2025 Predictions: Tech, Business, Media, Politics! · 苏菲拉底
字幕 字幕位置
--:--
点击播放,这里会跟随视频显示当前句的中英字幕。

2025 Predictions: Tech, Business, Media, Politics!

节目发布 2025-01-03 · All-In Podcast
加文·贝克 查马斯·帕里哈皮蒂亚 大大卫·弗里德伯格
EDITED TRANSCRIPT · 依据现场录音编译整理,可划线生成便签
2025年伊始,All-In播客录制了年度预测特辑。主持人杰森·卡拉卡尼斯与联合主持人查马斯·帕里哈皮蒂亚、大卫·弗里德伯格一起,请来科技股投资人、Atreides Management创始人加文·贝克第二次做客,逐项预测新一年的政治、商业、资产与媒体走向,常驻主持人大卫·萨克斯本期缺席。节目同时宣布与预测市场Polymarket合作,把当期的若干预测挂牌供公众交易。本文依据现场录音编译整理。

嘉宾登场

卡拉卡尼斯:欢迎回到All-In播客。现在是2025年,今天是我们的年度预测特辑。再次和我们坐在一起的,是一位真正了不起的老朋友,来自Atreides Management的加文·贝克。加文,欢迎回来。简单介绍一下,Atreides是做什么的?

贝克:谢谢杰森、查马斯和大卫。Atreides是一家跨越公开市场和私募市场的投资机构,专注消费和科技领域,从A轮一直投到超大市值公司。

卡拉卡尼斯:所以你是资本配置者,把赌注押在技术和全世界最重要的新公司上。你的基金规模多大?

查马斯:你不能刚见到人家就问基金多大。

卡拉卡尼斯:观众想知道这位是不是重量级选手。

贝克:大约四十亿美元。

卡拉卡尼斯:好,那比查马斯的大了那么一点点,他现在真有点紧张了。

与Polymarket合作

卡拉卡尼斯:今年每位老友都要做一两个「超级预测」,我们会把这些超级预测放到Polymarket上。不了解Polymarket的听众,它是一个预测市场,人们可以在任何一方下注、投资。弗里德伯格,你参与谈了这项合作,介绍一下细节。

弗里德伯格:我们和Polymarket的谢恩谈了很久,最后定下一个安排:我们可以在他们的网站上开设自己的市场。今天我们会讲几个预测,Polymarket随后会把它们挂上去,美国以外的用户可以交易。我们很兴奋,因为这样就能追踪我们预测的准确率,也能持续不断地开新市场。Polymarket上会有一个All-In专区,进去就能看到我们的所有市场,会很有意思。

雪季战绩

卡拉卡尼斯:我们几个刚结束滑雪之旅,尼克找到了一段大家滑雪的录像。加文,请你给每个人打分。我先下来了,姿势很好,然后是弗里德伯格,也不错,接着是第三位,我想乔·朗斯代尔给这类人起的技术术语叫「倒车嘀嘀声」。看查马斯切的这些S弯,比上个雪季进步太多了。

查马斯:听我说,我做了一个非常有策略性的决定:放弃单板,改学双板,为了能和孩子们一起滑。人到四十多岁再从头学一样东西,真的太难了。不过我很感谢你和你弟弟上个雪季对我的指点,我学到了几招。这是我的第三个雪季,滑得很开心。

卡拉卡尼斯:你的进步让我惊讶。弗里德伯格,你能上山我就已经很惊讶了,你滑了好几天,干得漂亮。

弗里德伯格:我需要一双新雪靴,上个雪季我的内侧副韧带撕裂了,这次出去又扭了一下,但没大碍,很开心,滑了四天。

卡拉卡尼斯:我昨天刚滑完第十六天,今天休息,专门为节目做准备。加文,我滑的是「高管套餐」:出去一个半到三小时,十趟,只滑上午。

贝克:我也一样,只滑上午。十二点一刻吃午饭,放松一下,处理点工作。

卡拉卡尼斯:加文,看了这段录像,你会把自己排在哪里?显然排序是我、查马斯、弗里德伯格。

贝克:我说两点。第一,论天生的运动能力,我排在最底层的百分位;第二,我确实有成千上万小时的滑雪时间。我就说到这里。

卡拉卡尼斯:那明年我们就和你一起去滑。

2025年政治赢家

卡拉卡尼斯:先从政治开始。去年的「最大政治赢家」预测,我们大体上猜错了。弗里德伯格说的是美国的独立第三党,我们确实看到了一些苗头,比如罗伯特·肯尼迪的崛起;查马斯说的是独立中间派,没实现;我说的是黑马总统候选人,或许能拿四分之一分。

查马斯:等等,什么叫没实现?独立中间派赢下了这场大选。

卡拉卡尼斯:你真这么认为?

查马斯:我认为独立中间派在2020年把选举摆向了拜登,又在2024年把选举摆向了特朗普。

卡拉卡尼斯:加文,从统计上看是这样吗?

贝克:我认为是对的。而且不只是中间派,很多一辈子投民主党的人这次投了特朗普。所以特朗普确实赢下了中间派。

卡拉卡尼斯:特朗普基本上赢下了一切。查马斯,你预测2025年最大的政治赢家是谁?

查马斯:财政保守派。我认为2025年我们会真正检验一个非常重要的概念,我也希望它能成功,那就是紧缩(austerity)。紧缩必须成功,因为紧缩之后剩下的唯一手段就是削减福利。在这个过程中,我们会弄清楚美国联邦政府里到底有多少浪费、欺诈和滥用。我认为这会外溢到许多州的选举。那些一直呼吁克制开支的财政保守派,2025年将迎来属于他们的日子。

卡拉卡尼斯:弗里德伯格,你的预测?

弗里德伯格:我也是按群体来选,我选年轻候选人。特朗普内阁人选的平均年龄在四十到四十五岁之间,拜登内阁的平均年龄是五十九岁多,接近六十。我认为这标志着一个新趋势的开端:政治候选人的年龄段整体向年轻转移。到2025年底,随着中期选举的候选人陆续出现,我们会看到一批年轻的新面孔,他们传递的信息能引起共鸣,而且不属于那个老化的政治阶层。

卡拉卡尼斯:加文,你的政治赢家?

贝克:特朗普和中间路线。

卡拉卡尼斯:我的选择和弗里德伯格接近:X世代和年长的千禧一代。看看X世代的重要任命:埃隆负责政府效率部(DOGE),萨克斯,马可·卢比奥,名单很长。再看年长的千禧一代:J·D·万斯、维韦克、图尔西·加巴德,一大批年轻人。我认为这会非常好,因为他们的思考不再只围绕自己,不像婴儿潮一代那样,围绕社会保障、税收、房地产这些对自己有利的议题选边站。他们会开始为自己的孩子考虑。一场天翻地覆的变化正在进行。

2025年政治输家

卡拉卡尼斯:接下来是政治输家,这次倒过来,加文先说。

贝克:普京。我认为普京会输得很惨。设想你是习近平,俄罗斯如今已经是中国的附庸国,正在发生的事情对你来说是一场灾难:欧洲开始重新武装,今年只会加速,这让美国可以把资源从欧洲抽出来,投到日本、韩国和整个太平洋,这让你最想做的事,也就是「统一」台湾,我们不妨直说,入侵台湾,变得困难得多。所以我认为习近平会开始和普京脱钩。而如果你是特朗普,你想证明自己是独立的,不受普京任何牵制,所以我认为特朗普对普京会比人们想的强硬得多,他会促成一个对俄罗斯非常糟糕的协议。俄罗斯损失了五十万人,为了什么?到底为了什么?

卡拉卡尼斯:战争打了一千多天,北约和美国没有一名士兵阵亡,我们只是提供了武器。对普京来说,这至今是一场耻辱性的失败。我同意你。弗里德伯格,你的政治输家?

弗里德伯格:主战的新保守派(neocons)。他们会和J·D·万斯、埃隆这些人正面交锋,而且我认为他们会输。这个多年来在世界各地鼓吹冲突的新保守派建制会出现一道大裂缝。我们已经从万斯等人的演讲和评论里听到了苗头,今年会是摊牌的一年,他们会站到失败的一边。

贝克:我能稍微唱一下反调吗?现实层面我认为你是对的。但特朗普说过一句关于约翰·博尔顿的话很有意思:这家伙绝对是疯的,可谈判的时候让他坐在屋里太好用了,对方看着这个人满脸通红、恨不得马上按核按钮、恨不得立刻开战,最后你就能拿到好得多的协议。所以我认为特朗普政府表现出来的好战姿态会比任何人预期的都多,目的只是在俄乌之间拿到一个好协议,再让中国和俄罗斯脱钩。现实层面你是对的,但在你被证明正确之前,会有大量与你所说相悖的言辞。

卡拉卡尼斯:我们已经看到了:对加拿大、对北约、对台湾,还有关税,谈判之前先摆出非常强硬的姿态,但愿能换来对美国有利的协议。查马斯,你的政治输家?

查马斯:进步主义(progressivism)。去年11月大选刚结束,我飞到伦敦,去牛津辩论社演讲。那场演讲是对MAGA的全力辩护,但更多是在解释MAGA,把它当作进步主义本能的解药,而这种本能已经渗透了整个西方八国集团,并且开始动摇它们。看看2025年将要发生的事:加拿大,贾斯汀·特鲁多会惨败给皮埃尔·博励治;德国,另类选择党(AfD)看起来会赢;法国,如果僵局导致重新选举,玛丽娜·勒庞很可能胜出;英国正在发酵一桩儿童性侵丑闻,据称二十多年间,成千上万的女孩被一些巴基斯坦裔穆斯林男子团伙强奸,而当局出于「担心煽动伊斯兰恐惧症」不予起诉,当年的决策者正是现任首相基尔·斯塔默。如果这一切在英国发酵下去,我认为工党政府会倒台,奈杰尔·法拉奇会赢。到2025年底,这些国家会是什么样?那是对阶级化身份政治的全面否定,会在全世界激起巨大的涟漪。所以2025年最大的政治输家,是我们所贴上「进步主义」标签的那套东西。

卡拉卡尼斯:我的选择很接近:每个政党里那一小撮出声的种族主义者。两边都有一点:一边是DEI,另一边是赤裸裸的种族主义,大概各占MAGA和左派的百分之五。

科赫家族与新金主

卡拉卡尼斯:回顾一下去年的政治输家预测:我说的是内塔尼亚胡,弗里德伯格说的是乌克兰,查马斯说的是科赫家族和共和党。

查马斯:这个我说中了。

卡拉卡尼斯:给不熟悉的听众解释一下。

查马斯:本质上,2024年是我们所熟知的共和党的终结。取而代之的是一个MAGA联盟,它被安放在「共和主义」的标签之下,也就是说,这些人未必是共和党人,他们信奉的是MAGA哲学,只是借共和党这个容器来推举候选人、赢得选举。在这场对现状的颠覆中,有一个家族过去几十年一直处于政治机器的中心,决定候选人、决定议程、决定政策,那就是科赫家族。他们为此花了巨额资金,机器几乎是在最高法院「公民联合」案判决的第二天就搭建起来的。从这个角度看,那几十亿美元的投入基本上归零了,因为它已经不再有任何意义。看看新一代将决定所谓「共和党政策」的金主,会是米里亚姆·阿德尔森、埃隆·马斯克这样的人,和科赫家族过去的决策方式完全不同。

卡拉卡尼斯:加文,你觉得现在政治里的钱太多了吗?

贝克:好问题。如果我不是如此深切地认同这个周期里最大的出资人,我的感受可能会很不一样。我的看法是,只要两边筹到的钱大体相当,我不认为这真的有什么影响。

商业赢家:机器人元年

卡拉卡尼斯:2025年最大的商业赢家。弗里德伯格先来。

弗里德伯格:我觉得我们正处在一个非常有意思的拐点,它会让2025年成为自主硬件(autonomous hardware)或者说机器人之年。如果说2024年是算力建设和AI软件系统铺开的一年,那2025年就是机器人之年。有一家中国公司叫宇树科技(Unitree),我们今天刚下了一笔订单。加文,你研究过这家公司吗?

贝克:看过,真的很疯狂。生意和产品都极其出色。

弗里德伯格:他们的Go2机器狗售价一千六百美元,带API,可以挂载荷,装了激光雷达和智能导航系统。今年早些时候我们看过一些视频,机器狗背上架着机枪,那就是这款系统,本质上是一种新型的野战士兵,自主作战的士兵。但它在科研上也大有用处,我们正考虑在试验农场用它巡田、拍照、回传数据。成本这么低,不到三千美元就能干出许多了不起的事。这家公司上个月刚融了几亿美元,主要来自中国投资者。我认为还会出现更多类似的企业。很多事情就是这样:很长时间看似不动,然后突然以你无法想象的速度发生。今年我们会看着人形机器人和自主系统,惊叹「天哪,我不敢相信它已经来了」。

卡拉卡尼斯:你要引的那句话是海明威的:「你是怎么破产的?先是慢慢地,然后一下子。」技术变革也是这样。我真想在牧场弄一台,让它绕着围栏巡逻。尼克,放一下那段带轮子的Go2越野视频,当时所有人都以为是特效或者AI生成的,但这是真实录像。加文,这是真的吗?

贝克:绝对是真的。我想买G1,他们的人形机器人。

卡拉卡尼斯:一万六千美元,和一辆二手普锐斯差不多,离「买得起」只有一步之遥。这东西可以在家里、工厂里听你指挥干活。

查马斯:还记得我们节目圣诞特辑舞台上那台机器人吗?我弟弟当时说了一个观察:机器人一出来,所有人都想虐待它。人有一种很有意思的天性,机器人一冒头,人类想做的就是伤害它,仿佛在说「这是我能支配的东西」。

弗里德伯格:我们太受威胁了,想确认自己还是物种顶端,可这种优越感不会维持太久了。这些东西很快就能把我们打趴下。我的预测就是机器人之年。

大企业与算力

贝克:我深深认同,机器人之年,而且包括全自动驾驶(FSD)。我认为FSD今天已经能用了,今年会跨入主流。我已经注意到,尤其是深夜叫Uber,我明显更愿意坐特斯拉,因为有时司机很疲惫,如果他开着FSD,我会觉得安全得多。它会以加速度持续复合增长。更广泛地讲,我认为在一段时间内赢家会是大企业,那些审慎地使用AI的大企业。原因在于OpenAI的新模型o3向我们展示的东西:推理(reasoning)和测试时算力(test-time compute)的结合。如果你是一家大企业,可以花一百万美元让AI就你业务里最重要的问题思考六个星期,这相对于付不起这笔钱的小企业是一种深刻的优势。顺带一提,推理算力会是这个趋势的衍生赢家。2025年我们会像2023年一样,把GPU、加速器和算力用光。

卡拉卡尼斯:所以你也看多英伟达和芯片厂商,比如Groq,因为我们在为算力找到新用途。

贝克:绝对如此。

稳定币爆发

卡拉卡尼斯:查马斯,你的商业赢家?

查马斯:美元计价的稳定币(stablecoins)。2024年发生了两件关键的事。第一,稳定币基本上和加密货币的波动脱钩了,开始被用于运营企业的批发性实用功能。有一张图可以看出,无论加密货币怎么波动,稳定币的使用量一路向右上方攀升。这是2024年发生的一次极其重要的脱钩。第二,数据还在陆续出来,但已经有一个惊人的数字:截至2024年第二季度末,稳定币的交易笔数约为十一亿笔,交易总额八万五千亿美元,是同期Visa交易量的两倍多。所以我认为它已经越过了不可逆转的临界点。去年我认为2024年的大趋势是比特币,今年我认为2025年的大趋势是稳定币的使用。我们终于要向Visa和万事达的双寡头发起进攻了。你会看到数不清的应用场景建在稳定币轨道上。特朗普就任总统后,我认为他会向极高的信用卡交易费用开刀,而由于高得离谱的年化利率,消费信贷已经出现裂痕。所有这些都会在2025年汇聚到一起。我认为到2025年底,稳定币规模可以翻四倍甚至五倍,会是一个巨大的市场。

卡拉卡尼斯:它们该被监管吗?在国会听证会上,Tether被指为恐怖主义、规避制裁和人口贩运的主要转账和支付工具。该怎么监管?

查马斯:我没资格说该怎么监管。但要点是,所有这些交易的两端和中间都有账本,信息是存在的,而且有一大批第三方服务提供这一层情报分析。真正该问的问题是:如果你从全球经济里拿掉三百个基点的摩擦损耗,价值有多大?我认为仅在美国就相当于一万亿美元。到了这个地步还不去做,在我看来有点匪夷所思。经济上的理由已经太充分了。

卡拉卡尼斯:加文,问你一个更难的问题。美国政府对美元有一种垄断,稳定币是这种垄断的竞争者。过去几年我们在节目里谈论金砖国家要建立自己的货币,稳定币会不会在某种意义上也是同类事物,会把人从美元本位上拉走?这是一些当权者提出的批评。

贝克:我认为稳定币可能对世界很好,但如果某种稳定币组合取代美元成为新的储备货币,那对美国会非常糟糕。能用本国货币借债是巨大的优势,轻易放弃是不智的。这意味着你随时可以控制自己债务的真实规模。哦,我们的债务很大?那就像我们刚做过的那样,让通胀跑热一点,然后债务实际上就变小了。所以我认为放弃会是一个大错误。

卡拉卡尼斯:弗里德伯格,你对稳定币有什么想法?

弗里德伯格:我对稳定币了解不够。

卡拉卡尼斯:那这会是今年的重要话题之一,我们应该请USDC的杰里米·阿莱尔来。

查马斯:顺便说一件事。我之前提过,我做了一个小产品,把我们的研究放在Substack上面对外开放。那些工具不太好用,我们用Stripe收款。杰里米联系我说,他可以用美元稳定币重建我全部的支付轨道。这是不用想的决定,因为我在交易成本上花的是几十万美元,现在可以省下来多雇一个人,或者多给团队发钱。

卡拉卡尼斯:你付给Substack百分之十,再加百分之三。

查马斯:对我来说是显而易见的事,对其他所有人也一样。

特斯拉与谷歌回归

卡拉卡尼斯:我的商业赢家预测是特斯拉和谷歌。今年显然是AI和机器人之年,但谷歌在AI上的回归让我惊叹,埃隆用Colossus做的事更让我敬畏。黄仁勋、迈克尔·戴尔都参与了xAI的建设,他们震惊于xAI在不到四十五天里立起十万块GPU。埃隆的特斯拉和xAI,加上谷歌,这些重金投入AI的公司,会是明年最大的赢家,他们会拿出人们意料之外的东西。我不认为人们预料到了谷歌接下来会推出什么。我一直和人说Gemini的深度研究(Deep Research)功能,我天天在用,它太惊人了。现在就暂停播客,下载Gemini,付二十美元,在桌面上玩一玩深度研究。它把市面上任何同类产品甩在身后,是一次台阶式的跃升。

查马斯:杰森,你不该顺着节目的调性解释一下它是什么吗?

卡拉卡尼斯:好提醒。比方说你想建一个模型,我就这么做过:「帮我建一个模型,测算把美国所有的汽车和出行全部替换成全自动驾驶汽车需要多少成本。」如果你把这个问题问ChatGPT、Grok或者Claude,它会给你一个不错的回答。而深度研究做的是:先弄清这个问题可以拆成哪些子问题,然后并行发出多条线程,实时搜索网络,这正是谷歌极其擅长的事,接着做摘要、给出引用,最后生成一份报告。这份报告看起来像高德纳、麦肯锡、波士顿咨询花三十天、收五十万到一百万美元才能做出来的东西。它好得离谱。但它消耗的算力也极其惊人,所以ChatGPT推出了类似功能,叫o1 Pro,每月两百美元,而Gemini只要二十美元。再想想谷歌能触达的东西:YouTube、你的Google Drive、Google Docs、Gmail。它会做出你意想不到的事,比如在回答你的问题时到你的Gmail里找线索,或者去翻你订阅频道的YouTube视频。它会解决人们不知道自己有的问题。加文,这个解释还行吗?

贝克:很好。

卡拉卡尼斯:回顾去年的商业赢家:查马斯说的是自力更生或者盈利的创业公司,这个说中了,我们看到一些公司上市了,或者以更高估值融资,至少没倒闭。弗里德伯格说的是大宗商品企业。

弗里德伯格:基本持平。当时我预设的通胀压力并没有出现,相对于其他指数算是输家。

卡拉卡尼斯:我说的是训练数据的拥有者,Reddit、《纽约时报》、谷歌。Reddit股价2024年涨了百分之二十四,还推出了自己的Reddit Answers;《纽约时报》股价也涨了,还向OpenAI提起了诉讼。

职业体育估值之争

卡拉卡尼斯:去年的商业输家预测:查马斯说的是职业体育球队,因为估值见顶,现在看是个不错的预测;弗里德伯格说的是垂直SaaS公司,因为AI的冲击,也像是个好预测;我说的是智能手机厂商,苹果股价2024年涨了百分之三十,不过手机确实没什么进步,人们换机周期拉长,销量下滑,靠服务收入撑着。

贝克:职业体育球队我要唱反调。这个市场正要被机构化,基金会开始买职业球队,而资本市场的资源远远超过迄今为止买球队的那些富人,能追逐职业球队的钱会是过去的十倍、一百倍。

查马斯:你完全正确。我去年那么说,是因为当时我已经很清楚,职业体育的收视存在问题,NBA是矿井里的金丝雀。NBA的比赛已经退化成篮板加扣篮,或者三分球,问题是这样的比赛看起来明显没那么有意思了。与此同时,由于电视转播合同决定了这些球队的折现价值,而转播合同大得离谱,联盟里不再有可靠的宿敌关系,因为球员几乎每年都在换队,他们能拿到的报酬实在太夸张了。我认为体育会有一段不错的行情,直到下一轮转播合同签订。举个例子,如果把制药广告从电视上拿掉,广告池就缩小了;收视下降,剩余的广告就更难卖,因为广告位少了、买家也少了,池子再缩一轮。那么电视网和流媒体愿意为体育付的钱就会下降。最价格敏感的群体,正是你说的那类人,非「战利品」买家。我当年入股勇士队,纯粹是把它当作一件战利品资产,我并不在乎价格。但我同意你的看法,现在这些球队的股东名单上有了私募基金,那些人全靠折现现金流模型和Excel表格,他们不是为情感买单,他们也不是从小就想着「我要拥有这支球队」,何况用的是机构出资人的钱。这就是我去年那么说的原因。

卡拉卡尼斯:我要代表菲尔·海尔穆斯和我自己,感谢你做了那笔虚荣投资,我们在你的场边座位坐了那么多次。

查马斯:我自己从来不去看比赛,我早该多去几场。你去了我的比赛,还差点被从我的座位上赶出去,你吃了一张红牌。

卡拉卡尼斯:是真的。

查马斯:我接到勇士队或者NBA的电话,说「查马斯,坐你座位的那个人差点被赶出去」。我说「什么?」我知道是他。他在冲我们自己的球员喊话。

卡拉卡尼斯:我和博古特、大卫·李吵起来了,因为我对史蒂夫·科尔喊:「你们领先我的尼克斯三十分了,让这些人下场,这是业余联赛的做法,史蒂芬·库里要是在场上受伤怎么办?」博古特让我闭嘴,说得难听得多。然后他们过来拍拍我的肩膀,我太太在旁边,她尴尬得要死。那个人递给我一张卡片,上面写着:「关于辱骂行为,我们只警告你一次,再警告一次就请你离场。」我看着他,他说「把卡片读完」,我读完了,他冲我竖了个大拇指,就完了。三年后我坐在同一个座位上,总决赛期间还和德雷蒙德互动,大家都相安无事。

贝克:我想回到查马斯的话。这些私募的人确实全靠折现现金流模型。NBA的管理糟糕透顶,连勒布朗都说他们有大问题。我会把NBA列为管理最差的联盟,NFL大概是管理最好的。但关于转播权我有一个反面观点:谷歌买了NFL的Sunday Ticket,非常满意;亚马逊和奈飞也都买了NFL的比赛,也非常满意。对这三家公司来说,制药广告消失与否根本无所谓。十八个月之内,你就能为每一个观众动态生成一条专属广告:如果谷歌知道你是高价值用户,正准备订度假行程,它会动态生成你想去的酒店和目的地,让你点击。所以只要体育还能像今天这样吸引眼球,我就不担心。当然,如果NBA不解决问题,球队的价值会开始下滑。但所有最大的科技公司都对自己买下的体育版权如此满意,他们会把所有版权都买下来,每一项运动。

卡拉卡尼斯:加文,你可能离得更近:为什么苹果、亚马逊或者谷歌不干脆把NBA整个买下来?所有iPhone用户免费看NBA,其他人付费,后台自己算账,亏五十亿、赚二十亿,谁在乎,对苹果来说是四舍五入的误差。

贝克:我同意。除了苹果,其他几家都已经试了水,苹果只有一个美职联(MLS)的合同。我想他们的感受是:哇,水很暖,跳进去吧。

卡拉卡尼斯:NBA新签的是一份十一年的合同,机会不常有,不过说不定谁会收购一家媒体公司,把版权一并接过来。

商业输家

卡拉卡尼斯:2025年最大的商业输家。加文?

贝克:政府服务承包商。你绝对不希望美国政府,任何层级的政府,占你收入的百分之三十五以上。

卡拉卡尼斯:好选择。在DOGE的时代,政府可能不再大手大脚花钱,可能真的会看一眼账单。多么疯狂的想法:核对账单。查马斯?

查马斯:这个有点冒险。我不知道从这里往下的跌幅百分比是多少,但我认为回头看,「七巨头」(Mag 7)的绝对美元回撤会以万亿计。原因不一定在这些公司的基本面,我几天前发过推文,我有点担心指数里前七、八、九、十家公司的集中度,已经接近百分之四十。从历史上看,这种程度的集中通常预示着一次大幅回撤,而且我看不出你能怎样给自己免疫这种风险。所以撇开公司质量不谈,我认为这些都是极其出色的企业,但集中度实在太高了,这是重新定价和回吐的前兆。哪怕只跌百分之十,放在七巨头身上就是两万亿美元。

卡拉卡尼斯:弗里德伯格?

弗里德伯格:我和加文思路相近,我选的是老牌国防和航空航天承包商,波音、洛克希德·马丁、雷神,也就是「国防1.0」。原因一是中国的主导地位,二是美国的国防预算需要转向更技术导向、更依靠机器人的思路,对采购和定价进行合理化。Palantir和Anduril这样的公司显然会受益。而且这些老牌企业在规模上有大量失败,太笨重,太官僚,波音2024年从航天项目的失败到飞机业务的困境都是例子。随着新鲜血液进入政府,整个政府承包行业今年都会受到严峻挑战。

卡拉卡尼斯:你的意思是「成本加成」(cost-plus)模式的公司会受苦,Anduril们会兴旺。

查马斯:你们会不会把传统咨询公司也加进来?

弗里德伯格:这个补充很好。

查马斯:Wipro、塔塔、HCL、埃森哲、Cognizant、Infosys,这些公司如果那一幕发生,都会有大麻烦。

弗里德伯格:政府提升效率,客户也会感受到,客户可能就不再雇他们了。当一批软件创业者开始给联邦政府提供运营建议,我很确定其中一些服务商会被淘汰。

查马斯:要记住一点:「成本加成」是「按工时和材料计费」的漂亮说法,而「按工时和材料计费」又是「人力套利」的漂亮说法。2025年所有这些都会受到极其严格的审视。如果你买的是八百美元一个的废纸篓、九千美元一个的雨伞挂钩,因为系统里有工时计费和各种疯狂的东西,这些一旦被清洗掉,很多公司和它们的商业模式会承受巨大压力。

卡拉卡尼斯:你给我看激励,我给你看结果。成本加成的激励就是保持低效和浪费。这些军工承包商在每一件弹药、每一辆车上加价六成、七成、八成。成本加成糟糕透顶。

查马斯:不,是「成本」本身糟糕透顶。

看空OpenAI

卡拉卡尼斯:我的商业输家,我得先过一遍过去一年里我批评过、吵过架的对象。首先想到的是商业地产,租约会持续到期,问题只是被往后踢了一段。然后我看了MicroStrategy,它的市值是所持比特币账面价值的两三四倍,我觉得说不通,不可持续。我看了Truth Social,年收入四五百万美元,估值七十亿,完全不合逻辑。但我认为定价最离谱、今年会看到估值见顶的,是OpenAI。OpenAI面临的逆风被严重低估了。就像我前面说的,谷歌势头凶猛,xAI刚刚起步就在建那么多重型基础设施,微软基本上在另一档播客上近乎嘲笑OpenAI把源代码卖给了他们,说他们已经不需要OpenAI了,他们手里什么都有,而且他们拥有全部的大型基础设施。OpenAI的估值说不通,我不认为他们能维持现在的定价。AWS、苹果、谷歌、xAI这一批公司会让那一千五百七十亿美元成为OpenAI的估值顶点。而且我认为存在不为零的可能性,他们会在把一千五百七十亿美元的价值从非营利组织转到营利公司的法律程序中败诉,整件事可能炸掉。

查马斯:这是你的超级预测吗?「OpenAI在2025年无法从非营利转为营利」?

卡拉卡尼斯:那只是一小部分,但作为超级预测确实很有意思。

弗里德伯格:顺便说一句,泄露的财务数据显示他们预计明年收入一百二十亿美元。和你刚才提到的其他几个不同,这是一家真实的企业,有真实的收入、规模、增长和技术,它不是一只梗股。

卡拉卡尼斯:没错,所以我才选它。挑一只梗股太容易了,挑一家真实的公司才难。但那部分收入很多来自消费者,很多来自开发者。我认识的开发者都想用开源,不想受制于OpenAI和萨姆·奥特曼。他们已经把查询分散到多个模型栈上试来试去。我不认为对OpenAI、对Gemini、对任何一家服务存在忠诚度,很多情况下他们最终会走向开源。

年度大交易

卡拉卡尼斯:2025年最大的商业交易。查马斯,去年你说星链会上市。

查马斯:完全落空。今年我认为我们会看到传统汽车厂商的崩塌。2024年底本田和日产的交易是一个信号,说明这个行业必须做的事:经历一波大规模整合。特斯拉凭借车辆、软件和FSD自动驾驶的品质处在一个不可思议的位置。我猜再有几次有分量的产品发布,公开资本市场就会意识到这些汽车厂商已经不可投资,结果就是一波汽车业的超级合并。本田和日产签了合并协议,三菱也参与其中,因为它属于日产的联盟,这显然能去掉所有的冗余。欧洲厂商的麻烦更大:大众怎么办?不清楚。Stellantis怎么办?不清楚。这些企业实际上都是正在融化的冰山,当融化的冰山企业被加文这样精明的投资者施压时,通常就会被迫合并。

卡拉卡尼斯:加文,你在做多空价差交易吗?你做空这些公司了吗?

贝克:我不想谈具体持仓,我只说我百分之百同意查马斯。如果没有大规模的保护主义,它们会失去中国市场的业务,因为它们已经造不出有竞争力的产品,我们已经看到迹象,它们会被夹在特斯拉和中国厂商之间。查马斯预测的唯一风险是政府干预,因为这些公司雇了那么多人,又常被视为国家冠军企业。但没有真正大力度的政府支持,它们全都深陷困境。

卡拉卡尼斯:弗里德伯格,去年你说会有版权方向AI训练授权数据的重磅交易,Reddit确实签了两笔。看看今年能不能连中两元。

弗里德伯格:我延续机器人之年的主题。今年会出现大规模融资交易,类似去年算力建设的那些,只是这次投向美国的硬件制造能力建设。这些交易可能是传统的私募股权形式,也可能包含政府支持的成分,用来推动和加速制造业回流。美国不会回去制造上个世纪的东西,我们必须把制造业推进到下一个十年、下一个世纪的生产方式,这意味着要造那些自主的、机器人化的系统,它们对我们至关重要,尤其是在中国正大规模扩产无人机、机器人和自动驾驶汽车的时候。我们需要把大量这类产能搬回国内,那会是巨额资本的流动,所以你会看到一批面向美国硬件建设的重磅交易。

卡拉卡尼斯:加文,你认为2025年最大的交易是什么?

贝克:最大的交易就是「会有交易」。经过四年什么都做不成的日子,你会看到一波并购的滔天巨浪,积压的需求极其庞大,这是第一点。第二点,英特尔会发生点什么,会很大,希望对美国是好事。第三,我认为许多独立的前沿AI实验室会被收购。杰森刚才的观点我很认同:谷歌手里的牌,xAI自有的算力,你们可以在这三家里随便选,但最终的AI赢家会是基础设施成本最低、算力成本最低的那家。而从定义上说,如果你的算力是从别人那里租来的,你就不可能是最低成本的供应商,因为中间有加价。你从Azure、AWS或者谷歌买算力,长期来看就比它们的内部服务处于劣势。全栈者胜。

卡拉卡尼斯:去年我的预测是字节跳动上市或者TikTok被剥离,再过十八天我们就知道最高法院会不会这么判了。今年,我看了所有媒体公司,可能性太多,华纳兄弟、苹果和迪士尼我们几年前就讨论过。但我认为自动驾驶的时代已经到来,亚马逊、DoorDash、Uber、特斯拉、Waymo这一批公司之间会出现合作。特斯拉现在只需要百分之十的市值就能买下Uber,我不会感到意外;Waymo可以分拆出来和Uber合作;亚马逊可以轻松买下DoorDash。随着莉娜·汗的怒火过去,这样的超级交易是有可能通过的。如果几笔这样的交易成功,联手的一方就能赢下自动驾驶、外卖配送和电商,这是一个巨大的领域。我没和我的朋友谈过这件事,但特斯拉买下DoorDash和Uber,或者亚马逊买下DoorDash和Uber,如果你想做一个超级应用,那可能是有史以来最伟大的服务。

自动驾驶并购猜想

贝克:我补充一点,和你说的以及弗里德伯格的机器人之年都能衔接:自主无人机。有一家公司叫Zipline,我的基金投了。对美国郊区来说,自主无人机确实是配送几乎任何东西的最佳方式,假以时日,城市里的配送问题也会解决。它可能是这些配送服务里的一张外卡。

卡拉卡尼斯:亚马逊已经在得克萨斯州做了,据我所知有六万个商品可以无人机配送,直接扔进你的后院,四十五分钟送达。中国的美团也在用无人机送外卖。

弗里德伯格:接着这个话题,我认为2025年另一个大交易的潜在对象是Waymo。Waymo在2023年8月以开放市场的方式在旧金山上线,当时Uber和Lyft的市场份额是百分之六十六和百分之三十四。十五个月后,到2024年11月,Waymo已经拿到旧金山百分之二十二的出行份额,和Lyft持平,Uber降到了百分之五十五。也就是说,Waymo在短短十五个月里从Uber和Lyft手中各吃掉了大约十二个百分点。现在他们正在洛杉矶、奥斯汀上线,遍布全国,凤凰城早就有了。而且他们刚把硬件平台换成一款新设备,据说能大幅降低新城市上线的资本开支,投入资本回报率会明显改善,意味着有了一种更高效的资本扩张方式。系统跑通了,正在规模化,正在开新市场。今年Waymo可能发生大事,也许是一笔巨额融资,也许是IPO,也许是和某家大型网约车公司合并或被收购。仔细想想,我不确定后者在战略上是否合理,但我认为Waymo会有一笔大交易。如果你们还没坐过Waymo,那是一种不可思议的体验。我认识的每个年龄段的人,坐过之后出来都说:这就是未来,它会彻底主导我以后的出行方式。

卡拉卡尼斯:大多数人说它现在慢、单调,路线也怪,但这些会随时间解决。声明一下,免得有人说我在为自己的持仓吆喝:我几乎在所有这些公司里都有可观的敞口,因为我看好整个领域。在美国,网约车只占全部出行的百分之一点五,全球不到百分之一,这个市场很快会涨到百分之二十,而且赢家会遍布各处。我最喜欢的Uber和八家公司签了合作,这是一个全球市场。比亚迪造车的成本只有其他厂商的一半,还有全自动驾驶,据我所知相当不错。所以这会是一场全球竞争,要想赢,Uber、DoorDash、Waymo、亚马逊、比亚迪之间会很快出现有意思的联手,因为赌注太大了。我的大交易预测:出行领域的整合。

最反共识预测

卡拉卡尼斯:去年的「最反共识观点」:查马斯说OpenAI的企业价值会下降,结果大致翻倍了,不过我今年把它选成了自己的预测,长期看我们是一致的。弗里德伯格说的是二战以来首次动用核武器的概率上升,谢天谢地没发生。

弗里德伯格:我没预测它会发生,我只是说概率上升了。

卡拉卡尼斯:末日博士又来了。我说的是苹果会成为AI玩家,他们确实推出了Apple Intelligence,但它糟透了。我刚买了谷歌的Pixel 9 Pro Fold,Gemini运行得完美,Siri该做的事它全能做:播放这首歌、下载这个应用、加到日历,一句话就做到,用时只有苹果的十分之一,而苹果每次都做错。它是一坨垃圾,一个耻辱。

贝克:Apple Intelligence比Copilot还差,这已经很说明问题了。

卡拉卡尼斯:查马斯,你2025年最反共识的观点?这个环节允许放飞,越辛辣越好。

查马斯:我认为某一家主要的大型银行会爆发一场银行危机。概率很小,但这是一个反共识的观点。如果要论证,思路大致如下:把「美利坚治下的和平」的总负债加起来,也就是美国政府债务加企业债务加抵押贷款债务,然后对大约百分之五的利率做敏感性测算,你很快会意识到,由于债务总量已经爆炸性增长,今天七十多万亿美元上的百分之五利率,按美元计算相当于二十五到三十年前的百分之十,因为当时的债务只有现在的一小部分。所以百分之五、六的利率带来的痛苦,可以很快像当年百分之十的利率那样在经济中扩散。人们看利率时忘记了绝对美元的冲击:当我们整体需要拿出三四万亿美元来付息时,怎么拿?因此我认为存在一个不可忽视的风险,很小,所以才反共识。一旦出现按市值计价的问题,或者企业、足够多消费者的信用违约问题,我认为最可能触发的是一场准备金危机。而这场准备金危机会落在某一家大型银行身上。我心里有两家比其他更明显,但我不想点名。

卡拉卡尼斯:我很喜欢这个反共识观点。

查马斯:我认为有一个小而合理的概率。

贝克:首先,可以让Gemini深度研究算一下「美利坚治下的和平」的未偿债务总额,套上当前市场利率,再看利息支出相对GDP的比例,做一张历史图,会很有意思。我不反对查马斯,什么都有可能,大银行确实可能出问题,我不会说「很可能」,但一切皆有可能。我最反共识的观点是:未来四年里的某一年,美国会录得超过百分之五的实际GDP增长。由于AI和放松管制,生产率会垂直起飞,这可能是又一个九十年代末。百分之五、六和百分之二、三听起来差别不大,其实天差地别:百分之五、六的增速下,经济大约每十二年翻一番,百分之三则要二十四年,对国家和每个美国人的财富都是巨大的差异。至于2025年的具体预测,因为我不知道那一年会在何时出现,我认为前沿实验室会停止发布最尖端的模型,以防止知识蒸馏,防止他们的知识产权实际上被偷走。中国的DeepSeek令人印象深刻,但它以为自己是GPT-4。所以实验室会把最好的模型留在内部,蒸馏之后陆续放出小模型。

卡拉卡尼斯:我刚把问题输进了Gemini Advanced 1.5 Pro的深度研究:「哪些银行最有可能资不抵债或者爆发金融危机?」它现在正在研究,大概十分钟后出结果,可见它做的研究量有多大。弗里德伯格,你的反共识观点?

社会主义回潮

弗里德伯格:官方口径是,社会主义在这次选举中被击败了,美国民众用选票响亮地否定了社会主义。我的反共识观点是,2025年我们会看到美国社会主义运动的急剧兴起。正如加文所说,我们会看到进步的加速,放松管制和AI会释放经济增长。但我们也谈到了美国汽车制造和其他一些行业的衰落,2025年会发生真正重大的转变,一些行业和公司会成为大赢家,一些会成为大输家,经济的一部分会大赢,另一部分会大输。当变化来得这么快时,往往有大批人被甩在后面,一旦如此,社会主义政策和运动就会壮大。庇隆在二十世纪四十年代中期在阿根廷上台时,那个国家的GDP增速是百分之八。增长并不意味着人人平等受益。有人会看着一些人从亿万富翁变成千亿富翁,看着世界上第一个万亿富翁出现,这同样会助燃这股潮流。所以我认为美国社会主义运动的广度和深度都会增加,尤其是当DOGE削减惠及个人的政府项目、联邦政府和联邦承包商裁员的时候。大量的剧变正要惹恼很多人。

贝克:我完全同意。人们喜欢说,在通用人工智能(AGI)或者超级人工智能(ASI)的世界里,钱将毫无意义。但在一段短暂的时期里,钱会比以往任何时候都重要,因为你能花在AI上、花在测试时算力上的钱,会给你巨大的优势,无论你是公司还是个人。所以我认为在一段时间内,AI会真正放大不平等。

弗里德伯格:我希望我是错的,希望AI能为大家创造各种机会。但2025年会有大量的就业和收入动荡,它会助燃社会主义运动,这会是更艰难的一年。我们都在硅谷科技行业工作,所以觉得前景一片玫瑰色,但对大多数美国人来说,地面上的现实可能比我们任何人预想的都严酷,这会造就一个非常艰难的政治和社会环境。

卡拉卡尼斯:这是一个很好的反共识预测,我想补充一点:这不只会打击蓝领。我在风投行业和创业者中间到处看到,那些资历极好、年薪六位数甚至六位数中段的人找不到工作,或者找不到和之前薪酬相当的工作。为什么?因为大家在用更少的人做更多的事。投资于深度研究和AI、把事情自动化、废弃掉或者外包到其他地区,在某些情况下比雇美国人更划算。这种思路不只会打击卡车司机,它会打击开发者、设计师、写作者。

查马斯:你们说的这些,反共识的部分在哪里?

弗里德伯格:官方口径是社会主义今年被打回去了,这次选举是一份反对社会主义及其政策的授权。反共识的观点是,我们其实完全搞错了,社会主义会大幅上升。

贝克:接着你刚才说的,「觉醒主义」和进步主义会和社会主义脱钩。觉醒主义和进步主义会走下坡路,但社会主义,也就是政府扮演更大角色的政策,会上升。

卡拉卡尼斯:为什么我们没有全民医保?为什么不是每个州都有学前教育?我认为美国人有权问这些问题。其他那些行为当然令人厌恶,但你确实有权问,为什么我们搞不定这些,为什么政府在课后项目、全民医保、全民托儿这样基本的事情上辜负了我们。这些事情很容易做到。当你看到所有人都在发财、财富两极分化,我能理解有人会说:别的国家都有这些基本的东西,我们为什么没有?这是一个合理的问题。

弗里德伯格:你做了一个预设,说这些事情很容易做到。当年关于教育也是这么说的:通过联邦学生贷款项目让每个人都能上大学。结果呢,这些项目一推出,学校就开始涨价,学费年年上涨,最终教育成本膨胀到超过了它带来的收益。这种事在医疗上发生了,在住房上发生了,在教育上发生了,在每一个政府出手提供资金支持的市场里都发生了:市场基本上不再自由运转。

卡拉卡尼斯:我接受你的质疑,也解释一下我为什么认为容易。如果把这些事下放到各州,引入竞争,我认为问题可以解决。引入学券制,制造竞争;把全民医保拆开,做一些实验,让不同的州从联邦政府领钱,跑五十个不同的实验,我认为能解决其中一些问题。但你说得对,联邦政府做的任何事情最终都会腐败、低效。我最反共识的观点是:OpenAI失去领先地位,非营利转营利失败,沦为AI领域的第四名。OpenAI的全面崩塌,这是我2025年最反共识的预测。

查马斯:你这是在加倍、三倍下注。

卡拉卡尼斯:我在努力让这个环节辣一点,这本来就是辛辣的环节。也许萨姆·奥特曼很了不起,我不知道,他的朋友全都离开了公司,也许他是个正派人。

查马斯:加文,你知道杰森预测过杰夫·贝索斯会竞选总统吗?

卡拉卡尼斯:那更像是我的愿望。他买了《华盛顿邮报》,辞掉了亚马逊的工作,在华盛顿买了房子,去了海湖庄园。等贝索斯度过他的中年危机,去完科切拉音乐节、尽情狂欢之后,如果他回来说「我想为国家效力」,别感到意外。

查马斯:你还坚持《华盛顿邮报》的预测吗?你觉得他会留着它吗?

卡拉卡尼斯:那就定一条:杰夫·贝索斯将在2025年出售《华盛顿邮报》。

查马斯:这条适合放到Polymarket上。卖给卡拉·斯威舍,让她把它开进沟里。

最佳资产

卡拉卡尼斯:表现最佳的资产,大家都喜欢这个环节,因为可以直接看价格。去年查马斯做了一笔价差交易:做多公开市场科技股,做空私募后期科技股,纳斯达克100科技股ETF在2024年涨了百分之十。弗里德伯格选的是铀矿ETF,2024年跌了百分之一。

弗里德伯格:横盘。上周准备节目时我看了那只ETF的成分股,里面全是垃圾,它不是交易铀的正确方式。但不管怎样,那是我去年的结果。

卡拉卡尼斯:我选的是按需经济,Uber、Airbnb和DoorDash。Uber一度涨了百分之三十,但受Robotaxi的逆风影响,现在只剩百分之三;Airbnb基本持平;DoorDash大涨百分之七十四,向斯坦利·唐和团队致敬。今年的最佳资产,让嘉宾先说。

贝克:高带宽内存(HBM)的制造商,回到我说的「我们会把算力用光」。有一个相当惊人的数据:在GPU的销售成本里,高带宽内存的占比比台积电还大。今天只有两家公司能造,SK海力士和美光,三星能不能振作起来还得看。HBM用在英伟达的GPU里,AMD的GPU里,亚马逊的Trainium里。尤其在测试时算力和推理如此重要的世界里,高带宽内存可以说比以往任何时候都重要,而且过去两年一直处于售罄状态。所以我选高带宽内存。

卡拉卡尼斯:好选择。查马斯?

查马斯:先做个铺垫:这个选择一百次里有九十二次会归零,剩下八次里有六次赚十倍,最后两次赚一百到一千倍。

卡拉卡尼斯:听起来像比特币。

查马斯:所以这是一笔会亏的交易。但我会做多信用违约互换(CDS)。我买的是什么?是保险。我用信用违约互换买所谓的「保护」,赌2025年不会有违约事件。我不会说是哪些公司、什么期限,但总体的想法是:我想在2025年买一份小小的保险,让我们选出来的那些男男女女有机会安安静静地把工作做完。我认为明年有很小的概率出现波动,我希望它不发生,我希望这笔交易像我说的那样,一百次里有九十二次亏钱。但如果它命中,它就会是2025年表现最佳的资产,相当于阿克曼在新冠危机刚开始时买入CDS。

卡拉卡尼斯:怎么买?

查马斯:你需要一份ISDA协议,去找大投行,他们会给你报价。我想说清楚,这不是我认为会发生的事,也不是我希望发生的事。但从能赚到的美元吨位和它呈现的巨大风险不对称来看,当你看到标普指数的集中度,看到债务总额的规模,看到利率飙升,所有这些都在说:手里有一点保险未必是坏事。

贝克:如果查马斯关于银行倒闭的预测成真,你绝对想持有CDS,那就不是一百倍,是一千倍、一万倍。

卡拉卡尼斯:弗里德伯格?

弗里德伯格:我选中国科技股,或者中国科技股ETF。过去几年所有人都抛售了中国科技股,都采取隔离的立场来配置组合,说「我们不能和中国做生意,结束了」。但我认为特朗普政府,尤其从他们最近要求暂缓TikTok禁令来看,正在筹划一个我称之为「对华大协议」的东西。我不知道他们具体想做什么,但我认为特朗普和美国政府的领导层想向美国公司打开中国市场,作为交换,让中国公司进入美国市场,而以中国目前的处境,他们很可能达成协议。这是第一个驱动因素。第二,中国电力产能建设的规模难以想象,他们最近批准了一座一千三百七十亿美元的水电大坝,会再增加几百吉瓦的发电能力,更不用说我们过去谈到的核电建设。每度电的成本本来就更低,可用的电力还在增加。第三,中国共产党有一种不可思议的能力,可以调节自由市场和创业活动的油门大小,眼下这个时刻,你可能会看到他们把油门推向鼓励更多创新、更多自由市场活动的方向,这也是他们达成协议的动力之一。把这些放在一起,很多中国科技公司是在「未来几年中美会陷入艰难冲突」的假设下被打压的,而进入2025年,情况可能并非如此。这些股票看起来相当便宜,我刚看了阿里巴巴,估值倍数很合理。如果宏观走对了,2025年可能会有一波不错的行情。

卡拉卡尼斯:这像是反共识的选择。我不认为特朗普能很好地平衡对华关系,我也不认为在上一次把所有人的膝盖都打断之后,还能重新煽起创业热情。而且我不信任任何来自中国的会计报表,他们想怎么改就怎么改。但作为「低买高卖」的机会,确实很有潜力。

弗里德伯格:好处是一年后你就知道我对不对。

卡拉卡尼斯:加文,你怎么看弗里德伯格这个热门观点?

贝克:真的很便宜。十五年多前的东南融通财务造假之后,我就给自己定了一条「不碰中国」的准则。那件事太惊人了,很多杰出的投资者都持有它,公司说得头头是道,还有西方审计师,结果发现造假发生在当地的邮局:他们把当地审计师签好字的文件拆开,改掉,再寄出不同的文件。所以我认为,如果你不是中国人,那是一个很难赚钱的地方。但大卫说的很多我都同意。特朗普和习近平都想要一份协议,而且我前面提到过,有一份可以把普京晾在寒风里的协议可做。如果成了,那就挡都挡不住。那里有一批质量极高的公司,有些估值倍数只有个位数中段,而且那是一个大市场。

弗里德伯格:而且中国公司服务的是全球市场,不只是美国,还有非洲、南美,整个南半球。

卡拉卡尼斯:看看比亚迪,遍布欧洲。

弗里德伯格:中国公司有最好的单位经济,最低的生产成本,各方面都占优。尤其是如果你相信机器人之年,全世界会对自动化和重建制造能力产生巨大需求,中国服务这些市场的效率会高于任何其他原产国。宏观驱动因素相当强劲。

卡拉卡尼斯:我认为七巨头会是表现最佳的资产,我站查马斯的对面。他们过去几年学到的是,通过不招人、把工作外包到世界其他地方、把事情自动化,可以实现惊人的利润扩张。他们为其他公司和消费者生产AI收益,但会先在内部应用。AI的内部应用会让这些公司在未来几年实现人们无法理解的利润增长。

最差资产

卡拉卡尼斯:表现最差的资产。去年查马斯说的是后期科技股指数,主要是SaaS;弗里德伯格做了一笔漂亮的价差交易:做空垂直SaaS,做多AI云服务商,SaaS有一点反弹,但谷歌涨了百分之三十六,微软百分之十四,亚马逊百分之四十六。我说的是OpenAI、Anthropic这类大模型创业公司,因为玩家太多、开源会杀掉定价,能追踪的只有OpenAI的估值,它翻倍了,但我认为长期看这个判断会成立。2025年的最差资产,加文?

贝克:企业应用软件。我认为2025年会是智能体(agents)之年,尤其是下半年。智能体就是能替你采取行动的AI模型,人能在网上做的任何事它都能做。如果实验室和大型云服务商主导了智能体,这看起来很可能,回到「低成本生产者获胜」的逻辑,企业应用软件会非常痛苦。这些公司里有些在大谈特谈智能体,但归根结底,他们没有自己的模型,也不拥有自己的算力,我看不出他们能在智能体的世界里成为最终赢家。我也可能错,他们确实有一些客户数据,也许有一点数据护城河,也有牢固的客户关系。但大多数公司和AWS、谷歌或者微软同样有关系。所以我认为企业应用软件会很难受。

查马斯:加文说得完全对,我要在他的基础上加码。有一个词我们会越来越多地使用,2024年我在8090内部已经开始用了,叫「软件工业复合体」(software industrial complex)。这些庞大臃肿的企业软件公司,实际上是说服了无数组织花巨额资金,去买一堆包在增删改查数据库外面的启发式规则和业务规则。他们真正练到极致的是市场推广和销售动作:高尔夫之旅、牛排晚宴,我们之前提过亚历克斯·卡普痛斥的那些,这些东西没有一样等同于产品价值。而在智能体软件和AI的世界里,这些工作流可以用极高效的方式重建。我认为市场推广会由CEO和CFO驱动,他们会开始向CIO施加更大的压力去管控开支。在那样的世界里,新一代AI企业的建造成本比它们的竞争对手低整整一个数量级,所以哪怕功能一模一样,这些老牌的传统企业软件公司,也就是软件工业复合体,2025年会开始出现裂痕。我和加文用词不同,结论一样。

卡拉卡尼斯:而且你在这件事上投了时间和钱。十一个月前你创办了8090,不到一年就拿下这样的市场牵引力,让我震惊。

查马斯:我不认为这一定说明我们有多厉害。我们不错,很不错,有很好的工程师。但我那个三十人的工程团队,我认为干的是三百人的活,明年会干三千人的活,而人数也许只增加百分之十或二十。原因是我们使用这些工具,生产率高到一个我原来认为不可能的数量级。结果就是我们可以用不同的方式定价:哪怕产品完全相同,我们的成本结构就是明显更低。这就是我说的,投资回报率计算器被炸飞了,所有传统的市场推广模式都被炸飞了:在招标或者任何销售场景里,对方报一百美元,你出现了,报十美元,而十美元对你来说仍然利润丰厚。

贝克:最后补充一点。从根本上说,这些企业应用软件公司,这个软件工业复合体,我同意查马斯说的一切,它们的根基是让白领员工更高效。而AI公司要做的是说:我们干脆把那个员工替换掉。这是一种根本不同的思维方式。

弗里德伯格:这也回到了大卫说的社会主义回潮,很多人会因此受苦。

卡拉卡尼斯:弗里德伯格,你的最差资产?

弗里德伯格:我大概会给「垂直SaaS」再画三道下划线:按席位收费的定价模式受到挑战,价格被压缩,因为企业在探索用AI自建工具来取代这些传统业务流程。这在我的清单上太显而易见了。

卡拉卡尼斯:你有别的吗?OpenAI?

弗里德伯格:OpenAI很棒。所有有才华的人都走了,不过我不知道他们去了哪里。

卡拉卡尼斯:他们都去创业和OpenAI竞争了,「复仇创业」,这是个有意思的趋势。收入还有一百二十亿美元的预测。

弗里德伯格:那是预测,预测有时会撞上现实。对应那一百二十亿收入的,还有六十亿美元的亏损,我想是明年的数字。

卡拉卡尼斯:这就说得通为什么他们现在想收每月两百美元了,有逆风。我的最差资产是这样想的:消费者一生要做五个大决定,大学、配偶、孩子,这三个显然没法交易,剩下的是汽车和房子。而消费者背着创纪录的债务。所以我认为传统车企和房地产会继续面临逆风,是糟糕的资产。汽车在某些地方已经过剩,经销商的停车场上堆满了车;而以现在的房贷利率,人们买不起房。得克萨斯这样的地方,房价和房租连续两年下跌,允许建房的州都在发生同样的事,而人们正在离开不允许建房的州。企业软件是我的另一个选择,但我选这个略有不同的。

最期待趋势

卡拉卡尼斯:2025年最期待的趋势。先回顾去年:查马斯,恭喜,你说比特币会首次冲上十万美元,说中了,半场投篮命中。

查马斯:当时感觉像个上篮。

卡拉卡尼斯:涨了百分之一百一十二。弗里德伯格说的是制药和生物工程领域的预测模型和AI驱动的发现。

弗里德伯格:那个领域拿到了大量融资,德米斯·哈萨比斯还得了诺贝尔奖,所以成立了。

卡拉卡尼斯:我选的是效率,也就是AI在劳动力和外包上的进步,我们投的Athena让我看到了很多。今年的最期待趋势,查马斯?

查马斯:我不想滥用「矿井里的金丝雀」这个词,但确实有几只金丝雀预示着「深层政府」的终结。我想指出其中一只,一个很冷门的东西,叫补充杠杆率(Supplementary Leverage Ratio)。它本质上是一种机制,决定银行在计算准备金时是否把国债计入。为什么有意思?对多数人来说它并不有趣,非常晦涩。但如果2025年我们无法管好债务问题,你会看到在这些晦涩法规的边缘做手脚,把问题往后踢。这有可能发生,目的是帮银行一把,因为我们,也就是整个美国,可能认为DOGE的效果不会达到它需要达到的程度,我们仍然需要发行一二十万亿美元的债务,来为今年到期的十万亿再融资,并且填补未来几年的窟窿。我的意思是,我们不需要弄懂细节,只需要知道:如果我们不去挪动这道球门,那就是一个很好的信号,说明掌舵的是我们选出来的人。这就是我最期待的趋势:那些让我们把问题往后踢的、细小晦涩的监管变动,被就地叫停。

弗里德伯格:我最期待的趋势,是2025年美国核电建设的宣布,作为放松管制和一些新技术的结果。我认为新政府会宽容得多,而且正如我过去说的,这是必需的,因为为了在竞争中跟上中国,我们必须以极快的速度建设电力,美国需要新增的发电能力超过任何其他可再生能源能扩张的规模。所以我相信核电是必然的,放松管制会在2025年发生。我认识很多非常聪明的人,他们辞掉了很好的工作去创办核电公司,就是押注这件事会在今年发生。我非常看好。

卡拉卡尼斯:聪明人把时间投在哪里,绝对是一个领先指标。加文?

贝克:AI。尼克,我发了你一张图,是ARC-AGI的成绩。我认为2025年AI每个季度取得的进展,会超过2023和2024年每一年的进展。原因就在o1和o3。我们不断挪动图灵测试和AGI的球门,我确信我们还会再挪,因为我们会把它们全都冲过去。发生的事情是:我们原来只沿着一条轴扩展,就是预训练;然后开始沿着推理时算力扩展;现在很清楚,我们加上了第三条性能扩展轴,那就是推理。互联网是由答案组成的,人们给出答案,而模型真正受益的是一个人抵达答案过程中的内心独白,AI术语里叫「推理轨迹」(reasoning trace)。这就是为什么十八个月前大家都发现,模型训练的代码越多,它在许多看似与代码无关的事情上表现越好,因为在代码里你能看到推理、内心独白、思考过程、一步一步的推导。现在人们用模型生成含有这些推理轨迹的合成数据:让模型解一道题,这道题必须是可以验证的、有答案或者我们知道答案的,然后说「展示你的过程」,让它用很多不同的方式做很多遍,挑出最好的,喂回模型,再做一些强化学习。于是你就在三条彼此相乘的轴上同时扩展。谷歌团队里有个人大约五天前发推说,「这会是直通ASI的一条直线」,我认为他可能是对的。伊利亚·苏茨克维,这个领域的先驱之一,在NeurIPS上做了一个演讲,说了一句我觉得可怕的话:会推理的模型天生不可预测。今天世界上最好的推理模型是下棋的那些,AlphaGo、AlphaZero一类,它们不断走出人类大师从未想到的不可预测的棋。现在这些模型会在各种领域做出同样不可预测的跳跃,希望是好事,但也可能不是。

卡拉卡尼斯:它们会绕过人们不曾考虑的、反直觉的弯角。回到刚才那个问题:哪些银行最有可能资不抵债或爆发金融危机,我打字时错字连篇。你们看深度研究做了什么:它先列出计划,研究一批网站,找出美国的银行名单,按资产找出最大的银行,找每家的最新财务报表,找每家的资本充足率,找贷款损失拨备,你们明白意思。然后分析,然后生成报告。整个过程大约十分钟。最终报告列出了摩根大通、美国银行、花旗集团等等,逐家分析,末尾列出了它抓取的全部网站,超过一百个,是实时数据,一百六十二个网站。它的结论是:这份分析给出了美国最大几家银行财务健康状况的快照,所有银行都面临固有风险,但与摩根大通、美国银行和高盛相比,花旗集团和富国银行的资不抵债或金融崩溃风险似乎最高,花旗最近的净亏损、更低的一级资本比率和高敞口是原因。我不知道这里面有多少是对的,我不是这方面的专家,但作为起点看起来相当不错。如果你还没用过1.5 Pro的深度研究,去gemini.com,没人付我钱,我只是认为它是市场上最好的产品。我最期待的趋势,加文自己前面提到过:随着莉娜·汗的怒火结束,并购和IPO激增,退出和向出资人的分配(DPI)会像雨一样落下来。

最期待的媒体内容

卡拉卡尼斯:最期待的媒体内容。我有两个候选:詹姆斯·古恩的DC宇宙,《超人》今年上映,我觉得会很精彩;《安多》第二季也可能很精彩,我预测他们会做真人版《克隆人战争》,如果你不知道那部动画,和孩子一起看,它讲的是前传时期的故事。但我最后选了一个有点奇怪的:看那些由亿万富翁拥有的传统媒体会发生什么。《华盛顿邮报》、CNN和《洛杉矶时报》正在转向中间,试图回归经典新闻业,编辑们在造反,卡拉·斯威舍很不高兴,他们在增加一些右翼声音。这将是所有人的爆米花时刻。查马斯?

查马斯:特朗普政府将解密和公布的文件,规模之大。我认为那会极其有趣、耸动、有用、真诚,以上全部。肯尼迪遇刺档案、爱泼斯坦档案、吹牛老爹的档案、登月,以及在所有那些「阴谋论」里可能真有几分真相的边缘话题,谁知道他们会找到什么。所有这些在2025年由特朗普政府公布的「内容」,姑且这么叫,会非常有意思。

弗里德伯格:我关注的是AI视频游戏。用AI之后制作成本大幅下降,你可以有动态的故事线,可以有今天不存在的新玩法概念。游戏开发工作室里常见的创意人才和技术人才,可以被最近上市的工具释放出来。我们已经看到很多生成式视频,但在结构上,视频游戏也可以被重建:游戏引擎可以在本地运行,生成参数,再交给现有的渲染引擎,于是可以有全新的故事线和情节序列。我认为游戏和游戏行业会被现在涌入市场的各种AI能力重写。那会是不可思议的娱乐,不玩游戏的人也会找到他们喜欢的东西。

贝克:毫无疑问,《1923》第二季。看电视我通常偏爱科幻、奇幻或者间谍片,但《1883》和《1923》是第一批像《权力的游戏》那样打动我的剧集,我等第二季等得发疯。

卡拉卡尼斯:你在看《Landman》吗?

贝克:还没看,很期待。泰勒·谢里丹这个人太厉害了,正处在巅峰状态。我也强烈推荐《豺狼的日子》。

查马斯:《豺狼的日子》真的很好。还有《边境杀手》。

卡拉卡尼斯:去年的媒体预测:查马斯说的是MrBeast,弗里德伯格说的是AI生成的新闻,我说的是《角斗士2》和《三体》,《角斗士2》平平,《三体》我没看完,也算平平。MrBeast在亚马逊的节目,在一百八十个国家里的一百四十个登顶无剧本类第一,了不起。

预测市场题目

卡拉卡尼斯:现在来出我们的预测市场题目。我的题目关于移民,特朗普承诺在第一年驱逐一千五百万人,我把标准线设在这个数字的百分之五,也就是七十五万。上任一年后,特朗普驱逐的人数会多于还是少于七十五万?

弗里德伯格:怎么衡量?得有一个事实来源。

卡拉卡尼斯:以白宫公布的驱逐数据为准,我们和Polymarket团队确认能不能这样设。这是这次大选最大的议题,所以我只按第一年的百分之五来设。

贝克:有意思的是,奥巴马驱逐了我想超过两百万人,人们好像忘了,他在边境问题上其实非常强硬。

卡拉卡尼斯:查马斯,你的题目?

查马斯:七巨头在标普500中的权重萎缩到百分之三十以下。

卡拡拉卡尼斯:好题目,意思是行情会分散到其他股票上,集中度下降。弗里德伯格?

弗里德伯格:我想玩的是微软、AWS和谷歌云的收入增长:2025年谁的增长最快?我感觉谷歌在加速领先。加文,你怎么看?

贝克:谷歌规模小得多,所以更容易增长得更快。

弗里德伯格:那就改成:2025年云收入的美元增量谁最大,谷歌、亚马逊还是微软?

贝克:大概不会是谷歌,因为如果它的美元增量最大,就意味着它在以疯狂的速度增长。不过Azure对AWS会很有意思。

卡拉卡尼斯:我还有一个题目,关于国债。前两届政府任内,国债每年增长大约两万亿美元,八年增加了超过十六万亿。特朗普说他不会增加国债,我把线设在明年国债增加一万亿。还是应该设两万亿?

弗里德伯格:我的做法是,以美国财政部2025年12月的联邦债务报告为准,设在三十七万亿美元。

卡拉卡尼斯:那肯定会超过,干脆设三十八万亿。

弗里德伯格:设线的时候,得让两边都有人愿意下注,不能所有人都站一边。那就定为:美国财政部2025年12月的报告,联邦债务高于还是低于三十八万亿美元。

卡拉卡尼斯:也就是一年增加大约一万五千亿,完美。看看他能不能控制住开支,还是一切照旧。加文,你怎么看明年的联邦债务?

贝克:和大多数事情一样,像移民一样,需要时间才能起势,但我认为他们会取得进展。顺便说一句,如果我下次再来All-In,我想聊聊UFO和无人机。

UFO与金字塔

卡拉卡尼斯:现在就聊。你要走全面阴谋论路线吗?你相信那些无人机是外星人?

贝克:我不知道。但如果你看看新泽西州那些市长和州长的说法,他们见过警察、联邦调查局和国防部,他们显然也不知道。特朗普说政府里有人知道那是什么,他还说自己短期内不会去贝德明斯特的度假村。过去大约八年里,每隔十八个月,《纽约时报》、《华盛顿邮报》、《纽约客》、《大西洋月刊》这些非常可信的媒体就会有一篇大报道,采访几十位战斗机飞行员和民航飞行员,讲他们用先进传感器看到的、无法解释的东西。曾经的禁忌消失了。《纽约客》有一篇文章引述了一批二十世纪五十年代在「臭鼬工厂」实验室工作过的人,他们说绝对见过从坠毁物中回收的地外材料,这是美国在材料科学上取得巨大飞跃的原因之一。还有「奥陌陌」(Oumuamua),当时很多天文学家说这显然是某种UFO,包括哈佛天文学系的主任。至于新泽西,如果那些不是UFO,而是美国或中国政府的无人机,最可能的解释是某种演习。而且一旦歇斯底里蔓延开来,人们会把民航客机误认为无人机,UFO为什么会有一闪一闪的红绿灯呢?但有一件事很有意思:1945到1960年我们向核技术扩张的时候,出现过一大批这类报告的集中期;而现在AI起步了,这是人类下一次技术相变,又出现了一大波报告。

卡拉卡尼斯:你可以下周再来了。恭喜,All-In偶像的赢家诞生了,加文现在是第四位老友。

弗里德伯格:你觉得有哪些关于UFO的文件是特朗普会公布的?肯定有一大堆可以走信息自由法申请的东西。

贝克:取决于保密级别,如果政府不想交出来,就不会交出来。他们现在把这叫做「不明异常现象」(UAP),不叫飞碟了,但感觉背后确实有些什么。

卡拉卡尼斯:你们认为美国政府手里握有地外生命证据的概率是多少?

查马斯:至少百分之二十。

贝克:我说百分之二十五,我买二十的上方。

卡拉卡尼斯:弗里德伯格,科学苏丹,你认为概率多大?

弗里德伯格:这需要更长的谈话,现在没时间。但说实话,我一般不认同这样的想法:以我们对技术和生物极其狭窄的理解,去设想什么东西会来拜访我们。我认为信息采集有一种延伸方式,不需要把物理的生物形态从银河系的一端搬到另一端。「UFO运送肉身」这整套前提,植根于人类当前的技术水平。亚瑟·克拉克为《2001太空漫游》写的故事大纲把这一点讲得最好,那份大纲早于剧本,也早于小说,现在可以买到。它的意思是,每一个文明最终都会到达足够先进的技术水平,不再需要物理地移动生物有机体的身体,而只是采集信息、影响信息。一旦你能把任何分子变成任何其他分子,并且拥有足够的能量,你最终会进化到这一步,你就能把你所在的那一小片宇宙变成任何你想要的样子,然后只需要从宇宙的其他各处收集信息。所以我认为,外星人的身体要满银河系跑,在基本逻辑上说不通。

卡拉卡尼斯:我现在就放话:如果特朗普公布地外生命的信息,我投票支持他连任第三届,修改宪法也行。最后一个问题给加文:你认为金字塔是地外生命建的吗?以它的精度,这是最合理的解释吗?

贝克:我不知道,但确实很奇怪。我们离克利奥帕特拉的时间,比她离大金字塔建成的时间还近,那是非常久远的事。奇怪的是,世界各地那么多文化里都有看起来像宇航员的描绘,还有那么多文艺复兴时期的画作,清楚地画着城市上空我们今天会称为UFO的东西。所以我不知道。

查马斯:我和你一样喜欢阴谋论。如果你是一个爱追问的人,这些东西会留下奇怪的、无法闭合的开放问题,就像模拟宇宙假说一样。

卡拉卡尼斯:关于金字塔,我看过的最可信的说法是他们用了水:把石块放在木排上,开凿水渠,像巴拿马运河那样分级抬升水位,把石块浮上去,放到位,再放水。我觉得挺有道理。这是All-In播客又一期精彩的节目,感谢各位,祝大家有一个了不起的2025年。

本期讲者
加文·贝克Atreides Management 创始人兼 CIO,管理约 40 亿美元,曾在富达任科技基金经理 18 年,是本期唯一的外部嘉宾。
查马斯·帕里哈皮蒂亚Social Capital 创始人,前 Facebook 增长副总裁,2024 年创立 AI 企业软件公司 8090,All-In 播客四位常驻主播之一。
大卫·弗里德伯格农业科技公司 Ohalo 创始人,此前创办 Climate Corporation 并以约 10 亿美元卖给孟山都;本期主持人 Jason Calacanis 是天使投资人、Launch 基金创始人。
章节 · 点击跳转视频
0:00 开场:嘉宾 Gavin Baker 与 Polymarket 合作 ▶ 正在看
5:10 2025 政治赢家与输家预测 ▶ 正在看
17:29 商业赢家:机器人、AI 与稳定币 ▶ 正在看
31:57 商业输家:体育、承包商与 OpenAI ▶ 正在看
46:38 年度大交易:车企合并与自动驾驶整合 ▶ 正在看
56:11 最反共识:银行危机与社会主义回潮 ▶ 正在看
1:12:30 最佳与最差资产:HBM、CDS、企业软件 ▶ 正在看
1:30:06 最期待趋势:核电与 AI 三轴扩展 ▶ 正在看
1:38:51 最期待媒体与 Polymarket 盘口 ▶ 正在看
1:46:34 UFO、无人机与金字塔闲谈 ▶ 正在看
本期论点
本期回应
其他论点
18:07
2025年将是机器人和自主硬件之年,正如2024年是算力建设之年 大卫·弗里德伯格
23:13
美元稳定币是2025年最大的商业赢家,使用量到年底可能翻四到五倍 查马斯·帕里哈皮蒂亚
33:13
职业体育球队的买家正从富有个人转向机构资本,入场资金将增长10到100倍 加文·贝克
40:06
「七巨头」在指数中集中度已近40%,历史上这种集中往往预示万亿级回撤 查马斯·帕里哈皮蒂亚
41:27
传统国防承包商将败给Palantir、Anduril这类科技导向的国防公司 David Friedberg
48:57
传统车企做不出有竞争力的产品,将失去中国市场并被特斯拉和中国车企夹击 加文·贝克
51:38
AI最终赢家是算力成本最低者,租用算力者无法成为低成本供应商 加文·贝克
1:01:30
在70多万亿美元债务规模下,5%利率对经济的压力相当于25年前的10% 查马斯·帕里哈皮蒂亚
1:04:45
前沿实验室将不再发布最领先的模型以防蒸馏窃取,只逐步放出蒸馏后的小模型 加文·贝克
1:11:34
OpenAI会失去领先地位、搞砸营利化转型,沦为AI领域第四名 查马斯·帕里哈皮蒂亚
1:14:57
高带宽内存是2025年最好的资产:占GPU成本比台积电还高且只有两家供应商 加文·贝克
1:18:32
特朗普政府将与中国谈成互相开放市场的大交易,被打压的中国科技股被严重低估 David Friedberg
1:24:01
智能体时代由大模型实验室和云厂商主导,企业应用软件将是2025年最差资产 加文·贝克
1:34:08
AI在2025年每个季度的进展将超过2023和2024年各自全年的进展 查马斯·帕里哈皮蒂亚
01开场:嘉宾 Gavin Baker 与 Polymarket 合作
0:00
welcome back to the all-in podcast everybody I'm your host Jason calacanis I'll put in a bunch of plugs for the projects I'm working on throughout the show to annoy my co-hosts and continue the grift but it is 2025 and we are doing our bestie Awards it's going to be amazing today we've got so much to do today with us again for the second time a truly amazing bestie Gavin Baker from at trees Capital am I correct is it a treaty Capital you put management tradies management oh a tradi management okay and uh just call it House of treaties oh no let's not get him in trouble I P Gavin welcome back to the program let us know um just briefly what does a treaties do thanks for having me here Jason chamama and Dave a tradies we're a we're crossover firm we invest publicly and privately in consumer and Tech and we go from series a to Mega cap got it okay so you are a capital allocator you're place bets on technology on the most important new companies in the world we'll get into that and this is our prediction
欢迎回到 All-In 播客,各位,我是主持人 Jason Calacanis,整期节目我会不停给自己在做的项目打广告,气一气我的联合主持人,把这个套路继续下去不过现在是 2025 年,我们要办我们的 Bestie 大奖了今天会非常精彩,今天有一大堆内容要聊,第二次和我们同台的是一位真正了不起的 bestie来自 Atreides Capital 的 Gavin Baker,我念对了吗?是 Atreides Capital 吗?你们叫 Atreides Management,哦,AtreidesManagement,好的,呃,干脆叫 House of Atreides 得了,哦不,别害他惹麻烦,Gavin 欢迎回到节目,简单跟我们说说,Atreides 是做什么的?谢谢你邀请我,Jason、Chamath 和 David,Atreides 是一家跨界基金,我们在消费和科技领域同时做公开市场和私募投资,从 A 轮一路投到超大市值公司,明白了,好所以你是资本配置者,你在科技领域下注,投世界上最重要的新兴公司,我们待会儿再展开聊,这期是我们的预测节目,我们已经办过 Bestie 大奖了,Gavin,你们基金规模多大?哦,你才刚认识人家
便签引用
1:04
show we did our besti Awards how big is your firm Gavin oh well you just met the guy free brg you don't just ask him how big his firm is the audience wants to know is this guy a player like what's the deal roughly $4 billion okay all right so that's about a half an inch bigger than shth he's really worried now's not that he's the size queen or anything let your winners ride Rainman David and instead we open sources to the fans and they've just gone crazy [Music] with we're going to do a really great prediction show today and we're going to do some super predictions this year each bestie gets to make a super prediction or two during the show and we're going to take those super predictions shamal and we're going to put them on poly Market yes that's right if you don't know polyk Market there a prediction Market where people can place a wager a bet an investment in one or the other side and uh they're p is my image good it feels like it's blurry it's a little blurry that's the Avalanche coming and so I guess is that
Freeberg,你不能一上来就问人家基金多大,观众想知道嘛,这哥们儿是不是个大玩家,到底什么情况?大概 40 亿美元,好的,行,那也就比 Chamath 大了半英寸,他现在可紧张了,倒不是说他是尺寸控还是怎么样,让赢家继续跑,Rainman David,然后我们把它开源给粉丝,他们已经彻底疯了[音乐] 今天我们要做一期非常棒的预测节目,而且今年我们要做一些超级预测,每位bestie 可以在节目中做一到两个超级预测,然后我们要把这些超级预测,Chamath放到 Polymarket 上去,没错,如果你还不了解 Polymarket,那是一个预测市场,人们可以在上面下注押其中一方,做一笔投资,然后他们的——我的画面还好吗?感觉有点糊,是有点糊那是雪崩要来了,那我猜这是 MAGA 雪崩?这是什么?雪崩是指人群尖叫着越过边境吗?这里的比喻是
便签引用
2:15
the Maga Avalanche what is this Avalanche is that people screaming across the border what's the metaphor here TR up is that the huge orgasm uh that you're about to lay down in 2025 rocking my friend no no this is a shout out to my friend Ruben Oland who's the director of a bunch of really amazing movies um including the one behind me Force measure there you go all right we did an amazing partnership with poly Market freeberg you want to detail the partnership and some way you I think you helped work on it well we obviously have talked to Shane at Poly market for some time and we set up a a deal with them where we could put our own markets up on the on the site so we're going to talk about a couple of our predictions this year and then poly Market will post them up on on their site and people can can trade them obviously xus and we're really excited about it because it'll allow us to kind of track how we're doing and give us the opportunity on an ongoing basis to create market so there'll be an
什么?还是说这是你 2025 年即将释放的那场大高潮,牛啊兄弟,不不,这是向我朋友 Ruben Östlund 致敬,他导演过一堆非常棒的电影,包括我身后这部《游客》,就是这个,好了,我们和 Polymarket 达成了一次很棒的合作,Freeberg,你来讲讲这个合作吧,我记得是你参与推动的,嗯,我们显然已经和 Polymarket 的 Shane 聊了一段时间了,我们和他们谈成了一个合作,我们可以在他们网站上开设自己的市场,所以我们今天会聊我们今年的几个预测,然后 Polymarket 会把它们挂到他们网站上,人们就可以去交易这些预测,当然是在美国境外,我们对此非常兴奋,因为这样就能追踪我们的预测表现也让我们今后能持续创建新的市场,所以 Polymarket 上会有一个 All-In 专区,你
便签引用
3:12
all-in section on poly Market where you can kind of go in and see the all-in market it's going to be really cool awesome fantastic and so uh we just got we're all wrapping up our our uh ski trip here Nick I understand there was some footage some found footage from the bestie skiing is that true oh here we go look at this here I come down the mountain Gavin I want you to rate everybody skiing here I come looking great there comes looking good and then Here Comes somebody I think the technical term Joe Lonsdale does for that third person is a beep beep okay here we go and look at this technique coming down look at chamama chamat you really Advanced since last season look at him cutting those S turns Gavin what do you think listen to me I I made a very strategic decision to stop snowboarding so I could learn to ski with my kids it is so hard to learn how to do something in your 40s no I mean if you've never done it but I do appreciate what you and your brother did for me last season I learned a couple
可以进去看 All-In 的市场,会非常酷太棒了,好极了,那么呃,我们刚刚结束了我们的滑雪之旅,Nick,我听说有一些片段,一些 bestie滑雪的“失踪录像”,是真的吗?哦,来了,看这个,我从山上滑下来了,Gavin,我要你给每个人的滑雪打个分,我来了看起来不错,来了,姿势挺好,然后来了一位,我觉得 Joe Lonsdale 给第三位这种滑法起的专业术语叫“哔哔”,好,来了,看这个技术动作,看 Chamath 滑下来,Chamath 你真的进步了不少,自从上个赛季 看他滑的那些S形弯道 Gavin你怎么看 听我说 我做了一个非常有战略性的决定就是不再玩单板 这样我就能跟孩子们一起学双板 四十多岁才开始学一样新东西真的太难了 我是说如果你以前从没做过的话 不过我真的很感谢你和你哥上个赛季对我的帮助 我学到了几招
便签引用
4:14
things so it's been my thir there with the black bomber my third season it was great I am absolutely amazed at the progress you're making freeberg I'm amazed you made it onto the mountain you you were out there a number of days you're getting it done I needed new boot and I tore my MCL last season so I'm in it got really tweaked when I went out there but I'm good it was great I had a lot of fun I mean I four days yeah you ski four days okay I finished my 16th day yesterday but I took today off just to get prepared here yeah I got 16 I do the executive program Gavin I go off for an hour and a half to three hours Zip Zip Zip 10 runs and I'm done I only ski in the morning yeah that's the way to do same same same and and at 1215 have lunch relax get a little work done need a healthy healthy mix of AA ski and ski Gavin how would you compare yourself to what you just witnessed in that video your skiing ability where would you rank yourself clearly it's me chamath freeberg in that ranking where would you
所以这是我的第三个赛季了 带着那件黑色羽绒服 感觉很棒 我对你的进步简直惊叹Freeberg 我惊讶的是你居然真的上山了 你确实滑了好几天你还是滑下来了 我需要换双雪靴 而且上个赛季我拉伤了内侧副韧带 所以我出去的时候伤又被扯到了 不过我没事 挺好的 玩得很开心 我滑了四天 对 你滑了四天 好 我昨天滑完了第16天但今天休息一下 为这期节目做准备 对 我滑了16天 我走的是高管路线 Gavin 我就出去一个半到三个小时 嗖嗖嗖 滑十趟就收工 我只在上午滑 对 就该这么滑一样一样 然后12点15吃午饭 放松一下 干点活 得健康地把滑雪和工作搭配起来Gavin 拿你自己跟你刚才在视频里看到的比 你的滑雪水平能排第几排名里显然有我、Chamath和Freeberg 你觉得自己排在哪 我 我会说
便签引用
022025 政治赢家与输家预测
5:10
land in the ranking I well I would say i' make two observations first like I'm in the bottom percentile in terms of natural athleticism but I do have many many thousands of hours skiing great okay and I would just leave I would just leave that yes well then we'll be going out with you next year new best all right we got to get to it we're going to start off with politics gentlemen we got to keep this moving because there's so many predictions we're going to do our 2025 prediction for biggest political winner now last year well yeah I think we we got it kind of wrong freedberg you said independent third party in the US and we did see some of that with you know the breakout of Robert Kennedy maybe chamat you said independent Centrist we didn't get there with that one and I said Darkhorse presidential candidates maybe I get a quarter Point Credit there with wait hold on what do you mean we didn't get there independent centrists won the election for mean independent Centrist won the election really that's what you
我想说两点 第一 论天生的运动细胞我排在最后几个百分位 但我确实有非常非常多个小时的滑雪经验 很好 好 我就说到这吧 好 那明年我们跟你一起滑 新最佳 好 我们开始吧 先从政治说起 各位 我们得快点推进因为要预测的东西太多了 我们来做2025年最大政治赢家的预测 去年嗯 我觉得我们算是猜错了 Freeberg 你说的是美国的独立第三党 我们确实看到了一些苗头 比如小罗伯特·肯尼迪的崛起 Chamath你说的是独立中间派 这个没实现我说的是黑马总统候选人 也许我能拿四分之一分 等等 你说的没实现是什么意思独立中间派赢了这次大选啊 独立中间派赢了大选 真的吗 你是这么
便签引用
6:09
believe yeah okay I didn't think that it was the Independence but okay I think independent Centrist swung the election for Biden in 2020 and then I think they swung the election for Trump in 2024 huh is that statistically correct you think freeberg you about that independent centrus did the election or Gavin what do you think I think it's right like I think a lot of there were a lot of people not just cists but a lot of people who had been lifelong Democratic voters who voted for Trump so I do think Trump won the cists I mean Trump trump won everything essentially yeah okay shabat who's your prediction for the biggest political winner of 2025 my biggest political winner for 2025 are fiscal conservatives I think that we are going to Ted a very important Concept in 2025 and I hope it works which is that of austerity and the reason why austerity has to work is that the only thing left after austerity is to cut entitlements and I think that in doing this we are going to figure out how much waste Fraud and Abuse exists in the
认为的 是啊 好吧 我不觉得是独立选民 不过行吧 我认为独立中间派在2020年把选票倒向了拜登 然后2024年又把选票倒向了特朗普 呃 从统计上看是这样吗 Freeberg你怎么看是独立中间派决定了选举结果吗 或者Gavin你怎么看 我觉得这个说法对 我觉得有很多人 不只是中间派 还有很多一辈子都投民主党的人这次投了特朗普 所以我确实认为特朗普赢下了中间派 我是说特朗普基本上什么都赢了 对 好 Chamath 你预测谁是2025年最大的政治赢家 我心目中2025年最大的政治赢家是财政保守派 我认为我们将在2025年检验一个非常重要的概念 我希望它能奏效 那就是财政紧缩 紧缩之所以必须奏效 是因为紧缩之后唯一剩下的选项就是削减福利项目 我认为在这个过程中我们会搞清楚美国联邦政府里到底存在多少浪费、欺诈和滥用 我认为这会外溢到很多
便签引用
7:22
United States federal government I think that that's going to spill over to a lot of State elections and I think that the fiscal conservatives that have been clamoring for a more restrained approach to spending will have their Day in 2025 okay fiscal conservatives from schoth freeberg who's your prediction for the biggest political winner of 2025 I also took a a class-based approach I chose young candidates so Trump cabinet picks have an average age of 40 to 45 years old compared to the Biden cabinet the average age is a little over 59 almost 60 years old and I do think that this marks the beginning of a new trend in the kind of age range of political candidates shifting younger so I think that this is something we should expect as candidates start to emerge for the midterms by the end of 2025 we'll start to see younger new names start to pop up that deliver resident messages and aren't part of kind of the Old Guard of you know the Aging political class so I think that's a trend kind of underway
州级选举 我认为那些一直呼吁更克制的支出政策的财政保守派会在2025年迎来他们的高光时刻 好 Chamath选的是财政保守派 Freeberg 你预测谁是2025年最大的政治赢家 我也是从一个群体的角度出发 我选的是年轻候选人 特朗普内阁人选的平均年龄是40到45岁 相比之下拜登内阁的平均年龄是59岁出头 差不多60岁 我确实认为这标志着一个新趋势的开始 政治候选人的年龄区间正在向年轻化转移 所以我觉得随着中期选举的候选人陆续浮出水面 到2025年底我们应该会看到更多年轻的新面孔冒出来 他们传递的信息能引起共鸣 而且不属于那种老派的日渐老去的政治阶层 所以我认为这个趋势已经在酝酿了 很棒 年轻候选人 所以我们有了
便签引用
8:30
now excellent young candidates so we got fiscal conservatives young candidates Gavin what do you got for the 2025 prediction for biggest political winner I would say Trump and centrism for my choice for 202's biggest political winner I went with something similar to you freedberg I said Gen X and the Elder Millennials if you look at the notable Gen X appointments you got Elon with do you got saxs obviously Marco Rubio my God it just goes down and then if you look at the Elder Millennials JD Vance V fake Tuli just a lot of young people and this is going to be absolutely fantastic I think because they're going to start thinking not just about themselves uh as the boomers are doing with Social Security taxes real estate all the different issues that they tend to pick for themselves they're going to start thinking about maybe their own kids and themselves so yeah it's obviously a sea change is underway so they have a folks for our predictions for political winner let's go with political loser here
财政保守派、年轻候选人 Gavin 你对2025年最大政治赢家的预测是什么我会说是特朗普和中间主义 这是我选的2025年最大政治赢家 Freeberg 我选的跟你有点像 我说的是X世代和年长的千禧一代 看看那些引人注目的X世代任命 埃隆和DOGE还有Sacks 当然还有马尔科·卢比奥 天呐 名单可以一直往下列 再看年长的千禧一代 JD·万斯维韦克·拉马斯瓦米 一大批年轻人 我觉得这会特别棒因为他们不会只考虑自己 不像婴儿潮一代那样 满脑子都是社保、税收、房地产 各种他们为自己挑的议题 他们会开始考虑自己的孩子 也考虑自己 所以是的 这显然是一场正在发生的巨变 好 这就是我们对政治赢家的预测 接下来说政治输家
便签引用
9:27
prediction for political loser we start off with You Gavin we'll do this a little round robin here we'll snake it around Gavin what do you think who'll be the biggest political loser of 2025 I think Putin I think Putin is going to lose bigly um so if you are xiin ping you know xim Russia is a client state of China at this point and if you're xiin ping what is happening is now a disaster for you because Europe is starting to rearm and which is which will only accelerate this year that will allow America to take resources out of Europe and put them in uh Japan and South Korea and all over the Pacific and that makes it a lot harder for you to do what you most want to do which is reunify um China and Taiwan or invade Taiwan uh let's call it what it is and I just think X is going to begin decoupling from Putin if you're Trump you want to show that you're independent that you're not um inthrall to Putin in any way so I think Trump is going to be a lot tougher on Putin than people think and I think he's going to get a very
政治输家的预测 我们从你开始 Gavin 我们轮着来 绕一圈Gavin 你觉得谁会是2025年最大的政治输家 我觉得是普京 我觉得普京会输得很惨 嗯 如果你是习近平 你知道 现在俄罗斯基本上是中国的附庸国 如果你是习近平 现在发生的事对你来说是场灾难 因为欧洲开始重新武装 而且今年只会加速 这会让美国把资源从欧洲抽出来 投到日本、韩国和整个太平洋地区 这就让你更难做你最想做的事 也就是统一中国和台湾 或者说入侵台湾 咱们就直说吧 我觉得习会开始跟普京脱钩 如果你是特朗普 你想表明自己是独立的 不受普京的任何摆布 所以我认为特朗普对普京的态度会比人们想象的强硬得多
便签引用
10:42
very uh a deal that's very very bad for Russia and Ukraine and you know you've lost half a million people and for what for what exactly yeah and it's a thousand over a thousand days into this nobody from the West H you know in the NATO or America no no soldiers have lost their lives from our side we've just given them weapons and so it's a humiliating defeat so far for Putin I agree with you freeberg your prediction for the biggest political loser of 2025 friberg's prediction I'm gonna predict the proar neocons who are going to go head-to-head with the JD Vance and Elon and others of the world and I think that they're going to lose and I think that there's going to be this kind of big crack in the establishment of this this neocon movement that's been very Pro conflict around the world and we've heard it in the speeches and in the commentary from JD and others and I think this is going to be the year it's all gonna kind of come to a head I think they're gonna end up on the losing side can iav can I take
我认为他会谈成一个对俄罗斯非常非常不利的协议 而乌克兰 你知道 你损失了五十万人 换来了什么到底换来了什么 是啊 而且这场战争已经打了一千多天 西方没有一个人北约或者美国 没有一个士兵在我方阵亡 我们只是给他们提供武器 所以到目前为止这对普京来说是场屈辱的失败 我同意你的看法 Freeberg 你对2025年最大政治输家的预测是什么 Freeberg的预测 我要预测的是那些主战的新保守派 他们会跟JD·万斯、埃隆这些人正面交锋 我认为他们会输 我认为建制派内部会出现一道大裂缝 就是这个一直在全球到处鼓吹冲突的新保守派运动 我们已经在JD和其他人的演讲和评论里听到了苗头 我认为今年就是一切摊牌的一年 我觉得他们最终会站在失败的一方 我能不能唱个反调 就稍微反驳一点 我觉得这就是你来这儿的意义 对 对
便签引用
11:44
the other side just just a little bit so I think that's why you're here yeah yeah I think it's right in reality but I think you know Trump said something very interesting about John Bolton he said the guy was absolutely crazy but it was awesome having him in the room when you negoti deals people looked at this guy looking angry red in the face so excited to hit the nuclear button so excited to go to war and you ended up with much better deals so I think you're going to see probably a lot more bellicosity from the Trump Administration than anyone expects and that's just to get a good deal for you know between Russia and Ukraine and then it's to get China to kind of decouple from Russia so I think the um in reality I think you're right David but I think there will be a lot of rhetoric that is at odds with what you're saying before you end up being right we already see it with Canada with NATO with Taiwan there's a lot you know we're gonna do this or this the tariffs obviously there's just a lot of very
我觉得从现实层面看你说得对 但我觉得 你知道 特朗普说过一句很有意思的话 关于约翰·博尔顿 他说这家伙绝对是个疯子 但谈判的时候有他在场太棒了 人家一看这人气得满脸通红 恨不得马上按核按钮 恨不得马上开战 结果你就能谈到好得多的条件 所以我觉得特朗普政府展现出的好战姿态会比任何人预期的都多 而这只是为了拿到好条件 比如在俄罗斯和乌克兰之间 然后是让中国跟俄罗斯脱钩 所以我觉得 呃 从现实层面看我认为你是对的 David 但我觉得会有大量跟你说的相反的言辞 然后你才会被证明是对的 我们已经在加拿大的问题上看到了北约、台湾 有很多 你知道 我们要干这个干那个 还有关税 显然会有大量非常
便签引用
12:42
aggressive posturing leading up to the negotiations that are hopefully going to get the US good deals we'll see shamat you have a prediction for your 2025 biggest political loser predictions hard to do the uh biggest political loser of 2025 is going to be progress iism so November of last year right after the election I flew to London and went up to Oxford and I spoke at the Oxford Union and very impressive my speech was a full-throated defense of Maga but it was mostly an explanation of Maga and it was sort of the antidote to Progressive instincts that had been riddling the Western G8 countries and was starting to basically come on gun and when you look at what's about to happen in 2025 in Canada Justin Trudeau is going to lose massively to Pier POV in Germany afd looks like they will win in France if there's a deadlock and it goes into an election more than likely Marine Leen is going to win and then in the UK where you see this unfolding child rape scandal where allegedly upwards of hundreds of
强硬的姿态 为接下来的谈判做铺垫 希望能给美国谈到好条件 我们走着瞧 Chamath你对2025年最大政治输家的预测是什么 这个挺难的 呃 2025年最大的政治输家会是进步主义 去年11月 大选刚结束之后 我飞去了伦敦 又去了牛津 在牛津辩论社演讲 非常厉害 我的演讲是对MAGA毫无保留的辩护 但主要还是在解释MAGA 它算是一剂解药 针对的是那种一直在侵蚀西方G8国家的进步主义思潮 而这股思潮当时才刚开始显现 看看2025年即将发生的事 在加拿大 贾斯汀·特鲁多会惨败给皮埃尔·普瓦列夫 在德国 德国选择党看样子会赢 在法国 如果出现僵局并走向大选 玛丽娜·勒庞极有可能获胜 然后是英国 那里正在曝出儿童性侵丑闻 据称二十多年间有数十万名
便签引用
13:58
thousands of young girls over the course of 20 plus years were being raped by organizations of Pakistani Muslim men who were then not prosecuted for Fears of stoking islamophobia as it turns out by the current prime minister Kier starmer and if all of that comes to pass in the UK I think you're going to see the labor government fall and I think you're going to see Nigel farage win so what do all of these countries look like like by the end of 25 it's very much a repudiation of this class-based identity politics and I think that has enormous Ripple effects all throughout the world and so I think the biggest political loser for 2025 I think stands to be progressivism quote unquote what they what we labeled progressivism I I took something very similar I said the racist vocal minority of each one of these parties and there's a little bit of this on either side you have Dei on one side and then you just have outright racism on the other side and it's probably 5% of Maga and 5% of the left and just to recap last year's
年轻女孩遭到巴基斯坦裔穆斯林男性团伙的性侵而这些人因为担心煽动伊斯兰恐惧症而没有被起诉 结果发现当时负责的正是现任首相基尔·斯塔默 如果这一切在英国成真 我认为你会看到工党政府垮台 我认为你会看到奈杰尔·法拉奇获胜 所以到2025年底 这些国家会是什么样子 这非常明显是对这种基于群体身份的身份政治的否定 我认为这会在全世界产生巨大的连锁反应 所以我认为2025年最大的政治输家很可能就是进步主义 打引号的 我们所说的进步主义 我选的也很类似 我说的是每一方里那些喧嚣的种族主义少数派 这两边都有一点 一边是DEI 另一边是赤裸裸的种族主义
便签引用
15:07
political predictions for 2024 our predictions for political loser was I said Netanyahu freeberg you said Ukraine and shath you said the coke family the GOP nailed that one I nailed that one freeberg any thoughts on last year's or anything on the Republican side more than the Koch family do you think I mean explain for the audience who maybe aren't as into it well I think essentially 2024 was the end of the Republican party as we knew it and I think what stands in its place is what I would call the Maga reflection of a coalition of people that will be housed under the label of republicanism which is to say that these folks aren't necessarily Republicans these folks are believers of the Maga philosophy they're just using the Republican vessel in order to run their candidates and get elected and in that overturning of the status quo what you had was this one family who was at the center for the last few Decades of political Machinery that essentially decided candidates that it decided agenda that
大概占MAGA的5% 也占左翼的5% 顺便回顾一下去年对2024年的政治预测 我们对政治输家的预测是 我说的是内塔尼亚胡 Freeberg你说的是乌克兰 Chamath你说的是科赫家族和共和党 那个说中了 我那个也说中了 Freeberg 对去年的预测有什么想法吗 或者共和党那边除了科赫家族还有什么 你怎么看 我是说 给可能不太了解的观众解释一下 我觉得本质上2024年是我们所认识的那个共和党的终结 我认为取而代之的是我称之为MAGA式的联盟 一群人被装在共和主义这个标签之下 也就是说这些人未必是共和党人 这些人信奉的是MAGA理念 他们只是在利用共和党这个容器来推自己的候选人参选并当选 而在这场对现状的颠覆中 有那么一个
便签引用
16:17
it decided policy and that was the coch family and they spent an enormous amount of money to get that which they stood up frankly literally the day after citizens united happened in the Supreme Court and I think that in that lens those billions of dollars of investment have essentially gone to zero because I don't think it means much of anything anymore and if you look at the new class of donors who will decide quote unquote Republican policy it's going to be the the muriam Adon of the world the Elon musks of the world and that's very different than how I think the cokes used to decide things do you think there's too much money in politics now Gavin do I think there's too much money in politics now um it's it's a good question you know I might feel very very differently if I didn't agree so profoundly with with the largest owner in this political cycle yeah I I would say the reality is is like the the the money as long as the money raised on each side is roughly equivalent I don't think it really
家族 过去几十年一直处在政治机器的中心 基本上由它决定候选人、决定议程、决定政策 那就是科赫家族 他们花了巨额的钱来换取这些 坦白说 他们几乎是在最高法院作出联合公民案判决的第二天就把这套机器建立起来了 我认为从这个角度看 那几十亿美元的投入基本上已经归零了 因为我觉得它现在已经没什么意义了 再看看决定所谓共和党政策的新一代金主 会是米里亚姆·阿德尔森这样的人 埃隆·马斯克这样的人这跟我认为科赫家族过去决定事情的方式非常不同 你觉得现在政治里的钱太多了吗Gavin 我觉得现在政治里的钱太多了吗 呃 这是个好问题 你知道 如果我不是这么彻底地认同这轮政治周期里最大的那个出资人 我的感受可能会非常非常不同 是啊 我想说现实情况是
便签引用
03商业赢家:机器人、AI 与稳定币
17:29
matters that would that would be my take you just want some equivalence that's all all right let's do our 2025 predictions for biggest business winner freeberg why don't you start us off biggest business winner for 2025 what's your prediction so I feel like we're at a really interesting inflection point that is going to make 2025 the year of autonomous Hardware or robotics if 2024 was the year of kind of compute build out and the roll out of AI systems in software I think 2025 will be the year of the robot you know there's a company I we just placed an order today actually out of China called unry Gavin have you looked at this company yeah it's pretty wild actually it's an incredible business incredible product so their go2 robot is $1600 has an API here it is you can run a payload on it it's got lar on it it's got kind of intelligence guidance systems on it this is the robot system that was used on some of those videos we looked at earlier this year where there were machine guns mounted to the back and it
只要双方筹到的钱大致相当 我觉得这真的不太要紧 这就是我的看法你只是希望双方对等 就这样 好 我们来做2025年最大商业赢家的预测 Freeberg 你先来吧2025年最大的商业赢家 你的预测是什么 我觉得我们正处在一个非常有意思的拐点上 它会让2025年成为自主硬件或者说机器人的一年 如果说2024年是算力建设和AI系统在软件端铺开的一年 我认为2025年会是机器人的一年 你知道 有家公司 我们今天刚下了单 是中国的 叫宇树 Gavin你关注过这家公司吗 关注过 其实挺疯狂的这生意非常了不起 产品也了不起 他们的Go2机器人1600美元 有API 就是这个 你可以在上面挂载负载 装了激光雷达 还有各种智能导航系统 就是这套机器人系统 出现在
便签引用
18:37
was basically a new kind of field Soldier it's an autonomous field Soldier but really you can use it in scientific applications we're looking at using them in our uh on our on our test Farms where it can wander the farm and take images and Report data back to us and it's such a low cost that like less than $3,000 you can do some incredible things with it so this business just raised a couple hundred million dollar last month from Mostly Chinese investors it's a Chinese company and I think that there's going to be you know other similar type businesses you know it kind of takes a long time for things to work and then all of a sudden they happen faster than you could have ever imagined I think this is going to be the year where we're all going to look at humanoid robots and autonomous systems and be like oh my God I can't believe this is here um so this could be the year such a great strong choice for the biggest winner and I think the quote you're looking for is how did you go bankrupt slowly and then all at once
我们今年早些时候看的那些视频里 背上架着机枪基本上就是一种新型的战地士兵 一个自主战地士兵 不过你其实也可以用在科研场景 我们正考虑把它用在我们的试验农场上 让它在农场里四处走动 拍照并把数据传回给我们 而且成本特别低 不到3000美元 你就能用它做一些非常了不起的事所以这家公司上个月刚融了几亿美元 投资方主要是中国投资人 它是一家中国公司 我觉得还会有其他类似的企业 你知道 事情往往需要很长时间才见效 然后突然之间 它们发生的速度快到你根本想象不到 我觉得今年就是我们会看着人形机器人和自主系统然后说 我的天简直不敢相信这已经成真了 嗯 今年可能就是那一年 这个最大赢家的选择太棒了 我觉得你想引用的那句话
便签引用
19:28
that's how these technology changes and man I would love to have one of these for the ranch to just run around and do the perimeter security sorry Nick pull up the you guys got to watch this if you haven't seen it Nick find the video of the go2 with the wheels on it U managing terrain and everyone thought this was a BS video that it was like CGI wasn't real it was AI generated but like this thing is just incredible here it is this is a real video of this thing yeah and it could learn from an llm too right like a large language model could teach it how to do so yeah Gavin is this real like what's your abely real yeah yeah I want to get the G1 which is their humanoid robot yeah so pull up the humanoid it's really cool yeah a little more expensive yeah yeah I mean $166,000 is the cost of a used Prius let's be honest that's in spitting distance of being affordable mid look at this thing 16 Grand and this thing you can you know you can basically command it to do things for you in your house or in your
天啊,我真想给农场搞一台这种东西,让它到处跑,做周界安防。抱歉,Nick,放一下那个视频,你们一定要看看,如果还没看过的话。Nick 找一下那个装了轮子的 Go2 的视频,它在应对复杂地形,当时大家都以为那是假视频,是 CGI、不是真的、是 AI 生成的,但是这东西真的太不可思议了。就是这个,这是这东西的真实视频。对,而且它还能从大语言模型学习,对吧?就是说大语言模型可以教它怎么做。对,Gavin,这是真的吗?你觉得——完全是真的,对对,我想买他们的 G1,就是他们的人形机器人。对,放一下那个人形的,真的很酷。对,贵一点。对对,我是说 16.6 万美元(原话)——一辆二手普锐斯的价格,说实话,这已经离“买得起”不远了。看看这东西1.6 万美元,这东西你基本上可以指挥它在你家里、在你的
便签引用
20:27
factory or place oh come on I knew you would do it oh come on I was going to say maybe we could get a poker dealer maybe B you know I wonder if this thing could deal do you guys remember that robot that was on stage for our all-in Holiday Spectacular One so my my brother made an observation he said as soon as that robot came out it's like everyone just wanted to abuse it he's like humans have this very interesting nature where like the robot emerges and all that humans want to do is like hurt the robot it's like ration of behavior yeah like just it's like oh here's something I dominate here's something I can tell we're so threatened by it we want to make sure that we're the top of the species here and it doesn't feel like it's going to be for much longer I mean these things will kick our ass pretty soon anyway these things are amazing I think this is going to be the year of the robots that's my pred year of the robots is your prediction I love it uh Gavin what do you got profoundly agree I think
工厂或者别的地方帮你干活。哦拜托,我就知道你会这么干。哦拜托,我本来想说也许我们可以搞一个发牌员机器人,也许吧。你知道吗,我在想这东西能不能发牌。你们还记得我们那场 All-In节日特别秀上台的那个机器人吗?我哥哥有个观察,他说那个机器人一出来,好像所有人就想欺负它。他说人类有个很有意思的天性,就是机器人一出现,人类想做的就是伤害那个机器人,就像是一种行为模式。对,就像“哦,这是个我能支配的东西”,“这是个我能——”我们太受威胁了,我们想确认自己还是物种里的老大,而感觉这也维持不了太久了。我是说这些东西很快就能把我们揍趴下。这些东西太惊人了。我觉得今年会是机器人之年,这就是我的预测。机器人之年是你的预测,我喜欢。呃,Gavin,你的预测是什么?我非常同意,我觉得就是机器人,也包括 FSD。
便签引用
21:23
you're the robots um inclusive of FSD you know I think FSD works today and it's gonna it's it's going to cross into mainstream adoption where I already notice particularly if I'm taking an Uber late at night I really really prefer to have a Tesla just sometimes you get an Uber driver who's who's tired and I just feel a lot safer if they have the FSD running and then obviously using it for yourself is amazing but I think it's going to continue compounding at an accelerating rate I think the in a lot of ways what I'd say broadly is I think for a while it's going to be big businesses are winners big businesses that use Ai thoughtfully and the reason is is that with what is happening what what 03 showed us open ai's new model the combination of reasoning and test time compute like I think you're if you're a big business and you can pay a million dollars to let let an AI think for 6 weeks about the most important question for your business that's going to be a profound Advantage relative to small businesses
你知道,我觉得 FSD 今天已经能用了,而且它会进入主流普及阶段。我已经注意到,尤其是我深夜打 Uber 的时候,我真的非常希望是特斯拉,因为有时候你会碰到一个很疲惫的 Uber 司机,如果他们开着 FSD,我会觉得安全得多。当然自己用也很棒,但我觉得它会以加速的速度持续复合增长。我觉得,从大方向上说,我认为有一段时间赢家会是大企业,是那些善用 AI 的大企业。原因在于,看看现在正在发生的事,o3给我们展示的,OpenAI 的新模型,推理加上测试时计算的结合——我觉得如果你是一家大企业,你能花一百万美元让 AI 针对你业务上最重要的问题思考六周,那相对于负担不起的小企业来说,这会是一个巨大的优势。顺便说一句,我们
便签引用
22:39
that can't afford that and by the way we are the other like inference compute is going to be the the kind of derivative when of that we're going to run out of gpus accelerators compute in 20125 the same way we did in 23 okay so you're also long Nvidia and chip makers grock Etc because we're finding new uses for all that compute absolutely okay very good chth do you have a prediction for your biggest business winner of 2025 sir I think the biggest business winner of 2025 are going to be dollar denominated stable coins oh I'll make two points in 2024 two critical things happened the first is that stable coins essentially became uncoupled from crypto volatility and it started to be used for wholesale useful functions in running businesses and there's an image here that starts to show that so independent of crypto volatility what you saw were stable coin usage just rising up and to the right that's an incredibly important decoupling that happened in 24 the second and this data is still just
另外——推理算力会是这件事的衍生结果,我们在 2025 年会耗尽GPU、加速器、算力,就像 2023 年那样。好,所以你也看多英伟达和芯片厂商、Groq 等等,因为我们在给这些算力找到新的用途。绝对是。好,非常好。Chamath,你对 2025 年最大的商业赢家有什么预测吗?我觉得2025 年最大的商业赢家会是美元稳定币。哦,我要说两点。2024 年发生了两件关键的事。第一,稳定币基本上和加密货币的波动性脱钩了,并且开始被用于企业运营中批量的、实用的功能。这里有张图开始能说明这点,也就是说,独立于加密货币的波动,你看到的是稳定币使用量一路向右上方攀升。这是 2024 年发生的一个极其重要的脱钩。第二点,这个数据还在陆续出来,但这是个惊人的数字:稳定币
便签引用
24:01
trickling in but it's an incredible stat stablecoin usage at the end of the second quarter of 2024 was about 1.1 billion transactions that summed to 8.5 trillion do of transaction volume if you compare that to Visa over the same period it was more than double visas transaction volume so I think what we have now is something that has fundamentally crossed a point of no return similar to how last year I thought that the big Trend was going to be Bitcoin in 2024 I would say that the big Trend in 25 is stable coin usage I think we're going to finally attack the duopoly of visa and MasterCard I think you're going to see an inal number of use cases that sit and use stable coin rails I think when Donald Trump becomes president I think you're going to see him go after incredibly High credit card transactions and costs you're already seeing cacks in consumer credit anyways because of these high aprs all of this stuff is going to come to a head in 2025 I think and I think stable coins could quadruple or
在 2024 年第二季度末的使用量约为 11 亿笔交易,累计 8.5 万亿美元的交易额。如果你把它和同期的 Visa 相比,这是 Visa 交易额的两倍多。所以我认为我们现在有了一个从根本上跨过了不可逆点的东西,类似于去年我认为 2024 年的大趋势会是比特币。我会说,25 年的大趋势是稳定币的使用。我觉得我们终于要冲击 Visa 和万事达的双寡头格局了。我觉得你会看到数不清的用例建立在稳定币的轨道上。我觉得当唐纳德·特朗普成为总统,你会看到他去开刀那些高得离谱的信用卡交易和费用。你已经能看到消费信贷里的坏账在上升,反正就是因为这些高年化利率。我觉得所有这些事情都会在 2025 年集中爆发,而且我觉得稳定币到 25 年底可能翻四倍或者五倍。我觉得这会是一个巨大的市场。
便签引用
25:28
quintuple by the end of 25 I think it's just going to be an enormous Market great choice should they be regulated chamath we saw at some of these Congressional hearings tether was kind of dragged for being the primary transfer and and monetary tool for terrorism for getting around sanctions and for human trafficking should it be regulated and how I I don't think I'm qualified to say how it should but I think the point is that there are these IM logs that sit in between and on either side of all of these transactions and so I think the knowledge is there and there are a whole bunch of third party services that add that intelligence layer I think the the thing to really question is if you just took 300 basis points of drag yes out of the global economy massive gain how valuable would that be and I think it would be valuable just in the United States alone to the tune of a trillion dollars so the idea that you wouldn't do it at this point is somewhat quizzical to me so I just think that the economic justification
很好的选择。Chamath,它们应该被监管吗?我们在一些国会听证会上看到,Tether 被狠狠地批评,说它是恐怖主义、规避制裁和人口贩卖的主要转账和货币工具。它应该被监管吗?怎么监管?我觉得我没资格说该怎么监管,但我想说的是,在所有这些交易的中间和两端,都存在这些日志记录,所以相关的信息是存在的,而且有一大堆第三方服务提供那一层情报分析。我觉得真正该问的问题是:如果你能从全球经济里把 300 个基点的摩擦成本拿掉,那是巨大的收益——那会有多大价值?我认为光在美国国内,这个价值就能达到一万亿美元的量级。所以到了这个地步还不去做,在我看来有点让人费解。所以我就是觉得,现在这件事的经济理由
便签引用
26:39
for this now is just so profound Gavin let me ask you a hard question there as well should the United States government give up the let's face it you know they they have a bit of a strangle hold they've got a bit of a monopoly on US Dollars these things are a competitor and we were sitting here on this program over the last couple years talking about the bricks and them starting their own currency aren't these in some way similar to that and they would take people off the USD standard how do you look at that because that is the criticism that some people in power have put forward I do I do think that they might be very good for the world but I do think they would be very bad for America if they replaced if some constellation of stable coins became the new Reserve currency it is such a an advantage to be able to borrow in your own currency and to give that up lightly that just that means you control you have you you control the real size of your debt at all times oh wow we have a big big debt I mean this
实在太充分了。Gavin,我也问你一个难题。美国政府应该放弃——说实话,你知道,他们对美元有点垄断地位,有点掐着脖子。这些东西是竞争者,我们过去几年就坐在这个节目上讨论过金砖国家和他们要搞自己的货币。这些东西在某种意义上不是和那个类似吗?它们会让人们脱离美元体系。你怎么看这个问题?因为这正是某些当权者提出的批评。我确实觉得它们可能对世界很有好处,但我确实认为,如果它们取代了美元,如果某种稳定币组合变成了新的储备货币,那对美国会非常不利。能用自己的货币借债是一个巨大的优势,轻易放弃它——那意味着你随时都掌控着自己债务的真实规模。哇,我们有一大笔债务。我是说这
便签引用
27:46
is a little bit like what we just did let's just have let's just let inflation run hot and oh wow that debts a lot smaller than it was effectively so I think it would be a big mistake freeberg you have any thoughts on it we we've been talking about currency here and a whole bunch over the last couple years any thoughts on stable coins and and just the uh Monopoly we have on it I I don't know enough about stable coins well that'll be one of our big topics for the year and we should uh definitely invite jir from USD on or usdc usdc Geremy air I mentioned by the way that I had this small little product that I exposed to the world which is just our research and we sit it on top of substack again the tools aren't very good and we use stripe and Jeremy reached out to me and he's like I'll just rebuild all of your payment rails on using US dollar stable coins and it's a no-brainer because I'm spending hundreds of thousands of dollars in transaction costs that now I can just save and use to hire somebody else or
有点像我们刚刚做的事:就让通胀跑热一点,然后哦哇,那笔债实际上比原来小多了。所以我觉得那会是个大错误。Freeberg,你有什么想法吗?我们过去几年在这儿聊了很多关于货币的话题。对稳定币,还有我们对它的垄断,有什么看法吗?我对稳定币了解得不够多。那这会是我们今年的大话题之一,我们应该一定要请 USDC 的 Jeremy 来,Jeremy Allaire。顺便提一句,我之前说过我有个小产品,就是我们的研究内容,我把它发布出来了,架在 Substack 上。还是那句话,工具不太好用,我们用 Stripe。Jeremy联系我,他说我可以把你所有的支付通道都用美元稳定币重建一遍,而且这是显而易见的选择,因为我现在在交易成本上花了几十万美元,而这些钱我可以省下来去多雇一个人,或者
便签引用
28:53
just pay the team more and so I you're paying 10% to suback in the 3% yeah if it's obvious for me it's obvious for everybody else all right my biggest business winner prediction for 2025 is Tesla and Google I think this is going to be the year of AI and Robotics obviously but I am just absolutely amazed at the comeback that Google has made with AI and I'm absolutely in awe of what Elon did with Colossus if those of you don't know Jensen Wong Michael D they all obviously participated in building out xai and they were shocked they couldn't believe that they stood up a 100,000 gpus in under 45 days was it Gavin something in that neighborhood it was a tremendous effort and uh you that's not Tesla right is that Tesla well Elon Tesla xai and Google that cohort who are investing heavily in AI those two specifically those two three names I think are going to be the biggest winners next year they're going to come out with some things that people didn't anticipate I don't I don't think people are anticipating what Google
给团队多发点钱。所以我——你付给 Substack 10%,再加 3%。对,如果这对我来说是显而易见的,那对其他所有人也是。好,我对 2025 年最大商业赢家的预测是特斯拉和谷歌。我觉得今年显然会是 AI 和机器人之年,但我对谷歌在 AI 上的翻身仗感到绝对惊叹,而且我对埃隆用 Colossus 做的事情肃然起敬。如果你们不知道的话,黄仁勋、Michael Dell 他们显然都参与了xAI 的建设,他们都震惊了,他们不敢相信这些人在不到 45 天内就搭起了 10 万块 GPU。是这样吗,Gavin?差不多是这个数。那是一个了不起的壮举。呃,那不是特斯拉吧?那算特斯拉吗?嗯,埃隆、特斯拉、xAI 和谷歌,这一批重金投入 AI 的公司。具体就是这两家——这两三个名字,我觉得会是明年最大的赢家。他们会拿出一些人们没有预料到的东西。我觉得人们没有预料到谷歌会拿出什么,因为
便签引用
29:58
Google is going to come out with because Gemini deep research I keep telling people this and I'm using the app is so impressive drop what you're doing right now pause the podcast download Gemini and just start playing with deep resarch on your desktop pay the 20 bucks for it it is mindblowing it blows anything else in the market out of the water period full stop it is a step function higher all right I'll just leave it at that Jason shouldn't you uh just go with the vibe of the podcast shouldn't you explain what it is what sorry okay so what it does is good call good call if you wanted to create a model let's say and I did this where I said hey make me a model based upon what it would take to replace all the cars and trips in the United States with full self-driving cars what would that cost to do and if you ask that to chat GPT or grock or you know Claude it's going to give you a decent answer what deep research does is it goes and it figures out what the subcomponents of each of those questions
Gemini 的 Deep Research——我一直跟人说这个,我一直在用这个 app,太惊艳了。放下你手里的事,现在就暂停播客,下载 Gemini,在你的电脑上玩一下 Deep Research,花那 20 美元,简直让人脑洞大开。它把市面上其他任何东西都甩开好几条街,句号,就这样。这是一个台阶式的跃升。好吧,我就说到这儿。Jason,你不是应该顺着播客的调性来吗?你不是应该解释一下它是什么吗?什么?抱歉,好吧,那我说说它是干什么的——问得好,问得好。比如说你想做一个模型,我就做过,我说,嘿,给我做一个模型,基于要把美国所有的车和出行都替换成全自动驾驶汽车需要什么条件,这要花多少钱。如果你把这个问题问 ChatGPT 或者 Grok 或者,你知道,Claude,它会给你一个还不错的答案。而 Deep Research 做的是,它会去搞清楚这些问题各自的子组成部分
便签引用
30:57
would be and then it fires off multiple threads and then it searches the web in real time which Google is extremely good at and then it summarizes it gives citations and it produces a report that looks like something that Gartner group McKenzie Boston Consulting Group would have spent I don't know 30 days on Gavin and a million dollars a half million dollars it's Bonkers how good this is but the amount of compute it takes is extraordinary that's why chat GPT released something similar that's 200 bucks a month the 01 Pro I think they call it and then this one is like 20 bucks from Gemini anyway given what Google has access to YouTube your G drive your Google Docs Etc your Gmail it's going to do things that you wouldn't anticipate like if you were asking this question it might look in your Gmail for possibilities to answer those questions or YouTube videos of channels you subscribe to so this is going to I think knock people's socks off and solve problems that people didn't know they had is that a pretty good explanation
是什么,然后它会并行发起多个线程,接着实时搜索网络——这正是谷歌极其擅长的——然后它做总结、给出引用来源,并生成一份报告,那种报告就像 Gartner、麦肯锡、波士顿咨询要花,我不知道,30 天,Gavin,还有一百万美元、五十万美元才能做出来的。这东西好得离谱。但它消耗的算力是惊人的,所以 ChatGPT 才推出了一个类似的东西,每月 200 美元,我记得他们叫 o1 Pro,而 Gemini 这个才 20 美元左右。总之,考虑到谷歌能接触到 YouTube、你的 Google Drive、你的 Google 文档等等、你的 Gmail,它会做出一些你预料不到的事,比如你问这个问题时,它可能会去你的 Gmail 里找能回答这些问题的线索,或者你订阅的频道的YouTube 视频。所以我觉得这会让人惊掉下巴,会解决一些人们自己都不知道自己有的问题。这个解释还算不错吧,Gavin?是的,我觉得很棒。去年我们对商业赢家的预测:
便签引用
04商业输家:体育、承包商与 OpenAI
31:57
Gavin yeah think it's great for last year our predictions for business winner chamat said bootstrapped or Endor profitable startups I think that's a really great pick and that came to pass we're seeing some of those go public and raise money at higher valuations or not go out of business freeberg you said Commodities businesses I'm not sure how that did do you know did you Che pretty flat yeah we didn't have a lot of the inflationary pressures that I think were underlying that pie so got it it ended up being I'd say a loser relative to other indices you could bought this year I said training data owners like Reddit New York Times Google Etc Reddit stock up 24% in 2024 and they launched their own Reddit answer call yeah I think I got that one uh New York Times stock is up as well and they uh fired off that lawsuit with open AI all right now we're going to move on to the 2025 biggest business loser predictions in 2024 shamat said pro sports teams because they hit Peak valuations that's looking
Chamath 说是自力更生或者已经盈利的初创公司。我觉得那是个很棒的选择,而且它应验了,我们看到其中一些上市了、以更高估值融资了,或者至少没有倒闭。Freeberg,你说的是大宗商品业务。我不太确定那个表现如何,你知道吗?相当平淡。对,我们并没有出现我认为支撑那个判断的通胀压力,所以,明白了。最后我会说,相对于今年你能买的其他指数,它算是个输家。我说的是训练数据的拥有者,比如 Reddit、《纽约时报》、谷歌等等。Reddit 股票在 2024 年涨了 24%,而且他们推出了自己的 Reddit Answers。对,我觉得这个我押对了。呃,《纽约时报》的股票也涨了,而且他们还发起了那场对 OpenAI 的诉讼。好,现在我们进入 2025 年最大商业输家的预测。2024 年Chamath 说的是职业体育球队,因为它们的估值已经见顶。那看起来是个相当不错的预测。Freeberg,
便签引用
32:53
like a pretty good prediction free perg you said vertical SAS companies because of AI disrupt that's another one that seems like a great prediction I said smartphone manufacturers Apple stock was up 30% in 2024 but I think we've all seen that these uh smartphones aren't advancing and people are taking their times the sales of those are down services are up I would take the other side on pro sports teams it feels like that market is about to be institutionalized and funds are going to start buying pro sports teams okay that ultimately and ultimately Capital markets have way more resources than kind of the wealthy individual ual who've done it for to date so I just think it's the amount of money that can go after pro sports teams is about uh you know 10x 100x no I think I think you're totally right the the reason I said that last year was it was pretty clear to me at the time and the NBA was the canary and the coal mine that there was a viewership problem in professional sports and specifically in the NBA the
你说的是垂直 SaaS 公司,因为 AI 的颠覆。那也是一个看起来很棒的预测。我说的是智能手机制造商。苹果股票在 2024 年涨了 30%,但我觉得我们都看到了,这些智能手机没什么进步,人们换机的周期变长了,手机销量在下滑,服务收入在上涨。关于职业体育球队,我会站在相反的一边。感觉这个市场即将被机构化,基金会开始收购职业体育球队。好,最终——而且最终资本市场拥有的资源,远远超过那些迄今为止一直在做这件事的富有个人,所以我就是觉得能涌向职业体育球队的资金规模大概会是,呃,10 倍、100 倍。不,我觉得我觉得你说得完全对。我去年那么说的原因是,当时我很清楚地看到,NBA 就是煤矿里的金丝雀——职业体育、特别是 NBA 存在收视率问题。
便签引用
33:53
game has devolved into essentially rebounds and dunks or three point ERS and the issue with that is that it becomes just meaningfully less interesting to watch at the same time because of the the fact that the TV deals are really what determines the discounted value of these sports franchises the TV deal was so enormous that it creates no reliable rivalries anymore because people will hop scotch teams almost every year because the compensation that they can get is just so obscene quite honestly and I think what happens is sports will have a decent run until the next TV deals get done and I think if you for example take Pharma ads outside of TV so like that pool shrinks if you have less viewership and so you can sell the remaining ads less effectively because there's just fewer of them and then there's fewer buyers and it shrinks yet again then the dollar pool that the television networks and the Stream are going to be willing to pay for sports will go down and the group that will be the most price sensitive are exactly who
比赛已经退化成基本上只剩篮板加扣篮,或者三分球,而问题在于这让比赛变得明显没那么好看了。同时,因为真正决定这些体育特许经营权折现价值的其实是电视转播合同,而转播合同金额太庞大了,以至于现在不再有稳定的宿敌对抗了,因为球员几乎每年都在换队,说实话,因为他们能拿到的薪酬实在太夸张了。我觉得接下来会发生的是,体育会有一段不错的行情,直到下一轮电视合同签完。我觉得,举个例子,如果你把药企广告从电视上拿掉,那么这个盘子就缩小了;如果收视率下降,你剩下的广告位也卖不动了,因为广告位变少了,买家也变少了,然后盘子又进一步缩水。那么电视网络和流媒体愿意为体育支付的资金总额就会下降,而对价格最敏感的那群人正是你说的那类,Gavin,也就是非“奖杯型”买家。比如说,你知道,我当初买入
便签引用
35:05
you said Gavin meaning the non trophy buyer so like you know when I bought into the Warriors I bought it purely as a trophy asset and I was pricing sensitive but I agree with you that now that you have the PE firms on the cap tables of these sports franchises those folks are all DCFS those folks are all Excel models I don't think that they're buying things for emotion I don't think they've grown up thinking I want to own this thing not using institutional LP dollars so I just think that's why that's why I said that in 24 so I just on behal of Phil Huth and I we appreciate you made that vanity investment for the number of times we got to sit in your court side seats yeah because I never we made with NBA players I never went to the games I should have gone to more games I went to your games and almost got thrown out of your seats I got a red card one time it's true is it true you did oh you don't I got a I got a call I got a call I think it was from the Warriors or it's from from the NBA something to the effect of chouth
勇士队的时候,我纯粹是把它当作奖杯资产来买的,我当时对价格不敏感。但我同意你的看法,现在这些体育特许经营权的股东名册上有了 PE 机构,那些人全都是做 DCF 的,那些人全都是Excel 模型。我不认为他们是出于感情去买这些东西,我不认为他们是从小就想着我要用机构 LP 的钱拥有这支球队的。所以我就是觉得,这就是我在 24 年那么说的原因。所以我代表 Phil Hellmuth 和我自己,感谢你做了那笔虚荣投资,让我们有那么多次能坐在你的场边座位上。对。因为我从来没——我们跟NBA球员打交道,我自己却从来不去看比赛,我真该多去几场。我去看过你的比赛结果差点被从你的座位上赶出去。我还吃过一张红牌,真的。真的假的?你还真有过?——我接到过一个电话,我接到一个电话,我记得是勇士队打来的,或者是NBA打来的,大概意思就是——Chamath
便签引用
36:00
this guy that was sitting in your seats was almost kicked out and I said excuse me because I knew that it was Game In fairness and and he was like jawboning our own players I was getting into it with bogot and with David Lee because I told it was unbelievable I told Steve cerr listen you're up 30 on my Knicks sit these guys down this is Bush League what if Steph Curry gets hurt out there like and uh you know Boga told me to shut up Andre boa told him to basically go pound in a basically did it was the funniest thing ever Gavin this is what happens when you get's one of our friends he's one of our good friends so that's why it's so funny but they can't they come over freeberg and they tap you on the shoulder I'm with my wife she's mortified and the guy hands you a card and the card says you've been warned one time and one time only about abusive you know uh Behavior if we have to warn you one more time you will be escorted out rid so I look at the guy and I say but he goes read the card I read the card I
坐在你座位上的那个人差点被请出去。我说,不好意思啊,因为我知道那是比赛嘛说句公道话,他当时一直在跟我们自己的球员斗嘴。我跟博古特还有大卫·李杠上了,因为我跟他们说——简直难以置信——我跟史蒂夫·科尔说,听着,你都领先我的尼克斯30分了,把这些人换下去吧,这太不讲究了万一斯蒂芬·库里在场上受伤了怎么办。然后你知道吗,博古特让我闭嘴,安德烈·伊戈达拉基本上是让他滚一边去,他基本上就是那么干的。那是我见过最搞笑的事。Gavin,这就是当你——他是我们的一个朋友,他是我们的好朋友,所以才这么好笑。但他们不能——他们走过来,Friedberg,然后拍拍你的肩膀。我当时跟我太太在一起,她尴尬死了,然后那人递给你一张卡片,卡片上写着:你已被警告一次,仅此一次,关于辱骂性的,你知道,行为。如果我们还得再警告你一次,你就会被请出去所以我看着那个人,我说——他说,读卡片。我读了卡片,我说——他说,给我个大拇指。就
便签引用
37:00
say he goes giv me a thumbs up that was it I never that was it you're not supposed to then three years later I'm sitting in the same seats I'm interacting with Draymond during the finals games and everybody's hokey doie with it when you say interacting what do you mean when you say interacting wa hold on G Gavin wanted to make a point about what I no I just want to just just come back to chath because for sure these private Equity guys there DCFS the NBA has been terribly managed I think even LeBron said they have a big problem you know and I'd separate the NBA is the worst managed NFL is probably the best managed but the one kind of CounterPoint I would say on the TV rights is Google bought Sunday Ticket and are extremely happy with it Amazon and Netflix also bought you know both bought NFL games and are really happy with them and so I just think those three companies it doesn't really matter to them if Pharma ads go away in 18 months you're going to be able to dynamically create an ad
这样,我再也没有——就这样。你是不该——结果三年后,我坐在同样的座位上,在总决赛期间跟追梦格林互动,大家都觉得没什么问题。你说互动,是什么意思?你说互动的时候——等一下,Gavin想就我刚才说的发表点看法。不,我只是想回到Chamath那个话题,因为这些私募股权的家伙肯定做过现金流折现模型。NBA的管理糟糕透顶,我记得连勒布朗都说过他们有个大问题你知道的。我会把它们分开看:NBA是管理最差的,NFL大概是管理最好的。但有一点反方观点我想在转播权这件事上说的是,谷歌买下了《周日门票》,而且他们对此非常满意。亚马逊和Netflix也都买了NFL的比赛,也非常满意。所以我觉得对这三家公司来说,医药广告在18个月后消不消失,其实无所谓。你将来能够为每一个人动态生成一条广告并展示给他,而如果你是高价值用户
便签引用
37:58
for each individual person and show it to them and if you're a high value user with a you know that Google knows that you're about to book a vacation you know they'll dynamically generate you know whatever you know Hotel destinations you want and show click here so I just think Sports as long as they command the eyeballs that they do and I for sure if the NBA doesn't fix it the value of those franchises will start to decline but um just the fact that all of the biggest tech companies are so happy with the sports kind of rights that they bought they're going to buy them all every sport understand Gavin maybe maybe you're closer to this but why wouldn't Apple Amazon or Google just back up the truck to the NBA and say well take the whole thing and then anybody who's got an iPhone gets the NBA for free and then everybody else has to pay and they figure it out in the back and they lose a you know 5 billion they make two billion who cares it's a rounding error for Apple I agree and they you know I mean they all all
谷歌知道你正准备订一趟度假,他们就会动态生成,你知道你想要的那些酒店目的地,然后显示“点击这里”。所以我只是觉得,体育只要还能吸引住那么多眼球——当然,如果NBA不把问题解决好,那些球队的估值会开始下滑但是,光是所有这些最大的科技公司都对他们买下的体育转播权这么满意这一点,就说明他们会把所有体育项目的转播权都买下来。我明白。Gavin,也许你更了解这块,但为什么苹果、亚马逊或者谷歌不直接把车倒到NBA门口说:整个打包我全要了,然后任何有iPhone的人都能免费看NBA,其他人就得付费,他们在后端把账算平,就算亏个50亿、赚20亿,谁在乎呢,对苹果来说那就是个零头。我同意,而且他们,你知道,我是说他们全都
便签引用
39:00
except for Apple really and apple they have I think an MLs deal but they all kind of stuck their toe in the water and I think they feel like wow the water's warm let's dive in um okay yeah and the New Deal is an 11year deal for the NBA so these things don't come up that often but we'll see over time maybe somebody will buy one of these media companies and inherit them okay biggest business loser for 2025 what's your prediction Gavin Baker government service providers you do not you do not want to have the United States government at any level is over 35% of your Revenue good choice yes obviously in the age of Doge they might not be spending wildly and they might actually look at the bill they might check the bill crazy thought crazy thought yeah crazy thought to actually check the bill you never know chth who do you predict will be the biggest business loser of 2025 well this is a a little bit precarious but I don't know what the percentage draw down from here will be but I think when we look
——除了苹果之外——苹果我记得有个MLS的合作,但他们都算是试探性地把脚伸进水里我觉得他们现在的感受是:哇,水挺暖和的,那就跳进去吧。嗯,好的。而NBA的新合同是11年的所以这种机会不会经常出现。但我们走着瞧,也许有一天会有人收购这些媒体公司顺带把转播权也继承过去。好,2025年最大的商业输家,你的预测是什么,Gavin?给政府提供服务的供应商你绝对不想让美国政府——不管哪一级——占到你收入的35%以上。选得好。是的,显然在DOGE(政府效率部)的时代,他们可能不会再大手大脚花钱了,他们可能真的会去看账单,他们会去核对账单,这想法太疯狂了。太疯狂了。是啊,真去核对账单,这想法太疯狂了。说不准哦。Chamath,你预测谁会是2025年最大的商业输家?嗯,这个说起来有点悬,我不知道从现在起回撤的百分比会是多少,但我认为当我们回头看的时候,“七巨头”的绝对美元回撤规模会达到数万亿美元。好的。原因倒不一定是
便签引用
40:07
back the absolute dollar draw down of the mag s will be in the trillions of dollars okay and the reason is not necessarily because of the underlying fundamentals of these companies but I am a little bit worried and I tweeted this a few days ago about just the general concentration of the top seven 8 nine 10 companies in the indices I think it's approaching 40% and I think that when you look at these historic concentrations they've generally foreshadowed a big draw down okay and unfortunately I don't see how you can sort of you know inoculate yourself from that risk so independent of the quality of these companies because I think these companies are exceptional businesses I do think you've just had too much concentration and I think it it is a setup to retrade and give back and in that it could even be 10% but 10% in the mag 8 will be a couple trillion bucks and so on absolute would be a lot yeah an absolute dollar a lot okay freeberg who do you predict you predict Max 7 government service providers for
因为这些公司的基本面有问题,但我确实有点担心,几天前我还发过推特讲这个就是指数里排名前七、八、九、十的公司的整体集中度问题。我觉得已经接近40%了,而我认为当你回顾这些历史上的高集中度时期,它们通常都预示着一次大幅回撤。好的。不幸的是,我看不出你有什么办法能让自己免受这种风险的影响。所以这跟这些公司的质量无关,因为我认为这些公司都是非常优秀的企业。我确实认为集中度太高了,我觉得这为回吐涨幅埋下了伏笔。就算只跌10%,但“八巨头”跌10%就是好几万亿美元,所以绝对金额会非常大。是的,绝对美元金额很大。好,Friedberg,你预测谁——你预测“七巨头”,Gavin预测政府服务供应商。Friedberg,你预测谁会是2025年最大的商业输家?我
便签引用
41:23
You Gavin Free who do you predict will be the biggest business loser of 202 I'm sort of on the same vein as Gavin I went with the kind of old defense and Aerospace providers Boeing loed Martin Ron defense 1.0 driven by kind of China dominance driven by us defense budgets that I think are going to need to shift towards a more kind of tech oriented and Roid driven rationalization and pricing and spend so the Palante and Andals obviously the world will benefit and I also just think like there's a lot of failure at scale with these businesses they've gotten too clunky and too bureaucratic as we've seen with boing from their space program failure in 2024 uh to the challenges with their their airplane business but I think that this government contracting business across the board is going to be deeply challenged this year with all the new blood so the Cost Plus are going to suffer is what you're saying and the andrs will Thrive would you guys add just a question would you guys add the traditional Consulting companies then to
跟Gavin的思路差不多,我选的是那些老牌国防和航空航天供应商:波音、洛克希德·马丁雷神——也就是国防1.0。在中国的主导优势推动下,在美国国防预算的推动下,我认为预算会需要转向更偏科技导向、更由机器人驱动的合理化改造,以及重新定价和重新分配支出。所以Palantir和Anduril显然会受益。另外我也觉得,这些企业存在大量的规模化失败它们变得太笨重、太官僚了,就像我们从波音身上看到的,从他们2024年太空项目的失败到他们飞机业务面临的种种挑战。但我认为,整个政府承包这个行业今年都会面临严峻挑战,因为来了这么多新血。所以你的意思是“成本加成”模式的公司会受苦,而Anduril这类公司会蓬勃发展。我问一句,你们会不会把传统咨询公司也
便签引用
42:22
the mix as well like if you that's a good point too that's a great one yeah we think what we talked like the Y Pros the tatas the hcls the accent cogn the Emy all these folks in are in real trouble if that if that also happens that's right yeah you the efficiency that they'll gain will also be felt by their customers and their customers might not hire them when you got a bunch of software entrepreneurs that are now advising and informing the federal government on how to run their their operations I'm pretty sure that some of these service providers are going to get washed out so it's a good idea the thing to keep in mind is like plus is a fancier way of saying time and materials and time and materials is a fancy way of saying human labor Arbitrage and I think that all of that stuff will get extremely scrutinized in 2025 because if you are buying you know $800 waste baskets and you know $9,000 uh umbrella hangers because of time and materials and all of this other crazy stuff that's in the system and all of
加进这个名单里?比如说——这也是个好观点,这个点很棒。是啊,我们说的那些,像Wipro、塔塔、HCL埃森哲、高知特、安永,所有这些人都会陷入真正的麻烦。如果那种情况真的发生的话。没错。是的,你说的他们获得的效率提升也会被他们的客户感受到,而客户可能就不再雇他们了。当你有一帮软件创业者现在在给联邦政府出谋划策、告诉他们怎么运营他们的业务,我很确定这些服务供应商里有一些会被冲刷出局。所以这是个好想法。要记住的一点是,“成本加成”其实就是“计时计料”的高级说法,而“计时计料”又是“人力套利”的高级说法。我认为所有这些东西在2025年都会受到极其严格的审查,因为如果你在买800美元的垃圾桶、9000美元的雨伞挂架,就因为计时计料,以及所有其他这些疯狂的东西还留在这个体系里——如果这些全都被清理掉,那这些人和他们的商业模式就会
便签引用
43:25
that gets washed out then a lot of these folks in their business models will will get put under severe stress the alignments you know you show me an incentive I'll show you the outcome the alignment of Cost Plus is to just keep being inefficient and and waste and and these defense contractors are taking 60 70 80% off the price of each piece of Munitions or vehicle yeah Cost Plus is terrible no just cost is terrible that's all all right for my biggest business loser prediction I had to go through a lot of the uh people I've uh criticized on in started fights with over the past year commercial real estate came to mind because these leases are going to keep coming up and I think like they kick the can down the road a little bit so I started there then I looked at micro strategies and that made no sense to me that they were trading at two three four times the book value of their Bitcoin and I think that's not sustainable I looked at truth social which is like doing four or5 million in revenue and is
承受巨大压力。这种激励机制的问题,你知道,你给我看激励机制,我就能告诉你结果。成本加成的激励方向就是让你一直低效下去、一直浪费下去。而这些国防承包商在每一批弹药或车辆的价格上要抽走60%、70%、80%。是啊,成本加成太糟糕了。不,是“成本”本身就很糟糕。就这样。好,说到我预测的最大商业输家,我得把过去一年里我批评过的、吵过架的那些人都过一遍我先想到的是商业地产,因为这些租约会陆续到期,而我觉得他们只是把问题往后拖了一阵子。所以我从那儿开始想。然后我看了MicroStrategy,我怎么也想不通他们的交易价格居然是其持有比特币账面价值的两三四倍,我认为这不可持续。我还看了Truth Social,它的收入大概就四五百万美元,估值却有70亿,这完全没有
便签引用
44:25
valued at 7 billion that makes Absol no logical sense but I think the one that is the most overpriced of of all of this and I think is going to see their Peak valuation is open AI I think the headwinds for open AI are being absolutely underappreciated like I said before in my previous one uh my previous prediction Google is kicking ass xai is just getting started and building out so much big Iron Microsoft was basically on another podcast almost like laughing at open a for selling them the source code and that they didn't even need open AI anymore because they had it all they don't need them and they own all the Big Iron I think open ai's valuation made no sense I don't think they're going to be able to keep charging the prices are charging and I think AWS Apple Google xai this cohort are going to really take that $57 billion valuation and make it the peak valuation of the company and I do I do think that there's a nonzero chance they could lose their process and these court cases of transferring $157 billion dollar in
任何逻辑可言。但我觉得这一切里面定价最离谱的一个,而且我认为会在今年见到其估值巅峰的,是OpenAI。我认为OpenAI面临的逆风被严重低估了。就像我之前在我上一个预测里说的,谷歌正在大杀四方,xAI才刚刚起步,正在搭建大量的算力硬件。微软基本上在另一档播客里几乎是在嘲笑OpenAI把源代码卖给了他们,还说他们甚至已经不再需要OpenAI了,因为他们全都有了,他们不需要他们,而且大型算力设备都是他们自己的。我觉得OpenAI的估值毫无道理。我不认为他们能够一直维持现在收取的价格,而我认为AWS、苹果、谷歌xAI这一批公司,真的会把那个1570亿美元的估值变成这家公司的历史最高估值。而且我确实认为他们有不小的可能会输掉这些官司——就是把1570亿美元的
便签引用
45:41
value from a nonprofit into a for-profit I think that whole thing could blow up is that your super prediction is open AI will not be able to convert from nonprofit to for-profit in 2025 I think that that's a small piece of it but that's a really interesting super prediction insert poly Market graphic here super prediction will open AI convert is that's a good in 2025 yeah by the way you know they're projecting I think in the the the leaked financials 12 billion Revenue next year jkl so unlike a lot of the other things you were talking about this is a real business with real Revenue real scale real growth real technology it's it's a little bit distinct from kind of being a meme stock or being a exactly that's why I'm picking it because I think it's easy to pick a meme stock it's harder to pick a real company but I do think think that that Revenue which you know a lot of it is consumer and a lot of it is developers the developers I know all want to do open source they don't want to have to be beholden to open Ai and
价值从非营利实体转移到营利实体的那些案子。我觉得整件事可能会炸掉。这就是你的“超级预测”吗?OpenAI在2025年将无法完成从非营利到营利的转型?我觉得那只是其中一小部分,但那确实是个很有意思的超级预测。这里插入Polymarket的图表:超级预测——OpenAI会不会转型?那是个不错的——在2025年内。对了,顺便说一句你知道他们的预测,我记得在泄露的财务数据里,明年收入是120亿。所以Jason,跟你刚才说的很多其他东西不一样,这是一家真正的公司,有真实的收入、真实的规模、真实的增长、真实的技术。它跟那种迷因股之类的还是有点——正是如此,所以我才选它,因为我觉得挑一只迷因股很容易挑一家真正的公司才更难。但我确实认为,那部分收入——你知道,其中很大一块来自消费者,很大一块来自开发者——我认识的开发者都想用开源的,他们不想受制于OpenAI和
便签引用
05年度大交易:车企合并与自动驾驶整合
46:38
Sam mman they would much rather and they're already setting their queries across multiple different sacks and trying different ones I don't think there's any loyalty to open AI or gemini or any of these Services I think they'll just go with eventually open source in many cases so anyway that's my thought on open AI 2025 biggest business deal what do you got shth what's your biggest business deal for 2025 prediction shth what do you got you said starlink to go public last year that one yeah totally whiffed on that one I think that this is the year that we will see the collapse of the traditional Auto oems and I think that the the deal that happened at the end of 2024 with Honda Nissan I think is a bit of a signal to what the industry has to do which is to go through a massive wave of consolidation I think that Tesla is just in an incredible position with the quality of their vehicles and the quality of their software and the qu quality of their autonomy with FSD so I suspect that after a couple of
Sam Altman。他们更愿意——而且他们已经在把自己的请求分散到不同的技术栈上尝试不同的方案。我不觉得对OpenAI、Gemini或者其他这些服务有什么忠诚度可言。我认为很多情况下他们最终会转向开源。总之,这就是我对OpenAI的看法。2025年最大的商业交易你有什么,Chamath?你对2025年最大商业交易的预测是什么?Chamath,你说什么?你去年说星链会上市,那个完全没说中。我认为今年将是我们看到传统汽车原始设备制造商(OEM)崩塌的一年。我认为2024年底本田和日产之间发生的那笔交易,在某种程度上预示了这个行业必须做的事,那就是经历一轮大规模的整合浪潮。我认为特斯拉的处境好得惊人,他们的车质量好,软件质量好,FSD的自动驾驶质量也好。所以我猜在再发布几款有分量的产品之后,就会触发公开资本市场的一种觉悟:这些汽车
便签引用
47:48
more meaningful product releases it's just going to trigger the realization by the public capital markets that these Auto oems are uninvestable and I think the result of that will be a wave of Auto Mega mergers and for people who don't know Honda and Nissan signed an agreement to merge and Mitsubishi is involved because they're part of I think Nissan's Alliance this would obviously get rid of all of the redundancy I think this is touch like the European oems are in real trouble you know what does Volkswagen do not clear what does stantis do it's not clear these are all businesses that are effectively melting icebergs and so you know typically melting Iceberg businesses when they get put under pressure from Smart investors like Gavin and and and is ilk they're forced to merge short gav Gavin you got a spread trade are you short these names um I I would rather not talk about specific positions I would just say I agree 100% with chamath I think there you go they're going to lose their Chinese
OEM厂商是不值得投资的。而我认为这样做的结果将是一波汽车业的超级合并潮。对于不知情的人说明一下本田和日产已经签署了合并协议,三菱也牵涉其中,因为他们是日产联盟的一部分。这显然会消除所有的冗余。我认为这有点——欧洲的那些OEM厂商麻烦大了你知道,大众要怎么办?不清楚。Stellantis要怎么办?不清楚。这些企业实质上都是正在融化的冰山。所以你知道,通常融化中的冰山型企业,当它们被Gavin这类精明的投资者施加压力时,就会被迫合并。做空吗?Gavin,你有没有做价差交易?你有没有做空这些股票?我还是不谈具体持仓吧,我只想说我100%同意Chamath的看法。我认为那就对了
便签引用
48:57
business because they don't make competitive products anymore if there's not massive protectionism and we're seeing signs that they'll be caught between Tesla and the Chinese oems and I think the only way that this doesn't H chat's the only risk to chat's prediction is just government intervention because they's such big employers and you know often seen as national champions but absent really significant government support they all in deep trouble okay well done freeberg last year you said there'd be some Blockbuster deals for right holders licensing data for AI training and Reddit did in fact sign two of those so uh a great prediction from you for last year on biggest business deal what do you got this year see if you can go two for two was it just Reddit I should probably look it up what you I'll just follow on my year of the robot theme I do think that there's going to be massive funding deals similar to what we saw this past year for compute buildout I think we're going to see massive funding deals for
他们会失去中国市场,因为他们做不出有竞争力的产品了。如果没有大规模的保护主义,而我们已经看到一些迹象表明他们会被夹在特斯拉和中国车企之间。我认为唯一能让这件事不发生的——Chamath这个预测唯一的风险,就是政府出手干预,因为它们是很大的雇主,而且常常被视为民族工业的代表。但如果没有真正重大的政府支持,它们全都深陷困境。好的,说得好。Friedberg,你去年说会有一些重磅交易,版权方把数据授权用于AI训练,而Reddit确实签了两笔这样的交易。所以你去年的预测非常准。关于今年最大的商业交易,你有什么?看看你能不能连中两次。就只有Reddit吗?我大概该去查一下。我就接着讲我那个“机器人元年”的主题吧。我确实认为会出现大规模的融资交易,类似我们过去这一年在算力基础设施建设上看到的那种。我认为我们会看到针对美国境内硬件制造产能建设的
便签引用
50:03
Hardware based Manufacturing buildout in the United States and I think that those deals may take the form of kind of traditional Equity from private markets or they may have some component that includes government support to to kind of motivate and drive and accelerate onshore manufacturing we're not going to go in the US to making stuff that is like last century I think we're going to need to move manufacturing to the next decade Next Century of of production and I think that's going to mean making some of these autonomous and robotic type systems that are going to become really critical for us particularly with China doing this massive ramp up and buildout for both drones and robots and autonomous vehicles so I think we're going to need to kind of onshore a lot of this and so that'll be a big amount of capital that's going to move and so you'll see a bunch of these big Blockbuster deals for Hardware buildout in the US Gavin what do you think to be the biggest business deal in 2025 I think the biggest business deal is that
大规模融资交易。而我认为这些交易可能采取来自私募市场的传统股权形式,也可能包含某些政府支持的成分,用来激励、推动和加速制造业回流。我们在美国不会回去做那种上个世纪的东西,我认为我们需要把制造业推进到下一个十年、下一个世纪的生产方式。而我认为这意味着要生产一些自主和机器人类型的系统,它们将对我们变得至关重要,尤其是在中国大规模扩产、大举建设无人机、机器人和自动驾驶车辆的背景下。所以我认为我们需要把很多这类产能迁回国内,因此会有大量资本流动起来,你会看到一批针对美国硬件产能建设的重磅交易。Gavin,你认为
便签引用
51:01
they're going to be deals I think there's you're going to see a Title Wave of m&a after four years of not being able to get anything done I think there's an enormous amount of pinup demand so kind of Point number one point number two something will happen with Intel and that will be big and hopefully it's good for America and then I do think you will see a lot of these Frontier AI Labs that are independent be quiet you know I thought your points were well taken Jason about you know everything that uh you know Google is bringing to Bear an ex xai owns their own compute and you guys can choose whichever one of these three you like but the ultimate AI winner will be the one with the lowest cost of infrastructure cost the lowest cost of compute and definitionally you can't be the lowcost provider if you're ring your compute from someone else because of markups yeah because there's a markup if you're buy your compute from Azure or AWS or Google you were at a disadvantage relative to their internal Services over
2025年最大的商业交易会是什么?我认为最大的商业交易就是——会有交易。我认为你会看到一波并购的浪潮,在四年什么都做不成之后。我认为积压的需求量极其庞大。这算是第一点。第二这些独立的前沿 AI 实验室安静点吧,你知道,我觉得你说的很有道理,Jason,关于你知道,Google 拿出来的那些东西,xAI 拥有自己的算力,你们可以从这三家里随便挑一家但最终的 AI 赢家将是基础设施成本最低的那一个,算力成本最低的那一个而从定义上讲,如果你的算力是从别人那里租来的,你就不可能成为低成本供应商因为有加价,对,因为如果你从 Azure、AWS 或 Google 买算力,就会有加价,随着时间推移,你相对于他们的内部业务就处于劣势所以全栈者胜。我去年的预测很不错,我当时预测字节
便签引用
52:05
time so the full stack wins great prediction for me last year my prediction was that bite dance would go public or Tik Tok would get divested or some version of this we are but 18 days away from figuring out if the Supreme Court will do just that my prediction for this year I was looking at all the media companies gentlemen there were so many possibilities there Warner Brothers we talked about Apple and Disney you know couple years ago but I think the age of autonomy is here and I think there is going to be some Partnerships that will happen between Amazon door- Uber Tesla weo and that cohort and I would not be surprised to see Tesla could buy Uber right now for but 10% of its market cap and weo could spin out and partner with Uber Amazon could buy door Dash fairly easily and with the wrath of KH being over the wrath of lenina con being over Gavin it is possible that Mega deals like these could go through and if a couple of Mega deals go through like this whoever uh you know teams up here
跳动会上市,或者 TikTok 会被剥离,或者类似的某种情况,我们现在距离揭晓答案只有 18 天了看最高法院会不会这么判。至于我今年的预测,先生们,我看了所有的媒体公司,有太多的可能性了,华纳兄弟,我们几年前聊过苹果和迪士尼,但我认为自动驾驶时代已经到来,我认为亚马逊、DoorDash、Uber、特斯拉、Waymo 这一批公司之间会出现一些合作。而且如果特斯拉现在用大约 10% 的市值就把 Uber 买下来,我一点都不会意外Waymo 可以分拆出来和 Uber 合作,亚马逊也可以相当轻松地收购 DoorDash。随着 Lina Khan 的铁腕时代结束——Lina Khan 的铁腕时代结束了,Gavin,这类超级并购是有可能通过的,如果有几笔这样的超级并购成了,那么联手的这几家就可能同时赢下自动驾驶、配送、外卖和电商,这是一个
便签引用
53:13
could win autonomy delivery food delivery and e-commerce this is a ginormous space and I think listen I don't I haven't spoken to uh obviously my friend about it but Tesla buying door Dash and Uber or Amazon buying door Dash and Uber could be the greatest service ever created if you wanted to build a super app the only thing I would add and I think dovet tales with what you said and and um friberg's kind of year of the robot is autonomous drones there's a company called zipline which my firm is an investor in fantastic yeah but autonomous drones they really are the best way to deliver almost anything to Suburban America and then I think over time it will be sorted out so that they could deliver in cities as well so I just think that that might be a little bit of a wild card for some of these delivery services well and Amazon in I believe it's Texas is actually doing these now they've got 60,000 SKS is my understanding already being delivered and they just drop it in your backyard and it's there in 45 minutes I
巨大无比的市场。我想说,我显然没跟我那位朋友聊过这事,但特斯拉收购 DoorDash和 Uber,或者亚马逊收购 DoorDash 和 Uber,如果你想打造一个超级 App,那可能会是有史以来最棒的服务我唯一想补充的一点,我觉得跟你说的很契合,也跟 Friedberg 说的「机器人之年」很契合就是自动驾驶无人机。有家公司叫 Zipline,我的基金是它的投资人,非常棒。是的,自动驾驶无人机真的是向美国郊区配送几乎任何东西的最佳方式,然后我觉得随着时间推移,问题会逐步解决让它们也能在城市里配送。所以我只是觉得,对这些配送服务来说,这可能算是一个小小的变数。而且亚马逊,我记得是在德州,现在已经在这么干了,据我了解他们已经有 6 万个 SKU在用无人机配送,直接扔到你家后院,45 分钟就送到了
便签引用
54:20
do this like two weeks ago I um you know Matan in China's got this and they they F have food delivery happening with these Dr by the way um just kind of dovetailing off of this I think the other big deal potential in 2025 just to be very specific is a deal with weo you know weo launched n SF in August 2023 in kind of an open market way Uber and Lyft were 66 and 34% market share at the time and in 15 months in November of 2024 weo is now a 22% market share of rides in SF which is the same as Lyft and Uber is now down to 55 so weo Aid in both Uber and lifts market share by on the order of 12 points um in just 15 months and now they're launching in LA in Austin all over the country they're already in Phoenix so I think you could see a massive and by the way you know they also just moved the the hardware platform over to a new d new device that supposedly going to bring the capex significantly down for new launches so you'll see much improved roic metrics which means that there's a much kind of
我大概两周前还试过。你知道,中国的美团也有这个,他们已经在用这些无人机送餐了。顺便说一句,接着这个话题往下讲,我觉得 2025 年另一个可能的大交易,说得具体一点就是跟 Waymo 有关的交易。Waymo 是 2023 年 8 月在旧金山以一种开放市场的方式上线的,当时 Uber 和 Lyft 的市场份额分别是 66%和 34%。15 个月后,到 2024 年 11 月,Waymo 在旧金山的打车市场份额已经达到 22%跟 Lyft 持平,而 Uber 掉到了 55%。所以 Waymo 大概从 Uber 和 Lyft 的市场份额里各吃掉了 12 个百分点,而且只用了 15 个月。现在他们又在洛杉矶、奥斯汀,在全国各地上线,凤凰城早就有了,所以我觉得你会看到一个巨大的……顺便说一句,他们刚把硬件平台换成了一套新设备,据说能让新城市上线的资本开支大幅下降,所以你会看到投入资本回报率指标大幅改善,这意味着
便签引用
55:20
more efficient way to use Capital to scale so the system works it is scaling they are opening up a new markets you could see something happen with weo this year that could either be something like a massive financing an IPO or a merger or acquisition with one of the big right sharing companies I don't know if that makes as much strategic sense as I've thought about it a little bit but I do think you could see a big deal with wayo this company is if you guys have not been in AO uh I think we all talked about this but it's an incredible experience and everyone I know of every age group that has been for a ride in wh comes out of it saying that is the future this is going to absolutely dominate how I'm going to get around I'm most people say it's slow right now and monotonous and they don't like it but you know because it does take weird routes but that'll be fixed over time obviously tot and uh just in case anybody thinks I'm talking my book here I have exposure to all these almost all these companies in a in a significant
扩张时资金的使用效率高得多。所以这个模式跑通了,规模在扩大,他们在开拓新市场,你今年可能会看到 Waymo 那边发生点什么,可能是一轮巨额融资、一次 IPO,或者跟某家大型网约车公司合并或被收购。我稍微想了想,不确定那在战略上是不是那么说得通但我确实觉得你可能会看到 Waymo 有一笔大交易。这家公司——如果你们还没坐过Waymo 的话,我想我们都聊过这个,那是种非常棒的体验,我认识的各个年龄层的人,只要坐过一次 Waymo出来都会说:这就是未来,这绝对会彻底改变我以后的出行方式。不过大多数人会说它现在开得慢而且很单调,他们不喜欢这点,因为它确实会绕些奇怪的路线,但这些问题随着时间推移肯定会解决。另外声明一下,免得有人觉得我在为自己的持仓说话——我对这些公司几乎都有不小的敞口,因为我看好整个赛道。在美国,网约车只占全部出行的 1.5%
便签引用
06最反共识:银行危机与社会主义回潮
56:11
way because I believe in the entire space one and a half% of rides in the US and globally it's less than 1% are done by ride sharing the Tam is going to go to 20% in a very short period of time and it's going to be across the board there's going to be a lot of winners here and to my favorite Uber obviously they have deals with eight people and it's a global market byd produces cars for half the price of any other car manufacturer and they have full self driving and it's pretty darn good from what I understand so this is going to be a global competition and that means if you want to win that Global competition something like uber and doores weo Amazon byd you're going to see some interesting Partnerships happen very quickly I believe because there's so much at stake all right and final thoughts on biggest business deals I think I maybe have a good one here consolidation in the transportation space okay moving on to 2025 most contrarian belief last year Enterprise value of open AI goes down from chth it roughly
全球来看还不到 1%,这个总市场规模会在很短时间内涨到 20%而且是全方位的,这里会出现很多赢家。我最看好的当然是 Uber他们跟八家都有合作,而且这是个全球市场。比亚迪造车的成本只有其他任何车企的一半而且他们有完全自动驾驶,据我了解还相当不错。所以这将是一场全球竞争,这意味着如果你想赢下这场全球竞争,像 Uber、DoorDash、Waymo、亚马逊、比亚迪这些公司之间,我相信很快就会出现一些有意思的合作,因为赌注实在太大了。好,关于最大商业交易的最后想法,我觉得我这个可能不错:交通出行领域的整合。好,进入 2025 年最反共识的观点。去年 chamath 说 OpenAI 的企业价值会下降,结果它大概翻了一倍。不过我当时就是把这个当成我的
便签引用
57:12
doubled but I just picked it as my prediction so I think we're s we're sympatico on uh what we think long term freeberg you said increased probability of a nuclear weapon being used for the first time since World War II thank the Lord that that didn't happen again this is contrarian beliefs people are going out there I I didn't predict it was gonna happen I just said the probability went up it was it's a tough it's a tough thing to call well this is most contrarian belief I think it's fine Dr Doom strikes again it's all good I picked Apple would become the player in Ai and they did launch Apple intelligence but it sucks I mean it sucks it's terrible I just bought the Google pixel fold n and the Gemini works perfectly it does everything that Ser is supposed to do if you say please play this song If you say please download this app if you say add this to my calendar it does it and it does it in one or tenth the amount of time that Apple takes and apple gets it wrong every time it's a piece of garbage
预测选的,所以我觉得我们在长期看法上是一致的。Friedberg,你说的是二战以来首次使用核武器的概率上升。谢天谢地这没有发生。再说一遍,这是反共识观点,大家都在放狠话。我并没有预测它一定会发生,我只是说概率上升了,这很难判断,真的很难判断。这个板块叫「最反共识的观点」,我觉得没问题,末日博士又出手了,都挺好。我当时选的是苹果会成为 AI 领域的玩家,他们确实推出了 Apple Intelligence,但它烂透了,我是说真的烂,太糟糕了。我刚买了 Google Pixel Fold,Gemini 用起来完美无缺,Siri 该做到的事它全都能做到你说「请播放这首歌」,你说「请下载这个 App」,你说「把这个加到我的日历里」,它都能做到,而且只花苹果十分之一的时间,而苹果每次都搞错,简直就是个垃圾
便签引用
58:15
it's a disgrace disad apple apple intelligence is even worse than co-pilot which is saying something Jason would you like to announce you B the domain dad.com I own it I'm not selling it right now it it redirects to J Paul Gavin Gavin do you have any idea what word he's even saying when he says that disad disad it's disg graad is what people used to say in Brooklyn for somebody who is disgraceful it is a Italian American slang how do you spell it you don't even know you own the domain you don't even know d i s g r a z i a d disg graad yes and if you type in disg graad it goes to Jake Paul I bought it when I was watching the Tyson fight it was so disgraceful I was saying to my brother Josh the black bomber this is and he just said disg graad and I just said I wonder if that's available as a domain I'm redirecting disg graad right now Tim Cook to Apple intelligence the website do you guys think Tyson through that that it was part of the deal th% I would love I think somebody said Saudi Arabia was offering to host a
是种耻辱,disgraziad(丢人现眼)。Apple Intelligence 甚至比 Copilot 还差,这可太说明问题了。Jason,你要不要宣布一下你买了那个域名?dad.com?我拥有它,现在不卖,它会跳转到 Jake Paul。Gavin,Gavin,你知道他说那个词到底是什么意思吗?disad、disad——是 disgraziad,以前布鲁克林人用来形容那些丢人现眼的人,这是意大利裔美国人的俚语。怎么拼?你都不知道自己买的域名怎么拼。d-i-s-g-r-a-z-i-a-d,disgraziad,对,你输入 disgraziad 就会跳到 Jake Paul。我是在看泰森那场比赛的时候买的,实在太丢人了,我当时跟我哥 Josh 说,「黑色轰炸机」怎么成这样了,然后他就说了句 disgraziad,我就想,这个词做域名还有没有人注册。我现在就把 disgraziad 重定向到 TimCook 或者 Apple Intelligence 的网站去。你们觉得泰森是故意放水的吗?是不是交易的一部分?百分之百是。我很想……我记得有人说沙特阿拉伯提出要举办一场翻拍,或者说重赛,一场复赛,赢的人
便签引用
59:34
remake or or whatever a redo a rematch a rematch of the fight where the winner would make 50 million and then we'll see what's what and the loser gets zero yes and the loser gets zero this is a good use of the Kingdom's money the people want to know we will all fly out there and do an Allin episode from you guys see this did you see this thing where is it is it Logan Paul is fighting Conor McGregor in India and Conor McGregor is gonna get $250 million what yeah it's like a whole stunt I I think it's boxing and it's meant to sort of like bring a bunch of tourists into India to show them the country but it's a quarter of a billion dollars to Conor MC how do we get in on this grift freeberg versus Baker poly hoaa versus kakanis oh my gosh we need to get our own celebrity box and who would we do come on keep going I got to drive home let's go come on too much fun this episode chth what's your most contrarian belief prediction for 2025 we're going out on a limb here this is going out in a limb folks don't judge us
拿 5000 万,然后咱们看看到底谁行,输的人一分钱没有。对,输的人一分钱没有,这才是沙特王国该花钱的地方,大家都想知道真相。我们都飞过去,在那儿录一期 All-In。你们看到这个了吗?你们看到那个消息了吗,是 Logan Paul 要在印度对阵 Conor McGregor,而 Conor McGregor能拿到 2.5 亿美元。什么?对,就是一场大型噱头。我想是拳击,而且目的好像是要把一大批游客吸引到印度,让他们看看这个国家,但给 Conor McGregor 的是 2.5 亿美元。我们怎么才能分一杯羹?Friedberg 对阵 Brad Gerstner,Chamath 对阵 Calacanis,天哪,我们得搞我们自己的名人拳击赛,那我们怎么配对?来嘛继续说,我得开车回家了,来吧,这期节目太好玩了。Chamath,你 2025 年最反共识的观点或预测是什么?我们这是在冒险放话,各位这是在放狠话,别评判我们
便签引用
1:00:40
the the spicier to take the better here this is where you can go freestyle I think that you're going to see a banking crisis in one of the major Mainline Banks a banking crisis wow there is a it's a small chance it happens but that's a contrar belief okay why Jam why Black Swan well I think if I had to kind of build the the case for this it would be along the following lines if you added up the total indebtedness of pax America which is US Government debt plus corporates plus Mortgage Debt and you actually sensitize it to rates that are around call it 5% what you quickly realize is on a dollar basis that because the amount of debt that we have has just massively exploded that 5% rates today on 70 odd trillion dollars is actually equivalent to 10% rates 25 or 30 years ago because we only had a fraction of that debt on a dollar basis and so the pain that you feel at 5 or 6% can very quickly Ripple through the economy the way that it did when rates 25 or 30 years ago were 10% and so I think that when people look
越劲爆越好,这是你可以自由发挥的地方。我认为你会看到一家主流大银行爆发银行危机。银行危机,哇。发生的概率不大,但这才叫反共识观点。好,为什么,Chamath?为什么是黑天鹅?我觉得如果我要论证这一点,思路大致是这样的:如果你把「美利坚治世」的总负债加起来,也就是美国政府债务加上企业债再加上房贷债务,然后按大概5% 的利率去做敏感性分析,你很快就会发现,从绝对金额来看,因为我们的债务规模已经爆炸式增长,今天 70 多万亿美元债务上的 5% 利率,实际上相当于 25 或 30 年前的 10% 利率,因为当时按绝对金额算我们的债务只有现在的一小部分。所以 5% 或 6% 带来的痛苦会很快传导到整个经济,就像 25 或 30 年前利率在 10% 时那样。所以我觉得人们看利率的时候
便签引用
1:02:12
at rates they forget the actual total dollar impact because when somebody or collectively when we have to come up with three or four trillion dollars how do you do that and so I think that there is a non-trivial risk I think it's small this is why it's contrarian but I think it's a it's a risk where if you have a marketto market problem if you have a credit default problem amongst the corporates or amongst enough individual consumers what I think that happens more than anything else is that it triggers a reserve issue and I think that the reserve issue in one of these Mainline Banks I have two that I think are more obvious than others but I don't want to name them okay fair enough I love this contrarian take this is great I think that there's a there's a small but reasonable chance that that happens I love it it's a great great contrarian take small chance but big impact Gavin what do you got well first Gemini deep research to um just ask it to look at the total you know Pax Americana debt
会忘掉实际的总金额影响,因为当某个人、或者整体上我们需要拿出三四万亿美元的时候,你怎么拿得出来?所以我认为这个风险并非微不足道。我觉得概率很小,这也正是它反共识的原因,但我认为这是个风险:如果你出现按市值计价的问题,如果在企业端或者在足够多的个人消费者中出现信用违约问题,我觉得最可能发生的是,它会触发准备金问题。我认为这个准备金问题会出现在某家主流大银行身上,我心里有两家比其他家更明显的候选,但我不想点名。好吧,可以理解。我喜欢这个反共识观点,太棒了。我认为这种情况发生的概率虽小但并非不合理。我很喜欢,非常棒的反共识观点,概率小但影响大。Gavin你有什么?首先,可以用 Gemini 深度研究,让它去查一下「美利坚治世」的总债务
便签引用
1:03:22
outstanding apply the current Market interest rate to it and then look at that you know interest expense relative to GDP and generate that chart over time that would be a cool chart that would be interesting I don't I I don't I don't disagree with chamath you know any anything is possible and um you know there there could for sure be be a problem at a big bank I don't know that I would say it's likely but you anything is possible my most cont belief so I think that America over the next at some point over the next four years will print at least one year of greater than 5% GDP growth real GDP growth I think productivity is going to go vertical because of AI and deregulation and I think there could be a world where this is the late 90s and it doesn't sound like a big difference you know five or 6% versus 2 or 3% but it is a massive difference you know at five or 6% you know the economy is doubling you'll call it every 12 years roughly versus 24 years at 3% I mean it's just it's a massive difference in
存量,套上当前的市场利率,然后看看这个利息支出相对 GDP 的比例,再生成一张随时间变化的图表,那张图会很酷,会很有意思。我并不反对 Chamath 的看法,你知道,一切皆有可能,而且确实完全有可能在某家大银行出问题。我不确定我会说这很可能发生,但一切皆有可能。我最反共识的观点是:我认为美国在未来四年里的某个时点,至少会有一年实现超过 5% 的 GDP 增长,是实际GDP 增长。我认为由于 AI 和放松管制,生产率会垂直起飞,我觉得有可能出现类似 90 年代末的局面。听起来 5% 或 6% 对比 2% 或 3% 好像差别不大,但这是巨大的差别。在 5% 或 6% 的增速下,经济大概每 12 年就翻一番而 3% 的话要 24 年。我是说,这在国家财富和美国个人财富方面是巨大的差别。至于
便签引用
1:04:38
terms of kind of the wealth of the country and individual Americans as far as a specific prediction for 25 uh because I don't know when when that will happen I think you will see the frontier Labs stop releasing their Leading Edge models to prevent knowledge distillation and their IP effectively being stolen you know seek from China it was really impressive but it thinks it's gp4 you know yeah so I just think you will you'll see the labs keep their best models in house distill them and put small small models out over time fber do you have a contrarian belief and I did put into Gemini Advanced 1.5 Pro with deep research the question which banks have the great biggest chance of being insolvent or having a financial collapse or crisis so it's doing its research right now it'll be done in about 10 minutes I think which gives you an idea of how much research it does it's nuts freeberg what do you got I think that the the party line is that socialism was defeated in this election cycle and that there was a resounding
2025 年的具体预测,因为我不知道那会在什么时候发生——我认为你会看到前沿实验室停止发布他们最领先的模型,以防止知识蒸馏和他们的知识产权实质上被窃取你知道,中国的 DeepSeek 确实很厉害,但它以为自己是 GPT-4。是的,所以我觉得你会看到各大实验室把最好的模型留在内部,蒸馏之后再逐步放出小模型。Friedberg,你有反共识观点吗?我刚才在 Gemini Advanced 1.5 Pro 里用深度研究提了个问题:哪些银行最有可能资不抵债或者发生金融崩溃或危机。它现在正在做研究,大概10 分钟后会完成,这能让你感受到它做了多少研究,太夸张了。Friedberg,你有什么?我觉得现在的官方说法是,社会主义在这轮选举中被击败了,美国民众发出了响亮的一票
便签引用
1:05:41
kind of vote from the American populace against socialism and I actually think my my contrarian belief is that we'll see a rise a dramatic rise in socialist movements in 2025 in the United States okay and I think that we going to see as Gavin pointed out an acceleration of progress we're going to see an unleashing of economic growth because of deregulation and AI but I do think that some markets we've also talked about the the the the downfall of us autom manufacturing and some other Industries there's going to be a real significant shift in 2025 some Industries and some companies are going to be huge winners and some companies are going to be huge losers there's going to be some parts of the economy that are going to be big Winners and some parts of the economy that are going to be big losers when you have this sort of a change this fast there are often large contingents of people that are left behind and when that happens I do think that the Socialist policies and the Socialist movements gain steam when um Peron came
反对社会主义。而我的反共识观点其实是,我们会看到社会主义运动在 2025 年的美国急剧兴起。好。而且我认为,正如 Gavin 指出的,我们会看到进步加速,会因为放松管制和 AI 而释放经济增长,但我也确实认为某些市场——我们也聊过美国汽车制造业和其他一些行业的衰落——2025 年会出现真正重大的变化,有些行业和有些公司会成为巨大的赢家有些公司会成为巨大的输家。经济中会有一部分是大赢家,另一部分是大输家。当变化来得这么快的时候往往会有大批人被落下,而一旦出现这种情况,我确实认为社会主义政策和社会主义运动就会壮大。当庇隆在 1940 年代中期在阿根廷上台时,那个国家正经历 8% 的 GDP 增长,这呼应了
便签引用
1:06:39
to power in Argentina in the mid1 1940s that country was experiencing 8% GDP growth to Gavin's point about accelerating growth growth does not mean that it benefits everyone equally and I think that some folks will see people go from being billionaires to 100 billionaires to the world's first trillionaire and it will also start to fuel this rise so I think that we will see an increase in the breadth and depth of socialist movements in the United States by the way particularly with Doge cutting Government funding to programs that benefit individuals employment being cut in in the federal government and through Federal contractors there's just a lot of Rapid change that's about to kind of really upset a lot of people I totally agree I think AI you know people are fond of saying that what in a world of AGI or ASI money will be meaningless but for a short period of time money will matter more than it's ever mattered before because the amount of money you can spend on AI on test time compute is going to give you a
Gavin 说的增长加速——增长并不意味着所有人都能平等受益。我认为有些人会看到,有人从亿万富翁变成千亿富翁,再到世界上第一个万亿富翁,这也会开始助推这种情绪上升。所以我认为,我们会看到社会主义运动在美国的广度和深度都有所增加顺便说一句,尤其是在 Doge 削减惠及个人的政府项目资金、联邦政府和联邦承包商岗位被裁减的背景下,即将到来的剧烈变化实在太多,会真的惹恼很多人。我完全同意。我觉得 AI……你知道,人们很爱说,在一个到了 AGI 或者 ASI 的时代,钱会变得毫无意义,但在此之前的一小段时间里,钱会比以往任何时候都更重要,因为你能在 AI、在测试时计算(test time compute)上花多少钱将给你带来巨大的优势,不管你是一家公司还是个人。所以我觉得
便签引用
1:07:43
massive Advantage whether you're a company or an individual so I just think AI is going to really amplify inequality for some period of time yeah I hope I'm wrong like I hope you know AI leads to you know all sorts of you know opportunities being created for low there's a lot of there's just going to be a lot of employment and income disruption in 2025 and it's going to fuel socialist movements and I think that this is going to be a more difficult Year everyone thinks it's kind of rosy red because we're all working in the tech industry in Silicon Valley but the reality on the ground for most Americans could be a lot harsher than any of us anticipate and that could make for a very difficult political environment and social environment this is a great contrarian prediction free I would build on it that I don't think this is just going to hit blue collar I seeing in the Venture industry and in entrepreneurs all over the place people who are super qualified who had six figure salaries even to Mid six figure
AI 在一段时间内真的会放大不平等。是的,我希望我是错的,我希望 AI 能带来各种各样的机会。总之,2025 年会出现大量的就业和收入动荡,这会助推社会主义运动。我认为明年会是更艰难的一年。大家都觉得形势一片大好,因为我们都在硅谷的科技行业里工作,但对大多数美国人来说,现实处境可能比我们任何人预想的都要严酷得多,这可能会造就一个非常艰难的政治环境和社会环境。这是个很棒的反共识预测。我想在这上面再加一点:我不认为这只会冲击蓝领。我在风投圈、在各地的创业者中都看到,那些条件特别优秀、原本拿六位数甚至中六位数薪水的人,找不到工作,或者找不到跟以前薪酬
便签引用
1:08:39
salaries not be able to find work we not find work at previous compensation levels why because people are doing more with less it's much better to invest and do deep research and Ai and automate stuff or deprecate stuff or delegate stuff to to other regions than to hire Americans in some cases and that philosophy that's happening isn't just going to hit truck drivers it's going to hit developers potentially designers writers what the contrar part of what you guys are saying what's the contrarian part socialism socialism ISM On The Rise I it's pretty interesting because I think you don't think it's contrarian you think it's obvious I think the party line has been that socialism was getting knocked back this year with this it was a mandate against socialism and some of the Socialist policies that were being put forth and I think that the contrarian view is that we're actually got this really wrong and socialism is going to be on a big rise but just to build on what you said earlier like w
水平相当的工作。为什么?因为大家都在用更少的人做更多的事。与其在某些情况下雇美国人,不如去投资、去做深度研究和 AI、去做自动化,或者砍掉一些业务,或者把活外包到其他地区。而这种正在发生的思路,不只会冲击卡车司机,还会冲击开发者,可能还有设计师、写作者。你们说的这些,反共识的部分在哪?反共识的部分是什么?社会主义。社会主义的兴起。我觉得挺有意思的,因为我以为你不觉得这是反共识的,你觉得这是显而易见的。我觉得主流说法一直是,今年社会主义被打回去了,这(次选举)是对社会主义、对那些被提出来的社会主义政策的一次否决。而我认为反共识的观点是,其实我们完全看错了,社会主义会迎来一次大幅上升。不过接着你刚才说的那点:觉醒主义(wokeism)和进步主义会和社会主义脱钩。我觉得觉醒主义和
便签引用
1:09:39
wokeism and progressivism will decouple from socialism I think wokeism and progressivism you know will be on a declining Trend but socialism in terms of the government like policy more role yeah yeah why have Universal why don't we have Universal healthare why don't we have prek or you know in every state and and you know these kind of things I think Americans are right to ask those questions now other behaviors are obviously appor but you do have the right to ask why we can't figure this out and why our government has failed Us in something as basic as you know providing after school programs or universal healthcare or Universal Child Care these things are easy to do and when you see everybody getting rich and and the polarization of wealth I can understand people saying why don't we have these basic things when other countries have them it's a reasonable you're making the assumptive statement that these things are easy to do which was what was said about education it's like let's give everyone access to
进步主义会呈下行趋势,但社会主义——就政府在政策上扮演更多角色而言——是的,为什么不搞全民医保?为什么我们没有全民医保?为什么不在每个州都有学前教育,诸如此类的事情?我觉得美国人问这些问题是有道理的。当然另一些行为显然不可取,但你确实有权利问:为什么我们连这个都搞不定?为什么我们的政府在这么基本的事情上让我们失望——比如提供课后项目、全民医保或者全民托育?这些事情很容易做到,而当你看到所有人都在发财、财富两极分化的时候,我完全能理解人们会说:为什么我们连这些最基本的东西都没有,而别的国家都有?这是个合理的……你这是在做一个假设性的论断,说这些事情很容易做到。教育当年也是这么说的:让所有人都能上
便签引用
1:10:37
college education with the uh the federal student loan programs and what happened was when we introduced those programs those schools started to charge more and the tuition went up every year and eventually the cost of Education inflated away from the benefit you were getting from it and this has happened universally in healthcare it has happened happed in housing it has happened in education it has happened in every Market where the unit states where the government has stepped in to provide Capital to support that market is that the market basically no longer operates in a free way I appreciate the challenge to it and I'll explain to you why I believe it's easy if you just put this down to the States and you make it more competitive I think you solve the problem if you introduce School vouchers and you create competition if you take universal healthcare and maybe you run some experiments where different states get that money from the federal government and let them run 50 different experiments I think we could actually
大学,于是搞了联邦助学贷款项目。结果呢?我们推出这些项目之后,那些学校开始收更多钱,学费年年上涨,最终教育的成本膨胀到远超你从中获得的收益。这种情况在医疗领域普遍发生过,在住房领域发生过,在教育领域发生过,在每一个政府出手提供资本去支撑的市场里都发生过,结果就是这个市场基本上不再以自由的方式运转了。我很感谢你的质疑,我来解释一下为什么我认为这很容易。如果你把这件事下放到各州,让它更有竞争性,我觉得就能解决问题。如果你引入教育券、制造竞争;如果你拿全民医保来做一些实验,让不同的州从联邦政府拿到这笔钱,让它们跑 50 种不同的实验,我觉得我们真的能解决其中一些问题。但你说得对,凡是联邦
便签引用
1:11:28
solve some of these problems but you're correct anything that the federal government does it eventually becomes corrupt and inefficient my most contrarian belief was open AI loses its lead loses its nonprofit to for-profit transition and becomes the number four player in AI the total collapse of open AI is my most contrarian prediction for 20 that's a good one the total collapse of dble double doubling and tripling down hey man I'm trying to make this spicy it's a spicy category let's do I got got it could still run for president Jal maybe he's great I don't know everybody who all his friends left the company maybe he's great maybe he's a standup guy I mean everybody quit but maybe he's gav are you are you familiar with J K's prediction that Jeff Bezos is gonna run for president it was that was my I was predicting I wanted him to I wanted him to and he bought Washington Post and he quit the Amazon job and he bought a house in Washington d and he went down to marago don't be surprised if Bezos after he gets through
政府做的事,最终都会变得腐败和低效。我最反共识的看法是:OpenAI 会失去领先地位,会搞砸从非营利到营利的转型,沦为 AI 领域的第四名。OpenAI 的全面崩盘,就是我对 2025 年最反共识的预测。这个不错。全面崩盘,加倍下注、三倍下注。嘿,我是想让这个环节够辣,这本来就是个辛辣的环节。来吧,我懂了。他还是可以竞选总统的,山姆(Sam Altman)也许很棒,我不知道,所有朋友都离开了公司。也许他很棒,也许他是个正派人。我是说所有人都辞职了,但也许他……Gavin,你知不知道 J·Cal 那个杰夫·贝索斯要竞选总统的预测?那是我……我当时是在预测,我是希望他去选。他买下了《华盛顿邮报》,辞掉了亚马逊的职务,还在华盛顿买了房子,又跑去了海湖庄园。所以别觉得意外:等贝索斯度过他的中年危机——去
便签引用
07最佳与最差资产:HBM、CDS、企业软件
1:12:30
his uh midlife crisis and going to Coachella and partying and having a great time which he deserves if he comes back and says you know what I want to serve my country I still stand by Washington Post or do you think he's gonna hold on to it well let's stick with that one because that's one prediction Jeff Bezos will sell the Washington Post in 2025 that's a good poly Market yeah sell it to Caren swier sell it to Karen swier run it right into the ground they'll Double Down uh okay best performing asset everybody loves this one best performing asset hey we can you can put a price on this best performing asset last year chamat did a spread trade he was long public tech stocks short private late stage Tech stock index NASDAQ top 100 Tech stock ETF was up 10% 2024 freeberg you went with your uranium ETF not Uranus uranium ETF and you're a UR are a was down 1% in 2024 I know sideway you know what I I looked at the components of that ETF last week when we were preparing for this it has absolute junk
科切拉音乐节、开派对、玩得很开心,这些他都配得上——如果他回来说:你知道吗,我想报效国家。我依然坚持《华盛顿邮报》那个预测,还是说你觉得他会一直留着它?我们就说这个吧,因为这是一个预测:杰夫·贝索斯将在 2025 年卖掉《华盛顿邮报》。这在 Polymarket 上会是个好盘口。是啊,卖给 Kara Swisher。卖给 Kara Swisher,让她把它彻底搞垮。他们会加倍下注。好,表现最好的资产,大家都喜欢这个环节。表现最好的资产,嘿,这个是可以定价的。去年表现最好的资产:Chamath 做了个价差交易,他做多公开市场科技股,做空私募后期科技股指数。纳斯达克 100 科技股 ETF 在 2024 年涨了 10%。Freeberg,你选的是铀 ETF——不是天王星(Uranus),是铀 ETF——你的 URA 在 2024 年跌了 1%。我知道,横盘。你知道吗,上周我们准备这期节目时我看了那只 ETF 的成分股,里面全是垃圾,
便签引用
1:13:35
in it like it it was not the right the right right way to kind of trade uranium but anyway that was my My outcome for the year I went with the on demand economy with Uber Airbnb and door Dash Uber was up 30% but now it's only up 3% with the robo taxi hadwin Airbnb is basically flat and doorish up big 74% shout out to Stanley Tang great job to the team over there so jam get your flowers there let's do our best performing asset let's let our guests go first Gavin what would be the best performing asset of 2025 I think the companies that make high bandwidth memory going back to we're we're going to run out of compute it's it's um it's actually a pretty shocking stat high band with memory is a bigger part of in cogs on gpus than Taiwan Simi is and today there's two companies that can make it h HX and Micron we'll see if Samsung gets their act together but hbm memory it's in Nvidia gpus AMD gpus Amazon terraniums and particularly in a world of test time compute and inference being so important H been with memory is
它根本不是交易铀的正确方式。不过无论如何,那就是我那年的结果。我选的是按需经济:优步、Airbnb 和 DoorDash。优步一度涨了 30%,但现在只涨了 3%,因为 Robotaxi 那阵风头。Airbnb 基本持平,DoorDash 大涨 74%,为 Stanley Tang 点赞,那边的团队干得漂亮。所以 Jamath,你该拿花。我们来说说表现最好的资产,先让我们的嘉宾来。Gavin,你觉得2025 年表现最好的资产是什么?我觉得是生产高带宽内存(HBM)的公司。回到我们说的那点:我们会面临算力短缺。有个数据其实挺让人震惊的:在 GPU 的成本构成里,高带宽内存占的比重比台积电还大。而今天能生产它的只有两家公司:SK 海力士和美光。我们再看看三星能不能振作起来。但 HBM内存用在英伟达的 GPU、AMD 的 GPU、亚马逊的 Trainium 里,特别是在一个测试时计算和推理如此重要的世界里,高带宽内存的重要性可以说前所未有,而且它
便签引用
1:14:53
arguably more important than it ever was and it's been um it's been sold out for the last two years so high band with memory would be my pick okay good pick great pick in fact what do you got Jam so let me preface this by saying that this is a pick that 92 times out of 100 goes to absolute zero okay and six out of the remaining eight times you make 10 extra money and then the final two times you make anywhere between a 100 to a thousand extra money okay sounds like Bitcoin okay so this is a loser trade okay but I would be long CDs so what am I buying I am buying insurance I'm buying Insurance using credit default swaps I'm buying what's called protection that there is no default event in 25 again I'm not going to tell you which companies or what maturities but just the general idea for me is I would like a little bit of an insurance policy in 2025 so that the men and the women that we have voted in have the chance to do their work in peace I think that there is a small chance of some volatility next year I
过去两年一直处于售罄状态。所以高带宽内存会是我的选择。好,好选择,确实是好选择。Chamath,你选什么?先说个前提:这个选择在 100 次里有 92 次会归零,好吧;剩下八次里有六次你能赚 10 倍;最后那两次,你能赚 100 倍到 1000 倍不等。听起来像比特币。好,所以这是个亏钱的交易。但我会做多 CDS。那我在买什么呢?我在买保险。我用信用违约互换买保险,买的是所谓的“保护”——赌 2025 年不会出现违约事件。同样,我不会告诉你具体是哪些公司、什么期限,但对我来说大致的想法就是,我想在 2025 年买一点保险,好让我们选上来的那些男男女女能安心把工作做完。我觉得
便签引用
1:16:24
hope it doesn't happen I hope that this trade like I said 92 times out of 100 loses money I hope it loses money but if it hits it will be the best performing asset of 2025 it will be the equivalent of acman buying CDs right at the right at the beginning of the covid crisis how does one buy those you have to have an isda you talk to the big investment Banks and they'll price it out for you but again I and and I just want to be clear this is not something I think will happen it's not something I want to happen but I do think that if you look back in terms of just the tonnage of dollars you can make and the massive risk asymmetry that it presents to you when you look at the concentration of the S&P when you look at just the total gross amount of debt that we have when you look at rate spiking all of these things say having a little Insurance may not be a bad thing so I hope you're nodding Gavin you're nodding you want to ask no no I was just I was just thinking I mean if if you know chat's prediction of a bank
明年出现一些波动的概率不大,我希望它别发生,我也希望这个交易就像我说的那样,100 次里有 92 次是亏钱的。我希望它亏钱。但如果它真的命中了,它会是 2025 年表现最好的资产,相当于 Ackman 在新冠危机刚开始时买 CDS 那一笔。这个东西怎么买?你得有 ISDA 协议,去跟那些大投行谈,他们会给你报价。但我还是想说清楚,这不是我认为会发生的事,也不是我希望发生的事。但我确实觉得,如果你回头看你能赚到的钱的绝对量级,以及它给你的巨大风险不对称性;当你看看标普的集中度,看看我们债务的总规模,看看利率飙升,所有这些因素都说明,买一点保险可能不是坏事。所以我希望你在点头,Gavin,你在点头。你想说什么?不不,我只是
便签引用
1:17:27
failure is true you absolutely want to own CDs you I mean you'll you forget 100x you'll a th000 x 10,000x what do you got freeberg so I went with Chinese tech stocks or Chinese Tech ETFs wow I think that the market everyone's kind of dumped Chinese tech stocks over the last couple of years everyone's taken this isolation stance on positioning portfolios and saying hey we can't do business with China it's over but I do think that the Trump Administration particularly with their recent request on Tik Tok being the ban on Tik Tok being kind of halted they're trying to line up what I would call kind of like the great deal with China so I don't know anything about what they're actually trying to do but I think that Trump and leadership in the US government want to kind of open up the Chinese market to American companies in order to give Chinese companies access to the American market and I think that they're going to get a deal done given the position with China I think that they're very likely to kind of get a
在想,如果 Chamath 那个银行倒闭的预测成真,你绝对会想持有 CDS。别说 100 倍了,你会赚 1000 倍、10000 倍。Freeberg,你选什么?我选的是中国科技股,或者中国科技 ETF。哇。我觉得市场……过去几年大家基本都把中国科技股抛掉了,大家在组合配置上都采取了这种脱钩立场,说我们没法跟中国做生意了,结束了。但我确实认为特朗普政府——尤其是他们最近要求暂缓 TikTok 禁令这件事——他们在试图促成我所谓的“与中国的大交易”。我并不知道他们实际上想做什么,但我认为特朗普和美国政府的领导层想把中国市场向美国公司打开,以此换取中国公司进入美国市场。我觉得他们会谈成一笔交易。考虑到
便签引用
1:18:32
deal done that's one driver I think there's three drivers the second driver is that obviously the cost of build out of electricity production in China is unfathomable they recently approved $1 137 billion hydroelectric Dam facility which is going to add another I think couple hundred gigawatts of electricity production in that country not to mention all the nuclear build that we've talked about in the past so the cost per hours already lower the amount of electricity available is going up and then I think that the Chinese Communist Party have this incredible ability to throttle up and down free markets and Entrepreneurship and this is a moment where you could kind of see them making the throttle go the way towards enabling more Innovation more free market activity and it's going to be one of the motivators for them to do a deal so when you put all of this together I do think that there's a lot of Chinese tech companies that have been beat up under the assumption that this is going to be
中国目前的处境,我认为他们很可能谈成。这是第一个驱动因素。我觉得有三个驱动因素。第二个是,中国在电力建设上的成本低得难以想象。他们最近批准了一个 1370 亿美元的水电站项目,我记得这会再增加几百吉瓦的发电能力,更别提我们过去聊过的那些核电建设。所以每(度电)的成本已经更低了,可用电量还在往上走。然后我觉得中国共产党有一种惊人的能力,可以给自由市场和创业活动踩油门或者踩刹车,而现在这个时刻你可能会看到他们把油门往鼓励更多创新、更多市场化活动的方向推,这也会成为他们促成交易的动机之一。所以把这些放在一起看,我确实觉得
便签引用
1:19:23
a a very difficult conflict with the US in the years ahead and I think that that may not be the case going into 25 and I think that their these stocks look pretty cheap I was just looking at Alibaba and it trades at a pretty decent multiple it seems like these stocks could be poised for a pretty good run in 25 if the macro works out this feels like a contrarian take I mean Trump I don't think is going to find uh a way to balance the relationship with China very well and I don't think you can rattle up entrepreneurship after you cut everybody's knees out the last time plus I don't trust any accounting statement coming out of China because they could fudge whatever they want but wow what a great uh potential you know Buy Low sell High the good thing is you'll know if I'm right you'll know if I'm right or wrong in a year you can exactly we'll know exactly that's right pull the ETF absolutely I mean Gaffin what do you think of this hot take here from from Mr freeberg really cheap I've had a I call
有很多中国科技公司被打压得很惨,因为大家假定未来几年中美会陷入非常严重的冲突。而我认为进入 2025 年后情况可能并非如此,我觉得这些股票看起来相当便宜。我刚才在看阿里巴巴,它的估值倍数相当不错,这些股票看起来可能会在 2025 年迎来一波不错的行情,如果宏观配合的话。这听着确实像个反共识观点。我是说,特朗普我觉得不会很好地找到平衡对华关系的办法,而且我不认为你能在把所有人的腿都打断之后,再把创业热情重新催起来。再加上我不相信中国出来的任何财务报表,因为他们想怎么做假账都行。不过哇,这确实是个很棒的潜在“低买高卖”机会。好在你会知道结果——一年之内你就会知道我是对是错。没错,我们会知道得清清楚楚。把那只 ETF 调出来,
便签引用
1:20:18
it a no China guideline ever since the longtop financial fraud more than 15 years ago or it was just amazing it was you know uh a lot of great investors owned it and you know they they talked a great game and they had a western auditor but then it turned out that the fraud was happening at the local post office where they were opening up the documents that the local Auditors had signed off on changing them and then sending different documents you know you know whatever fedexing them yeah so I just think it's it's a hard place if you're I think if you're not Chinese to make money but I tend to agree with a a lot of what David said I think Trump and G both want to deal I think there's a deal to be done that leaves Putin out in the cold as I referenced and if that happens Katie bar theor I mean there are really really high quality companies you know that are some of them are mid single digit multiples and there and it's a big market so if if you did get it right and you by the way the Chinese the Chinese
当然。Gavin,你怎么看 Freeberg 先生这个火辣观点?非常便宜。自从 15 年多前的东南融通财务造假案之后,我就给自己定了个叫“不碰中国”的准则。那事真是让人叹为观止,很多顶级投资人都持有它,他们把故事讲得很好,也请了西方的审计师,但后来发现造假发生在当地邮局——他们在那里拆开当地审计师已经签字的文件,把内容改掉,再把另一份文件寄出去,就用联邦快递寄走。所以我觉得,如果你不是中国人,在那里赚钱是件难事。不过我基本同意David 说的很多内容。我认为特朗普和习都想要一笔交易,我认为存在一笔能把普京晾在一边的交易,就像我之前提到的。如果那真的发生了,那就放手干吧。我是说那边确实有一些质量非常非常高的公司,
便签引用
1:21:26
companies serve a global market it's not just the US they serve uh you know Africa South America a mean the whole southern hemisphere Well yeah if you look at byd that it's all over Europe it's yeah it's it's hug look I mean I I mean the the Chinese companies and they have the best unit economics they have the best cost of production I mean everything is advantaged particularly if you believe it's the year of the robot there's going to be a kind of a massive demand around the world for Automation and for rebuilding manufacturing capacity all over the world and I think that China could service those markets more efficiently than any other kind of country of origin so it's a pretty pretty powerful set of macro drivers as well I think the max 7 is going to be the best performing asset I'm taking the other side chamat I think in an earlier prediction said maybe not so much I think that what they've learned in the last couple years is that there is an incredible earnings expansion you can do
其中一些的估值只有中个位数倍数,而且市场很大。所以如果你真的赌对了……顺便说一句,中国公司服务的是全球市场,不只是美国,他们还服务非洲、南美,基本上整个南半球。是啊,你看比亚迪,它在整个欧洲都是。对,规模很大。听着,我是说中国公司有最好的单位经济效益,有最低的生产成本,方方面面都有优势。尤其是如果你相信明年是机器人之年,全世界会出现巨大的自动化需求,以及在全球各地重建制造产能的需求,我认为中国能比任何其他产地更高效地服务这些市场。所以这是一组相当有力的宏观驱动因素。我觉得“七巨头”(Mag 7)会是表现最好的资产,我站对立面。Chamath 我记得在前面的预测里说不太看好。我认为他们过去几年学到的是,
便签引用
1:22:25
by not hiring people and Outsourcing jobs to other parts of the world and automating and the gains we'll see from AI which they're producing for other companies and for consumers they're applying first internally so the internal application of AI will allow these companies to have earnings growth that people will not be able to comprehend over the next couple of years so I'm going with Max s we got a lot of different takes here going different directions what a great episode so far all right worst performing asset last year chamat said late stage Tech stock index mostly SAS freeberg you put in a brilliant spread trade short vertical SAS long AI Cloud providers SAS has come had a little bit of a rebound but Google up 36% for AI Cloud Microsoft 14% Amazon 46% all in 2024 I said llm startups like open AI anthropic too many players open sourcing will kill pricing obviously the one valuation we can track is open Ai and it doubled but I do think I'll longterm have that one work worst performing asset of 2025 Gavin
通过不招人、把岗位外包到世界其他地方、搞自动化,可以实现惊人的利润扩张。而他们为其他公司和消费者生产的 AI 带来的收益,他们会先在内部应用。所以 AI 的内部应用会让这些公司实现未来几年人们难以方向……到目前为止这一期真是太棒了。好,去年表现最差的资产,Chamath 说是后期科技股指数,主要是 SaaS。Friedberg,你做了一个很漂亮的价差交易:做空垂直 SaaS,做多 AI 云服务商。SaaS 确实有一点反弹,但谷歌因为 AI 云涨了 36%,微软涨 14%,亚马逊涨 46%,都是 2024 年的数据。我说的是像OpenAI、Anthropic 这样的大模型创业公司,玩家太多,开源会把定价打下去。显然我们唯一能追踪到估值的是 OpenAI,它翻倍了。但我确实认为长期来看这个判断会成立。2025 年表现最差的资产——Gavin,企业应用软件。好。
便签引用
1:23:35
Enterprise application software okay this is basically just about like 2025 is going to be I think the year of Agents particularly in the second half agents just being AI models that can take action on your behalf that can do anything online that a human can do online and if the lab ABS and the big cloud providers dominate agents which seems likely going back to you know the the lowcost producers going to win Enterprise application software I think is going to be in a lot of pain and you know some of these companies are talking a big big game about agents but at the end of the day they don't have their own models they don't own their own compute and I just don't see them being ultimate winners in the world of Agents you know I could easily I could be wrong because they do you know they they they have some customer data maybe a little bit of a data mode and they do you know have strong customer relationships but you know most companies have relationships with AWS Google or Microsoft as well so I think Enterprise
这基本上就是说,我认为 2025 年会是智能体(Agent)之年,尤其是下半年。所谓智能体就是能代表你采取行动的 AI 模型,能在网上做任何人类在网上能做的事。如果那些实验室和大型云厂商主导了智能体——这看起来很有可能——回到你说的那句,低成本生产者会赢,那我认为企业应用软件会非常难受。你知道,有些公司在智能体这件事上把话说得特别大,但说到底他们没有自己的模型,也没有自己的算力,我实在看不出他们能成为智能体时代的最终赢家。当然我也可能看错,因为他们确实有一些客户数据,也许有一点数据护城河,而且他们确实有很强的客户关系。但大多数公司跟 AWS、谷歌或微软也都有关系,所以我认为企业
便签引用
1:24:40
application software will be in pain okay and chth what do you got I think Gavin just absolutely nailed it I'm going to double down on what he said I think that there's a term that that we will start to use more I started to use it internally at 890 over this last year in 2024 which is the software industrial complex these are these large bloated in many cases enterprise software companies that effectively have convinced incredible numbers of organizations to spend a tremendous amount of money essentially wrapping a bunch of heuristics and business rules around the crud database and along with that what they have perfected really is a go to market and sales motion the golf trips the steak dinners you know we mentioned this before what Alex karp railed against none of those things equate to product value and increasingly in the world of agentic software and AI I think that you can rebuild a lot of these workflows in very efficient ways M and I think the go to market is going to be driven by CEOs and CFOs Who start to
应用软件会很难受。好,Chamath 你的答案是什么?我觉得 Gavin 说得太到位了,我要在他的基础上加码。我觉得有一个词我们会越来越常用,我在 2024 年这一年里开始在 8090 内部用这个词,叫“软件工业复合体”。就是那些庞大臃肿的、很多时候是企业级的软件公司,它们实际上说服了数量惊人的组织花掉巨额的钱,本质上只是在一个增删改查数据库外面包了一堆启发式规则和业务规则。与此同时,它们真正做到极致的是市场进入和销售打法——打高尔夫、请吃牛排,我们之前提过,就是 Alex Karp 痛批的那些东西,这些都不等同于产品价值。而在智能体软件和 AI 的世界里,我认为你可以用非常高效的方式重建很多这样的工作流。而且我认为市场进入会由 CEO 和 CFO 来推动,他们开始
便签引用
1:26:00
exert a little bit more pressure on their cios to manage spend and in that world I just think that these next Generation AI businesses are built frankly an order of magnitude cheaper than the companies that they compete against and so even if it's just feature for feature the same I think the software industrial complex these old Mainline traditional enterprise software companies I think you're going to start to see fissures in those businesses in 2025 so I I agree with Gavin different words but same result and you put your time and money where your mouth is on this one 11 months ago you found it 88090 to get 80% the the traction in in less than a year to me is shocking and I don't think it speaks necessarily I I think we're decent we're very good we have great Engineers but my 30% engineer in team I think does the work of 300 people by next year it'll be doing the work of 3,000 people maybe it will only grow by 10 or 20% so and the reason is we use these tools we're productive to an order
给 CIO 施加更多压力去管控支出。在那样的世界里,我认为这些新一代 AI 公司的构建成本,坦白说比它们的竞争对手要低一个数量级。所以哪怕功能完全一样,我认为软件工业复合体——这些老牌主流传统企业软件公司——你会开始在 2025 年看到它们的业务出现裂缝。所以我同意 Gavin,说法不同但结论一样。而且你在这件事上是真金白银地下注,11 个月前你创立了 8090,不到一年就拿到了 80% 的进展,这对我来说太震撼了。我觉得这并不一定说明……我觉得我们还不错,我们很强,我们有很棒的工程师,但我那 30 人的工程团队,我认为干的是 300 个人的活,到明年就会干3000 人的活,而团队可能只增长 10% 或 20%。原因就是我们用了这些工具,我们的生产力提高了一个
便签引用
1:27:06
of magnitude that I didn't think was possible and as a result we're just able to price it differently so even if it's the exact same product its cost structure is just meaningfully lower and so this is what I mean where the ROI calculators get blown up all of the traditional go to market mod get blown up because in an RFP or in any other sales environment what you do is somebody says it's $100 and you show up and you say 10 okay and at 10 it's still hugely profitable for you and and just to add just one last thing is just you know fundamentally these Enterprise application software companies the software industrial complex and I agree with everything jamat said you know they're they're fundamentally based on making white collar human employees more efficient and what the AI companies are going to do is just say hey we're just going to replace that worker and it's just a fundamentally different mentality par shift yeah and it goes to like there may may be a lot of pain that a companies this and you know David's
我以前觉得不可能的数量级,结果就是我们能用完全不同的方式定价。所以哪怕是一模一样的产品,它的成本结构就是显著更低。这就是我说的——ROI 计算器会被炸掉,所有传统的市场进入模式都会被炸掉。因为在招标或者任何销售场景里,别人报价 100 美元,你出现然后说 10 美元。而在 10 美元这个价位,你依然利润极高。再补充最后一点,就是从根本上说,这些企业应用软件公司、这个软件工业复合体——我同意 Chamath 说的一切——它们的根基是让白领员工更高效,而 AI 公司要做的是直接说:嘿,我们干脆把那个员工替换掉。这是根本不同的心态,是范式转变。是的,这也引出了可能会有很多公司很痛苦,还有 David 提到的那些关于社会主义可能
便签引用
1:28:07
points around the you know potential rise of of socialism what do you got for worst performing asset Mr freeberg 2020 I'm probably just going to Triple underline vertical sass again perc pricing model being challenged pricing being compressed as companies explore inhouse tools built with AI that replaces these kind of traditional business practices be destroyed this is it was an obvious it was an yeah it was an obvious one on my list but I want to make a different one I want to make a different one how about do you have anything open Ai No open is great all the talented people left but you know I don't know where they are where do all the where are they all the other open AI employees are they on vacation they're skiing too right no they've all started competitors oh right that's what happened correct yeah they all left to compete and have Revenge startups wow that's an interesting Trend okay so consumers make five big decisions in their life as we all know college still doing 12 billion Revenue I
抬头的观点。Friedberg 先生,你 2025 年表现最差的资产选什么?我大概只是要再给垂直 SaaS 划三道下划线。定价模式受到挑战,价格被压缩,因为企业开始尝试用 AI 自建内部工具,替代这些传统业务做法,直接被摧毁。这是个很明显的……对,这在我的清单上是个很明显的选择,但我想说点不一样的,我想说点不一样的。要不……你有没有什么关于 OpenAI 的?OpenAI 没事,OpenAI很好,所有有才华的人都走了。但你知道,我不知道他们在哪儿,那些人都去哪儿了?其他 OpenAI 员工呢?他们在度假吗?他们也在滑雪?不,他们全都创立竞争对手了。哦对,原来是这样。没错,他们都离职去竞争,搞“复仇式创业”了。哇,这倒是个有意思的趋势。好,众所周知,消费者一生中会做五个重大决定。大学……还在做 120 亿美元的营收,我
便签引用
1:29:05
mean well that's that's a projection sometimes meet reality six billion in losses too David against that 12 billion in Revenue yeah six billion this year or six billion next year I think the six is for next year but I yeah I don't know I guess it makes sense why they're trying to charge 200 bucks a month now they got some headwinds okay listen consumers make five big decisions in their life we know college spouse kids are uh the obvious ones and then cars and homes and the consumer is up against it with record debt so since you can't trade College spouse kids really I think Legacy car companies in real estate are going to face continued headwinds and be terrible assets because listen we've overbuilt in some cases there's tons of cars on lots and people can't afford homes with these mortgages so I think these are going to be the two worst performing asset people in the Legacy oems as chth pointed out in a previous prediction and then I just think real estate is the same if you look at a place like Texas two years in a row
是说,那只是个预测,有时候会撞上现实。而且还有 60 亿美元的亏损,David,对着那 120 亿的营收。对,60 亿是今年还是明年?我觉得那 60 亿是明年的。但我也不确定,我猜这也说得通,为什么他们现在想收每月 200 美元,他们确实遇到逆风了。好,听着,消费者一生中会做五个重大决定:大学、配偶、孩子,这些是显而易见的,然后是车和房。而现在消费者面临创纪录的债务压力。既然你没法交易大学、配偶和孩子,我认为传统车企和房地产会继续面临逆风,会是很糟糕的资产。因为听着,我们在某些地方建得太多了,车场上停满了卖不掉的车,而人们在这种房贷利率下买不起房。所以我认为这会是表现最差的两类资产——传统整车厂,就像 Chamath 之前那条预测里指出的;然后房地产我觉得也一样。如果你看看德州这样的地方,房价已经连续两年下跌,租金也连续两年下跌,同样的事
便签引用
08最期待趋势:核电与 AI 三轴扩展
1:30:06
housing values have gone down rent has gone down two years in a row same things happening in other states where they allowed you to build and people are leaving states where they don't allow you to build so that's my worst performing asset trade Enterprise was my other choice but I'll go with something slightly different most anticipated trend for 2025 last year chamath congratulations you said last year that Bitcoin would hit 100K for the first time nailed it half court shot well done I think at the time you made that prediction it was probably trading at it felt like a layup felt like a layup to you up 112% uh freeberg you said predictive models in AI driven Discovery and farma and bioengineering how did that do that prediction well there was a lot of funding and Demus won the Nobel priz Noel prize there you go so it worked out I picked efficiency in the form of AI advancements in labor and Outsourcing and I've seen that Bunch with our investment in Athena go to Athena w.com okay got my plugin there
也在其他允许建房的州发生,而人们正在离开那些不允许建房的州。所以这就是我表现最差的资产交易。企业软件本来是我另一个选项,但我要说点略微不同的。2025 年最期待的趋势。去年 Chamath,恭喜你,你去年说比特币会首次突破 10 万美元,你说中了,半场投篮命中,漂亮。我记得你做那个预测的时候,它大概交易在……感觉就是个上篮,感觉对你来说就是个上篮,涨了 112%。Friedberg,你说的是 AI 驱动的预测模型,在制药和生物工程领域的发现。这个预测表现如何?确实有很多融资,而且 Demis 拿了诺贝尔奖。这不就成了。我选的是效率,以AI 在劳动力和外包方面进展的形式,我们投资 Athena 之后我看到了一大堆这样的案例,去 Athena 官网看看。好,我的植入广告打完了。
便签引用
1:30:59
let's go for most anticipated 2025 Trend what do he got Chim off I think that there are a handful of not to overuse an overuse term but canaries in the coal mine for the end of the deep State and I would like to point to one which is this obscure thing called the supplemental loss ratio and essentially what it is is a mechanism that the banks can use to include or not include treasuries and how they calculate reserves etc etc now why is this an interesting thing it's not interesting for many people it's very Arcane but if we are unable to manage the debt situation in 2025 I think what you're going going to see is maneuvering at the edges of these Arcane regulations that effectively Kick the Can down the road and there is a chance that that could happen to help the banks because you know we believe that we meaning collectively America that maybe Doge won't be as effective as it needs to be that there's still going to need to be you know 1020 trillion dollars of debt issuance to refinance
我们来说 2025 年最期待的趋势。Chamath 你说什么?我认为有那么几个——不是要滥用一个被用滥的词——深层政府(deep state)终结的“煤矿里的金丝雀”。我想指出一个,就是这个很冷门的东西,叫补充杠杆率(supplementary leverage ratio)。本质上它是一种机制,让银行可以在计算准备金等等的时候,把国债计入或者不计入。那这为什么有意思呢?对很多人来说并不有意思,它非常晦涩。但如果我们在 2025 年没能管住债务局面,我认为你会看到的是在这些晦涩法规的边缘做腾挪,实际效果就是把问题往后拖。而且有可能真的会这么干,来帮银行一把。因为我们——我说的“我们”是指整个美国——相信也许 DOGE 不会像它需要的那样有效,仍然会需要发行 10 到 20 万亿美元的债务,来为今年到期的 10 万亿再融资,并填补未来几年的窟窿。我说这一切的重点是,我们
便签引用
1:32:21
the 10 that's maturing this year and to plug holes in the coming years my point in all of this is we don't need to understand the details except that if we don't move the goalpost here it is a great sign that the folks that are running the show are the ones that we all elected okay so that is my most anticipated Trend small Arcane regulatory changes that allow us to Kick the Can down the road stop in its tracks this is an example okay freeberg you got a most anticipated trend for 2025 my most anticipated trend is around the announcement of buildout of nuclear power in the United States in 2025 as a function of deregulation and some new technologies I do think that this new government is going to be U much more accommodating and it's going to as I've said in the past I think it's a necessity just because of the the the rate at which we have to do power build outs to meet kind of competitive demand against China the United States is going to need to add more power electricity production capacity than we can scale up
不需要理解这些细节,只需要知道:如果我们不移动球门柱,那就是一个很好的迹象,说明真正在掌舵的是我们大家选出来的那些人。好,这就是我最期待的趋势:那些让我们能够继续把问题往后拖的小而晦涩的监管改动,戛然而止。这就是个例子。好,Friedberg,你 2025 年最期待的趋势是什么?我最期待的趋势是关于 2025 年美国宣布大规模建设核电,这是放松管制和一些新技术带来的结果。我确实认为这届新政府会宽松得多,而且正如我过去说过的,我认为这是必需品,仅仅因为我们必须建设电力的速度,要跟上对标中国的竞争性需求。美国需要新增的发电产能,超过我们用任何其他
便签引用
1:33:34
with any other renewable source so I do believe that nuclear is an inevitability I think that the deregulation will happen in 25 and I do know a lot of very smart people who are actually starting nuclear power companies and have left very good jobs to go and do this in anticipation of this happening in 25 so I'm very bullish that's a leading indicator for sure when smart people do something with their time that's a great indicator Gavin what do you got uh I think Ai and Nick I sent you a chart I don't know if you can Flash it up but um I think AI is going to make more progress per quarter in 2025 than it did per year in 23 and 24 and the reason is just with um with 01 and 03 this is rcgi it's designed you we keep we keep changing the goalposts for the touring test and AGI and I'm sure we're going to change them again as we're going to blow through this and I just think what has happened is we were scaling around around one on one axis which was pre-training and then we started scaling around in Fritz time compute and it's
可再生能源所能扩张的规模。所以我确实相信核电是必然。我认为放松管制会在 25 年发生。而且我确实认识很多非常聪明的人,他们真的在创办核电公司,辞掉了很好的工作去做这件事,就是预期这会在 25 年发生。所以我非常看好。这肯定是个先行指标——当聪明人把时间投到某件事上,那就是很好的指标。Gavin,你选什么?呃,我觉得是 AI。Nick,我发给你一张图,不知道你能不能放出来。但我认为 AI 在 2025 年每个季度的进展,会超过 23 年和 24 年每一整年的进展。原因就是有了 o1 和 o3 这些。这就是递归自我改进,它是被设计出来的。我们一直在改图灵测试和 AGI 的球门柱,我敢肯定我们还会再改一次,因为我们要冲破这一关了。我觉得发生的事情是:我们原本只在一个轴上做规模扩张,也就是预训练,然后我们开始在推理时计算(inference-time compute)上扩张。而现在非常清楚,我们又加上了第三个性能扩张轴,
便签引用
1:34:44
very clear that we have now added a third axis of scaling performance and that is reasoning and what this is is these models the internet is composed of answers people giving answers and what the models really benefit from is kind of the internal monologue of somebody getting to that answer and this is you know in in AI terms they call it a reasoning trace and it's one reason you know 18 months ago everybody was like oh wow the more code you train a model on the better it does and all sorts of things that have seemingly nothing to do with code but all code you see kind of the reasoning the internal monologue the thought process the step by step and so what's happening is you're using models to generate synthetic data that contains these reasoning traces so you ask a model you know solve this problem it has to be a problem that is functionally verifiable that has an answer or we know the answer and we say show your work and we have it do that many many different ways and times and then you pick the best ones
那就是推理(reasoning)。这是什么意思呢?互联网是由答案组成的,人们给出答案。而模型真正受益的,是某个人得出那个答案过程中的内心独白。用 AI 的术语说,这叫“推理轨迹”(reasoning trace)。这也是为什么 18 个月前大家都在说:哇,你用越多代码去训练一个模型,它在各种看起来跟代码毫无关系的事情上表现就越好。但所有代码里你都能看到那种推理、内心独白、思考过程、一步一步的推导。所以现在发生的是,你用模型去生成包含这些推理轨迹的合成数据。你让模型解一道题——必须是一道功能上可验证的题,有答案,或者我们知道答案——我们说“把过程写出来”,然后让它用很多很多不同的方式做很多遍,接着你挑出最好的那些,把它们喂回
便签引用
1:35:48
and you kind of feed those back into the model you apply some reinforcement learning to it and so now you're scaling along three axes that are multiplicative with each other and I just I you know a guy on the Google team said tweeted maybe five days ago it's going to be a straight shot to ASI artificial super intelligence and I think that might be right Ilia gave a talk at NPS Ilia Suk um one of kind of the original pioneers of this field and he said something that I thought was scary and he said these models that reason are inherently unpredictable so the best reasoning models in the world today are the ones that play games the kind of alphago alpha zero style models and they are constantly making unpredictable moves that no human Grandmaster ever could have come up with and now these models are going to be making similarly unpredictable leaps in all sorts of domains which you know hopefully will be awesome but you know might not be yeah well they're going to go around corners that people might not have considered
模型,再施加一些强化学习。于是现在你是在三个轴上扩张,而这三个轴是相乘的关系。我……你知道,谷歌团队有个人大概五天前发推说,这将是一路直奔 ASI,也就是人工超级智能。我觉得这可能是对的。Ilya 在 NeurIPS 上做了个演讲,Ilya Sutskever,这个领域最初的开拓者之一。他说了一句我觉得挺吓人的话,他说这些会推理的模型本质上是不可预测的。今天世界上最好的推理模型,是那些下棋玩游戏的模型,就是 AlphaGo、AlphaZero 那一类,它们不断走出人类特级大师根本想不到的不可预测的棋。而现在这些模型将会在各种各样的领域里做出类似的、不可预测的飞跃。这个嘛,希望会很棒,但也可能不是。对,它们会拐到人们可能没考虑过的
便签引用
1:36:55
that are non-intuitive and then just to Circle back and do a little call back here remember I asked which banks would have the biggest chance of being insolvent uh or having Financial collapse or crisis tons of misspellings in there as I typed it while we were talking and you can see what deep M uh deep research did here it went and said here's what we're going to do we're going to research a bunch of websites and find a list of banks in the US find top banks by us by asset for each one of these Banks find their latest Financial savings for each of these Banks find their Capital adequacy ratios for each of these Banks find their loan loss you get the idea and then it went and it analyzed and then it created a report this took about 10 minutes and when you look it created this final report here where it gave a list chth of JP Morgan Bank of America City Group yada yada it did a an analysis of each one and at the end you can see all the different websites it pulled it was well over a hundred websites that it pulled that's
弯道上,走出反直觉的路。然后回过头来,我这里稍微回顾一下:记得我问过哪些银行最有可能资不抵债,或者发生金融崩溃、危机——我边聊边打字,里面一堆拼写错误——你可以看看 Deep……深度研究(deep research)在这里做了什么。它说:这是我们要做的,我们要去检索一堆网站,找出美国的银行名单,按资产找出美国的顶级银行,对其中每一家银行找出它们最新的财务数据,对每一家银行找出它们的资本充足率,对每一家银行找出它们的贷款损失……你懂的。然后它就去分析,接着生成了一份报告。这大概花了 10 分钟。你看,它在这里生成了最终报告,给出了一个名单,Chamath,摩根大通、美国银行、花旗集团等等。它对每一家都做了分析,最后你可以看到它调取的所有网站,远超一百个网站,这是实时数据,162 个网站。而且你得付费,
便签引用
1:37:51
inred live data 162 websites and you have to pay for this but it's only $20 a month and it's conclusion this analysis has provided a snapshot of the financial health of some of the largest US Banks while all banks face inherent risk City Group in Wells Fargo appear to be the highest risk of insolvency of financial collapse compared to JP mortgage Chase Bank of America Goldman Sachs City groups recent net loss lower c one ratio and high exposure I mean just try looking at this now I don't know how much of this is correct I'm no expert on this but it does seem like it's a pretty good start of where it's getting to and if you haven't used 1.5 Pro with deep research just go to gemini.com I'm not being paid to say this I just think it's the best product in the market and it's pretty darn impressive okay my most anticipated trend for 2025 was alluded to in an earlier prediction I think by Gavin himself mine is that uh exits and DPI shower down uh and we'll have this incredible uh distribution as
但每月只要 20 美元。它的结论是:这项分析提供了几家最大美国银行财务健康状况的一个快照。虽然所有银行都面临固有风险,但花旗集团和富国银行看起来是资不抵债或金融崩溃风险最高的,相比摩根大通、美国银行、高盛。花旗集团近期的净亏损、较低的一级资本比率和高敞口……我是说,你现在就试着看看这个。我不知道这里面有多少是正确的,我不是这方面的专家,但看起来确实是个相当不错的起点。如果你还没用过带深度研究的 1.5 Pro,就去 gemini.com 试试。我说这个没拿钱,我就是觉得这是市场上最好的产品,而且相当惊艳。好,我 2025 年最期待的趋势,在前面某条预测里已经被提到了,我想是 Gavin 自己提的。我的答案是:退出和 DPI 会像下雨一样倾泻而下,我们会迎来惊人的分配,
便签引用
09最期待媒体与 Polymarket 盘口
1:38:51
the wrath of lenina con ends and m&a and and IPOs will surge that is my prediction my most anticipated Trend okay let's go to most anticipated media I had two that I was working back and forth from Jame James guns DC Universe with Superman coming out this year I think it's going to be amazing Andor season 2 could be amazing I predict they're going to do a clone wars live action series if you don't know the Clone Wars watch it with your kids it's amazing animated series that takes place during the prequels but I went with a a little of a weird Choice here I think my most anticipated media is seeing what happens with Legacy Media Outlets owned by billionaires Andor people who no longer want to pick aide Washington Post CNN and LA Times specifically are steering towards the middle and trying to get back to Classic journalism the editors are revolting Karen swisser is upset and uh they're adding some rightwing voices it is going to be popcorn time for everybody so enjoy whatever you want to enjoy Star Wars I
因为 Lina Khan 的怒火结束了,并购和 IPO 会激增。这就是我的预测,我最期待的趋势。好,我们来说最期待的媒体……我一直在两个之间摇摆——詹姆斯·古恩的 DC 宇宙,今年《超人》要上映,我觉得会很棒;《安多》第二季也可能很精彩。我预测他们会做一部《克隆人战争》真人剧集。如果你没看过《克隆人战争》,一定要和孩子一起看,这部动画剧集太棒了,故事发生在前传三部曲期间。不过我选了一个有点奇怪的答案。我觉得我最期待的"媒体",是看那些被亿万富翁收购的传统媒体机构会发生什么,还有那些不想再选边站的人。《华盛顿邮报》、CNN 和《洛杉矶时报》正在往中间靠,试图回归经典新闻业。编辑们在造反,凯拉·斯威舍很不爽,而且他们还在加入一些右翼声音。这将是全民吃瓜时刻。所以你想看什么就看什么吧——《星球大战》,我给你准备了《安多》第二季;也有《超人》;
便签引用
1:39:55
got a little Star Wars for you with Andro season 2 I got a little Superman or you can watch the chaos in the editorial newsrooms at Washington Post and LA Times what do you got from most anticipated media chath it is the enormity of the files that are going to get Declassified and released by the Trump Administration I think it's going to be unbelievably interesting salacious useful good Earnest all all of the above so the JFK files the Epstein FES Diddy files the moon landing who knows what they find across all of these other Fringe quote unquote conspiracy theories that may turn out actually to have some shred of truth but all of that let's call it content for lack of a better word that gets released in 2025 by the Trump Administration I think will be incredibly interesting what do you got Dave media any anticipated media for you what do you got I'm into AI video games the cost of production comes way down when you use Ai and you can have Dynamic story lines you can have new gameplay Concepts things that don't exist today I
或者你可以看《华盛顿邮报》和《洛杉矶时报》编辑部的混乱大戏。你最期待的媒体内容是什么,Chamath?是特朗普政府将要解密和公开的那一大批档案。我觉得那会极其有意思、极其劲爆、极其有用、极其真诚——以上皆是。肯尼迪遇刺档案、爱泼斯坦档案、Diddy 档案、登月档案,谁知道他们会在这些所谓的"边缘"阴谋论里发现什么,这些说法里可能真的有几分真相。总之,姑且叫它"内容"吧,找不到更好的词——这些在 2025 年由特朗普政府公开的东西,我觉得会非常有意思。你呢,Dave?有什么期待的媒体内容吗?我看好 AI 电子游戏。用 AI 之后制作成本大幅下降,而且你可以有动态的故事线,可以有全新的玩法概念,那些今天还不存在的东西。我认为开发公司里通常具备的创意人才
便签引用
1:41:05
think that the creative Talent Plus the technical talent that you typically find at development houses can be kind of Unleashed with tools that have come to come to Market recently we've seen a lot of the generative video stuff but there's also ways structurally that video games can kind of be rebuilt where the video game engine can run kind of locally it can be generating parameters that can then use an existing rendering engine so you can have entirely new story lines and entirely new kind of plot sequences so I do think that there's going to be a rewrite of video games in the video game industry with the V variety of AI kind of capabilities that are hitting market now and uh it's going to be incredible it's going to be incredible entertainment people are GNA people that don't play video games are going to find stuff that they're going to love Gavin most anticipating media 2025 unquestioned 1923 season 2 I'm normally a science fiction kind of fantasy or you spy kind of guy when it comes to TV but
加上技术人才,可以借助最近上市的工具被彻底释放出来。我们已经看到很多生成式视频的东西,但从结构上讲,电子游戏还有别的重构方式——游戏引擎可以在本地运行,生成参数,然后交给现有的渲染引擎去处理。所以你可以有全新的故事线、全新的情节走向。所以我确实认为,随着各种 AI 能力涌入市场,整个电子游戏行业会被重写,而且会非常了不起,会是非常了不起的娱乐体验。那些本来不玩游戏的人也会找到他们会爱上的东西。Gavin,2025 年最期待的媒体?毫无疑问是《1923》第二季。我看电视一般偏爱科幻、奇幻或者间谍类,但《1883》和《1923》是头两部
便签引用
1:42:01
1883 and 1923 were the first TV shows that kind of hit me the way Game of Thrones did and I'm crazy excited for that to come out fantastic are you watching land man I'm enjoying landman with I haven't watched it yet I'm excited to try it it's quite fun quite fun turns out this guy Taylor sharidan he's I mean it's incredible he's in the zone I highly Rec Day of the Jackal I know chamath also loved Day of the Jackal oh that was really good that was really was a good that was a good J Sugg for stre yeah I'm Jackal sakario yeah absolutely what was our predictions for for this last year last year shth predicted Mr Beast freeberg AI generated news I predicted Gladiator 2 and the three body problem Gladiator 2 was mid but okay three body problem I didn't finish so I guess mid Jimmy show on Amazon Prime the Mr Beast thing was the top unscripted drama in 140 of 180 countries wow incredible all right we're going to do our prediction markets here I'm going to put up a prediction Market which is based upon immigration
给我带来当年《权力的游戏》那种感觉的剧。我超级期待它上线。太棒了。你在看《山河血脉》吗?我挺喜欢《山河血脉》的。我还没看,很期待去试试。挺好看的,挺好看的。这个泰勒·谢里丹,我是说,太厉害了,他现在状态正佳。我强烈推荐《豺狼的日子》,我知道 Chamath 也很喜欢《豺狼的日子》。哦,那个真不错,那真的是——那是个好推荐。是的,《豺狼》,还有《边境杀手》,绝对的。我们去年的预测是什么来着?去年 Chamath 预测的是 Mr Beast,Freeberg 预测 AI 生成的新闻,我预测的是《角斗士 2》和《三体》。《角斗士 2》挺一般的,还行吧;《三体》我没看完,所以我猜也挺一般。Mr Beast 在亚马逊 Prime 上的节目,在 180 个国家中有 140 个是排名第一的非剧本类节目。哇,太厉害了。好,我们来说说预测市场。我要开一个基于移民的预测
便签引用
1:43:11
and the promise that Trump made to have 15 million people be deported from the country I think in the first year I'm going to set the over under at 5% of that stated number which is 750,000 so after one year in office will Trump have deported 750,000 less or more that's my first prediction how are you going to measure that you have to there has to be like a a source of truth on these things right well do it based on the White House's reporting of deportations yeah it's reasonable and we'll check with the poly Market team if there's a way to do that prediction Market obviously but I think that was the biggest issue of the election so I'm putting it up there with just 5% in the first year yeah it is interesting Obama deported I think more than two million people it's like people have forgotten but he was actually really really tough on the border yeah chamat do you have a prediction Market you want to put up the magate representation in the S&P 500 shrinks below 30% oh that's a good one that's a
市场——特朗普承诺要把 1500 万人驱逐出境。我觉得第一年,我要把盘口设在那个数字的 5%,也就是 75 万人。所以上任一年之后,特朗普驱逐的人数会低于还是高于 75 万?这是我的第一个预测。你打算怎么衡量?你得有一个——这些事得有个权威数据源,对吧?就用白宫公布的驱逐数据吧。嗯,这合理。我们会问问 Polymarket 团队有没有办法做这个预测市场。但我觉得这是这次大选最大的议题,所以我把它拿出来,第一年只算 5%。是的,有意思的是,奥巴马驱逐了我记得两百多万人。人们好像都忘了,但他其实在边境问题上非常非常强硬。是啊。Chamath,你有想开的预测市场吗?"七巨头"在标普 500 中的占比跌破 30%。哦,这个不错,这个不错。也就是说其他股票会出现分化,集中度下降。
便签引用
1:44:14
good one so dispersion would happen to the other stocks less concentration that's a good one so we'll finalize these you'll be able to I wanted to kind of play with was Microsoft AWS and Google Cloud Revenue growth who's going to win in growth in 2025 I don't know how closely you guys crack but um I get a sense that Google is kind of accelerating ahead Gavin I don't know how much you've SP I mean Google is I wouldn't say they're a lot smaller which makes it easier for them to grow faster yeah [Music] um we talk about yeah I mean I I guess um who has the largest gain dollar dollar gain in Revenue in Cloud Revenue in 2025 is it Google Amazon or Microsoft well it probably won't be Google because if they had the largest dollar gain it would mean they were growing at at insane rates but I think Azure versus AWS will be interesting oh I know the other one I want you do which was the national debt the national debt has grown about two trillion per year in each of the previous two presidential administrations right so
这个不错。那我们把这些敲定。我想玩的一个是微软、AWS 和谷歌云的营收增长——2025 年谁的增长会胜出?我不知道你们盯得多紧,但我感觉谷歌有点在加速追赶。Gavin,我不知道你——我是说谷歌,我不会说他们小很多,但体量小确实让它更容易增长得更快。是啊。嗯,我们说的是——我想我们可以说,谁在 2025 年的云业务营收里实现最大的绝对美元增量:是谷歌、亚马逊还是微软?那大概率不会是谷歌,因为如果他们拿下最大的美元增量,那意味着他们的增速高得离谱。不过我觉得 Azure 对 AWS 会很有看头。哦,我想到另一个我想做的,是国债。过去两届总统任期里,国债每年大约增长两万亿美元,对吧?所以过去八年一共涨了
便签引用
1:45:29
over the last eight years we've gone up over 16 trillion two trillion per year average so I'm going to set it at and Trump said he would not increase the national debt I'll just set it at the national debt increases 1 trillion in the next year over under is that a good prediction Market or is it should be two trillion um yeah so what I would do is I would set the Federal debt for December the US Treasury Market report on federal debt in 2025 and I would set it at 30 37 trillion and so you basically the US Treasury um market report 20 December 2025 above or below 37 trillion it's going to be above I mean let's just call it 38 trillion then yeah you want to basically in order to make this when you set a line Dave it's got to be one where people take either side of it it can't be everybody takes we'll we'll do the US Treasury Market market report December 2025 federal debt above 38 trillion or below 38 trillion perfect which would be how much added in the year about one and a half yeah I love that I love that
超过 16 万亿,平均每年两万亿。所以我打算设成——特朗普说他不会增加国债,那我就设成明年国债增加 1 万亿,赌高于还是低于。这个预测市场好吗,还是应该设成两万亿?嗯,我的做法是,我会用美国财政部 2025 年 12 月的联邦债务报告,然后把线设在 37 万亿。所以基本上是:美国财政部 2025 年 12 月的市场报告,高于还是低于 37 万亿?那肯定会高于。那我们就设成 38 万亿吧。对,你要——Dave,设盘口的时候,得设在一个两边都有人愿意下注的位置,不能所有人都押一边。那我们就用美国财政部 2025 年 12 月的报告,联邦债务高于 38 万亿还是
便签引用
10UFO、无人机与金字塔闲谈
1:46:34
that's a perfect one we'll see if he can make any he can control the spending or if it'll just be the same Gavin what's your sense on federal debt next year I think like most things it'll take like immigration will'll take up time to get going yeah it just it's going to take a little bit of time but I think they will make progress by the way at some point if I'm here on the Allin again I want to talk about UFOs and the drones let's talk about it now do it right now are you going full conspiracy theorist are you are you trying to get your seat here by being a little more Alex Jones do you believe those drones were aliens what's going on so I don't know and I think it's very clear if you look at all the um you know statements from the New Jersey Mayors and governors you know who've met with the police the FBI the defense department that they don't know now you know Trump said someone in the government knows what they are he also said he was not going to go to Bedminster anytime soon his Resort there
低于 38 万亿。完美。那这一年增加多少?大概一万五千亿。对。我喜欢这个,我喜欢这个,这个很完美。我们看看他能不能控制住支出,还是说一切照旧。Gavin,你对明年的联邦债务怎么看?我觉得跟大多数事情一样,就像移民问题,需要时间才能启动起来。是啊,就是需要一点时间,但我觉得他们会取得一些进展。对了,如果我以后还有机会上《All-In》,我想聊聊 UFO 和那些无人机。现在就聊吧。现在就来。你这是要全面转向阴谋论了吗?你是想靠扮演亚历克斯·琼斯来保住这个位置吗?你相信那些无人机是外星人吗?到底怎么回事?我不知道。而且我觉得很明显,你看所有那些新泽西市长和州长的表态,他们见过警方、FBI、国防部,他们自己也不知道。特朗普说政府里有人知道那是什么,
便签引用
1:47:32
I just think over the last I'd say eight years every 18 months there's a big story in the New York Times the Washington posts the New Yorker the Atlantic very credible media sources with dozens of interviews with fighter pilots and commercial pilots talking about seeing things with Advanced sensors that made no sense to them and there's kind of a taboo that went away there was an article in the New Yorker where they quoted from a bunch of people who were at you know kind of the the Skunk Works Laboratories in the 1950s and said they absolutely saw extraterrestrial materials that have been recovered from a crash it's one reason the US made such big leaps in Material Science and um you know there was Mona thing where essentially a lot of you know astronomist said hey this is this is clearly a UFO of some sort including the head of astronomy at Harvard and then just the conspiracy theory I would have is if if if if if these are actually UFOs and not you know government drones either from the
他还说他近期不会去贝德明斯特他那个度假村。我就是觉得,过去我想说八年里,每隔 18 个月就有一篇大报道出现在《纽约时报》、《华盛顿邮报》、《纽约客》、《大西洋月刊》这些非常可信的媒体上,采访了几十位战斗机飞行员和民航飞行员,讲他们用先进传感器看到一些完全无法解释的东西。而且那种禁忌感消失了。《纽约客》有篇文章引用了一群上世纪五十年代在臭鼬工厂实验室工作过的人的说法,他们说自己确实见过从坠毁现场回收的外星材料,这也是美国在材料科学上取得那么大飞跃的原因之一。还有"奥陌陌"那件事,本质上是很多天文学家说,嘿,这明显是某种不明飞行物,
便签引用
1:48:48
United States or China you know it seems like the most likely explanation in New Jersey is at some sort of a drill and obviously once this hysteria gets going people um you know misidentify commercial airplanes as drones and like why would UFOs have blinking green and red lights but it is just interesting that there was you know a big concentration of these reports as we were scaling into nuclear technology 1945 to 1960 and then now that AI is getting going which is the next kind of technological phase shift for Humanity there's another big b of reports you can come back come back next week you you won congratulations we have our winner for all in Idol Gavin is now the fourth best I love this kind of conspiracy but I mean come on I we how do we make a poly Market out of this freed I don't know we are there are there documents that you think Trump will release about the past in terms if UFOs there must be just an entire spectrum of stuff that that could be subject to Foya if we went after it
包括哈佛天文系的负责人。至于阴谋论层面,我会想:如果,如果这些真的是 UFO,而不是美国或中国的政府无人机——在新泽西,最可能的解释似乎是某种演习。而且显然一旦这种歇斯底里开始了,人们就会把民航飞机误认成无人机。还有,UFO 为什么会有一闪一闪的红绿灯呢?但有意思的是,在我们扩张核技术的那段时期,也就是 1945 到 1960 年,这类报告出现了一个大高峰,然后现在 AI 起飞了,这是人类下一个技术相变阶段,于是又出现了一波报告高峰。你可以下周再来。你赢了,恭喜,我们的《All-In》选秀冠军诞生了,Gavin 现在是第四位好兄弟。我爱这种阴谋论,但拜托——我们怎么把这个做成 Polymarket 市场呢,Freeberg?我不知道。有没有你觉得
便签引用
1:50:00
no yeah I mean I think it depends how deeply it's classified and you know I'm sure if the government doesn't want to give it up um they won't give it up but you know there's all the you know there's all but you think but look let's be honest you think that there are docks like there there's documentation that the US government has that there are UFOs that there have been we just can't explain it yeah and you know they they call it uaps now to make it you know not flying saucers but it just feels like something F what percentage chance do you put on we have actual knowledge on a percentage basis what percentage chance you put on the government US government is sitting on knowledge of extral 20% at least 20% from chath what do you say gav I would say 25 I would take the over on the 20 okay great so you think there's a a non a significant small chance that this is the case free ber sultant of science what chance do you think there is this government sitting on some extraterrestrial life or proof of it
特朗普会公开的关于 UFO 往事的文件?肯定有一整片领域的东西,如果我们去申请的话是可以走《信息自由法》的。不,是的,我觉得这要看它被列为多高的密级。而且我相信如果政府不想交出来,他们就不会交。但你知道,还有所有那些——还有所有那些——但你是不是觉得——说实话,你是不是觉得美国政府手里有文件,有文档,证明存在 UFO,证明确实出现过我们无法解释的东西?是的。而且你知道,他们现在管这叫 UAP,好让它听起来不像"飞碟"。但就是感觉有点什么。你觉得有多大概率——我们真的掌握了实证,按百分比算,你觉得美国政府握有外星存在证据的概率是多少?20%。Chamath 说至少 20%。你说呢,Gav?我会说 25%,我押高于 20% 那边。
便签引用
1:51:02
it's a longer conversation I don't have time for this but I I gotta be honest I don't um I I don't generally align with the idea that like our very narrow range of like understanding of technology and biology kind of is what visits us or would visit us I think that there's an extension of information gathering that doesn't require moving physical biological life forms from one part of the galaxy to another so I think that the whole premise of like UFOs moving bodies around is rooted in the current state of Technology of humanity which is the basis of Arthur Clark's treatment of 2001 A Space Odyssey not which was done before the screenplay which was done before the book and the treatment which you can buy I think highlights this the best which is eventually every civilization reaches a sufficiently advanced level of technology that you no longer need to physically move the bodies of the the biological organisms around that you simply are gathering information and affecting information because once you
好的,那你们都认为这事有一个不小但不大的可能性。Freeberg,科学界的代表,你觉得政府手里握有外星生命或其证据的概率有多大?这是个很长的话题,我现在没时间展开,但老实说,我不——我一般不认同这种想法:认为我们对技术和生物学极其狭隘的理解范围内的东西,就是来拜访我们、或者会来拜访我们的东西。我觉得存在一种信息采集的延伸方式,不需要把物理的生物体从银河系的一端搬到另一端。所以我觉得"UFO 搬运躯体"这整个前提,是建立在人类当前技术水平之上的,这也是亚瑟·克拉克那份《2001 太空漫游》treatment 的基础——那份 treatment 写在剧本之前,剧本又写在小说之前,那份 treatment 你现在能买到,我觉得它把这一点讲得最透:最终每个文明都会达到一个足够先进的
便签引用
1:52:06
have the ability to convert any molecule into any other molecule and you have access to sufficient energy which you will eventually evolve to be able to do you can basically turn your local part of the universe into anything you want it to turn into and then you're simply gathering information from all over the rest of the universe so I would argue that there doesn't really make a lot of basic sense in why alien bodies would want to move around the Galaxy I'm just saying it right now if Trump releases information on extraterrestrial life I voting for him to have a third ter I'm putting it out there I'm going full Maga he gets a third term for me with changing the Constitution this has been another amazing episode of the all in one last question for Gavin one oh well forg okay do you think extraterrestrial life built the pyramids is that the most plausible explanation to the accuracy I don't know it is it is so strange the way there are you know in so many I would say cultures all over the world
技术水平,让你不再需要把生物体的躯体搬来搬去,你只是在采集信息、影响信息。因为一旦你有能力把任何分子转换成任何别的分子,并且拥有足够的能量——而你最终会进化出这种能力——你基本上可以把你所在的那片宇宙变成任何你想要的样子,然后你只是从宇宙其他地方采集信息。所以我认为,外星躯体为什么要在银河系里到处跑,这在根本上说不太通。我现在就把话放这儿:如果特朗普公开关于外星生命的信息,我投票让他干第三任。我把话放这儿,我全面转向 MAGA。他可以拿
便签引用
1:53:07
pre you know we are closer in time to Cleopatra than she was to the time of the Great Pyramids like there was a long time ago but it's just weird that in so many cultures all over the world there are these depictions of what looked like astronauts and then there's all these Renaissance Era paintings that clearly show what we would call UFOs in the Skies over cities so I don't know I just think it's I'm with you I love conspiracy Gavin this has been if you're if you're Jason if you're a questioning person these things create these weird un you know open Loops that it's like hard simulation so fully I'm fully vested in this anyway Free's got to drive back free BG's driving home I think the pyramids the best most credible thing I saw on the pyramids was that they used water and they raised water in channels so they put them on wood slats or something and then they would raise like um in like the um Panama Canal they would do canals and Float them up and then place them and then release the water which you know
第三任期,修宪都行。这又是一期精彩的《All-In》。最后问 Gavin 一个问题。哦,好吧,行。你觉得金字塔是外星生命建的吗?考虑到那种精度,这是最合理的解释吗?我不知道。太奇怪了,世界各地那么多文化里——要知道我们在时间上离克娄巴特拉,比她离大金字塔的年代还要近,那确实是很久很久以前了。但就是很奇怪,世界各地那么多文化里,都有这些看起来像宇航员的形象描绘,还有那么多文艺复兴时期的画作,明明白白画着我们今天会称之为 UFO 的东西悬在城市上空。所以我不知道,我就是觉得——我跟你想的一样,我爱阴谋论。Gavin,这——如果你,Jason,如果你是个爱追问的人,这些东西会形成
便签引用
1:54:15
sounds great to me this has been another amazing episode of the Allin podcast and thank you to everybody let's have an amazing 2025 to everybody all kick let's kick ass 2025 let's take all we'll see you all next time byebye bye bye let your winners ride Rainman David and in said we open source it to the fans and they've just gone crazy with it love queen of [Music] besties [Music] are my dog taking your driveway man oh man myit will meet me we should all just get a room and just have one big huge orgy cuz they're all this useless it's like this like sexual tension that they just need to release
一些奇怪的、你知道的,闭不上的循环,就像是很难——模拟论。所以我完全信,我完全投入了。总之,Freeberg 得开车回去了,Freeberg 要开车回家了。关于金字塔,我看过最可信的说法是,他们用水,在渠道里把水位抬高,把石块放在木板之类的东西上,然后像巴拿马运河那样一级一级把水位提上去,把石块浮上去,再放到位,然后把水放掉。这听起来挺靠谱的。这又是一期精彩的《All-In》播客,谢谢大家。祝大家
便签引用
1:55:18
somehow we need to get mer [Music] I'm going all in
2025 年精彩万分。大家一起干翻 2025!下期再见。拜拜,拜拜。
便签引用
视频总结 · 一句话概括与核心要点

一句话概括

All-In 播客 2025 年预测特辑(嘉宾 Gavin Baker):四人围绕政治、商业、资产、趋势与媒体逐项押注,共识是"机器人/自主系统元年 + AI 推理算力爆发 + 财政紧缩与去监管",分歧集中在 Mag 7 涨跌、OpenAI 前景、中国科技股和银行系统性风险上。

核心要点

  • 政治赢家是"年轻化 + 财政保守派 + 中间派",输家是进步主义与普京。 Freeberg 指出特朗普内阁平均年龄 40–45 岁,拜登内阁近 60 岁,认为中期选举将涌现更年轻的候选人;Chamath 押财政保守派,逻辑是紧缩是削减福利前的最后一道防线,DOGE 会暴露联邦政府的浪费与欺诈;他同时预测加拿大特鲁多、德国、法国、英国的进步派政府将在 2025 年接连失势。Gavin 认为普京是最大输家:欧洲重新武装让美国得以把资源转向太平洋,习近平将开始与俄脱钩,特朗普对俄会比外界预期更强硬。
  • 2025 是"机器人之年",且大公司凭 AI 推理算力拉开与小企业的差距。 Freeberg 以中国宇树科技(Unitree)为例:Go2 四足机器人 1600 美元且带 API 和激光雷达,G1 人形机器人约 1.6 万美元("一辆二手 Prius 的价钱"),他的团队已下单用于试验农场巡检。Gavin 补充 FSD 已进入主流,并强调 o3 展示的"推理 + 测试时计算"意味着能花 100 万美元让 AI 思考 6 周的大企业将获得结构性优势,2025 年 GPU/加速器会像 2023 年那样再度短缺。
  • 美元稳定币将正面冲击 Visa/Mastercard 双寡头。 Chamath 的数据:2024 年二季度末稳定币交易量约 11 亿笔、金额 8.5 万亿美元,同期超过 Visa 交易额的两倍,且使用量已与加密币价波动脱钩;他估算若从全球经济中剔除 300 个基点的支付摩擦,仅美国就值约 1 万亿美元,预测稳定币规模到 2025 年底翻 4–5 倍。Gavin 反方观点:若稳定币组合取代美元成为储备货币,对世界或许好,对美国是灾难,因为本币借债能力让美国得以用通胀稀释债务。
  • 商业输家:政府服务商、Cost-Plus 国防承包商、咨询巨头,以及可能回撤数万亿美元的 Mag 7。 Gavin 警告政府收入占比超 35% 的公司在 DOGE 时代会被"真的核对账单";Freeberg 点名波音、洛马、雷神等"国防 1.0",受益者是 Palantir、Anduril;Chamath 补充 Cost-Plus 本质是"时间与材料 = 人力套利",800 美元垃圾桶之类的支出会被彻底清算,Wipro、TCS、埃森哲等也将受冲击。Chamath 另预测前 7–10 大公司在指数中占比逼近 40%,历史上此类集中度往往预示大回撤,哪怕仅跌 10% 也是数万亿美元的绝对损失。
  • Jason 双倍押注 OpenAI 见顶:估值 1570 亿美元是峰值,非营利转营利可能失败,最终沦为第四名。 理由是 Google 的 Gemini Deep Research(20 美元/月,能拆解问题、并发搜索上百个网页、10 分钟生成带引用的咨询级报告)远超对手;xAI 的 Colossus 在 45 天内立起 10 万块 GPU;微软已拿到源代码不再依赖 OpenAI;开发者不忠诚于任何一家模型,倾向多模型路由和开源。Gavin 提出更本质的判断:AI 最终赢家是算力成本最低者,从 Azure/AWS/Google 租算力必然背负加价,"全栈者胜"。反方 Freeberg 提醒 OpenAI 预计 2025 年营收 120 亿美元,是真实业务,但同时也有约 60 亿美元亏损。
  • 最大交易:传统车企大并购、Waymo 巨额交易、自动驾驶配送整合、并购潮全面回归。 Chamath 以本田-日产合并为信号,认为特斯拉的车、软件和 FSD 会让公开市场意识到传统 OEM"不可投资",大众、Stellantis 是"融化的冰山",唯一风险是政府以就业为由干预。Freeberg 给出 Waymo 数据:2023 年 8 月在旧金山开放时 Uber/Lyft 份额为 66%/34%,15 个月后 Waymo 拿到 22%,与 Lyft 持平,Uber 降至 55%。Jason 预测特斯拉/亚马逊/Uber/DoorDash/Waymo 之间会出现超级并购(特斯拉只需约 10% 市值即可买下 Uber),Gavin 补充 Lina Khan 时代结束后四年积压的并购需求将释放,且英特尔会有大动作。
  • 最佳资产分歧巨大:HBM 高带宽内存、CDS 信用违约保险、中国科技股、Mag 7 各执一词。 Gavin 押 HBM(SK 海力士、美光),因为它在 GPU 成本中占比超过台积电代工,且已连续两年售罄。Chamath 买 CDS 作为"92% 归零、2% 赚 100–1000 倍"的保险,配合他的逆向预测:美国政府+企业+房贷总债务约 70 万亿美元,5% 利率的美元利息负担相当于 25–30 年前 10% 利率,可能触发某家主流银行的准备金危机。Freeberg 押中国科技 ETF,理由是特朗普可能与中国达成大交易、中国电力成本极低(新批 1370 亿美元水电站)、阿里巴巴等估值极低;Gavin 因 15 年前 Longtop 财务造假而奉行"不碰中国",但承认逻辑成立。Jason 押 Mag 7,认为它们率先在内部用 AI 替代招聘,会带来市场无法理解的盈利扩张。
  • 最差资产共识:企业应用软件("软件工业复合体")。 Gavin 认为 2025 下半年是"Agent 之年",没有自有模型和算力的企业软件公司会输给实验室和云厂商;Chamath 称这类公司本质是"在 CRUD 数据库上包一层业务规则 + 高尔夫和牛排晚宴式销售",他的 8090 公司用 30 人工程团队干 300 人的活,能以 1/10 价格投标且仍高利润,ROI 计算器和传统销售模式全部失效;Gavin 点出范式差异:旧软件是让白领更高效,AI 公司是直接替换白领。
  • 逆向观点:AI 与去监管推动美国 GDP 增速超 5%,但同时社会主义思潮回潮,前沿实验室将停止发布顶级模型。 Gavin 预测未来四年内至少一年实际 GDP 增长超 5%(5–6% 增速下经济每 12 年翻倍,3% 需 24 年),并认为实验室会为防止知识蒸馏(如 DeepSeek 自称 GPT-4)只发布小模型。Freeberg 的逆向判断是主流认为大选击败了社会主义,但快速技术变革、联邦裁员、财富极化("首个万亿富翁")会像 1940 年代阿根廷庇隆上台时(GDP 增速 8%)那样催生社会主义运动;Gavin 补充在 AGI 前的过渡期"钱比任何时候都重要",因为测试时计算可以直接用钱买优势,AI 会短期放大不平等。Jason 观察到 VC 圈内六位数薪资的高技能人才也已难找工作。
  • AI 进展将从"按年"变为"按季",因为出现了第三个扩展轴。 Gavin 解释:预训练和推理时计算之外,推理能力(reasoning)成为新轴——用模型生成含"推理轨迹"的合成数据,针对可验证问题让模型多路"展示解题过程",挑选最优结果回灌并做强化学习,三轴相乘。他引用 Google 员工"直通 ASI"的说法,并引 Ilya Sutskever 在 NeurIPS 的警告:会推理的模型本质上不可预测,像 AlphaGo 一样会走出人类想不到的棋。

结论与值得注意的细节

  • Polymarket 合作与超级预测:本期与 Polymarket 合作上线"All-In 专区",拟设市场包括:特朗普首年遣返人数是否超过 75 万(承诺 1500 万的 5%);Mag 7 在标普 500 占比是否跌破 30%;2025 年 12 月美国联邦债务是否超过 38 万亿美元(意味年增约 1.5 万亿);OpenAI 能否在 2025 年完成营利转型;贝索斯是否出售《华盛顿邮报》。
  • 去年预测复盘:Chamath 去年押比特币破 10 万美元(涨 112%)和 Koch 家族政治机器失势("数十亿美元投入归零"),均命中;Freeberg 押垂直 SaaS 输、AI 云厂商赢(Google +36%、亚马逊 +46%)和 Reddit 类数据授权交易,命中;Freeberg 的铀 ETF 跌 1%,Jason 的 Uber 因 Robotaxi 预期从 +30% 回落到 +3%,DoorDash +74%。
  • 体育版权反转:Chamath 去年看空职业球队(NBA 沦为"三分与扣篮"、明星年年跳槽无宿敌),但 Gavin 指出 Google 买 NFL Sunday Ticket、亚马逊/Netflix 买 NFL 转播都非常满意,18 个月内可为每个观众动态生成广告,科技巨头会"买下所有体育版权",且 PE 进场后估值由 DCF 而非情感驱动。
  • 值得关注的趋势:Chamath 点名"补充杠杆率"(SLR)这类允许银行调整国债计入准备金方式的冷门规则——若 2025 年被悄悄修改以"踢皮球",说明深层政府仍在运作;Freeberg 看好核电去监管,指出许多聪明人已辞去好工作创办核电公司,是领先指标;Jason 预测退出与 DPI 大爆发。
  • 零星判断:Chamath 预测特朗普政府解密的 JFK、爱泼斯坦、Diddy 等档案将是 2025 年最"好看"的媒体内容;Gavin 对新泽西无人机事件持开放态度,Chamath 与 Gavin 分别给美国政府掌握外星证据 20% 和 25% 的概率,Freeberg 从物理学角度反驳"高级文明无需搬运生物体跨越星系";Jason 直言 Apple Intelligence "比 Copilot 还差",Pixel 上的 Gemini 完成任务只需苹果十分之一的时间。
核心句型 · 10
1. If A was the year of X, (I think) B will be the year of Y
“If 2024 was the year of kind of compute build out and the roll out of AI systems in software I think 2025 will be the year of the robot”
用对照结构给出年度判断:前半句是共识事实,后半句是预测。适合做趋势总结或演讲开头,仿写时确保 X 与 Y 同类。
2. You show me an incentive, I'll show you the outcome
“You show me an incentive I'll show you the outcome the alignment of Cost Plus is to just keep being inefficient”
芒格名言的口语化引用,用「给我看 A,我给你看 B」的平行结构说明因果必然性。可仿写为 You show me the data, I'll show you the decision。
3. I would take the other side on X / I'm taking the other side
“I would take the other side on pro sports teams it feels like that market is about to be institutionalized”
表达反对观点的礼貌说法,源自交易术语「做对手盘」。比 I disagree 更中性,适合讨论中提出相反判断。
4. X is a fancier way of saying Y
“Plus is a fancier way of saying time and materials and time and materials is a fancy way of saying human labor Arbitrage”
层层拆解术语,把光鲜说法还原为本质。连续使用两次形成递进,是揭穿包装、直指核心的犀利表达。
5. Let's call it what it is
“Reunify China and Taiwan or invade Taiwan let's call it what it is”
用于纠正委婉说法、直呼本质。放在两种表述之间,前面是官方说法,后面是直白说法,语气坚决。
6. Just in case anybody thinks I'm talking my book, ...
“Just in case anybody thinks I'm talking my book here I have exposure to all these almost all these companies”
投资圈用语,主动声明利益关联以增加可信度。仿写:Just in case anyone thinks I'm biased, I should say ...
7. N times out of 100, X happens; the remaining times, Y
“This is a pick that 92 times out of 100 goes to absolute zero okay and six out of the remaining eight times you make 10 extra money”
用具体频率拆分概率分布,比说 unlikely 更精确、更有说服力。适合描述高风险高回报的不对称赌注。
8. The only risk to X's prediction is Y
“The only risk to chat's prediction is just government intervention because they's such big employers”
先表示赞同,再点出唯一反例,是有建设性地补充他人观点的句式。仿写时 Y 应是具体、可识别的外部因素。
9. It's not something I think will happen, it's not something I want to happen, but ...
“This is not something I think will happen it's not something I want to happen but I do think that if you look back.”
双重否定铺垫后转折,用于提出敏感或悲观预测时降低攻击性。适合讨论风险对冲、危机情景。
10. We keep changing the goalposts for X
“We keep changing the goalposts for the touring test and AGI and I'm sure we're going to change them again”
「移动球门柱」喻事后改变评判标准。用 keep + V-ing 强调反复发生,常用于批评标准不断上调的现象。
词汇精讲 · 183 · 按出现顺序
grift /ɡrɪft/ n. 0:00
小骗局、捞钱套路(此处自嘲式用法)
crossover firm n. phr. 0:00
跨界基金:同时投资公开市场与私募市场
capital allocator n. phr. 0:00
资本配置者,决定钱投向哪里的人
let your winners ride phr. 1:04
让赢的仓位继续跑,不要过早止盈
tweaked /twiːkt/ v. 4:14
(口语)扭伤、拉伤
Darkhorse /ˈdɑːrkhɔːrs/ n. 5:10
黑马,出人意料的竞争者
swung the election phr. 6:09
左右了选举结果
austerity /ɔːˈsterəti/ n. 6:09
财政紧缩
entitlements /ɪnˈtaɪtlmənts/ n. 6:09
(美国)法定福利项目,如社保、医保
clamoring for phr. v. 7:22
大声疾呼、强烈要求
Old Guard n. phr. 7:22
老派、守旧势力
sea change n. phr. 8:30
根本性转变
client state n. phr. 9:27
附庸国,依附于大国的国家
rearm /ˌriːˈɑːrm/ v. 9:27
重新武装
bigly /ˈbɪɡli/ adv. 9:27
(特朗普式用语)大大地、惨重地
humiliating defeat n. phr. 10:42
屈辱性的失败
neocons /ˈniːoʊkɑːnz/ n. 10:42
新保守派,主张对外军事干预的美国右翼
come to a head phr. 10:42
(矛盾)达到顶点、摊牌
bellicosity /ˌbelɪˈkɑːsəti/ n. 11:44
好战姿态
at odds with phr. 11:44
与……相矛盾
aggressive posturing n. phr. 12:42
强硬作态、摆架势
full-throated /ˌfʊlˈθroʊtɪd/ adj. 12:42
毫无保留的、全力的
antidote /ˈæntidoʊt/ n. 12:42
解药、对策
deadlock /ˈdedlɑːk/ n. 12:42
僵局
stoking /ˈstoʊkɪŋ/ v. 13:58
煽动、助长(情绪)
repudiation /rɪˌpjuːdiˈeɪʃn/ n. 13:58
否定、拒斥
Ripple effects n. phr. 13:58
连锁反应、涟漪效应
vessel /ˈvesl/ n. 15:07
容器、载体(此处喻政党)
through that lens phr. 16:17
从那个角度看
inflection point n. phr. 17:29
拐点
payload /ˈpeɪloʊd/ n. 17:29
有效载荷,机器人可挂载的设备
all at once phr. 18:37
突然一下子
perimeter security n. phr. 19:28
周界安防
spitting distance n. phr. 19:28
近在咫尺,非常接近
mainstream adoption n. phr. 21:23
主流普及
compounding /kəmˈpaʊndɪŋ/ v. 21:23
复合增长
test time compute n. phr. 21:23
测试时计算:模型推理阶段投入的算力
uncoupled /ʌnˈkʌpld/ adj. 22:39
脱钩的、解除关联的
up and to the right phr. 22:39
(图表)一路向右上方增长
point of no return n. phr. 24:01
不可逆转点
duopoly /duˈɑːpəli/ n. 24:01
双寡头垄断
rails /reɪlz/ n. 24:01
(支付)轨道、底层通道
dragged /dræɡd/ v. 25:28
(俚)被公开痛批
basis points n. phr. 25:28
基点,万分之一
quizzical /ˈkwɪzɪkl/ adj. 25:28
令人费解的、疑惑的
to the tune of phr. 25:28
达到……之多(金额)
strangle hold n. phr. 26:39
扼制、垄断性控制
Reserve currency n. phr. 26:39
储备货币
no-brainer /ˌnoʊˈbreɪnər/ n. 27:46
无需思考的明智选择
in awe of phr. 28:53
对……肃然起敬
stood up phr. v. 28:53
(技术语)搭建并投入运行
cohort /ˈkoʊhɔːrt/ n. 28:53
一批、同类群体
blows anything else in the market out of the water phr. 29:58
把市场上其他产品远远甩开
step function n. phr. 29:58
阶跃式(提升),非渐进的飞跃
Bonkers /ˈbɑːŋkərz/ adj. 30:57
疯狂的、离谱的(此处为褒义)
knock people's socks off phr. 30:57
让人惊叹不已
bootstrapped /ˈbuːtstræpt/ adj. 31:57
(创业公司)靅自有资金、不靠外部融资的
came to pass phr. 31:57
(预言)应验、成为现实
institutionalized /ˌɪnstɪˈtuːʃənəlaɪzd/ v. 32:53
机构化,成为机构投资标的
viewership /ˈvjuːərʃɪp/ n. 32:53
收视人数、收视率
devolved into phr. v. 33:53
退化成
obscene /əbˈsiːn/ adj. 33:53
(金额)高得离谱的
trophy asset n. phr. 35:05
奖杯资产:为面子而非回报持有的资产
cap tables n. phr. 35:05
股权结构表
jawboning /ˈdʒɔːboʊnɪŋ/ v. 36:00
喋喋不休地说、斗嘴施压
Bush League adj. phr. 36:00
业余水准的、不专业的
mortified /ˈmɔːrtɪfaɪd/ adj. 36:00
羞愧难当的
back up the truck phr. 37:58
大手笔买入、整个打包
rounding error n. phr. 37:58
四舍五入误差,喻微不足道的数额
command the eyeballs phr. 37:58
吸引大量观众注意力
stuck their toe in the water phr. 39:00
试探性尝试
precarious /prɪˈkeriəs/ adj. 39:00
不稳的、有风险的
draw down n. phr. 40:07
(资产)回撤
foreshadowed /fɔːrˈʃædoʊd/ v. 40:07
预示
inoculate /ɪˈnɑːkjuleɪt/ v. 40:07
使免疫,喻规避风险
clunky /ˈklʌŋki/ adj. 41:23
笨重迟钝的
Cost Plus n. phr. 41:23
成本加成合同:按成本加固定利润结算
washed out phr. v. 42:22
被冲刷出局、被淘汰
human labor Arbitrage n. phr. 42:22
人力套利:利用地区间工资差赚差价
scrutinized /ˈskruːtənaɪzd/ v. 42:22
被严格审查
kick the can down the road phr. 43:25
把问题往后拖
book value n. phr. 43:25
账面价值
Munitions /mjuːˈnɪʃnz/ n. 43:25
军火弹药
headwinds /ˈhedwɪndz/ n. 44:25
逆风、不利因素
underappreciated /ˌʌndərəˈpriːʃieɪtɪd/ adj. 44:25
被低估的
Big Iron n. phr. 44:25
(行话)大型算力硬件
meme stock n. phr. 45:41
迷因股,靠社交炒作而非基本面的股票
beholden to phr. 45:41
受制于、依附于
whiffed /wɪft/ v. 46:38
(俚)完全没猜中、挥空
uninvestable /ˌʌnɪnˈvestəbl/ adj. 47:48
不值得投资的
melting icebergs n. phr. 47:48
融化的冰山,喻缓慢萎缩的业务
ilk /ɪlk/ n. 47:48
同类人(and his ilk:他那类人)
spread trade n. phr. 47:48
价差交易:同时做多一方做空另一方
protectionism /prəˈtekʃənɪzəm/ n. 48:57
贸易保护主义
national champions n. phr. 48:57
民族龙头企业
onshore manufacturing n. phr. 50:03
本土制造、制造业回流
pinup demand n. phr. 51:01
(应为 pent-up demand)积压的需求
markups /ˈmɑːrkʌps/ n. 51:01
加价、溢价
full stack n. phr. 51:01
全栈:从底层算力到应用全部自有
divested /daɪˈvestɪd/ v. 52:05
被剥离、被出售
spin out phr. v. 52:05
(业务)分拆独立
ginormous /dʒaɪˈnɔːrməs/ adj. 53:13
巨大无比的(gigantic+enormous)
wild card n. phr. 53:13
变数、不可预测因素
capex /ˈkæpeks/ n. 54:20
资本开支
monotonous /məˈnɑːtənəs/ adj. 55:20
单调乏味的
talking my book phr. 55:20
为自己的持仓说话、夹带私货
across the board phr. 56:11
全面地、无一例外
at stake phr. 56:11
利害攸关
sympatico /sɪmˈpɑːtɪkoʊ/ adj. 57:12
意见一致、合拍的
going out on a limb phr. 59:34
冒险说出大胆看法
Black Swan n. phr. 1:00:40
黑天鹅事件:极小概率、巨大影响
indebtedness /ɪnˈdetɪdnəs/ n. 1:00:40
负债总额
sensitize /ˈsensətaɪz/ v. 1:00:40
(金融)做敏感性分析
non-trivial /ˌnɑːnˈtrɪviəl/ adj. 1:02:12
不可忽视的、不小的
go vertical phr. 1:03:22
(曲线)垂直起飞、急剧上升
knowledge distillation n. phr. 1:04:38
知识蒸馏:用大模型输出训练小模型
party line n. phr. 1:04:38
官方说法、主流叙事
resounding /rɪˈzaʊndɪŋ/ adj. 1:04:38
响亮的、压倒性的
gain steam phr. 1:05:41
势头增强
rosy /ˈroʊzi/ adj. 1:07:43
乐观的、一片美好的
deprecate /ˈdeprəkeɪt/ v. 1:08:39
(技术语)弃用、砍掉
mandate /ˈmændeɪt/ n. 1:08:39
(选举给予的)授权、民意
decouple /diːˈkʌpl/ v. 1:09:39
脱钩
inflated away phr. v. 1:10:37
因通胀而膨胀到失去价值
School vouchers n. phr. 1:10:37
教育券:政府发给家庭自主择校的补贴
doubling and tripling down phr. 1:11:28
加倍、三倍下注,更坚定押注
standup guy n. phr. 1:11:28
正派可靠的人
midlife crisis n. phr. 1:12:30
中年危机
run it right into the ground phr. 1:12:30
把它彻底搞垮
cogs /kɑːɡz/ n. 1:13:35
COGS,销售成本
gets their act together phr. 1:13:35
振作起来、理顺自己
preface /ˈprefəs/ v. 1:14:53
先做说明、开场铺垫
maturities /məˈtʊrətiz/ n. 1:14:53
(债券)到期期限
tonnage /ˈtʌnɪdʒ/ n. 1:16:24
吨位,喻绝对数量之大
asymmetry /ˌeɪˈsɪmətri/ n. 1:16:24
不对称(风险收益)
dumped /dʌmpt/ v. 1:17:27
抛售
unfathomable /ʌnˈfæðəməbl/ adj. 1:18:32
难以想象的、深不可测的
throttle up and down phr. 1:18:32
加油门与踩刹车,喻放松或收紧
poised for phr. 1:19:23
蓄势待发
fudge /fʌdʒ/ v. 1:19:23
篡改、做假(数据)
talked a great game phr. 1:20:18
说得头头是道(但做不到)
out in the cold phr. 1:20:18
被冷落、被排除在外
unit economics n. phr. 1:21:26
单位经济效益
earnings expansion n. phr. 1:22:25
利润扩张
bloated /ˈbloʊtɪd/ adj. 1:24:40
臃肿的
heuristics /hjuˈrɪstɪks/ n. 1:24:40
启发式规则、经验法则
go to market n. phr. 1:24:40
市场进入与销售策略
railed against phr. v. 1:24:40
痛斥、猛烈抨击
order of magnitude n. phr. 1:26:00
一个数量级(十倍)
fissures /ˈfɪʃərz/ n. 1:26:00
裂缝
traction /ˈtrækʃn/ n. 1:26:00
(初创公司)市场进展、起势
cost structure n. phr. 1:27:06
成本结构
Revenge startups n. phr. 1:28:07
复仇式创业:离职后创办公司与老东家竞争
up against it phr. 1:29:05
处境艰难
layup /ˈleɪʌp/ n. 1:30:06
篮球上篮,喻轻而易举的事
canaries in the coal mine n. phr. 1:30:59
煤矿里的金丝雀,先行预警信号
Arcane /ɑːrˈkeɪn/ adj. 1:30:59
晦涩难懂的
move the goalpost phr. 1:32:21
移动球门柱,事后改变标准
accommodating /əˈkɑːmədeɪtɪŋ/ adj. 1:32:21
通融的、宽松的
inevitability /ɪnˌevɪtəˈbɪləti/ n. 1:33:34
必然性
bullish /ˈbʊlɪʃ/ adj. 1:33:34
看涨的、乐观的
leading indicator n. phr. 1:33:34
先行指标
internal monologue n. phr. 1:34:44
内心独白
reasoning trace n. phr. 1:34:44
推理轨迹:模型逐步思考的中间过程
synthetic data n. phr. 1:34:44
合成数据
reinforcement learning n. phr. 1:35:48
强化学习
multiplicative /ˌmʌltɪˈplɪkətɪv/ adj. 1:35:48
相乘的
straight shot n. phr. 1:35:48
一路直达
insolvent /ɪnˈsɑːlvənt/ adj. 1:36:55
资不抵债的
Capital adequacy ratios n. phr. 1:36:55
资本充足率
Declassified /ˌdiːˈklæsɪfaɪd/ v. 1:39:55
解密
salacious /səˈleɪʃəs/ adj. 1:39:55
劲爆猎奇的、有色情意味的
shred of truth n. phr. 1:39:55
一丝真相
in the zone phr. 1:42:01
状态正佳
unscripted /ʌnˈskrɪptɪd/ adj. 1:42:01
非剧本的、真人秀类的
over under n. phr. 1:43:11
(博彩)大小盘口线
source of truth n. phr. 1:43:11
权威数据源
dispersion /dɪˈspɜːrʒn/ n. 1:44:14
(股票)分化、收益离散
hysteria /hɪˈsteriə/ n. 1:48:48
集体歇斯底里
phase shift n. phr. 1:48:48
相变,喻根本性阶段跃迁
take the over phr. 1:50:00
(博彩)押高于盘口
premise /ˈpremɪs/ n. 1:51:02
前提
plausible /ˈplɔːzəbl/ adj. 1:52:06
合理可信的
精读便签
下载便签 手机:长按图片也可保存
← 上一期 · NO.201The Next Breakthroughs in AI with Yossi Matias | The Google Research Podcast 下一期 · NO.203 →Jared Isaacman: A New Era for NASA and American Space Exploration
苏菲周报 · THE WEEKLY 每周一封,
追问一个大问题。
苏菲拉底的每周来信,写这一周在追问的问题和看到的回应。
苏菲拉底
ASK THE BIG QUESTIONS · THINK DEEPLY · SEE THE WORLD DIFFERENTLY
苏菲拉底微信公众号二维码 微信公众号
© 2026 苏菲拉底 · 内容仅供学习 [email protected]