20VC with Harry Stebbings · 2026-05-09 · 双语整理

Why Tokens Will Outspend Salaries

Speechify CEO Cliff Weitzman 谈跑遍世界 100 个消费 CEO 学到的东西

"Don't even bother spending money on any platform that's not Meta until you reach $100,000 a month on Meta."
Cliff Weitzman 把 Speechify 做到 6000 万用户、100 万 5 星评价。他自费飞遍全球见了 100 个消费订阅 CEO,把他们的 playbook 全套学回来。这一集聊投流的反共识、AI 帮他诊断爸爸癌症、为什么 Speechify 从 Cursor 迁到 Claude Code,以及为什么"60 秒响应时间"是工程师能力最准的预测因子。

嘉宾 Cliff Weitzman(Speechify Founder & CEO)· 视频 1:57:54 · 原片 YouTube
TL;DR · 速读

跑遍世界 100 个消费 CEO 后,Cliff 总结的反共识

  1. 买更多彩票:申 26 个大学不是 8 个

    "It's a lottery. How do you win a lottery? You buy more tickets than everybody else."

    "For kids with perfect SAT scores and perfect grades, you have about a one in 10 chance of getting in. I wrote 48 drafts of my main essay and applied to 26 schools."

    Cliff 的人生原则就是 volume of work——更多反复、更多投篮、更多 free throw——这样才能落在钟形曲线的"意外好结果"那一端。

  2. 100K/月之前不要碰非 Meta 平台

    "Don't even bother spending money on any platform that's not Meta until you reach $100,000 a month on Meta."

    "Meta is the only platform that has the algorithm dialed in for the kind of personalization needed to find your buyer cheaply at small scale."

    在 100K/月 以下,其他平台(TikTok/Google/Reddit)CAC 都打不过 Meta——投流分散资源就是给自己挖坑。

  3. tokens 会超过工资支出

    "Soon we're going to spend more on tokens than we spend on actual salaries."

    "We test almost a thousand AI-generated ads a day. If you don't spend a thousand credits a day, I'm disappointed in you."

    这一集标题。Speechify 每天测 ~1000 个 AI 生成广告,token 支出已经在追上工资——这是 AI-first 公司的真实成本结构变化。

  4. 公司有"增肌期"和"减脂期",别永远紧绷

    "Companies have bulking and cutting cycles."

    "In the bulking cycle you can be a bit sloppy, hire fast, accept some waste. In the cutting cycle, you sharpen up."

    永远 cutting = 长不大,永远 bulking = 烧光。Cliff 把健身的周期化思维套到公司管理上,反对单一节奏。

  5. OpenAI 广告会非常大,Meta 是仍可买

    "OpenAI ads will be massive. If I had cash, I'd buy Meta today."

    "OpenAI sits between humans and intent in a way Google never did. And Meta still has the most accurate identity graph + the algo most dialed in for performance."

    两个看似矛盾的下注——OpenAI 是新平台,Meta 是旧霸权未失——核心都是 identity + intent 这两个东西的稀缺度。

  6. Speechify 从 Cursor 迁到 Claude Code

    "It actually drives me crazy that not everybody has moved over to Claude Code. If you don't spend a thousand credits a day, I'm disappointed in you."

    "We were huge adopters of Cursor about 2 years ago. I got the Anthropic team plan in December. I'm pushing people constantly to hit limits."

    Cliff 现在的态度是"24 小时内不交 Loom 用 AI 干了什么、就准备离开"。他用极端 push 来把团队从右边推到中间。

  7. AI 帮他推动医生改变方案、找到爸爸的癌症

    "I would have never had the confidence to push on the doctor so hard to get the scan if I didn't have LLMs open during the conversation with the doctor."

    "The doctor refused to do the PSMA scan. I hired people on Kaggle, found a higher-resolution machine called U Explorer at UC Davis, messaged the lab director. By Jan 7, my dad had the left seminal vesicle excised."

    他爸前列腺癌 PSA 持续上升,医生说"等",Cliff 用 LLM + Kaggle 在医生之前定位到病灶——在原本计划的二次扫描日期之前就完成了手术。

  8. 3 小时不回消息 = Titanic 规则违反

    "You got to respond to messages within 60 seconds cuz you want to unblock people. — If you don't respond within a 3-hour period, I call it the Titanic rule. The movie Titanic is 3 hours long."

    "That can be: 'Hey, I'm busy. I'm at a football game.' Cool. Fine. That's a good answer. But you need to respond within 3 hours, even on weekends. I got decimated for it."

    不是要求随时在线——是要求让人知道你接收到了。Cliff 因为这个 tweet 被骂惨,但他坚持。"速度是当今公司唯一的战略优势"。

  9. AI 时代最值钱的技能是 QA

    "In a world where software engineering is commoditized and design commoditized, QA is the most valuable skill in the world."

    "If you try to build stuff with Claude Code, even if you give it all the tools in the world, it will not succeed in QAing itself to perfection. You still need a human to do that."

    写代码变便宜,审视代码变贵。Cliff 的招聘信号:不擅长 outcome-owning 的工程师,先放 QA 试试再说。

  10. 2022 年他弟买了 NVDA 3x 杠杆,现在 36 倍

    "Tyler bought a 3x lever option that's now like 36x up."

    "We were trying to spend millions of dollars buying H100 GPUs and no one would let us buy them. The market was mispricing Nvidia. Tyler had even higher conviction than me."

    不是事后吹牛——是 ChatGPT 出来之前,他们买 H100 抢不到那一刻就看到了 inference 需求曲线。"diversify to concentrate"——为了重仓特定机会,你必须先广撒以挑出 10/10。

Chapter 01

Volume of Work + Applying to 26 Colleges

"工作量"哲学 + 申 26 所大学
00:00 — 03:20 · 冷开场金句 + 大学申请像彩票 + 48 稿 essay
CLIFF00:02

"Don't even bother spending money on any platform that's not Meta until you reach $100,000 a month and spend on Meta."

"在你 Meta 每月花到 $100,000 之前——别浪费钱去任何非 Meta 平台。"

HARRY00:06

"In our show today, we have Cliff Weitzman, founder and CEO of Speechify."

"今天我们的节目嘉宾是 Cliff Weitzman——Speechify 的 founder 和 CEO。"

"This is the guy who solved his dyslexia and ADHD by building a voice AI agent that's now used by over 60 million people and has over a million five-star ratings."

"他用一个 voice AI agent 解决了自己的阅读障碍和 ADHD——这个产品现在有 6000 万用户、超过 100 万 5 星评价。"

"We test almost a thousand AI-generated ads a day."

""我们一天测试将近 1000 个 AI 生成的广告。""

"Right now, we're getting to the point where soon we're going to spend more in tokens than we spend on actual salaries."

""我们快到那个临界点了——很快 token 支出会超过工资支出。""

"If you don't spend a thousand credits a day, I'm disappointed in you."

""你一天不花 1000 个 credits——我对你失望。""

"If you can really trust someone, you can afford to pay them significantly more."

""如果你真的能信任一个人——你完全可以付他多得多的钱。""

"The world around us was built by people no smarter than you are."

""我们周围的世界——是由不比你聪明的人建造的。""

"You sent me your topic suggestions, and I was like, oh my god — I actually sent them to the team and said, this is wild."

"你提前发给我的话题——我看完就跟团队说,这个太炸了。"

"First, thank you for making my life easier with the schedule. Second, thank you so much for joining me in person."

"第一,谢谢你的日程那么好排。第二,谢谢你愿意亲自来录。"

CLIFF01:16

"My pleasure. It's great to be here."

"我荣幸。来这里很棒。"

HARRY01:17

"You applied to 26 colleges when everyone else applied to eight. Can you talk to me about the strategy?"

"别人都申 8 所大学——你申了 26 所。能聊聊背后的策略吗?"

CLIFF01:29

"I moved to the US when I was 13. I didn't speak English."

"我 13 岁来美国——不会英语。"

"My parents knew about three schools — Harvard, MIT, and Stanford."

"我父母只听说过三所学校——Harvard、MIT、Stanford。"

"Our college counselor recommended applying to six schools."

"高中升学顾问建议申 6 所。"

"Quickly I realized that even for kids with perfect SAT scores and perfect grades, you have about a one in 10 chance of getting in."

"我很快意识到——即使是 SAT 满分、GPA 满分的孩子——录取率也只有 1/10 左右。"

"I was like: okay, it's a lottery. How do you win a lottery? You buy more tickets than everybody else."

"我心想:这就是一张彩票。怎么中彩票?买比别人多的票。"

"Most kids in my school didn't need to do seventh period when they were a senior."

"我们学校大部分高三学生第 7 节课不用上。"

"But I had academic workshop — basically the class you go to if you're on the short bus — because I just couldn't read very well."

"但我有 academic workshop——就是给"差班"留的那节课——因为我读不太了。"

"Instead of three or five variations of my main essay, I wrote 48 drafts."

"主文书,我没写 3 个或 5 个版本——我写了 48 稿。"

"I had to make a second Common App account because I ended up applying not to 20 schools (which is the max) but 26."

"我甚至开了第二个 Common App 账号——因为我申的不是上限 20 所,而是 26 所。"

"In the end, I didn't get into many schools I wanted, but I got into a lot of schools I didn't expect to."

"最后,很多想去的没录我——但很多没指望的录了我。"

"My overall principle in life is volume of work."

"我整个人生原则就是 volume of work。"

"More reps, more shots of goal, more free throws — that's how you get unexpected outcomes on the bell curve."

"更多反复、更多射门、更多罚球——这样你才能在钟形曲线的"意外好结果"那一端落点。"

HARRY02:55

"Alex Hormozi has this line: volume times leverage equals output."

"Alex Hormozi 有句话:volume × leverage = output。"

"I used to send hundreds of emails to sponsors with zero leverage. Now with more leverage, the output is significantly more."

"我以前给赞助商发几百封邮件——零杠杆。现在杠杆大了——产出大很多。"

"My girlfriend asked me last night: did you create a text-to-speech app because it took you a long time to read?"

"我女友昨晚问我:你做 TTS 是不是因为自己读得慢?"

CLIFF03:10

"Yes. I'm super dyslexic."

"是。我有严重阅读障碍。"

"First, third, fourth grade — couldn't figure out how to read. I wanted to read fast but I couldn't."

"小学一年级、三年级、四年级——根本搞不清楚怎么读。我想读快——但读不快。"

"So we built a deep-learning-based text-to-speech model in 2015 that now reads more than 10 million books worth of words to people every year."

"于是 2015 年我们做了一个深度学习的 TTS 模型——现在每年给人读的内容相当于 1000 万本书。"

Chapter 02

Speechify Origin: Dyslexia, Deep Learning, 10M Books/Year

Speechify 起源:阅读障碍 + 深度学习 + 每年读 1000 万本书
03:20 — 05:35 · 2015 年深度学习 TTS · 60M 用户 · 100 万 5 星评价
HARRY03:45

"Apparently the start in the yearbooks of 40 of your 14-year-old friends freshman year — you wrote: '[name], you are a great gay.'"

"据说你 14 岁那年——给 40 个同学的毕业纪念册上都写了:"[名字],你是个 great gay"。"

"How did that lead to Speechify?"

"这事怎么牵出 Speechify 的?"

CLIFF03:56

"When I was 14, freshman year of high school, I was very self-conscious about being very bad at spelling."

"我 14 岁高一,对自己拼写太烂特别自卑。"

"You get a yearbook and typically people sign the back with some nice message about you."

"美高有个习俗——大家在毕业纪念册背面互相写话。"

"I didn't want to misspell stuff in everybody's yearbooks. So I came up with a strategy."

"我不想给别人本子上拼错字。于是我想了个办法。"

"I would write the least number of words possible — the same thing in everyone's yearbook."

"我尽量写最少的字——而且每本写一样的话。"

"David, you are a great guy. Steve, you are a great guy. Bill, you are a great guy."

""David, you are a great guy. Steve, you are a great guy. Bill, you are a great guy.""

"After the last day of school, we went to sleep over at a friend's house. He goes upstairs to show his mom the yearbook."

"最后一天放学,我们去朋友家过夜——他上楼把本子给他妈看。"

"And I hear him scream from the top of the stairs: 'WHAT THE HELL, CLIFF? I'M NOT GAY.'"

"然后我听见他从楼梯上吼:"卧槽 Cliff?我不是 gay!""

"I'm like: oh my god, what have I done?"

"我心想:天啊,我干了什么。"

"Eight years later, I was graduating college. I wrote a 30-page paper for one of my philosophy classes."

"8 年后,我大学要毕业——给哲学课写了一篇 30 页的论文。"

"I submitted it to a friend before fully submitting it. They sent it back with three spelling mistakes — and I cried."

"我先发给朋友看——他改完发回来,只有 3 个拼写错——我哭了。"

"If you had told 14-year-old Cliff he'd go to college, he wouldn't be sure. Definitely not Brown. For sure would never expect to write a 30-page paper with only 3 spelling mistakes."

"如果你告诉 14 岁的 Cliff,他能上大学,他不敢信;能上 Brown,他想都不敢想;能写出只有 3 个拼写错的 30 页论文——绝对不可能。"

"The reason I didn't have that many is because I can hear the spelling mistakes with Speechify."

"我之所以错得这么少——是因为用 Speechify 听文章时,能听出拼写错。"

HARRY05:30

"Dude, I was listening to you and Ali Abdaal — great show."

"哥们,我听你和 Ali Abdaal 那期——超棒。"

"In 2020 you made a list of the top 100 consumer subscription companies in the world by revenue, flew all over the world to meet each of them, and tried to learn from them."

"2020 年你列了一份全球收入最高的 100 家消费订阅公司——飞遍世界去见每一家——然后向他们学习。"

"With the greatest of respects — I was frustrated listening, because you said that and I was like: amazing, what did you learn?"

"恕我直言——我听的时候有点 frustrated——你提了这事我心想:厉害啊,那你到底学到啥?"

Chapter 03

Flying Around the World to Meet 100 Consumer CEOs

飞遍全球见 100 个消费订阅 CEO
05:35 — 10:46 · 怎么搭上线 · 学到什么 · "100K 之前别碰非 Meta"
CLIFF05:41

"At every point in Speechify's history, I have led every department except for AI."

"Speechify 历史上每个阶段——除了 AI——其他部门都是我亲自带的。"

"Design, engineering, QA, growth."

"设计、工程、QA、增长。"

"My rule is: I will read 100 books about the topic — listen to them with Speechify."

"我的规则是:这个领域,读 100 本书——用 Speechify 听。"

"I will identify 100 experts and have a conversation with them."

"然后找到 100 个专家——一个个去聊。"

"First thing I learned: if you send a good email, people will respond — no matter how amazing they are."

"第一件学到的事:你写一封像样的邮件,别人会回——不管他多牛。"

"I did Mike Krieger from Instagram. Ev Williams from Twitter. The founders of Plaid, 23andMe, Honey, Grammarly. Vlad from Robinhood."

"我聊了 Instagram 的 Mike Krieger、Twitter 的 Ev Williams、Plaid / 23andMe / Honey / Grammarly 的 founder、Robinhood 的 Vlad。"

"All of those became Speechify investors very early on."

"他们后来都成了 Speechify 的早期投资人。"

"The second rule was again volume of work."

"第二条规则,还是 volume of work。"

"If the CEO didn't respond, I'd message the CMO. If the CMO didn't, the head of growth. Then politely continue until they told me to stop (less than 1% of the time) or agreed to chat."

"CEO 不回——我发 CMO;CMO 不回——发 growth 负责人;再不回,我继续礼貌地发——直到他们让我停(<1% 才会),或者答应聊。"

"I'd also DM them on Instagram. I'd also ask people for intros."

"我还会 Instagram 私信、找别人引荐。"

"Just be unlimited hungry to get the information."

"对获取信息——保持无限的饥饿。"

"Once we'd get on a Zoom call, there was one thing that always succeeded: I'd ask what's your favorite books?"

"上 Zoom 之后,有一招永远管用——我问"你最喜欢的书是什么?""

"I've listened to 100 books a year, every year, for the last 15 years."

"我过去 15 年——每年听 100 本书。"

"They'd tell me, and most likely I read at least one of them. So we'd connect on the book."

"他们说出书名——大概率我读过其中一本。我们就能在书上接上线。"

"Then I'd be like: oh my gosh, I'm going to be in Denmark this Sunday — do you want to hang out?"

"然后我说:哎我这周日在 Denmark——见一面?"

"And then I'd book a flight to Denmark and I'd show up."

"——然后我就订机票飞过去。"

"I did this with Anton at Lovable recently. I'm still doing the same thing."

"最近见 Lovable 的 Anton 也是这么干的——我现在还是这套打法。"

"The first thing I learned is: growth is just an arbitrage game."

"我学到的第一个东西是——growth 就是一场套利游戏。"

"You are competing with every other person in the world who wants to get their product in front of users."

"你的对手是全世界所有想把产品塞到用户面前的人。"

"Your lever points are either the content you make or how you distribute that content."

"你的杠杆点只有两个:做什么内容,以及怎么分发。"

"One person I learned from was the founder of Blinkist."

"我学到东西的人里有 Blinkist 的 founder。"

"He said as an aside: don't even bother spending money on any platform that's not Meta until you reach $100,000 a month on Meta."

"他随口一说:在 Meta 每月烧到 $100K 之前——别浪费钱去任何非 Meta 平台。"

"And then I stopped spending on other places."

"我立刻把别的平台的预算停了。"

"A lot of people who were senior in these companies were already rusty — not strong individual practitioners of the craft."

"很多公司里 senior 的人——其实已经 rusty 了——不是真正会上手干活的人。"

"So I'd go down a couple of levels to find the person who's actually buying the ads, editing the content, seeing what the CTA is."

"所以我会往下钻一两层——找真正在投广告、剪内容、设计 CTA 的人。"

"I made sure everybody in our company, no matter how senior, was still not a fat general sitting in the back — but a warrior who could take out their sword."

"我们公司的人——不管多 senior——不能是后方的胖将军——必须是能拔剑的战士。"

"We hired a head of growth. 30 days in: nothing. 60 days in: nothing."

"我们招过一个 head of growth——30 天没产出,60 天还是没产出。"

"I said: bro, source the people yourself, message the people yourself, get on the calls yourself."

"我跟他说:兄弟,自己 source、自己发消息、自己上电话。"

"You've got seven days. I need to see you making ads in CapCut, editing in Figma or Photoshop, and buying the ads."

"给你 7 天——我要看到你自己在 CapCut 做广告、在 Figma/Photoshop 编辑、自己投广告。"

"It reminds me of Keith Rabois: we want to hire barrels, not ammunition."

"这让我想起 Keith Rabois 说的:招"枪管",不要招"子弹"。"

"Someone who can go from one to 10 on their own and do full stack."

"一个人能从 1 走到 10、能 full stack 干所有事。"

"I don't want: oh, I can put the video on CapCut, but someone else needs to edit. I want: I take the video, edit it, post it."

"我不要"我能上传到 CapCut,但剪辑要别人做"——我要"我接视频、我剪、我发"。"

CLIFF09:50

"We look for three things when we hire: fire in the belly for the product, high loyalty to the team, and the ability to learn fast and ship to production."

"我们招人看三件事:对产品有 fire in the belly、对团队高忠诚、学得快并能交付到生产。"

"And we have a principle even inside of engineering: you've got to respond to messages within 60 seconds — because you want to unblock people."

"工程内部有个原则:消息要在 60 秒内回——因为你要给别人解封。"

"Just inform them whether you can unblock them or not. If you can't, that's fine — they'll go learn it on their own."

"哪怕你说"我现在解不开",也行——他自己去学就好。关键是要回。"

"If you ship a PR to a repo that your team is not responsible for, it gets fast-tracked — that team has to respond very quickly."

"如果你给非你团队负责的 repo 提 PR——那条 PR 会被快通道处理——那个团队必须快速响应。"

"Because I want to encourage people to learn how to contribute to other codebases."

"因为我想鼓励大家学着给别的 codebase 贡献。"

"The person in the company who learns the fastest is my brother Tyler, who leads our AI team."

"公司里学得最快的人——我弟 Tyler——他带我们的 AI team。"

"The whole principle inside our AI engineering team is: just learn everything and then you win."

"AI 工程团队的原则就是:把所有东西都学会——你就赢了。"

Chapter 04

Bulking & Cutting Cycles + Paid Ads + CAC

公司增肌期 / 减脂期 + 投流和 CAC
10:46 — 30:18 · 周期化管理 + 一天测 1000 个 AI 广告 + tokens 超过工资
CLIFF10:47

"Companies go through cutting and bulking cycles like bodybuilders do."

"公司像健身的人一样——有"减脂期"和"增肌期"。"

"There's a reason why cutting and bulking makes sense — you can't oscillate between the two in one week."

"为什么周期化有意义?——因为你不能一周之内来回切换。"

"You have to commit for like 6 months to doing one."

"你必须 commit 大概 6 个月只做一件事。"

"It took Speechify four and a half years to find real product market fit."

"Speechify 找到真正 PMF 花了 4 年半。"

"In 2020, we started growing really fast. Install-to-trial, trial-to-paid, renewals all went up."

"2020 年开始,我们涨得很快——install→trial、trial→paid、续费率都涨了。"

"All I was thinking about was growth — which is why it was a no-brainer to fly to all these places to meet people."

"我脑子里只有一件事:growth——所以飞遍全世界见人,不用想。"

"I started logging my time in 30-minute increments. Every day I'd have 48 units."

"我开始按 30 分钟一格记录时间——每天 48 格。"

"I noticed 30% of my time I spent helping other people. 30% I spent learning SEO."

"我发现 30% 时间在帮别人,30% 时间在学 SEO。"

"So I decided I'm not allowed to learn anything from that point on. I'm only allowed to execute."

"我决定:从此不许学习——只许执行。"

"The second I saw myself reading an article, I'd close it and go execute."

"看到自己在读文章——立刻关掉去执行。"

"Any Harvard MBA can cut costs if you're smart. It takes a genius to grow revenue."

"任何聪明的哈佛 MBA 都能砍成本——但增长收入需要天才。"

"That's why investors value exponential growth of revenue even for companies that are not making a profit."

"所以投资人愿意为指数级收入增长付钱——哪怕公司不赚钱。"

"It's very difficult to do both at the same time. You're clicking the gas and clicking the brake at the same time."

"两件事同时做非常难——等于油门和刹车一起踩。"

"Speechify just finished 4.5 years of running profitable. Now we're in a hyper growth phase."

"Speechify 刚跑完 4.5 年的盈利期——现在进入 hyper growth 阶段。"

"You have bulking years and you have cutting years. You need to decide which year it is for you."

"你有增肌年——也有减脂年。你必须决定自己今年是哪种。"

HARRY13:08

"In a hyper-growth phase, are we fully accepting CAC will go up, and we're happy with that?"

"hyper growth 期里,你完全接受 CAC 会涨——而且 OK?"

CLIFF13:15

"No. Not at all."

"不,完全不是。"

"You have to consider what is my LTV and what's the first payment period."

"你得算 LTV——以及第一付费期。"

"I'm never willing to burn CAC for no reason. The blended CAC can go up — not the direct CAC."

"我从不无意义地烧 CAC——blended CAC 可以涨,direct CAC 不行。"

"I'm okay spending $500,000 a month making 20,000 pieces of creative — because that's a creative investment."

"我可以一个月花 $50 万做 2 万条 creative——因为这是 creative 上的投资。"

"We're one of the 200 companies provisioned to test ads on OpenAI."

"我们是 200 家被批准在 OpenAI 上测广告的公司之一。"

"We're spending a ton of money on OpenAI right now and we're learning it faster than everybody else."

"我们现在在 OpenAI 投了很多钱——比别人学得快。"

"You think about everything as a test. If the test succeeds, do more. If it fails, that's fine."

"所有事都是 test——成了就放大,败了也行。"

"Kelly at Netflix told me: I know I'm wasting half my spend. I just don't know which half."

"Netflix 的 Kelly 告诉我:我知道我浪费了一半的钱——我只是不知道是哪一半。"

HARRY15:10

"Who had the most atypical / weird ad-buying process?"

"哪家公司投广告的方式最反常?"

CLIFF15:11

"Greg at Ladder. He's the CEO, but he buys the ads himself, and he found a lot of arbitrage plays."

"Ladder 的 Greg。他是 CEO——但广告他自己投——找到了很多套利机会。"

HARRY15:29

"What's the play today that people don't know?"

"今天大家还不知道的 play 是什么?"

CLIFF15:31

"Whitelisting ads on Instagram works really well."

"Instagram 的 whitelist ads——效果非常好。"

"You find a creator that is in your niche — maybe teachers, ADHD, dyslexia, an AI creator."

"找一个你 niche 里的 creator——可能是老师、ADHD、阅读障碍、AI creator。"

"They make a video for you. Then you go: hey, make me 20 videos. I don't need you to post them. I'll run them as ads myself."

"他给你做一条视频——然后你说:再做 20 条,你不用发——我以广告形式跑。"

"If some are good, we can post them from your account and increase the spend."

"如果有几条特别好——我们可以从他账号发出去,然后加预算。"

CLIFF16:10

"We test almost a thousand AI-generated ads a day right now, on top of the roughly 8,000 organic creatives we make with humans every single month."

"我们一天测大约 1000 个 AI 生成广告——另外每月还有 8000 条人手做的。"

"One thing we do is reskin. My sister Geffen leads our ads team."

"我们的一招叫 reskin。我妹 Geffen 带我们的 ads team。"

"We have great ads with Geffen, then we'll turn her into a 60-year-old lady, a 40-year-old man, African-American, Asian — change the background from a coffee shop to a library."

"我们用 AI 把 Geffen 改成 60 岁老太、40 岁男人、非裔、亚裔——把背景从咖啡馆换到图书馆。"

"So you just test all the variations."

"——然后所有 variation 全部测。"

"Highest converting demographics for us are people in their 30s and 40s in terms of paid."

"我们付费用户里转化最高的人群是 30-40 岁。"

"But we have crazy usage in the 18-28 market. And people 60+ need reading glasses — so they use Speechify a ton."

"但 18-28 岁用得疯狂。60+ 因为视力问题——也用得很多。"

CLIFF18:16

"It's an arbitrage game. If you use the tool everybody else uses, you typically don't win."

"这是套利游戏——你用别人都在用的工具,你赢不了。"

"So we just built our own platform."

"所以我们自己造了一套平台。"

"There was a point when my girlfriend went to Disneyland — I had 4 days free."

"有次我女友去 Disneyland——我空出 4 天。"

"I did a course from Alex Hormozi for AI-generated video ads, hired everybody from inside the community who was really good."

"我上了 Alex Hormozi 的 AI 视频广告课——直接在他社群里挖人。"

"In one week I hired 10 people. 24/7 all I was thinking about was AI-generated content."

"一周我招了 10 个人——24/7 脑子里只想 AI 生成内容。"

"My goal was to get to 1,300 ads tested every single day."

"目标是一天测 1300 条广告。"

"Initially I built it on n8n, but all the n8n instances would time out. So we built our own platform."

"最早搭在 n8n 上——n8n 实例老超时。于是我们自己造了一套。"

CLIFF23:49

"Rule of 100. I made a list of the top 100 best-performing ads in history."

"rule of 100——我列出史上表现最好的 100 条广告。"

"Every Super Bowl ad that was really funny, every ad that made me cry, every ad that won an award."

"每条让人笑的 Super Bowl、每条让我哭的、每条得奖的。"

"I rewrote the script for each one to be about Speechify, then upped the ante on the ad."

"我把剧本改写成 Speechify 版本——并把强度再往上推。"

"The rule I developed is: you win or lose by the first 1080 pixels — the first image of the video."

"我自己悟出的规则:广告输赢取决于前 1080 个像素——视频第一帧。"

"I put bright red headphones on my head, a suit on, and went into a hot tub. First shot: me with red headphones, suit, hot tub."

"我戴上鲜红耳机,穿西装,坐进按摩浴缸——第一帧:红耳机 + 西装 + 浴缸。"

"That YouTube ad basically launched the company into a whole new stratosphere — performed for like 3 years straight."

"这条 YouTube 广告把公司送进了新的轨道——连续跑了 3 年都还有效。"

"The one I did not expect to work that worked amazing: I wrote a rap song about my dyslexia and ADHD. That thing made us $3 million."

"我没指望会爆的反而爆了——我写了一首关于自己阅读障碍和 ADHD 的 rap——那条带来了 $300 万收入。"

CLIFF28:08

"Only do things that result in conversion."

"只做能带来转化的事。"

"50 million people use Speechify. If you don't download, don't start a trial, don't convert — there's two options."

"5000 万人在用 Speechify——不下载、不试用、不转化——我有两种心态。"

"The first one is: I don't care. The second one is: I actually do genuinely care that we solve dyslexia and ADHD."

"第一是:我不在乎。第二是:我真的想解决阅读障碍和 ADHD。"

"There's two billion people with ADHD, low vision, autism, anxiety, concussion, second language learners."

"全世界有 20 亿人有 ADHD、低视力、自闭、焦虑、脑震荡、或第二语言学习者。"

"Our philosophy is: we roll out the red carpet for the user. We give them the best experience. If you can't pay, that's fine."

"我们的理念是:给用户红毯待遇——付不起也没事。"

"I think about that as our form of charity. But the ad spend is covered by people who pay."

"我把这看作我们的慈善——但广告费由付费用户买单。"

CLIFF29:18

"Word of mouth is still the biggest way of getting people."

"word of mouth(口碑)还是我们最大的获客方式。"

"Someone telling their friend: hey, I see you read at a normal speed — don't you want to read 3x faster?"

"一个人告诉朋友:嘿,你只能正常速度阅读——不想 3 倍速吗?"

"We crush it on Apple — we won the design award last year, got featured 14 times."

"我们在 Apple 表现极好——去年拿了 design award,被 feature 14 次。"

"15% of our users now come organically from ChatGPT."

"现在 15% 的用户从 ChatGPT 自然来。"

"Inside performance marketing: Meta is still the best winner. Google is the most profitable. YouTube is pretty good, especially on desktop."

"投流里:Meta 仍是最强,Google 最赚钱,YouTube 不错(尤其桌面端)。"

Chapter 05

Why OpenAI Ads Will Be Massive + Why Buy Meta Today

OpenAI 广告会很大 + Meta 现在仍可买
30:18 — 34:57 · 新平台 vs 旧霸权 · identity graph + intent
HARRY30:18

"OpenAI ads — do you think it will be huge?"

"OpenAI 广告——你觉得会很大吗?"

CLIFF30:24

"Of course. Yeah. Massive. Massive. Massive."

"当然。会非常非常非常大。"

"Because it knows so much about you."

"因为它对你的了解太多。"

"Same thing as Meta — the reason why Meta ads are so good is because they know what you message on Instagram and on WhatsApp."

"和 Meta 一样——Meta 广告好的原因是它知道你在 Instagram 和 WhatsApp 上发什么。"

"OpenAI knows everything about your history and what you're interested in — inside your psyche."

"OpenAI 知道你历史聊天的一切——以及你内心想什么。"

"It's exactly the same way Google has built such an amazing business — there's relatively high intent when you talk to it."

"和 Google 起家的逻辑一样——你跟它说话的时候 intent 已经很高。"

"Right now it's extremely high CPMs. But if you find content that converts, the CPMs don't matter."

"现在 CPM 极高——但只要你的内容能转化,CPM 高也不要紧。"

"They just launched the SDK for tracking. That's really important."

"他们刚发布了 tracking SDK——这一步非常关键。"

HARRY31:08

"Would you buy Meta today?"

"你今天会买 Meta 吗?"

CLIFF31:08

"Absolutely. Yeah. I think Meta is one of the most underrated stocks right now."

"绝对。Meta 是现在被严重低估的股票之一。"

"Number one: Mark Zuckerberg is an absolute savage."

"第一:Mark Zuckerberg 是个绝对的 savage。"

"He will hire people if they're right, no matter what."

"只要看准人——他不计代价挖。"

"In a world where engineering is commoditized and design for the most part is commoditized — the most important thing becomes excellent QA and customer acquisition."

"在工程和设计大面积被商品化的世界——最重要的事变成 QA 和获客。"

"Google has pressure from all the LLMs, but Meta doesn't. The only real pressure on Meta now is TikTok — and TikTok has issues now that it's been separated."

"Google 被 LLM 压力压着——Meta 没有。Meta 现在唯一真压力是 TikTok——而 TikTok 自己分拆后问题一堆。"

"Zuck has proven he is the number one acquirer in the world. No one acquires as well as Meta."

"Zuck 证明了他是世界第一收购大师——没有公司收购像 Meta 这么强。"

"In an AI world, data is everything — and they have more proprietary data than everybody else."

"AI 世界里数据是一切——他们手上独家数据比所有人多。"

"They haven't figured out the model side yet, but I really do believe they will."

"模型这一块他们还没搞定——但我相信他们会搞定。"

"Look at all the people who are vibe coding new products right now. Where do you think they're going to get their users?"

"看看现在所有 vibe coding 的新产品——他们的用户从哪来?"

HARRY32:16

"Everyone talks about taste — it's the new word of 2026."

"现在大家都在聊 taste——是 2026 年的热词。"

"Is that bullshit?"

"这是不是扯淡?"

CLIFF32:27

"No. Not bullshit. Everything is in taste."

"不,不是扯淡。一切都在 taste 里。"

"When I say design, I mean the actual grind of sitting in Photoshop, Figma, Illustrator, Canva — pixel-perfect making things."

"我说设计被商品化——指的是坐在 Photoshop/Figma/Illustrator/Canva 里抠 pixel 那部分活。"

"Most of my leadership team is sending their emails from inside an IDE — Claude Code terminals."

"我们 leadership team 大部分人现在写邮件都是在 IDE 里——Claude Code 终端里。"

"When we launch a new product, we don't assign a designer in the beginning — just use the default Claude Code stuff."

"我们启动新产品时——不会一开始就配设计师——直接用 Claude Code 默认那套。"

"The hammer-and-chisel part of design is not as important. The thing that's left is taste — very few people have taste."

"设计里"锤子凿子"那部分不重要了——剩下的是 taste——而 taste 是稀缺的。"

HARRY33:32

"I'm down 40% on my Figma position. Hold or sell?"

"我 Figma 仓位已经 -40% 了——hold 还是卖?"

CLIFF33:36

"Hold. Dylan Field is a savage. Anything Dylan Field does, you should go for."

"Hold。Dylan Field 是个 savage——他做什么,你就跟。"

"Figma is a relatively anti-fragile company. It's a really good team. They're excellent at sales. They have really good taste."

"Figma 是相对反脆弱的公司——团队好、销售强、taste 极佳。"

"Figma traded at 40x revenue and now they're at 6x revenue. The market is making a correction — that's fine."

"Figma 之前 40x 收入交易——现在 6x。市场在修正——很正常。"

Chapter 06

Why Speechify Moved From Cursor to Claude Code

为什么 Speechify 从 Cursor 迁到 Claude Code
34:57 — 38:22 · 模型想 run 而不是被 type · 一周 > 一个月
HARRY34:32

"Claude Code or Cursor internally — what does everyone use engineering-wise?"

"Speechify 工程内部——大家用 Claude Code 还是 Cursor?"

CLIFF34:36

"Claude Code."

"Claude Code。"

"We were huge adopters of Cursor about 2 years ago. I got the Anthropic team plan in December."

"我们 2 年前是 Cursor 的重度用户——12 月我开了 Anthropic 的 team plan。"

"It actually drives me crazy that not everybody has moved over to Claude Code."

"现在让我很抓狂的是——还没所有人都迁到 Claude Code。"

"A lot of people use Claude Code from inside the terminal in Cursor. I prefer if they use Claude Code inside of the Claude Code IDE."

"有些人在 Cursor 终端里跑 Claude Code——我更希望他们直接在 Claude Code 的 IDE 里跑。"

"I'm pushing people constantly to hit limits."

"我一直在催大家把额度用满。"

"We're getting to the point where soon we're going to spend more in tokens than we spend on actual salaries."

"我们快到那个临界点——很快 token 支出会超过工资支出。"

HARRY35:25

"You'll spend more in tokens than on salaries. Is that atypical?"

"token 支出超过工资——这个 atypical 吗?"

CLIFF35:30

"It's atypical right now, but I don't think it'll be atypical in the long term."

"现在 atypical——但我觉得长期不会是。"

"You just need to be a really strong early adopter."

"你只要是非常坚定的 early adopter。"

"Across engineering — and I'd like to be spending more tokens on our growth team and creative."

"工程那一边——我希望 growth team 和 creative 也能花更多 token。"

CLIFF36:27

"Speechify is a voice agents ecosystem with 50 million users and 1.1 million five-star reviews."

"Speechify 是一个 voice agents 生态——5000 万用户、110 万 5 星评价。"

"We need to educate our users on how to use AI."

"我们需要教用户怎么用 AI。"

"Our first agent was a text-to-speech tool. The second was speech-to-text for dictation."

"第一个 agent 是 TTS——第二个是 STT 听写。"

"The third was a podcast generator where you can be in the podcast as well as two new hosts."

"第三个是 podcast 生成器——你可以加入到 podcast 里,和两个新主持人对话。"

"Then we launched Teacher Agent — it asks you questions about a document and teaches back the content gap."

"然后是 Teacher Agent——它就你的文档问你问题——并补上你不会的部分。"

"Now we just launched Jarvis — Speechify for work."

"现在我们刚发布 Jarvis——work 版的 Speechify。"

"It takes about 100 iterations on a feature before it gets adopted. But once it gets adopted, it gets adopted really well."

"一个 feature 大概要 100 次迭代才会被采纳——但一旦被采纳,粘性极高。"

"Internally I am constantly screaming from the rooftops that people have to use Claude Code."

"内部我一直在屋顶上呐喊——大家必须用 Claude Code。"

"I'm more extreme about it than it needs to be — because I need to move people from all the way over here to all the way over there."

"我比正常要求更极端——因为我要把人从这头推到那头——能在中间见我也 OK。"

"I'm okay if they meet me in the middle. So I'm just extreme."

"所以我刻意极端。"

"If you don't spend a thousand credits a day, I'm disappointed in you."

"一天花不到 1000 个 credits——我对你失望。"

"Send me screenshots and I will get people on a Zoom call, screen share the entire engineering — and I'll show them how I use it."

"给我截图——我会拉所有工程师上 Zoom——屏幕共享——演示我自己怎么用。"

"If you are not using this tool right now — comment in Slack and explain why."

"如果你现在没在用这个工具——在 Slack 里 comment 告诉我为什么。"

"You have 24 hours. Send me a Loom video with what you used AI to make."

"给你 24 小时——发我一条 Loom 视频,告诉我你用 AI 做了什么。"

Chapter 07

How AI Helped Cliff Diagnose His Dad's Cancer

AI 怎么帮 Cliff 诊断爸爸的癌症
38:22 — 41:01 · 比医院早 3 周 · 这件事改变了他对 AI 的紧迫感
HARRY38:17

"What single thing do you use Claude for that has been most impactful?"

"你用 Claude 做的一件事——影响最大的是什么?"

CLIFF38:23

"The personal side: medical stuff."

"个人这一边——医疗相关。"

"My brother Eric has Lyme neural inflammation and psychosis."

"我弟 Eric 有莱姆病引起的神经炎症和精神病症状。"

"We did proteomics, mapped all of his genome, put it on Supabase, wrote an MCP for it, and compared to every article that had ever been published about anomalies in the genome."

"我们做了蛋白质组学,把他基因组全 mapping,放进 Supabase,写了个 MCP——和所有发表过的基因组异常文章对比。"

"We found four anomalies in how his genes expressed. That was a breakthrough for us."

"我们发现了 4 个基因表达异常——这对我们是突破。"

"The second thing: my dad has had prostate cancer on and off for the last eight years."

"第二件事——我爸过去 8 年反复有前列腺癌。"

"His PSA continued to go up, which meant it came back. We just didn't know where it was in the body."

"PSA 值在涨——意味着癌复发——但不知道在哪里。"

"The next thing you need to do is a PSMA scan. The doctor refused — he said: the cancer is too small. We need to sit and wait."

"下一步该做 PSMA 扫描——医生拒绝了——他说"癌太小,先等"。"

"I said: that sounds like the worst advice ever. We know he has cancer — waiting 3 more years is terrible."

"我说:这是史上最烂的建议。我们知道他有癌——再等 3 年是灾难。"

"So I went on Kaggle — the largest online competition for data scientists — and started hiring people there to help diagnose my dad's cancer from the radiology reports."

"我上 Kaggle——最大的数据科学竞赛平台——在那里直接雇人——根据影像报告帮我诊断我爸的癌。"

"It has open-source repositories on time series of PSMA PET scans."

"Kaggle 上有 PSMA PET 扫描时间序列的开源 repo。"

"I came up with four theories — including 'tipping theory' — where you look at the level above and below where you think the cancer is, make the data noisy, and see where it actually is."

"我推了 4 个理论——其中一个叫 tipping theory——你看疑似病灶上下一层,人为加噪声,看真正在哪。"

"From talking to Gemini, Grok, Perplexity — I realized there was a machine that could help."

"我和 Gemini、Grok、Perplexity 聊——发现有一台机器能帮上忙。"

"UCSF's China Basin does a normal PSMA PET scan that scans a voxel of 4mm in diameter. But there's another machine called the U Explorer that can scan 2mm."

"UCSF 的 China Basin 标准 PSMA PET 体素是 4mm——但 U Explorer 这台机器能做到 2mm。"

"There were three of them in the United States, and one was at UC Davis."

"美国一共 3 台——其中一台在 UC Davis。"

"I messaged the guy responsible for that lab. He agreed to test my dad."

"我私信那个 lab 的负责人——他同意给我爸做。"

"The doctor I'd put pressure on to give my dad the exam had scheduled it for end of January. In December, we got the scan on the higher-quality machine — and we found it. Left seminal vesicle."

"被我推的医生原本把检查排在 1 月底。我们 12 月在那台高分辨率机器上做了——找到了——在左精囊。"

"I would have never had the confidence to push on the doctor so hard to get the scan if I didn't have LLMs open during the conversation with the doctor."

"如果不是和医生沟通时 LLM 一直开着——我永远没有那么大的信心去硬推他做扫描。"

"I definitely would not have found another machine, got my dad scanned, and figured it out."

"我也绝对不会找到那台机器、安排扫描、定位病灶。"

"By January 7th, he had the left seminal vesicle excised."

"1 月 7 日——他的左精囊被切除了。"

"Before he was even supposed to have the second scan — he had the surgery."

"在他原本应该做第二次扫描之前——手术已经完成。"

CLIFF41:00

"The problem is not the doctors. The problem is the system where doctors have to treat statistics and make sure they don't get sued."

"问题不在医生身上——问题在医疗体系——医生要对统计学负责、要防被告。"

"In our case, because we're educating ourselves and willing to take the risk ourselves, we can do things that result in better outcomes — but a little bit more risky."

"我们这种情况——我们自学、愿意自己承担风险——能拿到更好的结果——代价是承担更多风险。"

"My contention is: just learn as much as possible. The information is accessible. You are the quarterback of your own medical experience or that of your family members."

"我的主张是:能学多少学多少——信息触手可及——你是自己和家人医疗的 quarterback。"

Chapter 08

Building an Internal AI-First Engineering Culture + Layoff Worries

建一支 AI-first 的工程文化 + 担心裁员吗
41:01 — 58:44 · CLAUDE.md / 招人 / 工具 · 不担心因为前面工作只多不少
HARRY42:32

"You said adversity quotient matters more than IQ or EQ."

"你说过 AQ(adversity quotient,逆境商)比 IQ 和 EQ 都重要。"

CLIFF42:50

"People talk a lot about IQ or EQ — but the most important one is AQ, adversity quotient. How good are you when things get really tough?"

"大家都聊 IQ、EQ——但最重要的是 AQ——逆境商:事情真的变难时,你能撑多好。"

"Think of a Jeep going through the savanna. I don't want the Jeep that goes the fastest — I want the Jeep that's not going to have a flat tire."

"想象一辆开过大草原的 Jeep——我不要跑最快的那辆——我要不爆胎的那辆。"

"There's a heuristic: sometimes you look at an engineer who's willing to grapple with a really hard problem for 8 hours straight."

"有个启发式:有些工程师愿意死磕一个难题 8 个小时不放。"

"Most people, if it's really hard, they quit after 30 minutes."

"大部分人遇到难题 30 分钟就放弃了。"

"But the things that really move the company are on the other side of 5 hours of grappling with a hard problem."

"但真正推动公司前进的事——都在"啃难题 5 小时"那一头。"

"My friend Valentin Perez has a great line: a little bit of slope makes up for a lot of y-intercept."

"我朋友 Valentin Perez 有句好话:"一点点 slope(斜率)就能补很多 y 截距"。"

"It doesn't matter where you start. It matters how fast you grow."

"起点在哪不重要——重要的是你长得多快。"

HARRY45:12

"What if you were an investor looking for adversity quotient — what would you ask founders?"

"如果你是投资人,想筛 AQ——你会问 founder 什么问题?"

CLIFF45:19

"There's a great YC question: what is a non-computer-science system you've hacked to your advantage?"

"YC 有个好问题:你 hack 过什么非计算机的系统?"

"It shows your ability to (a) build a system, (b) learn something, (c) have an inhuman amount of desire, will, and grit."

"它能看出三件事:(a) 你能搭系统,(b) 你能学东西,(c) 你有非人级的欲望、意志和坚持。"

CLIFF45:56

"Red flags: any situation where you can discover dishonesty."

"red flag 之一:任何能查到不诚实的地方。"

"What I really dislike is when people give me a stat and say "I raised X by Y%". The more questions you ask, the more you realize they changed the definition of the denominator."

"我特别反感的是:有人扔出一个指标——我多问几句——发现他偷偷换了分母的定义。"

"Second: people who talk too much. The signal per sentence needs to be high."

"第二:话太多的人——每句话的信号密度要高。"

"Third: you don't want people who are too comfortable. You want people who will be loyal."

"第三:不要太舒服的人——要忠诚的人。"

"I always ask: what made you interested to interview with Speechify? If the answer is bad, we shouldn't hire."

"我必问:你为什么想来 Speechify 面试?——答案不好——不能招。"

HARRY46:54

"What is a bad answer? 'I want to earn more money' — is that a bad answer?"

""我想多挣钱"——这是不是坏答案?"

CLIFF46:58

"Terrible answer. Never would hire that person. Not in a million years."

"史诗级烂答案。我绝不招——百年都不招。"

"Because someone else can come offer you more money, and then you leave — after we spent a year coaching you up."

"因为有人出更多钱——你就走了——而我们已经花了一年训练你。"

"A good answer: 'I have a kid with dyslexia. I have a niece with ADHD. I use the product every day.'"

"好答案是:"我儿子有阅读障碍,我侄女有 ADHD,我自己每天用。""

CLIFF47:39

"40% of people who are incarcerated have dyslexia, even though 17% of the population has it."

"美国监狱里 40% 的人有阅读障碍——人口里只占 17%。"

"But 40% of billionaires also have dyslexia."

"但 40% 的亿万富翁也有阅读障碍。"

"You coach yourself that it's okay to fail. That's exactly what startups are."

"你训练自己接受"失败 OK"——这正是创业。"

"You're stacking the odds against you, but you develop this AQ to never give up."

"你在主动让赔率不利——但你的 AQ 让你不放弃。"

CLIFF49:00

"What role doesn't exist today that will be common in 5 years? Coach."

"5 年后会很普遍但今天还没有的岗位?Coach。"

"Last year, 178,000 people applied to open engineering positions at Speechify."

"去年有 17.8 万人申请 Speechify 的工程岗位。"

"We have an asynchronous technical challenge. 19,800 people finished it. We're very good at hiring."

"我们有异步技术挑战——1.98 万人完成了。我们招人挺好的。"

"Still, we focus much more on coaching the people inside the company."

"但我们更多精力放在 coaching 公司内部的人。"

"My friend Austin Ray at Ramp — his entire job is teaching the organization how to use AI tools."

"Ramp 的 Austin Ray——他全职就是教整个公司怎么用 AI 工具。"

CLIFF52:00

"I have the same philosophy about hiring: if you can really trust someone, you can afford to pay them significantly more."

"招人也一样:你真信一个人——可以付他多得多的钱。"

"Just hire the person for more money. If they don't work out, you part ways."

"高薪招进来——如果不合适——好聚好散。"

"Always be trigger-happy to hire really good people if you see they're really good. Ignore the fact that they're expensive."

"看到真好的人——手要快——别管贵。"

HARRY52:46

"You said before: why we never hire from Google."

"你之前说过——为什么我们从不从 Google 招人。"

CLIFF52:47

"When we hire folks from Google, they come from a place of a ton of abundance."

"从 Google 招的人——他们来自极大的舒适和丰盈。"

"The hunger is not the same hunger. You need to coach them out of essentially bad habits."

"那种饥饿感不一样——你得帮他们改掉很多坏习惯。"

"In general, hiring in the Bay Area — loyalty is a lot less high than hiring in Europe."

"湾区招人总体上——忠诚度比欧洲低很多。"

"In Europe people really value a good company and a good job. It's just not the same thing in New York or San Francisco."

"在欧洲,人们真心珍惜一份好工作——纽约和 SF 不是这样。"

CLIFF56:29

"Are you worried by the number of layoffs we're seeing? Of course. I'm super worried."

"你担心现在的裁员潮吗?当然——我超担心。"

"In a world where we create more value than ever, and the Gini index gets really messed up — value accrues to the top 0.1%, not to the rest of the population — what happens?"

"我们创造的价值前所未有地多——但基尼系数极度失衡——价值流向头部 0.1%——剩下的人怎么办?"

"How do you upskill people? I don't know the answer."

"怎么帮人升级技能?——我不知道答案。"

"If you look at this graduating class of computer science students, there's basically no jobs."

"看今年毕业的 CS 学生——基本没工作。"

"You need to really make your own side projects to look impressive."

"你必须自己做出真东西——才能让人 impressed。"

"More people should be entrepreneurs."

"更多人应该去创业。"

HARRY57:07

"Do you have a bent towards younger hires today given their proclivity to be more AI native?"

"你现在偏向招更年轻的人吗——因为他们更 AI native?"

CLIFF57:13

"No. I have a bent towards older hires."

"不——我更偏向招大龄。"

"They read textbooks about software engineering. When they see what's coming out of the LLMs — they understand it."

"他们读过软件工程教科书——LLM 输出什么——他们能看懂。"

"Not all of them are adopters of AI, but some are. So you can hire people who are both grizzled veterans and good at AI."

"不是所有大龄都拥抱 AI——但有些是。你能同时找到老兵 + AI 强者。"

"We have an entrance exam that's a gigantic code repo. You can't read it with your eyes. You have to use AI."

"我们有个入职测试——一个巨大的代码仓库——你不能用眼睛读——必须用 AI。"

"You get a test at the end — how much of the repo did you break? Then 30 minutes to fix what you broke."

"最后考你——你把这个 repo 弄坏了多少——给你 30 分钟修。"

Chapter 09

60-Second Response Time + QA as Top Skill

60 秒响应时间 + QA 是 AI 时代最值钱技能
58:44 — 82:19 · "回得快的人也产出多" + 写代码变便宜审代码变贵
CLIFF58:44

"Speechify is a remote-first team. 45 AI engineers on our research team. About 150 product-facing software engineers. People in 36 countries — 200 people in total."

"Speechify 是远程优先——研究团队 45 个 AI 工程师——产品工程师约 150 人——分布在 36 个国家——总共 200 人。"

"Everybody's really good at engineering. So communications become really important."

"每个人工程都很强——所以沟通变得极重要。"

"If someone else is blocking you, it's your responsibility to inform them they are your bottleneck — multiple times a day."

"如果别人 block 你——你的责任是告诉他他是你的瓶颈——而且一天提醒多次。"

"If that person doesn't give you an answer, call them on the phone. I don't know why people don't use WhatsApp calls more."

"如果对方不回——打电话。我不懂为什么大家不多用 WhatsApp 打电话。"

CLIFF60:10

"Young people are very scared to call."

"年轻人非常怕打电话。"

"At this point, people above 45 are willing to call, and below 45 are not — that skill has atrophied."

"现在 45 岁以上的人愿意打电话——45 岁以下的不愿意——这个能力萎缩了。"

"We had a viral video the other day. I said: if you don't respond within a 3-hour period, I call it the Titanic rule."

"前阵子我们有条 viral 视频。我说:"3 小时不回消息——我管这叫 Titanic 规则"。"

"The movie Titanic is 3 hours long. Everyone should be able to go to the cinema with their families."

"电影 Titanic 是 3 小时——任何人都能陪家人看完一场电影。"

"You don't need to respond within seconds — just within 3 hours. Even on weekends. I got decimated for it. Decimated."

"不用秒回——3 小时内即可——周末也算。我因此被骂得很惨。"

CLIFF60:55

"We are so privileged to live in a world where you can work from your house, remotely, take vacation, be anywhere."

"我们活在一个特别幸运的时代——能在家工作、远程、休假、随便去哪。"

"You don't need to work 9 to 5. Take all of Tuesday off — just get the work done and own your responsibility to the team."

"你不必朝九晚五——周二可以全休——但要把活干完、对团队负责。"

"The team is a living organism. Help the organism win."

"团队是一个活着的有机体——帮它赢。"

"You need a fire under your ass that everything needs to have been shipped yesterday."

"你屁股下面要有把火——所有东西都该"昨天就发了"。"

"Speed is the number one strategic advantage for any company, especially today."

"速度是任何公司当今最重要的战略优势——尤其今天。"

HARRY61:39

"What can I do today to increase the speed of my company?"

"我今天能做什么来提升公司速度?"

CLIFF61:44

"A dream is a goal without a deadline. Always add a deadline."

"没有 deadline 的目标——只是梦想。一定要加 deadline。"

"Put it on a calendar with multiple people on the invite — 'ship to production by 3 PM Tuesday'."

"挂到日历上,拉相关人——"周二下午 3 点之前上线"。"

"Sometimes we have calendar invites that start with 'this is not a meeting, this is a deadline'."

"我们有时候日历邀请直接写"这不是会议,这是 deadline"。"

"If you know it's not going to be released by that time — update everybody else."

"如果你知道赶不上——必须更新所有人。"

CLIFF65:36

"Performance reviews are a total waste of time."

"绩效评审完全是浪费时间。"

"The problem is you didn't communicate to them ahead of time what the goal was."

"问题是你事先没把目标说清楚。"

"Typically the reason people end up on a PIP is they didn't ship something to production."

"人之所以走到 PIP——通常是因为东西没上线。"

"I have a philosophy: if you made something amazing but it's not in production, it's a waste of time."

"我的理念是:做得再好——没上线——白做。"

"You're a farmer with the best milk in the world, but you didn't put it on my doorstep. The milk is spoiled."

"你是农民,你有全世界最好的奶——但你没送到我门口——奶馊了。"

"You're a football player who didn't put the ball in the goal — you put it right before the line. No credit."

"你是足球员——没把球踢进门——只在门前。零积分。"

HARRY67:31

"You said before: QA is now the most valuable skill in the world. Why?"

"你说过:QA 现在是世界上最值钱的技能。为什么?"

CLIFF67:39

"In a world where software engineering is commoditized and design is commoditized — if you try to build with Claude Code, even with all the tools, it will not succeed in QAing itself to perfection."

"在工程和设计大面积被商品化的世界——即使给 Claude Code 所有工具——它也没法完美 QA 自己。"

"Definitely not on different devices, different phones, when the WiFi goes down — you still need a human to do that."

"尤其是跨设备、跨手机、WiFi 断了——还得人来做。"

"I'm amazed how often even very good software engineers and designers fail to put themselves in the user's shoes and test edge cases."

"让我惊讶的是——很好的工程师和设计师——常常做不到站在用户角度、测边界情况。"

"I have a friend who works directly for Elon. Elon's the best QA. He'll always find your bugs — it's super embarrassing."

"我朋友直接给 Elon 干活——Elon 是顶级 QA——他总能挑出 bug——很尴尬。"

"Let's say you built a feature with one shot of a great Claude Code prompt. The next 48 hours is just you finding bugs."

"假设你用 Claude Code 一发 prompt 做出一个 feature——接下来 48 小时——就是你找 bug。"

"That's what separates a product from a great product. It's just QA."

"普通产品和伟大产品的区别——就是 QA。"

"If you have a person who's not an amazing outcome owner — that's okay. If they're a good software engineer, they'll be good at QA."

"如果有人不擅长 outcome ownership——没关系。如果他工程不错——他通常 QA 也好。"

"Before letting them go, I'd actually rather put them in QA for a while."

"开除之前——我会先让他去 QA 试一段时间。"

CLIFF69:40

"We hate meetings."

"我们讨厌会议。"

"I like having an asynchronous culture. If there's a 60-minute meeting on your calendar, 30 minutes before you still need to be home. 30 minutes after, same."

"我喜欢异步文化。日历上 60 分钟会议——前 30 分钟、后 30 分钟你都得在家附近——这就破坏了你做事情的状态。"

"Better if I just call you for 3 minutes or text you."

"还不如我直接打 3 分钟电话或发消息。"

"The places where meetings become a core of a company are companies that don't know their direction — they use meetings to find the direction."

"会议变成核心的公司——通常是没方向的公司——他们用会议找方向。"

"We don't have managers — we have leaders."

"我们没有 manager——只有 leader。"

"I don't like the fat general sitting at the back saying 'take that hill'. I want the guy who's running up the hill first, engaging the enemy first."

"我不喜欢坐后方的胖将军喊"拿下那座山"——我要冲在最前面、最早接敌的人。"

"You should be the strongest individual contributor on your team if you're leading it."

"如果你 lead 一个团队——你应该是这个团队最强的 IC。"

CLIFF73:07

"Everybody can be a doer."

"每个人都能成为 doer。"

"Steve Jobs has this great quote: the world around us was built by people no smarter than you or I."

"Steve Jobs 那句话:我们周围的世界——是由不比你我聪明的人造出来的。"

"You live your life trying not to bump into the world too much — but you could do things. Someone just needs to tell you that you can."

"你过日子,总在小心翼翼别去撞这个世界——其实你能动手做事——你只是需要有人告诉你"你可以"。"

"My favorite thing to do: I meet someone, see how talented they are — I believe in their ability more than they believe in their own."

"我最喜欢做的事:遇到一个特别有 talent 的人——我对他能力的相信比他对自己的相信更深。"

"I give them more responsibility and opportunity. They almost always rise to the occasion."

"我给他更多责任和机会——他几乎总能爬上来。"

CLIFF75:25

"Speechify has 1.1 million five-star reviews. We have 94% of the B2C voice agents market."

"Speechify 有 110 万条 5 星评价——在 B2C voice agent 市场我们占 94%。"

"We read more to people than any other AI product in the world."

"我们给人读的内容比世界上任何 AI 产品都多。"

"There are only three reasons people share fundraising numbers — bring users in, hire, get investors — and we don't need any of those."

"披露融资数字只有三个理由:拉新、招人、找投资人——三件事我们都不需要。"

Chapter 10

What Cliff Learned Spending Time With Mr Beast

和 Mr Beast 相处学到的
82:19 — 85:36 · 团队结构 · 取舍 · "他像一个公司"
HARRY82:00

"Mr Beast is another of these characters. You slept at Mr Beast's house for 3 weeks. What did you learn?"

"Mr Beast 是另一个这样的人——你在他家住了 3 周。你学到什么?"

CLIFF82:06

"I first met Jimmy at a party at a mutual friend's house."

"我第一次见 Jimmy 是在共同朋友家的派对上。"

"I had 100 people, 8 leaders on the leadership team at Speechify. He had 100 people, 8 leaders. We ended up talking for 3 hours."

"我那时 Speechify 100 人、8 个 leader——他那时也 100 人、8 个 leader——我们聊了 3 小时。"

"I was supposed to fly to New York. He said: don't fly to New York. Come back with me to North Carolina."

"我本来要飞 NYC——他说"别去 NYC,跟我回北卡"。"

"So I flew there and I just ended up staying. We worked out. I would see how he filmed videos."

"我就飞过去——结果就住下了。我们一起健身——我看他怎么拍视频。"

"I once saw him stop a shoot with 100 people on set just to repaint the door pink."

"我亲眼见过他叫停整个 100 人 set 的拍摄——只为把门重漆成粉色。"

"I say that no genius can figure out what's going to convert in ads. Jimmy knows what's going to convert in organic content."

"我说广告里没人能预测什么会爆——但 Jimmy 知道 organic 内容里什么会爆。"

"He has the algorithm in his head. He has spent between age 13 and today just studying what converts."

"算法在他脑子里——从 13 岁到今天——他一直在研究"什么能转化"。"

"His videos are so data driven. They figure out: here's a format that works. How do we up the ante?"

"他的视频极度数据驱动——找到一个 format——然后"如何把强度再推高一档"。"

"The content that converts best is content that doesn't require language — because that works in every continent."

"最易转化的内容——是不需要语言的——因为这种内容在所有大洲都管用。"

"You want the most simple concepts to be in the video."

"视频里要放最简单的概念。"

"For storm-proofing: a bucket, hole, duct tape, water. Everybody understands these concepts."

"卖防水的:桶、洞、胶带、水——所有人秒懂。"

"For us, PDF is a complicated concept. Book is not. So our best-performing ads include books, not PDFs."

"我们这里:PDF 是复杂概念,Book 不是。所以我们表现最好的广告里——都是"书",不是 PDF。"

"He's great at hiring. Hires a ton. Tests people. If they don't work, they don't work. But he'll always try and try and try."

"他招人极强——招得多、试得快——不合适就拜拜——但他总愿意试试试。"

"Maybe the most ambitious person I know. Insane level of ambition as an operator, business person, and creator."

"我认识的人里——可能是 ambition 最大的——operator、商人、creator 三个身份的野心都疯狂。"

Chapter 11

Buying a 3x Leveraged Nvidia Option in 2022

2022 年买 NVDA 3 倍杠杆 option
85:36 — 88:13 · 技术圈看到的事和资本市场的时差 · "size for wrong, bet for right"
HARRY85:00

"Another one I loved: what led us to buying a 3x leveraged option on Nvidia in 2022 before the market caught on?"

"另一个我特别想问的——2022 年市场没醒之前——你们怎么买了 NVDA 的 3 倍杠杆 option?"

CLIFF85:15

"It was not me. It was my brother Tyler."

"不是我——是我弟 Tyler。"

"We were very excited about A100 GPUs. H100s were coming out and they were three times faster than anything before."

"我们当时对 A100 GPU 很兴奋——H100 即将出来,比之前快 3 倍。"

"We spent so much time trying to spend millions of dollars buying H100 GPUs — and no one would let us buy them."

"我们花了很多时间想拿几百万美元买 H100——但根本没人卖。"

"There was a presentation that Jensen made as part of their earnings call. It was an amazing example of how they could render a world."

"Jensen 在那场财报电话会的演示——展示了 NVDA 能 render 出一个完整世界——震撼。"

"We were like: this is a completely different level of technology."

"我们想:这完全是另一个量级的技术。"

"It became very clear that the market was mispricing Nvidia."

"很清楚——市场低估了 NVDA。"

"Tyler had even higher conviction than me. I just bought a bunch of Nvidia stock directly. Tyler bought a 3x lever option that's now like 36x up."

"Tyler 信念比我还高——我只买了正股——Tyler 买了 3 倍杠杆 option——现在涨了 36 倍。"

"My undergrad was renewable energy engineering. I put one-third of my money into Tesla in 2015 because I had context."

"我本科念可再生能源工程——2015 年我把 1/3 的钱投进 Tesla——因为我有 context。"

"I did a capstone in photovoltaics engineering. I just understood what was going on."

"我的毕设是光伏工程——我真的懂在发生什么。"

"Charlie Munger and Warren Buffett have this philosophy of the 20 punch card investment idea."

"Charlie Munger 和 Warren Buffett 有那套"打孔卡"投资理念——一生只能 20 个机会。"

"Everybody would be a much better investor if they only had 20 opportunities."

"如果每个人一辈子只能投 20 次——大家都会成为更好的投资者。"

"But when you find the opportunity, lever as hard as you can into it in a responsible way."

"但当你真找到那个机会——在负责任的前提下,杠杆拉满。"

"The biggest lesson I have in 10 years of investing is: you have to diversify to concentrate."

"我 10 年投资最大的领悟是——你必须分散,才能集中。"

"You have to be broad to have the luxury to really know who to bet on."

"先广撒——才有奢侈余地真正认出该重仓谁。"

HARRY87:26

"Would you still say Nvidia is a buy today?"

"NVDA 今天还能买吗?"

CLIFF87:28

"I don't know of a company in the history of the world that has made as much profit as Nvidia today."

"我不知道历史上有哪家公司——单年利润能像今天的 NVDA 这么高。"

"At a 35 multiple on revenue, given how much they're growing — they have a lot more juice."

"35 倍收入估值——以现在的增速——还有不少空间。"

Chapter 12

Where to Invest in Energy for the AI Era

AI 时代该往哪投能源
88:13 — end · 数据中心 · 核 · 输电网 · 哪些 second-order beneficiary 还没被定价
HARRY87:54

"How do you think about energy requirements from AI? We're now 1% of energy utilization going towards AI — that will likely 10-12x in the next few years."

"你怎么看 AI 的能源需求?现在 AI 占全球能耗 1%——未来几年很可能涨 10-12 倍。"

CLIFF88:08

"One thing I was surprised by: hydro."

"让我意外的一件事是——水电。"

"All the dams that can be dammed have been dammed. Hydro is the most efficient and easiest way to create energy — we just don't have the ability to invest in hydro more."

"能拦的坝都拦了——水电最高效、最简单——但已经没空间继续投了。"

"So that leaves you with fossil fuels, solar, or nuclear."

"剩下的就是化石燃料、太阳能、核能。"

"I do think we are going to figure out fusion nuclear. Fission nuclear is also pretty great. But fusion is coming — we just don't know when."

"我相信聚变核能会被搞出来——裂变核能也挺好——聚变要来了——只是不知道什么时候。"

"One of my biggest side projects I really want to do is build a mini fusion reactor in my garage."

"我有个 side project 想做——在自家车库里造一个 mini 聚变反应堆。"

"Not one that creates energy — just one that is contained."

"不是产能型——只要把它"封住"就行。"

"In the meantime, while we haven't figured out fusion — I'm very much a solar maximalist. All in on solar."

"在聚变还没搞定之前——我是个 solar maximalist——全 in 太阳能。"

"The theoretical limit for solar, when I was in school, was 47-49% efficiency. We're getting very close to it."

"我上学时,太阳能的理论效率上限是 47-49%——我们已经很接近了。"

"The cost of solar typically is mostly borne in installation and manufacturing — not the cells."

"太阳能成本的大头在"安装"和"制造"——不在电池片本身。"

"China has done an amazing job in bringing the cost of solar down a lot."

"中国把太阳能成本压下来这件事做得极好。"

"We just have not installed enough solar in the United States. China is outdoing us tremendously."

"美国装得太少——中国甩我们一大截。"

"There's so much public land that nobody wants — could have solar installations. Just a lot of regulation and zoning that stops it."

"美国一大堆没人要的公共土地——本来可以装太阳能——只是被法规和分区挡住。"

CLIFF90:00

"Let's talk about Solar City. The reason I invested in Tesla way back in the day was because I found Elon from a Wikipedia article from the co-CEOs of Solar City."

"说说 Solar City——我当年投 Tesla,是因为我从 Solar City 联合 CEO 的 Wikipedia 文章里找到 Elon。"

"I was like: who is this guy walking with Obama with rockets in the background?"

"我当时想:这个跟奥巴马一起、背景有火箭的人是谁?"

"Solar City was not a technology company. They were a finance company."

"Solar City 其实不是科技公司——它是个金融公司。"

"They convinced JP Morgan to advertise the value of a solar panel over 30 years — and got them to underwrite that value."

"他们说服 JP Morgan,把太阳能板的 30 年价值打散广告化——并让 JPM 给这价值担保。"

"Therefore they could loan money to consumers to install solar panels on their house."

"因此他们能给消费者贷款装太阳能板。"

"They became the largest installer of solar panels to consumers in the United States by far."

"他们成为了美国消费者太阳能板安装的绝对第一名。"

"The innovation doesn't always have to be technological. It could be a go-to-market innovation around finance or customer acquisition."

"创新不一定是技术——也可以是"金融 + 获客"层面的 GTM 创新。"

HARRY90:41

"What do you know about product market fit that most founders don't think about?"

"关于 PMF——你知道、但大多数 founder 不想的事是什么?"

CLIFF90:46

"The first one is the importance of iterations."

"第一是——迭代的重要性。"

"When we started Speechify, I was living in Palo Alto. I was a solo founder. I had a bunch of interns."

"Speechify 开始时——我住在 Palo Alto——单人 founder——带一群实习生。"

"I would walk onto Stanford's campus every day with a roll of duct tape. I'd put it over my mouth and give my phone to users."

"我每天带着一卷胶带去 Stanford 校园——把胶带贴在自己嘴上——把手机递给用户。"

"I'd see where they'd click — and they'd always click on the wrong buttons. I'd go back and fix it and try again."

"看他们点哪——他们总是点错按钮——我回去修——再试。"

"I added a big red button in the bottom right of the app that said 'help — message us'."

"我在 App 右下加了个大红按钮"求助——给我们发消息"。"

"It didn't go to Intercom or email — it went to my personal iMessage."

"它不进 Intercom、不进邮件——直接进我个人 iMessage。"

"If it was 3 AM and you messaged me, I'd respond — and you didn't need to enable notifications because I was iMessaging you."

"凌晨 3 点你发——我回——你不用开通知——因为我用 iMessage 回。"

"Every day I'd export the people who had churned given that it was 30 days since they downloaded — and I would call them."

"每天我把下载后 30 天里流失的用户导出来——给他们打电话。"

"That took a ton of energy. It's like getting kicked in the teeth every time someone tells you your baby is ugly."

"非常消耗精力——每次有人说"你的孩子丑"——就像被一脚踹脸。"

"AB tests are abdicating decision-making to the user. Sometimes you want a clear vision, go for a big goal, and work towards that until it's in the market."

"AB 测试是把决策权让给用户。有时候你要的是清晰愿景——朝着一个大目标走——直到把它推上市。"

"It goes back to AQ. The most important thing for PMF is not repositioning. Just stay in the game. Keep iterating until it fits like a glove."

"还是回到 AQ——PMF 最重要的不是 repositioning——是坚持下去——一直迭代——直到它合身得像手套。"

CLIFF94:45

"We had a security vulnerability that didn't expose user data but cost us a lot of money."

"我们有过一个安全漏洞——没暴露用户数据,但烧了我们不少钱。"

"Someone could use our models for free. I was aware of this — and chose to launch new features instead."

"有人能免费用我们的模型。我知道这事——但我选择继续发新功能。"

"Then we got hacked. The amount we lost was a couple hundred thousand dollars — not crazy."

"然后我们被黑了——损失大概几十万美元——不算夸张。"

"I went on a search of figuring out who hacked us. Found the person. The attack came from a school domain."

"我去找黑我们的人——找到了——攻击来自某个学校的域。"

"I messaged the provost of the school. We got on a Zoom call. They fessed up."

"我给那个学校的副校长发消息——上 Zoom——对方坦白了。"

"They were a kid. They didn't know it was not okay. I asked them to stop. They stopped."

"那是个孩子——他不知道这事不该做——我请他停——他停了。"

"To be honest with you, this is pretty impressive. This is exactly what I would have loved to have done at your age — if I didn't know it was not okay."

"说真的——这挺厉害的——如果我那时候不知道这是错的——我也会这么干。"

"I hired him to do personal projects for me. Small apps. He was my personal engineer for a long time."

"我雇他帮我做个人项目——一些小 app——他长期是我的私人工程师。"

"Eventually he came and lived with me — first in LA, then in Miami. Then he joined the core team."

"后来他搬来和我住——先在 LA,再到 Miami——之后加入核心团队。"

"My philosophy: if you find someone that's talented, you better work with them."

"我的哲学是:发现 talent——你最好和他合作。"

HARRY102:15

"You can buy OpenAI at $850B or Anthropic at $850B. You can also short the other. What do you do?"

"OpenAI 估值 $850B、Anthropic 也 $850B——你选哪个买、哪个空?"

CLIFF103:20

"Oh damn. I like to be contrarian. Most people in my seat would say Anthropic right now."

"哎哟。我喜欢做反共识——和我同行的大多数人现在都会选 Anthropic。"

"I love Claude Code. Oh my god. I would marry Claude Code."

"我爱 Claude Code——我的天——我能娶 Claude Code。"

"Very hard decision. To be contrarian — I would say OpenAI."

"非常难选——为了反共识——我选 OpenAI。"

"I often look at the founders. Sam Altman comes off slimy and rubs people the wrong way — but he's really good as a company operator."

"我看 founder——Sam Altman 让人觉得油腻、招人不爽——但他作为公司 operator 真的很强。"

"No one has raised as much money as Sam Altman in the history of Silicon Valley. The recruiting for OpenAI was absolutely insane."

"硅谷历史上没人融过这么多钱——OpenAI 早期挖人也太疯。"

"There's a lot of AQ in that company. OpenAI started in 2015. ChatGPT didn't come out until 2022."

"那家公司里 AQ 巨高——OpenAI 2015 年创立——ChatGPT 2022 年才出。"

"You had Columbus going to the king of Spain saying 'I want to get to India, AGI' — and on the way he found America, ChatGPT."

"就像哥伦布去找西班牙国王说"我想去印度,AGI"——结果路上发现了美洲,ChatGPT。"

"Anthropic is running circles around them right now — especially in the coding arena."

"Anthropic 现在在 OpenAI 周围绕圈跑——尤其是 coding 这一块。"

"I right now use Anthropic for everything and I do not use ChatGPT for almost anything."

"我现在所有事都用 Anthropic——ChatGPT 几乎不用。"

"One of my least favorite things in the world is ChatGPT click-baiting me at the end of its responses."

"我最讨厌的一件事——ChatGPT 在回答末尾搞 click bait。"

CLIFF105:59

"Will Grok be a meaningful competitor? Yes."

"Grok 会成为有意义的对手吗?会。"

"There's a rule: never bet against Elon."

"有条铁律:别赌 Elon 会输。"

CLIFF106:11

"What have you changed your mind on in the last 12 months?"

"过去 12 个月你改变了什么看法?"

"Last year, Mr Beast wanted to do a video where someone starts off with no money, no phone, no computer in the middle of America and has to start a business."

"去年 Mr Beast 想拍一期——一个人在美国中部、零钱、零手机、零电脑——从零创业。"

"He said I'd be a hilarious person to do it. I'll do this many Speechify integrations. I said: that's a really good deal. Let's do it."

"他说我演会很好笑——并给 Speechify 多少集成换条件——我说:这交易划算,行。"

"I spent 11 days sleeping in a Walmart parking lot in the middle of America."

"我在美国中部 Walmart 停车场睡了 11 天。"

"It took me 3 days to make enough money to get an iPhone. 5 days to make enough for a laptop. Worked at a pizza restaurant, a gym."

"3 天攒到 iPhone 的钱——5 天攒到 laptop 的钱——在披萨店和健身房打过工。"

"The video essentially fell apart. I was really upset. I take 100% responsibility, 200% responsibility for everything I do."

"视频最后拍崩了——我很不爽——我对自己做的事一向 100%、200% 负责。"

"It took me a couple months to figure it out: I became resentful. I forgot to be grateful that I was doing something so cool — it could inspire hundreds of millions of people."

"我花几个月才想明白——我变得怨恨。我忘了感恩——这件事本身那么酷——可能影响几亿人。"

"I often tell myself my greatest advantage is AQ — I keep going when other people stop."

"我一直告诉自己,我最大的优势是 AQ——别人停了我还走。"

"But that's actually not my biggest advantage. My biggest advantage is I'm typically very grateful and really optimistic."

"但这其实不是我最大的优势——我最大的优势是——我通常很感恩、很乐观。"

"In that moment when things got really tough — around day 9, very hungry — I forgot to be grateful and optimistic."

"那个 11 天的过程中——第 9 天最难、我饿得很——我忘了感恩、忘了乐观。"

"The thing I changed my mind on: the most important thing is — even when things are shitty — you've got to find something to be grateful for."

"我改变了想法的事是——即使糟透了——你也必须找出一件值得感恩的事。"

CLIFF115:49

"What excites me most about the next 10 years? Humanity is going through a new form of evolution."

"接下来 10 年我最兴奋的事——人类在经历一种新的进化。"

"Evolution is no longer a matter of your physiology but of your brain's ability to intake information faster and use it differently."

"进化不再是生理上的——而是你大脑获取信息的速度、理解和使用方式。"

"In fantasy audiobooks there's this thing called a familiar — like a dire wolf that can connect with you mentally."

"奇幻小说里有种东西叫 familiar——像一只 dire wolf——能跟你心灵相通。"

"I feel the same way about my computer, my phone, the tools I get to use — and my team."

"我对我的电脑、手机、所有工具、还有团队——都是这种感觉。"

"You can achieve more. Look at what happened to my dad with cancer — we figured it out."

"你能做到更多——看看我爸的癌——我们搞定了。"

"When you talk to a doctor, you can easily educate yourself to be as sophisticated."

"和医生谈话时——你可以快速把自己学到同等专业。"

"My dream for the world: everybody believes they have a higher level of agency than we do today."

"我对世界的梦想是——每个人相信自己有比今天更大的能动性。"

"The negative side: people think they have less agency because they abdicate all decision-making to AI."

"反面是——人觉得自己能动性更少了——因为他们把决策全交给了 AI。"

"You need people to be in the driver's seat and use the tools — instead of having the tools use them."

"我们要让人坐在驾驶位上、使用工具——而不是被工具使用。"

HARRY117:24

"Sam Altman said it well: older people use it as a replacement to Google. Younger people use it as their operating system for life."

"Sam Altman 说得好:大龄人把它当 Google 替代品;年轻人把它当人生的操作系统。"

"Cliff, this has been fantastic. Thank you so much for putting up with my wandering questions — but you've been amazing."

"Cliff,这一集真的太棒了——谢谢你忍受我跳来跳去的问题——你太厉害了。"

CLIFF117:43

"Thanks for having me."

"谢谢邀请。"