00:20:36 就是不知道。好的。我可以問你嗎

00:20:36 just don't know. Okay. Can I ask you probably one of the toughest questions?
00:20:36 就是不知道。好的。我可以問你一個可能是最難的問題嗎？

00:20:37 probably one of the toughest questions?
00:20:37 這可能是最棘手的問題之一？

00:20:38 probably one of the toughest questions? I don't know if you saw Mark Andre
00:20:38 這可能是最難的問題之一？我不知道你有沒有見過馬克安德烈

00:20:40 I don't know if you saw Mark Andre
00:20:40 我不知道你是否見過馬克安德烈

00:20:40 I don't know if you saw Mark Andre uh he went and talked to the Biden
00:20:40 我不知道你是否看到馬克安德烈，他去和拜登談話

00:20:42 uh he went and talked to the Biden
00:20:42 呃，他去跟拜登談了

00:20:42 uh he went and talked to the Biden administration past administration and
00:20:42 呃，他去跟拜登政府談了談，以及

00:20:44 administration past administration and
00:20:44 政府過去政府和

00:20:44 administration past administration and said how are we going to deal with
00:20:44 政府上一屆政府說，我們將如何處理

00:20:45 said how are we going to deal with
00:20:45 說我們該如何處理

00:20:45 said how are we going to deal with exactly what you just talked about
00:20:45 我們該如何處理你剛才談到的問題

00:20:47 exactly what you just talked about
00:20:47 正是你剛才說的

00:20:47 exactly what you just talked about chemical and biological and radiological
00:20:47 正是你剛才談到的化學、生物和放射學

00:20:49 chemical and biological and radiological
00:20:49 化學、生物和放射學

00:20:49 chemical and biological and radiological and nuclear risks from big foundation
00:20:49 大基金會的化學、生物、放射與核子風險

00:20:51 and nuclear risks from big foundation
00:20:51 以及大基金會的核子風險

00:20:51 and nuclear risks from big foundation models being operated by foreign
00:20:51 以及外國營運的大型基礎模型的核子風險

00:20:53 models being operated by foreign
00:20:53 外國營運的模型

00:20:53 models being operated by foreign countries. And the Biden answer was you
00:20:53 外國正在運作的模型。拜登的答案是你

00:20:56 countries. And the Biden answer was you
00:20:56 國。拜登的答案是你

00:20:56 countries. And the Biden answer was you know we're going to keep it into the
00:20:56 國。拜登的回答是，你知道我們會把它保留到

00:20:57 know we're going to keep it into the
00:20:57 知道我們會把它保留到

00:20:57 know we're going to keep it into the three or four big companies like Google
00:20:57 我們知道我們會把它保留在谷歌這樣的三四家大公司中

00:21:00 three or four big companies like Google
00:21:00 像 Google 這樣的三、四家大公司

00:21:00 three or four big companies like Google and we'll just regulate them. And Mark
00:21:00 像 Google 這樣的三、四家大公司，我們會對它們進行監管。馬克

00:21:03 and we'll just regulate them. And Mark
00:21:03 我們會監管它們。馬克

00:21:03 and we'll just regulate them. And Mark was like, "That is a surefire way to
00:21:03 我們會監管它們。馬克說：「這肯定能

00:21:05 was like, "That is a surefire way to
00:21:05 就像是，「這是一個萬無一失的方法

00:21:05 was like, "That is a surefire way to lose the race with China because all
00:21:05 就像是，「這肯定會輸掉與中國的競爭，因為所有

00:21:08 lose the race with China because all
00:21:08 輸掉與中國的競爭，因為所有

00:21:08 lose the race with China because all innovation comes from a startup that you
00:21:08 輸掉與中國的競爭，因為所有的創新都來自於你

00:21:10 innovation comes from a startup that you
00:21:10 創新來自於你創辦的新創公司

00:21:10 innovation comes from a startup that you didn't anticipate or you know it's just
00:21:10 創新來自於你沒有預料到的創業，或者你知道這只是

00:21:12 didn't anticipate or you know it's just
00:21:12 沒有預料到，或者你知道這只是

00:21:12 didn't anticipate or you know it's just the American history and you're you're
00:21:12 沒有預料到，或者你知道這只是美國歷史，你

00:21:14 the American history and you're you're
00:21:14 美國史，你

00:21:14 the American history and you're you're cutting off the entrepreneur from
00:21:14 美國歷史，你正在切斷企業家與

00:21:16 cutting off the entrepreneur from
00:21:16 切斷創業家

00:21:16 cutting off the entrepreneur from participating in this." So as of right
00:21:16 阻止企業家參與此事。 」因此

00:21:18 participating in this." So as of right
00:21:18 參與其中。 」所以

00:21:18 participating in this." So as of right now with the open source models, the
00:21:18 參與其中。 」所以就目前開源模型而言，

00:21:20 now with the open source models, the
00:21:20 現在有了開源模型，

00:21:20 now with the open source models, the entrepreneurs are in great shape. But if
00:21:20 現在有了開源模式，創業者們的狀態很好。但如果

00:21:22 entrepreneurs are in great shape. But if
00:21:22 企業家們的狀態很好。但如果

00:21:22 entrepreneurs are in great shape. But if you think about the models getting crazy
00:21:22 企業家們狀態很好。但如果你想想那些模特兒變得瘋狂

00:21:24 you think about the models getting crazy
00:21:24 你認為模特兒會變得瘋狂

00:21:24 you think about the models getting crazy smart a year from now, how are we going
00:21:24 想像一年後模型會變得多麼聰明，我們該怎麼辦

00:21:26 smart a year from now, how are we going
00:21:26 一年後，我們會怎樣

00:21:26 smart a year from now, how are we going to have the the balance between startups
00:21:26 一年後，我們將如何在新創公司之間取得平衡

00:21:29 to have the the balance between startups
00:21:29 在新創企業之間取得平衡

00:21:29 to have the the balance between startups actually being able to work with the
00:21:29 找到新創公司與

00:21:31 actually being able to work with the
00:21:31 實際上能夠與

00:21:31 actually being able to work with the best technology but proliferation not
00:21:31 實際上能夠使用最好的技術，但擴散卻不是

00:21:34 best technology but proliferation not
00:21:34 最好的技術，但擴散不是

00:21:34 best technology but proliferation not percolating to every country in the
00:21:34 最好的技術，但擴散並沒有滲透到每個國家

00:21:36 percolating to every country in the
00:21:36 滲透到每個國家

00:21:36 percolating to every country in the world.
00:21:36 滲透到世界每個國家。

00:21:37 world.  00:21:37 世界。

00:21:37 world. Again, a set of unknown questions and
00:21:37 世界。又有一系列未知的問題和

00:21:38 Again, a set of unknown questions and
00:21:38 再次，一組未知的問題和

00:21:38 Again, a set of unknown questions and anybody who knows the answer to these
00:21:38 再一次，一連串未知的問題，任何知道答案的人

00:21:40 anybody who knows the answer to these
00:21:40 任何知道答案的人

00:21:40 anybody who knows the answer to these things is not telling the full truth.
00:21:40 任何知道這些問題答案的人都沒有說出全部真相。

00:21:42 things is not telling the full truth.
00:21:42 事情並沒有說出全部真相。

00:21:42 things is not telling the full truth. Um the doctrine in the B administration
00:21:42 事情並沒有完全真相。嗯，B 政府的信條

00:21:45 Um the doctrine in the B administration
00:21:45 嗯，B 政府的理論

00:21:45 Um the doctrine in the B administration was called 10 to the 26 flops. It was a
00:21:45 嗯，B 政府的理論被稱為「10 到 26 次失敗」。這是一個

00:21:48 was called 10 to the 26 flops. It was a
00:21:48 稱為 10 到 26 翻牌。這是

00:21:48 was called 10 to the 26 flops. It was a point that was a consensus above which
00:21:48 被稱為 10 比 26 的假摔。這是一個共識，高於這個點

00:21:51 point that was a consensus above which
00:21:51 這是一項共識，

00:21:51 point that was a consensus above which the models were powerful enough to cause
00:21:51 點，這是共識，高於這個點，模型就足夠強大，可以導致

00:21:54 the models were powerful enough to cause
00:21:54 這些模型夠強大，可以

00:21:54 the models were powerful enough to cause some damage. So the theory was that if
00:21:54 這些模型夠強大，足以造成一些損害。所以理論是，如果

00:21:56 some damage. So the theory was that if
00:21:56 一些損害。所以理論是，如果

00:21:56 some damage. So the theory was that if you stayed below 10 the 26 you didn't
00:21:56 一些損害。所以理論上，如果你保持在10以下，那麼26你就不會

00:21:59 you stayed below 10 the 26 you didn't
00:21:59 你保持在 10 以下，而 26 你沒有

00:21:59 you stayed below 10 the 26 you didn't need to be regulated.
00:21:59 如果你保持在 10 以下，那麼 26 你就不需要受到監管。

00:22:00 need to be regulated.
00:22:00 需要監管。

00:22:00 need to be regulated. But if you were above that you needed to
00:22:00 需要監管。但如果你的收入高於這個水平，你就需要

00:22:02 But if you were above that you needed to
00:22:02 但如果你高於這個水平，你需要

00:22:02 But if you were above that you needed to be regulated. And the proposal in the
00:22:02 但如果你超出了這一點，你就需要受到監管。而

00:22:04 be regulated. And the proposal in the
00:22:04 受到監管。

00:22:04 be regulated. And the proposal in the Biden administration was to regulate
00:22:04 受到監管。拜登政府的提案是監管

00:22:06 Biden administration was to regulate
00:22:06 拜登政府將監管

00:22:06 Biden administration was to regulate both the open source and the closed
00:22:06 拜登政府將同時監管開源和閉源

00:22:07 both the open source and the closed
00:22:07 開源和閉源

00:22:08 both the open source and the closed source.
00:22:08 開源和閉源。

00:22:08 source.  00:22:08 來源。

00:22:08 source. Okay that's that's the those are the the
00:22:08 來源。好的，這就是這些

00:22:11 Okay that's that's the those are the the
00:22:11 好的，就是這樣，這些就是

00:22:11 Okay that's that's the those are the the summary
00:22:11 好的，這就是摘要

00:22:11 summary  00:22:11 摘要

00:22:11 summary that of course has been ended by the
00:22:11 摘要當然已經結束了

00:22:13 that of course has been ended by the
00:22:13 當然已經結束了

00:22:13 that of course has been ended by the Trump administration. um they have not
00:22:13 當然，川普政府已經終止了這項進程。嗯，他們還沒有

00:22:16 Trump administration. um they have not
00:22:16 川普政府。嗯，他們沒有

00:22:16 Trump administration. um they have not yet produced their own thinking in this
00:22:16 川普政府。嗯，他們還沒有在這個問題上提出自己的想法

00:22:18 yet produced their own thinking in this
00:22:18 卻產生了自己的想法

00:22:18 yet produced their own thinking in this area. They're very concerned about China
00:22:18 但在這個領域他們也有自己的想法。他們非常擔心中國

00:22:20 area. They're very concerned about China
00:22:20 地區。他們非常擔心中國

00:22:20 area. They're very concerned about China and it getting forward. So, they'll come
00:22:20 地區。他們非常關心中國及其發展。所以，他們會來

00:22:22 and it getting forward. So, they'll come
00:22:22 並且它正在向前發展。所以他們會來

00:22:22 and it getting forward. So, they'll come out with something. From my perspective,
00:22:22 並且它正在推進。所以，他們會拿出一些東西。從我的角度來看，

00:22:25 out with something. From my perspective,
00:22:25 出了點問題。從我的角度來看，

00:22:25 out with something. From my perspective, the the core questions are the
00:22:25 提出一些問題。在我看來，核心問題是

00:22:27 the the core questions are the
00:22:27 核心問題是

00:22:27 the the core questions are the following. Will the Chinese be able to
00:22:27 核心問題如下。中國能不能

00:22:30 following. Will the Chinese be able to
00:22:30 中國人能

00:22:30 following. Will the Chinese be able to use even with um chip restrictions, will
00:22:30 如下。中國人是否能夠使用，即使有晶片限制，

00:22:33 use even with um chip restrictions, will
00:22:33 即使有晶片限制，也會使用

00:22:33 use even with um chip restrictions, will they use architectural changes that will
00:22:33 即使有晶片限制，他們也會使用架構變化來

00:22:34 they use architectural changes that will
00:22:34 他們使用架構變化來

00:22:34 they use architectural changes that will allow them to build models as powerful
00:22:34 他們使用架構變化，這將使他們能夠建立強大的模型

00:22:36 allow them to build models as powerful
00:22:36 讓他們能夠建立強大的模型

00:22:36 allow them to build models as powerful as ours?
00:22:36 允許他們建立和我們一樣強大的模型？

00:22:37 as ours?  00:22:37 和我們的一樣嗎？

00:22:37 as ours? And let's assume they're government
00:22:37 和我們一樣？假設他們是政府

00:22:38 And let's assume they're government
00:22:38 假設他們是政府

00:22:38 And let's assume they're government funded. That's the first question. The
00:22:38 我們假設它們是由政府資助的。這是第一個問題。

00:22:41 funded. That's the first question. The
00:22:41 資金到位。這是第一個問題。

00:22:41 funded. That's the first question. The next fun question is how will you raise
00:22:41 資金到位。這是第一個問題。下一個有趣的問題是，你將如何籌集資金

00:22:44 next fun question is how will you raise
00:22:44 下一個有趣的問題是，你將如何籌集

00:22:44 next fun question is how will you raise $50 billion for your data center if your
00:22:44 下一個有趣的問題是，如果你的

00:22:47 $50 billion for your data center if your
00:22:47 如果你的資料中心

00:22:47 $50 billion for your data center if your product is open source?
00:22:47 如果您的產品是開源的，您的資料中心將獲得 500 億美元的收益？

00:22:48 product is open source?
00:22:48 產品是開源的嗎？

00:22:48 product is open source? Yeah.
00:22:48 產品是開源的嗎？是的。

00:22:49 Yeah.  00:22:49 是的。

00:22:49 Yeah. In the American model, part of the
00:22:49 是的。在美國模式中，

00:22:51 In the American model, part of the
00:22:51 在美國模式中，

00:22:51 In the American model, part of the reason these models are closed is that
00:22:51 在美國模式中，這些模型封閉的部分原因是

00:22:52 reason these models are closed is that
00:22:52 這些模型關閉的原因是

00:22:52 reason these models are closed is that the business people and the lawyers
00:22:52 這些模型被關閉的原因是商人和律師

00:22:54 the business people and the lawyers
00:22:54 商人和律師

00:22:54 the business people and the lawyers correctly are saying I've got to sell
00:22:54 商人和律師正確地說我必須賣掉

00:22:57 correctly are saying I've got to sell
00:22:57 正確地說我必須賣掉

00:22:57 correctly are saying I've got to sell this thing because I've got to pay for
00:22:57 正確地說我必須賣掉這個東西，因為我必須支付

00:22:58 this thing because I've got to pay for
00:22:58 因為我得付錢

00:22:58 this thing because I've got to pay for my capital. These are not free goods.
00:22:58 因為我得支付我的資本。這不是免費的商品。

00:23:00 my capital. These are not free goods.
00:23:00 我的資本。這些不是免費的。

00:23:00 my capital. These are not free goods. And the US government correctly is not
00:23:00 我的資本。這些不是免費的商品。美國政府正確地

00:23:02 And the US government correctly is not
00:23:02 美國政府並沒有正確

00:23:02 And the US government correctly is not giving $50 billion to these companies.
00:23:02 美國政府沒有向這些公司提供 500 億美元，這是正確的。

00:23:05 giving $50 billion to these companies.
00:23:05 向這些公司提供 500 億美元。

00:23:05 giving $50 billion to these companies. So we don't know that. Um the to me the
00:23:05 給這些公司500億美元。所以我們不知道。嗯，對我來說

00:23:09 So we don't know that. Um the to me the
00:23:09 所以我們不知道。嗯，對我來說

00:23:09 So we don't know that. Um the to me the key question to watch is look at
00:23:09 所以我們不知道。嗯，對我來說，關鍵問題是觀察

00:23:11 key question to watch is look at
00:23:11 需要關注的關鍵問題是

00:23:11 key question to watch is look at Deepseek. So Deepseek um a week or so
00:23:11 關鍵問題是關注 Deepseek。 Deepseek 大概一週左右

00:23:14 Deepseek. So Deepseek um a week or so
00:23:14 Deepseek。所以 Deepseek 大概一週左右

00:23:14 Deepseek. So Deepseek um a week or so ago Gemini 2.5 Pro got to the top of the
00:23:14 Deepseek。大約一週前，Gemini 2.5 Pro 登上了

00:23:18 ago Gemini 2.5 Pro got to the top of the
00:23:18 前 Gemini 2.5 Pro 登上

00:23:18 ago Gemini 2.5 Pro got to the top of the leaderboards in intelligence. Great
00:23:18 前 Gemini 2.5 Pro 登上了智慧排行榜榜首。太棒了

00:23:21 leaderboards in intelligence. Great
00:23:21 情報排行榜。太棒了

00:23:21 leaderboards in intelligence. Great achievement for my friends at Gem at
00:23:21 智力排行榜。我的朋友們在 Gem 取得了巨大的成就

00:23:22 achievement for my friends at Gem at
00:23:22 為我的朋友們在 Gem at 所取得的成就

00:23:22 achievement for my friends at Gem at Gemini. A week later deepseek comes in
00:23:22 為我在 Gemini 的 Gem 朋友們取得了成就。一週後，Deepseek 也來了

00:23:27 Gemini. A week later deepseek comes in
00:23:27 雙子座。一週後，Deepseek 來了

00:23:27 Gemini. A week later deepseek comes in and is slightly better than Gemini. and
00:23:27 Gemini。一週後，Deepseek 上線，比 Gemini 略好一些。

00:23:29 and is slightly better than Gemini. and
00:23:29 比 Gemini 好一點。

00:23:29 and is slightly better than Gemini. and Deeps of course is trained on the
00:23:29 比 Gemini 略好一點。 Deeps 當然是在

00:23:31 Deeps of course is trained on the
00:23:31 Deeps 當然在

00:23:31 Deeps of course is trained on the existing hardware that's in China which
00:23:31 Deeps 當然是在中國現有的硬體上進行訓練的，

00:23:33 existing hardware that's in China which
00:23:33 在中國現有的硬體

00:23:33 existing hardware that's in China which includes stuff that's been Pilford and
00:23:33 在中國現有的硬件，包括 Pilford 和

00:23:35 includes stuff that's been Pilford and
00:23:35 包括 Pilford 和

00:23:35 includes stuff that's been Pilford and some of the Ascend it's called the
00:23:35 包括 Pilford 和一些 Ascend 的東西，它被稱為

00:23:37 some of the Ascend it's called the
00:23:37 一些 Ascend 被稱為

00:23:37 some of the Ascend it's called the Ascend Huawei chips and a few others
00:23:37 其中一些被稱為 Ascend Huawei 晶片，還有一些其他的

00:23:41 Ascend Huawei chips and a few others
00:23:41 Ascend 華為晶片和其他一些

00:23:41 Ascend Huawei chips and a few others what happens now the US people say well
00:23:41 華為晶片和其他一些晶片現在發生了什麼，美國人民說好

00:23:45 what happens now the US people say well
00:23:45 現在發生了什麼，美國人民說好

00:23:45 what happens now the US people say well you know the the deepseek people cheated
00:23:45 現在發生了什麼，美國人說，你知道 Deepseek 的人被騙了

00:23:48 you know the the deepseek people cheated
00:23:48 你知道 Deepseek 的人作弊了嗎

00:23:48 you know the the deepseek people cheated and they cheated by doing a technique
00:23:48 你知道深度搜尋的人作弊了，他們透過一種技術作弊

00:23:50 and they cheated by doing a technique
00:23:50 他們用一種技巧作弊

00:23:50 and they cheated by doing a technique called distillation where you take a
00:23:50 他們透過一種叫做蒸餾的技術作弊，你取一個

00:23:52 called distillation where you take a
00:23:52 蒸餾，你取一個

00:23:52 called distillation where you take a large model and you ask it 10,000
00:23:52 稱為蒸餾，你拿一個大模型，然後問它 10,000

00:23:54 large model and you ask it 10,000
00:23:54 大型模型，你問它 10,000

00:23:54 large model and you ask it 10,000 questions you get its answers and then
00:23:54 大型模型，你問它 10,000 個問題，它會給出答案，然後

00:23:55 questions you get its answers and then
00:23:55 問題你得到答案然後

00:23:55 questions you get its answers and then then you use that as your training
00:23:55 問題，你會得到答案，然後你用它作為你的訓練

00:23:57 then you use that as your training
00:23:57 然後你用它作為你的訓練

00:23:57 then you use that as your training material
00:23:57 然後你用它作為你的訓練材料

00:23:57 material  00:23:57 材料

00:23:57 material yep  00:23:57 材料是的

00:23:57 yep  00:23:57 是的

00:23:58 yep so the US companies will have to figure
00:23:58 是的，所以美國公司必須弄清楚

00:23:59 so the US companies will have to figure
00:23:59 所以美國公司必須弄清楚

00:23:59 so the US companies will have to figure out a way to make sure that their
00:23:59 因此美國公司必須想辦法確保他們的

00:24:01 out a way to make sure that their
00:24:01 找到一種方法來確保他們的

00:24:01 out a way to make sure that their proprietary information that they've
00:24:01 找到一種方法來確保他們擁有的專有資訊

00:24:02 proprietary information that they've
00:24:02 他們擁有的專有資訊

00:24:02 proprietary information that they've spent so much money on does not get
00:24:02 他們花了這麼多錢獲得的專有資訊沒有得到

00:24:05 spent so much money on does not get
00:24:05 花了那麼多錢卻沒有得到

00:24:05 spent so much money on does not get leaked into these open source things. Um
00:24:05 花了這麼多錢，但資訊卻沒有洩漏到開源專案上。嗯

00:24:08 leaked into these open source things. Um
00:24:08 洩漏到這些開源專案中。嗯

00:24:08 leaked into these open source things. Um I just don't know with respect to uh
00:24:08 洩漏到這些開源專案中。嗯，我只是不知道

00:24:11 I just don't know with respect to uh
00:24:11 我只是不知道

00:24:11 I just don't know with respect to uh nuclear, biological, chemical and so
00:24:11 我只是不知道關於核、生物、化學等等

00:24:13 nuclear, biological, chemical and so
00:24:13 核、生物、化學等

00:24:13 nuclear, biological, chemical and so forth issues. Um the US companies are
00:24:13 核子、生物、化學等問題。嗯，美國公司

00:24:15 forth issues. Um the US companies are
00:24:15 第四個問題。嗯，美國公司

00:24:15 forth issues. Um the US companies are doing a really good job of looking for
00:24:15 第四點。嗯，美國公司在尋找

00:24:17 doing a really good job of looking for
00:24:17 好好尋找

00:24:17 doing a really good job of looking for that. There's a great concern, for
00:24:17 在這方面做得很好。人們非常擔心，

00:24:19 that. There's a great concern, for
00:24:19 這一點。人們非常擔心

00:24:19 that. There's a great concern, for example, that nuclear information would
00:24:19 這一點。例如，人們非常擔心核子資訊會

00:24:21 example, that nuclear information would
00:24:21 例如，核資訊將

00:24:22 example, that nuclear information would leak into these models as they're
00:24:22 例如，核資訊可能會洩漏到這些模型中，因為它們

00:24:23 leak into these models as they're
00:24:23 洩漏到這些模型中

00:24:24 leak into these models as they're training without us knowing it. And by
00:24:24 在我們不知情的情況下，這些模型在訓練過程中會洩漏這些資訊。

00:24:25 training without us knowing it. And by
00:24:25 訓練過程中，我們卻渾然不知。

00:24:25 training without us knowing it. And by the way, that's a violation of law.
00:24:25 在我們不知情的情況下進行訓練。順便說一句，這是違法的。

00:24:27 the way, that's a violation of law.
00:24:27 順便說一句，這是違法的。

00:24:27 the way, that's a violation of law. Oh, really? they work and the whole
00:24:27 順便說一句，這是違法的。哦，真的嗎？他們工作，而且整個

00:24:29 Oh, really? they work and the whole
00:24:29 哦，真的嗎？它們有效，而且整個

00:24:29 Oh, really? they work and the whole nuclear information thing is is there's
00:24:29 哦，真的嗎？它們有效，而且整個核子資訊都是

00:24:31 nuclear information thing is is there's
00:24:31 核子訊息的事情是

00:24:31 nuclear information thing is is there's no free speech in that world for good
00:24:31 核子資訊的問題在於，這個世界根本沒有言論自由

00:24:33 no free speech in that world for good
00:24:33 這世界永遠沒有言論自由

00:24:33 no free speech in that world for good reasons
00:24:33 這個世界沒有言論自由，這是有原因的

00:24:34 reasons  00:24:34 原因

00:24:34 reasons and there's no free use and copyright
00:24:34 原因，沒有免費使用和版權

00:24:36 and there's no free use and copyright
00:24:36 並且沒有免費使用和版權

00:24:36 and there's no free use and copyright and all that kind of stuff. It's illegal
00:24:36 並且沒有自由使用和版權之類的權利。這是非法的

00:24:37 and all that kind of stuff. It's illegal
00:24:37 諸如此類。這是違法的

00:24:37 and all that kind of stuff. It's illegal to do it and so they're doing a really
00:24:37 諸如此類的事。這樣做是違法的，所以他們真的

00:24:39 to do it and so they're doing a really
00:24:39 這樣做，所以他們真的在做

00:24:39 to do it and so they're doing a really really good job of making sure that that
00:24:39 這樣做，所以他們做得非常非常好，確保

00:24:41 really good job of making sure that that
00:24:41 確實做得很好，確保了

00:24:41 really good job of making sure that that does not happen. They also put in very
00:24:41 確保這種情況不會發生，做得很好。他們也投入了非常

00:24:44 does not happen. They also put in very
00:24:44 不會發生。他們也投入了非常

00:24:44 does not happen. They also put in very significant tests for biological
00:24:44 不會發生。他們也對生物進行了非常重要的測試

00:24:45 significant tests for biological
00:24:45 生物顯著性檢驗

00:24:45 significant tests for biological information and certain kinds of cyber
00:24:45 生物資訊和某些網路資訊的重要測試

00:24:47 information and certain kinds of cyber
00:24:47 資訊和某些類型的網絡

00:24:47 information and certain kinds of cyber attacks. What happens there? Their
00:24:47 資訊和某些類型的網路攻擊。會發生什麼事？他們的

00:24:50 attacks. What happens there? Their
00:24:50 襲擊。那裡發生了什麼事？他們的

00:24:50 attacks. What happens there? Their incentive is their incentive to continue
00:24:50 攻擊。那裡發生了什麼事？他們的動機就是繼續

00:24:51 incentive is their incentive to continue
00:24:51 激勵是他們繼續下去的動力

00:24:51 incentive is their incentive to continue especially if it's not if it's not
00:24:51 激勵是他們繼續下去的動力，特別是如果不是

00:24:53 especially if it's not if it's not
00:24:53 尤其是如果它不是

00:24:53 especially if it's not if it's not required by law. The government has just
00:24:53 特別是如果法律沒有要求的話。政府只是

00:24:56 required by law. The government has just
00:24:56 法律要求。政府剛剛

00:24:56 required by law. The government has just gotten rid of the the safety institutes
00:24:56 法律要求。政府剛取消了安全機構

00:24:58 gotten rid of the the safety institutes
00:24:58 擺脫了安全機構

00:24:58 gotten rid of the the safety institutes that were in place in Biden and are
00:24:58 擺脫了拜登時期設立的安全機構，

00:25:00 that were in place in Biden and are
00:25:00 拜登時代就已經存在了

00:25:00 that were in place in Biden and are replacing it by a new term which is
00:25:00 拜登時代就已經存在了，現在用一個新字來取代它，那就是

00:25:02 replacing it by a new term which is
00:25:02 用一個新術語取代它

00:25:02 replacing it by a new term which is largely a safety assessment program
00:25:02 用一個新術語取代它，主要是安全評估程序

00:25:05 largely a safety assessment program
00:25:05 主要是安全評估項目

00:25:05 largely a safety assessment program which is a fine answer. I think
00:25:05 主要是安全評估項目，這是一個很好的答案。我認為

00:25:07 which is a fine answer. I think
00:25:07 答案很好。我認為

00:25:07 which is a fine answer. I think collectively we in the industry just
00:25:07 答案很好。我認為我們這個行業整體上

00:25:10 collectively we in the industry just
00:25:10 我們這個行業

00:25:10 collectively we in the industry just want the government at the secret and
00:25:10 我們這個行業的人都希望政府能保守秘密

00:25:11 want the government at the secret and
00:25:11 想要政府知道這個秘密

00:25:12 want the government at the secret and top secret level to have people who are
00:25:12 希望政府在秘密和最高機密級別上擁有

00:25:14 top secret level to have people who are
00:25:14 絕密級別，

00:25:14 top secret level to have people who are really studying what China and others
00:25:14 絕密級別，讓真正研究中國和其他國家

00:25:16 really studying what China and others
00:25:16 真正研究中國和其他國家

00:25:16 really studying what China and others are doing. You can be sure that China
00:25:16 真正研究中國和其他國家正在做的事情。你可以肯定中國

00:25:18 are doing. You can be sure that China
00:25:18 正在做。你可以肯定中國

00:25:18 are doing. You can be sure that China really has very smart people studying
00:25:18 正在做。你可以肯定中國確實有很多非常聰明的人在學習

00:25:20 really has very smart people studying
00:25:20 確實有非常聰明的人在學習

00:25:20 really has very smart people studying what we're doing. We at the secret and
00:25:20 確實有很多聰明人研究我們在做什麼。我們處於秘密之中

00:25:23 what we're doing. We at the secret and
00:25:23 我們正在做什麼。我們處於秘密之中

00:25:23 what we're doing. We at the secret and top secret level should have the same
00:25:23 我們正在做什麼。我們在秘密和絕密級別應該有相同的

00:25:25 top secret level should have the same
00:25:25 最高機密等級應該有相同的

00:25:25 top secret level should have the same thing.
00:25:25 絕密等級應該也有同樣的東西。

00:25:25 thing.  00:25:25 事。

00:25:25 thing. Have you read the uh AI27 paper?
00:25:25 這件事。你有讀過 AI27 論文嗎？

00:25:28 Have you read the uh AI27 paper?
00:25:28 你有讀過 AI27 論文嗎？

00:25:28 Have you read the uh AI27 paper? I have. Uh, and so for those listening
00:25:28 你有讀過 AI27 的論文嗎？我讀過。嗯，所以對於那些正在聽的人來說

00:25:31 I have. Uh, and so for those listening
00:25:31 是的。呃，所以對於那些正在聽的人來說

00:25:31 I have. Uh, and so for those listening who haven't read it, it's a it's a
00:25:31 是的。嗯，所以對於那些還沒讀過的人來說，這是一個

00:25:33 who haven't read it, it's a it's a
00:25:33 誰沒讀過，這是一個這是一個

00:25:33 who haven't read it, it's a it's a future vision of the AI and US and China
00:25:33 那些沒讀過的人，這是對美國和中國人工智慧的未來展望

00:25:36 future vision of the AI and US and China
00:25:36 人工智慧的未來願景以及美國和中國

00:25:36 future vision of the AI and US and China racing towards AI and at some point the
00:25:36 人工智慧的未來願景，美國和中國競相向人工智慧邁進，在某個時候

00:25:40 racing towards AI and at some point the
00:25:40 向人工智慧邁進，在某個時候

00:25:40 racing towards AI and at some point the story splits into a we're going to slow
00:25:40 向人工智慧邁進，在某個時刻，故事分裂成我們將放慢

00:25:42 story splits into a we're going to slow
00:25:42 故事分成了“我們要放慢速度”

00:25:42 story splits into a we're going to slow down and work on alignment or we're
00:25:42 故事分成兩個部分，一部分是我們要放慢速度，進行協調，另一部分是

00:25:45 down and work on alignment or we're
00:25:45 向下並進行對齊，否則我們

00:25:45 down and work on alignment or we're going full out and uh, you know, spoiler
00:25:45 向下調整，或者我們全力以赴，呃，你知道，劇透

00:25:48 going full out and uh, you know, spoiler
00:25:48 全力以赴，你知道，劇透

00:25:48 going full out and uh, you know, spoiler alert and the race to infinity uh,
00:25:48 全力以赴，你知道，劇透警告和無限競賽，

00:25:52 alert and the race to infinity uh,
00:25:52 警報和無限競賽呃，

00:25:52 alert and the race to infinity uh, humanity vanishes. So the right outcome
00:25:52 警報響起，人類的競賽無限，人類最終消失。所以正確的結果

00:25:55 humanity vanishes. So the right outcome
00:25:55 人類消失了。所以正確的結果

00:25:55 humanity vanishes. So the right outcome will ultimately be some form of
00:25:55 人類消失了。所以正確的結果最終會是某種形式的

00:25:58 will ultimately be some form of
00:25:58 最終會以某種形式

00:25:58 will ultimately be some form of deterrence and mutually assured
00:25:58 最終將是某種形式的威懾和相互保證

00:26:00 deterrence and mutually assured
00:26:00 威懾與相互保證

00:26:00 deterrence and mutually assured destruction. Uh I wrote a paper with two
00:26:00 威懾和相互確保摧毀。呃，我寫了一篇論文，裡面有兩個

00:26:03 destruction. Uh I wrote a paper with two
00:26:03 毀滅。呃，我寫了一篇論文，裡面有兩個

00:26:03 destruction. Uh I wrote a paper with two other authors Dan Hendricks and Alex
00:26:03 毀滅。呃，我和另外兩位作者丹‧亨德里克斯和亞歷克斯寫了一篇論文

00:26:05 other authors Dan Hendricks and Alex
00:26:05 其他作者 Dan Hendricks 與 Alex

00:26:05 other authors Dan Hendricks and Alex Wang where we named it mutual AI
00:26:05 其他作者 Dan Hendricks 和 Alex Wang 我們將其命名為相互人工智慧

00:26:09 Wang where we named it mutual AI
00:26:09 王，我們將其命名為相互人工智慧

00:26:09 Wang where we named it mutual AI malfunction.
00:26:09 王，我們將其命名為相互人工智慧故障。

00:26:10 malfunction.  00:26:10 故障。

00:26:10 malfunction. And the idea was goes something like
00:26:10 故障。這個想法是這樣的

00:26:12 And the idea was goes something like
00:26:12 這個想法是這樣的

00:26:12 And the idea was goes something like this. Um you're the United States, I'm
00:26:12 這個想法是這樣的。嗯，你是美國，我是

00:26:15 this. Um you're the United States, I'm
00:26:15 這個。嗯，你是美國，我是

00:26:15 this. Um you're the United States, I'm China, you're ahead of me. Um at some
00:26:15 這個。嗯，你是美國，我是中國，你領先我。嗯，在某種程度上

00:26:17 China, you're ahead of me. Um at some
00:26:17 中國，你領先我了。嗯，在某種程度上

00:26:17 China, you're ahead of me. Um at some point you cross a line. You know, you
00:26:17 中國，你領先我了。嗯，總有一天你會越界。你知道，你

00:26:19 point you cross a line. You know, you
00:26:19 點，你越界了。你知道，你

00:26:19 point you cross a line. You know, you Peter cross a line and I China go this
00:26:19 點，你越界了。你知道，你彼得越界了，我中國就這麼做了

00:26:22 Peter cross a line and I China go this
00:26:22 彼得越界了，我中國就這麼做了

00:26:22 Peter cross a line and I China go this is unacceptable.
00:26:22 彼得越界了，我中國認為這是不可接受的。

00:26:23 is unacceptable.
00:26:23 是不可接受的。

00:26:23 is unacceptable. At some point it becomes
00:26:23 是不可接受的。在某種程度上，它變成了

00:26:25 At some point it becomes
00:26:25 到了某個時候，它就變成了

00:26:25 At some point it becomes in terms of amount of compute and amount
00:26:25 在某種程度上，它變成了計算量和數量

00:26:26 in terms of amount of compute and amount
00:26:26 就計算量和金額而言

00:26:26 in terms of amount of compute and amount of
00:26:26 就計算量和數量而言

00:26:27 of  00:26:27 的

00:26:27 of it's it's something you're doing where
00:26:27 這是你正在做的事情

00:26:29 it's it's something you're doing where
00:26:29 這是你正在做的事情

00:26:29 it's it's something you're doing where it affects my sovereignty.
00:26:29 你所做的事情影響了我的主權。

00:26:31 it affects my sovereignty.
00:26:31 它影響了我的主權。

00:26:31 it affects my sovereignty. It's not just words and yelling and an
00:26:31 這影響了我的主權。這不僅僅是言語和叫喊，

00:26:34 It's not just words and yelling and an
00:26:34 這不只是言語和叫喊，

00:26:34 It's not just words and yelling and an occasional shooting down a jet. It's
00:26:34 這不只是言語和叫喊，偶爾還會擊落一架飛機。這是

00:26:35 occasional shooting down a jet. It's
00:26:35 偶爾會擊落一架噴射機。

00:26:35 occasional shooting down a jet. It's it's a real threat to the identity of my
00:26:35 偶爾擊落一架噴射機。這對我的身分構成了真正的威脅

00:26:40 it's a real threat to the identity of my
00:26:40 這對我的身分構成了真正的威脅

00:26:40 it's a real threat to the identity of my my country, my economic what have you.
00:26:40 這對我的國家的認同、我的經濟等等都構成了真正的威脅。

00:26:43 my country, my economic what have you.
00:26:43 我的國家，我的經濟怎麼了。

00:26:43 my country, my economic what have you. Under this scenario, I would be highly
00:26:43 我的國家，我的經濟狀況如何？在這種情況下，我會非常

00:26:45 Under this scenario, I would be highly
00:26:45 在這種情況下，我會非常

00:26:45 Under this scenario, I would be highly tempted to do a cyber attack to slow you
00:26:45 在這種情況下，我很想發動網路攻擊來拖慢你的速度

00:26:49 tempted to do a cyber attack to slow you
00:26:49 試圖透過網路攻擊來拖慢你的速度

00:26:49 tempted to do a cyber attack to slow you down. Okay? In mutually assured mal
00:26:49 試圖發動網路攻擊來拖慢你的速度。好嗎？在相互保證的惡意攻擊中

00:26:53 down. Okay? In mutually assured mal
00:26:53 下來。好嗎？在相互保證的

00:26:53 down. Okay? In mutually assured mal malfunction, if you will, we have to
00:26:53 下來。好嗎？在相互保證故障的情況下，如果你願意，我們必須

00:26:55 malfunction, if you will, we have to
00:26:55 故障，如果你願意，我們必須

00:26:56 malfunction, if you will, we have to engineer it so that you have the ability
00:26:56 故障，如果你願意，我們必須設計它，這樣你就有能力

00:26:57 engineer it so that you have the ability
00:26:57 設計它，這樣你就有能力

00:26:57 engineer it so that you have the ability to then do the same thing to me.
00:26:57 設計它，以便你有能力對我做同樣的事情。

00:27:00 to then do the same thing to me.
00:27:00 然後對我做同樣的事情。

00:27:00 to then do the same thing to me. And that causes both of us to be careful
00:27:00 然後對我做同樣的事情。這讓我們倆都小心

00:27:04 And that causes both of us to be careful
00:27:04 這讓我們兩個都要小心

00:27:04 And that causes both of us to be careful not to trigger the other.
00:27:04 這導致我們雙方都要小心，以免激怒對方。

00:27:05 not to trigger the other.
00:27:05 不會觸發另一個。

00:27:06 not to trigger the other. That's what mutual assured destruction
00:27:06 不觸發對方。這就是相互保證摧毀

00:27:07 That's what mutual assured destruction
00:27:07 這就是相互保證摧毀

00:27:07 That's what mutual assured destruction is. That's our best formulation right
00:27:07 這就是「相互保證摧毀」的精髓。這是我們最好的表述，對吧？

00:27:09 is. That's our best formulation right
00:27:09 是。這是我們最好的表述，對吧

00:27:09 is. That's our best formulation right now. We also recommend in our work, and
00:27:09 是。這是我們目前最好的方案。我們也建議在我們的工作中，

00:27:12 now. We also recommend in our work, and
00:27:12 現在。我們也建議在我們的工作中，

00:27:12 now. We also recommend in our work, and I think it's very strong, that the
00:27:12 現在。我們在工作中也建議，我認為非常有力的是，

00:27:14 I think it's very strong, that the
00:27:14 我認為這非常有力，

00:27:14 I think it's very strong, that the government require that we know where
00:27:14 我認為這非常有力，政府要求我們知道

00:27:16 government require that we know where
00:27:16 政府要求我們知道

00:27:16 government require that we know where all the chips are. And remember, the
00:27:16 政府要求我們知道所有晶片的存放位置。記住，

00:27:18 all the chips are. And remember, the
00:27:18 所有的籌碼都是。記住，

00:27:18 all the chips are. And remember, the chips can tell you where they are
00:27:18 所有晶片都是。記住，晶片可以告訴你它們在哪裡

00:27:19 chips can tell you where they are
00:27:19 晶片可以告訴你它們在哪裡

00:27:19 chips can tell you where they are because they're computers. Yeah.
00:27:19 晶片可以告訴你它們在哪裡，因為它們是計算機。是的。

00:27:21 because they're computers. Yeah.
00:27:21 因為它們是電腦。是的。

00:27:21 because they're computers. Yeah. And it would be easy to add a little
00:27:21 因為它們是電腦。是的。而且很容易添加一點

00:27:22 And it would be easy to add a little
00:27:22 很容易加一點

00:27:22 And it would be easy to add a little crypto thing, which would say, "Yeah,
00:27:22 並且很容易添加一些加密的東西，比如說，「是的，

00:27:24 crypto thing, which would say, "Yeah,
00:27:24 加密的東西，它會說，「是的，

00:27:24 crypto thing, which would say, "Yeah, here I am, and this is what I'm doing."
00:27:24 加密的東西，它會說，“是的，我在這裡，這就是我正在做的事情。”

00:27:26 here I am, and this is what I'm doing."
00:27:26 我在這裡，這就是我正在做的事情。 」

00:27:26 here I am, and this is what I'm doing." So, so knowing where the chips are,
00:27:26 我在這裡，這就是我正在做的事情。所以，知道晶片在哪裡，

00:27:28 So, so knowing where the chips are,
00:27:28 所以，知道晶片在哪裡，

00:27:28 So, so knowing where the chips are, knowing where the training runs are, and
00:27:28 所以，知道晶片在哪裡，知道訓練運作在哪裡，

00:27:30 knowing where the training runs are, and
00:27:30 知道訓練路線在哪裡，並且

00:27:30 knowing where the training runs are, and knowing what these fault lines are are
00:27:30 知道訓練運行在哪裡，知道這些斷層線在哪裡

00:27:33 knowing what these fault lines are are
00:27:33 知道這些斷層線是什麼

00:27:33 knowing what these fault lines are are very important. Now, there are a whole
00:27:33 知道這些斷層線在哪裡非常重要。現在，有一整套

00:27:35 very important. Now, there are a whole
00:27:35 非常重要。現在，

00:27:35 very important. Now, there are a whole bunch of assumptions in this scenario
00:27:35 非常重要。現在，這個場景中有很多假設

00:27:37 bunch of assumptions in this scenario
00:27:37 在這個場景中有很多假設

00:27:37 bunch of assumptions in this scenario that I described. The first is that
00:27:37 我描述的這個場景有很多假設。首先是

00:27:39 that I described. The first is that
00:27:39 我描述的。首先是

00:27:39 that I described. The first is that there was enough electricity. The second
00:27:39 就像我所描述的。第一，電力充足。第二

00:27:41 there was enough electricity. The second
00:27:41 電力充足。第二個

00:27:41 there was enough electricity. The second is that there was enough power. The
00:27:41 有足夠的電力。第二是有足夠的動力。

00:27:43 is that there was enough power. The
00:27:43 是有足夠的力量。

00:27:43 is that there was enough power. The third is the Chinese had enough
00:27:43 是有足夠的力量。第三是中國人有足夠的

00:27:44 third is the Chinese had enough
00:27:44 第三是中國受夠了

00:27:44 third is the Chinese had enough electricity, which they do, and enough
00:27:44 第三，中國人有足夠的電力，他們確實有，而且足夠

00:27:46 electricity, which they do, and enough
00:27:46 電力，他們確實這麼做了，而且足夠

00:27:46 electricity, which they do, and enough computing resources, which they may or
00:27:46 電力，他們有，還有足夠的計算資源，他們可能或

00:27:48 computing resources, which they may or
00:27:48 計算資源，它們可能或

00:27:48 computing resources, which they may or may not have
00:27:48 計算資源，他們可能有也可能沒有

00:27:49 may not have  00:27:49 可能沒有

00:27:49 may not have or may in the future have,
00:27:49 可能沒有，或將來可能會有，

00:27:50 or may in the future have,
00:27:50 或將來可能會有，

00:27:50 or may in the future have, and may in the future have. And also,
00:27:50 或將來可能會有，並且將來可能會有。而且，

00:27:52 and may in the future have. And also,
00:27:52 將來也可能會。而且，

00:27:52 and may in the future have. And also, I'm asserting that everyone arrives at
00:27:52 並且將來可能還會有。而且，我斷言每個人都會

00:27:55 I'm asserting that everyone arrives at
00:27:55 我斷言每個人都會到達

00:27:55 I'm asserting that everyone arrives at this eventual state of super
00:27:55 我斷言每個人最終都會達到超級

00:27:57 this eventual state of super
00:27:57 超級的最終狀態

00:27:57 this eventual state of super intelligence at a roughly the same time.
00:27:57 大約在同一時間，超級智慧的最終狀態。

00:27:59 intelligence at a roughly the same time.
00:27:59 情報大致在同一時間。

00:27:59 intelligence at a roughly the same time. Again, these are debatable points, but
00:27:59 情報大致相同。同樣，這些都是有爭議的觀點，但是

00:28:02 Again, these are debatable points, but
00:28:02 再次強調，這些都是有爭議的觀點，但是

00:28:02 Again, these are debatable points, but the most interesting scenario is we're
00:28:02 再次強調，這些都是有爭議的觀點，但最有趣的情況是，我們

00:28:04 the most interesting scenario is we're
00:28:04 最有趣的情況是我們

00:28:04 the most interesting scenario is we're saying it's 1938. the letter has come,
00:28:04 最有趣的場景是我們說這是1938年。這封信來了，

00:28:08 saying it's 1938. the letter has come,
00:28:08 說是1938年。信來了，

00:28:08 saying it's 1938. the letter has come, you know, from Einstein to the president
00:28:08 說是1938年。你知道，這封信是愛因斯坦寫給總統的

00:28:10 you know, from Einstein to the president
00:28:10 你知道，從愛因斯坦到總統

00:28:10 you know, from Einstein to the president and we're having a conversation and
00:28:10 你知道，從愛因斯坦到總統，我們正在進行對話，

00:28:12 and we're having a conversation and
00:28:12 我們正在交談

00:28:12 and we're having a conversation and we're saying,"Well, how does this end?"
00:28:12 我們正在進行對話，我們說，“好吧，這件事該如何結束？”

00:28:15 we're saying,"Well, how does this end?"
00:28:15 我們說，“好吧，這件事該如何結束？”

00:28:15 we're saying,"Well, how does this end?" Okay. So, if you were so brilliant in
00:28:15 我們會說，「好吧，這該怎麼收場？」 好吧。所以，如果你在

00:28:17 Okay. So, if you were so brilliant in
00:28:17 好的。所以，如果你在

00:28:17 Okay. So, if you were so brilliant in 38, what you would have said is this
00:28:17 好的。所以，如果你在38歲時如此聰明，你會說這樣的話

00:28:20 38, what you would have said is this
00:28:20 38，你會說的是

00:28:20 38, what you would have said is this ultimately ends with us having a bomb,
00:28:20 38，你會說這最終會以我們擁有一枚炸彈而告終，

00:28:22 ultimately ends with us having a bomb,
00:28:22 最後我們得到了一枚炸彈，

00:28:22 ultimately ends with us having a bomb, the other guys having a bomb, and then
00:28:22 最終結果是我們得到了一枚炸彈，其他人也得到了一枚炸彈，然後

00:28:24 the other guys having a bomb, and then
00:28:24 其他人有炸彈，然後

00:28:24 the other guys having a bomb, and then we're going to have one heck of a
00:28:24 其他人有炸彈，然後我們就會有一個

00:28:25 we're going to have one heck of a
00:28:25 我們將會有一個

00:28:25 we're going to have one heck of a negotiation to try to make sure that we
00:28:25 我們將進行一場艱苦的談判，以確保我們

00:28:28 negotiation to try to make sure that we
00:28:28 談判，試圖確保我們

00:28:28 negotiation to try to make sure that we don't end up destroying each other. And
00:28:28 談判，以確保我們最終不會互相毀滅。

00:28:30 don't end up destroying each other. And
00:28:30 不要最終互相毀滅。

00:28:30 don't end up destroying each other. And I think the same conversation needs to
00:28:30 不要最終互相毀滅。我認為同樣的對話需要

00:28:32 I think the same conversation needs to
00:28:32 我認為同樣的對話需要

00:28:32 I think the same conversation needs to get started now, well before the
00:28:32 我認為同樣的對話需要現在就開始，在

00:28:35 get started now, well before the
00:28:35 現在就開始，在

00:28:35 get started now, well before the Chernobyl events, well before the
00:28:35 現在就開始，早在切爾諾貝利事件之前，早在

00:28:37 Chernobyl events, well before the
00:28:37 切爾諾貝利事件，早在

00:28:37 Chernobyl events, well before the buildups.
00:28:37 切爾諾貝利事件，遠在核外洩漏之前。

00:28:38 buildups.  00:28:38 累積。

00:28:38 buildups. Can I just take that one more step? And
00:28:38 累積。我可以再走一步嗎？

00:28:40 Can I just take that one more step? And
00:28:40 我可以再走一步嗎？

00:28:40 Can I just take that one more step? And and don't answer if you don't want to,
00:28:40 我可以再走一步嗎？如果你不想回答，就不要回答。

00:28:42 and don't answer if you don't want to,
00:28:42 不想回答就不要回答，

00:28:42 and don't answer if you don't want to, but if it was 1947, 1948,
00:28:42 不想回答就不要回答，但如果是 1947 年、1948 年，

00:28:45 but if it was 1947, 1948,
00:28:45 但如果是 1947 年、1948 年，

00:28:45 but if it was 1947, 1948, so before the Cold War really took off,
00:28:45 但如果是1947年、1948年，也就是冷戰真正開始之前，

00:28:48 so before the Cold War really took off,
00:28:48 所以在冷戰真正開始之前，

00:28:48 so before the Cold War really took off, and you say, well, that's similar to
00:28:48 所以在冷戰真正爆發之前，你會說，嗯，這類似於

00:28:50 and you say, well, that's similar to
00:28:50 你會說，這類似於

00:28:50 and you say, well, that's similar to where we are with China right now. We
00:28:50 你會說，這和我們現在在中國的情況很相似。我們

00:28:51 where we are with China right now. We
00:28:51 我們目前與中國的關係如何。我們

00:28:51 where we are with China right now. We have a competitive lead, but it may or
00:28:51 我們目前與中國的差距在哪裡？我們擁有競爭優勢，但也可能

00:28:53 have a competitive lead, but it may or
00:28:53 擁有競爭優勢，但也可能

00:28:54 have a competitive lead, but it may or may not be fragile.
00:28:54 擁有競爭領先優勢，但這種領先優勢可能脆弱，也可能不脆弱。

00:28:56 may not be fragile.
00:28:56 可能並不脆弱。

00:28:56 may not be fragile. What would you do differently 1947 1940
00:28:56 可能並不脆弱。你會做什麼不同的事情？ 1947 1940

00:28:58 What would you do differently 1947 1940
00:28:58 你會做什麼不同的事 1947 1940

00:28:58 What would you do differently 1947 1940 or what would Kissinger do different
00:28:58 你會在 1947 年和 1940 年採取什麼不同的做法，或者基辛格會採取什麼不同的做法

00:28:59 or what would Kissinger do different
00:28:59 或是基辛格會做什麼不同的事情

00:28:59 or what would Kissinger do different 1947 1948 1949 than what we did do?
00:28:59 或者基辛格在 1947 年、1948 年和 1949 年會採取與我們不同的行動嗎？

00:29:03 1947 1948 1949 than what we did do?
00:29:03 1947 1948 1949 比我們做的要多嗎？

00:29:03 1947 1948 1949 than what we did do? You know I I wrote two books with Dr.
00:29:03 1947、1948、1949年比我們做的多嗎？你知道我和博士一起寫了兩本書。

00:29:05 You know I I wrote two books with Dr.
00:29:05 你知道我和博士一起寫了兩本書。

00:29:05 You know I I wrote two books with Dr. Kissinger and I miss him very much. He
00:29:05 你知道我和基辛格博士合寫了兩本書，我非常想念他。他

00:29:07 Kissinger and I miss him very much. He
00:29:07 基辛格和我都非常想念他。他

00:29:07 Kissinger and I miss him very much. He was my closest friend. Um and Henry was
00:29:07 基辛格，我非常想念他。他是我最親密的朋友。嗯，亨利

00:29:11 was my closest friend. Um and Henry was
00:29:11 是我最親密的朋友。嗯，亨利是

00:29:11 was my closest friend. Um and Henry was very much a realist in the sense that
00:29:11 是我最親密的朋友。嗯，亨利非常現實，因為他

00:29:14 very much a realist in the sense that
00:29:14 非常現實主義，因為

00:29:14 very much a realist in the sense that when you look at his history in uh
00:29:14 非常現實主義，當你回顧他的歷史時

00:29:17 when you look at his history in uh
00:29:17 當你回顧他的歷史時

00:29:17 when you look at his history in uh roughly 36 38 he and his uh I guess 37
00:29:17 當你回顧他的歷史時，大概是 36 38 他和他的呃我猜是 37

00:29:21 roughly 36 38 he and his uh I guess 37
00:29:21 大概 36 38 他和他的呃我猜 37

00:29:21 roughly 36 38 he and his uh I guess 37 38 his family were were Jewish were
00:29:21 大概 36 38 他和他的呃我猜 37 38 他的家人是猶太人

00:29:23 38 his family were were Jewish were
00:29:23 38 他的家人都是猶太人

00:29:24 38 his family were were Jewish were forced to immigrate from uh Germany
00:29:24 38 他的家人都是猶太人，被迫從德國移民

00:29:26 forced to immigrate from uh Germany
00:29:26 被迫從德國移民