How this OpenAI engineer uses Codex + ChatGPT Work to automate everything

频道: How I AI
视频: https://www.youtube.com/watch?v=EgrI2rUmM48
原文语言: en
统计: 共 155 轮 · Nick Baumann 80 · Claire Vo 58 · Nick 2 · Claire 3

[0:00] Nick Baumann

If we go back to, you know, 2022, 2023 around the release of chatbt, I think a lot of people felt the AI then a chatbot that I say things to and intelligently says things back to me. I think with the advent of coding agents and like the coding agent harness in particular, now this AI is going out and like reading files and creating files, like running commands and this is crazy. I sell it one thing and then it figures out all the other things on it own.

如果回到 2022、2023 年 ChatGPT 刚发布那会儿，我觉得很多人对 AI 的感受是：这是个聊天机器人，我说点什么，它挺聪明地回我点什么。而随着 coding agent、尤其是 coding agent 这套 harness 的出现，现在的 AI 会自己跑出去读文件、建文件、跑命令，这太疯狂了。我只交代一件事，剩下的它全都自己推演出来了。

[0:22] Claire Vo

I haven't gone deep with voice. I want to see it from the pro. How should we be using it? What's cool about it?

我还没深入玩过语音。我想看看高手是怎么用的。我们该怎么用它？它有意思的点在哪？

[0:27] Nick Baumann

So, I'm just going to trigger voice. I've got a hotkey for it.

那我直接把语音唤起来，我给它设了个快捷键。

[0:30]

I'm listening. Go ahead.

我在听，你说。

[0:32] Nick Baumann

Okay. There should be some Amazon receipt for a a mic I just bought recently. It's for DX uses. I'm hoping you can help me expense it. Could you spin up another thread to handle taking care of that expense report for me?

好。我最近买了个麦克风，应该有一张亚马逊的收据，是 DX 团队要用的，希望你能帮我报销。你能不能另开一个 thread，帮我把这份报销单搞定？

[0:42]

The expense task is running. It'll find the Amazon receipt and stage the Navon report, but not submit it until you confirm. I'll let you know when it's ready to review.

报销这个任务已经在跑了。它会找到那张亚马逊收据，把 Navan 上的报销单填好待提交，但在你确认之前不会真的提交。可以复核了我会通知你。

[0:51] Claire Vo

If anybody has been lucky enough to have an assistant, this is how it gets done. And it is just like such a nice delegation experience to be like, "Hey, can you take care of this? Hey, can you fix this? Tell me what's going on."

如果你有幸配过助理，事情就是这么办成的。这种委派体验特别舒服——「嘿，这个你帮我处理一下」「嘿，这个帮我修一下」「跟我说说现在什么情况」。

[1:02] Nick Baumann

Typically, voice is great. You can like talk back and forth, but it's usually with a less strong model that is less capable. Whereas this, it's fully able to delegate and manage these fully, you know, five, six threads on its own, which is great cuz then you can have it do things on your behalf.

一般来说语音体验都不错，你可以来回对话，但背后往往是一个偏弱、能力没那么强的模型。这个不一样，它是真能自己去委派、去管理五六个 thread 的，这就很棒——你可以让它替你把活儿干了。

[1:17] Claire Vo

Hot take question. What do you think matters more on voice experience latency or intelligence?

问个暴论向的问题：你觉得语音体验里，延迟和智能哪个更重要？

[1:24]

[music]

[音乐]

[1:26] Claire Vo

Welcome back to How I AI. I'm Clarvo, product leader and AI obsessive here on a mission to help you build better with these new tools. Today we have Nick Bowman at OpenAI and he's going to show us some of the advanced use cases of chattex and chatt work including how you can talk to your computer to book your flights, use chatd sites to build websites you can share with anybody or nobody and use my favorite workflow which is editing creator content in codecs. [music] Let's get to it. This episode is brought to you by bolt.new New the AI app builder for people who have ideas and want to ship them. Most AI tools spit out code that looks great in a demo and falls apart the [music] second you try to do anything real with it. Or they lock you into their own platform with no real way out. Bolt is [music] different. You describe what you want to build. A startup MVP, a landing page, an internal tool, a side project, and Bolt generates productionready code in minutes. Connect Stripe [music] or any other MCP. Hook up your domain and deploy it live. Founders are using Bolt to build businesses doing real revenue. Product managers are shipping prototypes their teams actually use. Designers and marketers are launching campaigns without waiting in line. Anyone can build. Engineering can ship. Everyone wins. You just need an idea and a weekend. Check it out at bolt.new/howi aai. Nick, welcome to How I AI. Hello Claire, thanks for having me.

欢迎回到 How I AI。我是 Claire Vo，产品人，也是个 AI 狂热分子，我在这儿的使命是帮你用这些新工具做出更好的东西。今天请到的是 OpenAI 的 Nick Baumann，他会带我们看几个 Codex 和 ChatGPT Work 的进阶用法，包括：怎么用嘴跟电脑说话就把机票订了、怎么用 ChatGPT Sites 做出可以分享给任何人（或者谁都不分享）的网站，还有我最喜欢的一条工作流——用 Codex 剪创作者内容。[音乐] 我们开始吧。本期由 bolt.new 赞助——面向「有想法、想把它做出来」的人的 AI 应用搭建工具。大多数 AI 工具吐出来的代码，demo 里看着很漂亮，[音乐] 一上真活儿就散架；要么就把你锁死在它自己的平台上，根本没有退出的路。Bolt 不一样。[音乐] 你描述你想做什么——创业公司的 MVP、落地页、内部工具、业余项目——Bolt 几分钟内生成可上生产的代码。接上 Stripe [音乐] 或者任何别的 MCP，绑上你的域名，直接 deploy 上线。已经有创始人用 Bolt 做出了真有收入的生意，产品经理用它交付团队真正在用的原型，设计师和市场同学不用排队等排期就能把 campaign 发出去。人人都能做，工程能交付，皆大欢喜。你只需要一个想法和一个周末。去 bolt.new/howiai 看看。Nick，欢迎来到 How I AI。——你好 Claire，谢谢邀请。

[3:01] Claire Vo

I am going to make you laugh because I might be the number one Codeex fan fanboy tech bro fanboy in the world. I am like constantly telling people, wait, but have you have you tried code? Have you tried Codeex yet? It is just my daily driver. So, I'm so excited to have you on because even though I use it all the time and for some things that you and I think both think are valuable, I have not touched most of the new features that have come out in the last week, it's really hard to keep up. And so, I'm really psyched that you're here to just show us some of the new stuff and maybe how you use Codeex both for work and life. So tell us what are the things in Codex you feel like people don't really know about or or brand new that you think are completely going to change how how we all work?

我要说件能把你逗笑的事：我可能是全世界头号 Codex 迷弟。我天天跟人念叨：等等，你试过 Codex 没有？你试过没？它就是我的日常主力工具。所以特别高兴请到你——虽然我天天在用，而且用在一些你我都觉得很有价值的场景上，但过去这一周出的新功能我基本都还没碰过，实在跟不上节奏。所以我特别兴奋你能来给我们看看新东西，还有你自己工作和生活里是怎么用 Codex 的。那说说看，Codex 里有哪些东西是你觉得大家其实不太知道的，或者刚出的、你觉得会彻底改变我们工作方式的？

[3:47] Nick Baumann

Yeah, I feel like you know with yesterday's release of voice um the chatbt codeex app, it's kind of like the amalgamation of all these primitives we've been like putting together kind of quietly and now we're super useful. So a lot of people don't know that the chatbt app you can ask it to create threads, you could ask it to message your existing threads and these threads actually talk to each other too. And it's like not super kind of forward in the app, but these are all capabilities. And now we've added this voice orchestration layer really where now you can, you know, the hotkey you can trigger this this orb that like pops in your screen and no matter where you are in your computer, you can talk to it. it'll talk back to and it can see your screen, can see what you're working on, and because we have those primitives of see your existing threads, create new ones, talk to them, it can like basically manage your entire chatb for you.

嗯，我觉得随着昨天语音功能、还有 ChatGPT Codex 这个 app 的发布，它有点像是我们这一路悄悄攒下来的那些基础能力（primitives）终于合体了，现在特别好用。很多人不知道，在 ChatGPT app 里你可以让它创建 thread、可以让它给你已有的 thread 发消息，而且这些 thread 之间还能互相对话。这些在 app 里不算显眼，但能力都在。现在我们又加了语音这层编排——按快捷键就会弹出一个光球，不管你在电脑上干什么，都能跟它说话，它也会说话回你，而且它能看见你的屏幕、看见你正在做什么。再加上底层那些 primitives——看你已有的 thread、创建新的、跟它们对话——它基本上可以替你把整个 ChatGPT 管起来。

[4:40] Claire Vo

Yeah. One of the things that I I don't want people to miss before we even get into like the fancy voice stuff is I really do feel like most people do not know the meta capabilities of codecs and even like we won't say but like the the other one in that like its ability to not only spin off sub agents but actually start completely new threads, fork threads, look through old data, look through its own memory. Um, it's it's a really good mechanism to let the AI manage context and decide when forking off different tasks makes a lot of sense. And so I've been doing that a lot. But I actually had to discover that capability sort of organically when one day I was sitting there and it was probably five six was like I'm going to kick off a new thread and I was like you're going to do what? Like what now? And so I do I do feel one of the challenges with these core kind of platforms like codeex's and chatbt is people just don't know how to discover the features and they're so rich and so do you have any you know like we'll talk about voice but are there any other like sort of edge features of codeex or chatgpt that you feel like people underuse not just this this thread one

对。在进到那些花哨的语音功能之前，有一点我不希望大家错过：我真觉得大多数人不知道 Codex 的那些「元能力」，甚至包括——我们就不点名了——另外那家的那个。它不光能拆出 sub agent，还能真的开全新的 thread、fork thread、翻旧数据、翻它自己的记忆。这套机制特别适合让 AI 自己管理 context、自己判断什么时候把任务 fork 出去更合适。所以我最近一直这么用。但这个能力其实是我误打误撞发现的：有一天我坐在那儿，大概是 5.6 那个版本吧，它突然说「我要开一个新 thread」，我心想「你要干嘛？什么玩意儿？」所以我确实觉得，Codex、ChatGPT 这类底座型平台有个难题——功能太丰富了，大家根本不知道怎么发现。那除了 thread 这个之外，你觉得 Codex 或 ChatGPT 里还有哪些偏边缘的功能是大家用得太少的？

[5:52] Nick Baumann

the chat GPT work experience you know people are talking about you know should we using chatbt work or codecs.

ChatGPT Work 的体验。大家老在讨论：我们到底该用 ChatGPT Work 还是 Codex。

[5:57] Nick Baumann

I think chat GBT work on the web specifically mobile. I mean that's where I use it mostly is so insanely powerful. It it's one it's it's essentially like the codeex harness which has this incredible compaction. You have all your plugins. So if you've like already connected to Gmail, Gcal, like whatever you use, it's already there

我觉得网页版的 ChatGPT Work，尤其是手机端——我主要就是在手机上用——强得离谱。首先，它本质上就是 Codex 那套 harness，带着非常出色的 compaction（上下文压缩）。你所有的插件都在，如果你已经连了 Gmail、Google Calendar、随便你用什么，它们都已经在那儿了。

[6:15] Nick Baumann

and then you have it all accessible in your phone. So you can like create these automations um also as heartbeats um and it's already there. So, like I mean a few times a day I'll just get like a notification saying, "Hey, like I got this email about this thing you bought. Are you sure you bought that?" Or like, "Hey, this package is arriving. Like, make sure you grab that." And I'm not watching my email or my finances, but that's just kind of being monitored for me. That's that's really magical experience.

然后这一整套在你手机上都能用。所以你可以把这些自动化做成 heartbeats，直接就跑在那儿。像我一天会收到好几次通知，比如「嘿，我看到一封邮件说你买了这个东西，你确定是你买的吗？」或者「嘿，这个包裹要到了，记得取一下」。我根本没在盯我的邮箱和账单，但这些事有人替我盯着。这体验真的很神奇。

[6:41] Claire Vo

Okay. I love it. All right. So, let's get into Chad Voice and what you what you use it for because I it came out yesterday. Um, I've been super busy and around humans. We're going to talk about that a little bit later. So, I haven't gone deep with voice. I want to see it from the pro. How should we be using it? What's cool about it?

好，我喜欢。那我们进到 ChatGPT 语音，聊聊你都拿它干什么。它昨天才发布，而我最近特别忙、还老得跟真人打交道——这个我们待会儿会聊到——所以我还没深入玩过语音。我想看看高手是怎么用的。我们该怎么用它？它有意思的点在哪？

[7:01] Nick Baumann

Yeah, I I think it's really magical. So, I'm just going to trigger voice. I've got a hotkey for it.

嗯，我觉得它真的挺神的。那我直接把语音唤起来，我给它设了个快捷键。

[7:06]

I'm listening. Go ahead.

我在听，你说。

[7:08] Nick Baumann

I'm hoping you can help me handle a few logistical things right now. Can you can you see the the offsite I've got pull up in Slack?

我希望你能帮我处理几件杂事。你能不能看到我 Slack 里打开的那个 offsite（团队异地活动）？

[7:16]

I'll take a look. Let me check that Slack view. Yep. It's the DX team offsite in Paris August 26th through 28th. You're due in Tuesday with the program starting Wednesday morning and wrapping Friday afternoon. Flights and hotel are supposed to be booked in Navon this week.

我看一下，让我看看那个 Slack 窗口。有了，是 DX 团队在巴黎的 offsite，8 月 26 到 28 号。你需要周二到，正式日程周三上午开始、周五下午结束。机票和酒店要在这周内通过 Navan 订好。

[7:35] Nick Baumann

All right. So, I I muted myself so that Codex doesn't hear me right now. But what happened there is it can already see my screen. You can see I've got Slack opened up. And what it did is it took an appshot. And so an appshot is something that you can do where you whatever window you have open, you just hit command command and it essentially takes like a smart screenshot loaded with context. I can use computer use or browser use on it instantly. But you know kind of a meta feature. Codex voice already has that built in. So I'm hoping you can help me book my flights and my hotel for the Parisoft s night. I'm also thinking about sticking around to the weekend and maybe a couple days after. Could you maybe like check my calendar, see what my availability is like and see if it's if it's feasible for me to stick around?

好，我现在把自己静音了，这样 Codex 听不到我说话。刚才发生的事是：它已经能看见我的屏幕了，你能看到我开着 Slack。它做的动作是拍了一个 App Shot。App Shot 是这么个功能——不管你当前开着哪个窗口，连按两下 Command，它就会拍一张带上下文的「智能截图」，我可以立刻对它用 computer use 或者 browser use。这本来算是个偏元的功能，但 Codex 语音里已经内置了。——所以我希望你能帮我订巴黎 offsite 的机票和酒店。我还在考虑顺便留到周末，可能再多待几天。你能不能查一下我的日历，看看我档期怎么样，留下来现不现实？

[8:20]

I'll check your calendar against the off-site dates and then look at travel options.

我会把你的日历跟 offsite 的日期比对一下，然后再看出行方案。

[8:24] Claire Vo

Is this using the connectors?

这是在用 connectors 吗？

[8:27] Nick Baumann

Good question. Yeah, so this is using the plugins underneath the hood right now. So, it's not having to actually go

好问题。是的，现在底层用的是 plugins，所以它不用真的跑去……

[8:32]

after the offsite through Monday, August 31st. A Tuesday, September 1st return would conflict with a late afternoon meeting. The simplest plan is returning Monday, August 31st. Do you want me to plan around that?

……offsite 结束之后一直到 8 月 31 号周一都空着。9 月 1 号周二返程会跟一个下午晚些时候的会议撞上。最省事的方案是 8 月 31 号周一回。要我按这个来安排吗？

[8:45] Nick Baumann

You know what? I'm thinking I might just take PTO on that Monday and Tuesday. I'm flying out of SFO. So, why don't you just go ahead and use Non. Um, spin up another task to figure out the travel. Pick a hotel that's reasonable and within our company policy, and then when it's ready, let me know and I'll I'll click book now.

这样吧，我想那个周一和周二干脆请 PTO 算了。我从 SFO 出发。你就直接用 Navan。再另开一个任务把行程搞定，挑一家价格合理、符合公司差旅政策的酒店，弄好了告诉我，我来点「立即预订」。

[9:06]

Sure, I'll sort that out. I'll find options for flying out of SFO, arriving Tuesday before the Wednesday start, and returning after Tuesday, September 1st, then bring you choices that fit policy for you to book. So, is this chat going to be in Oh, sorry. [laughter]

没问题，我来搞定。我会找从 SFO 出发、周三开始前的周二到达、9 月 1 号周二之后返程的方案，然后把符合政策的几个选项拿给你，由你来下单。——（Claire）那这个对话会不会……哦，抱歉。[笑]

[9:25] Claire Vo

I I I'm talking to I'm talking to his CEX. Invite Claire to the Paris offsite. Put her up in the rits.

我我我在跟——我在跟他的 Codex 说话：把 Claire 也邀请到巴黎 offsite，给她安排丽思酒店（the Ritz）。

[9:32] Nick Baumann

Get her in business class. And honestly, we we don't we don't really have a Paris off site scheduled yet, but so much of our DX team is French that I'm hoping that I can like, you know, start seeding the the ideas here.

给她订商务舱。说实话，我们其实还没定下巴黎的 off-site，但我们 DX 团队一大半都是法国人，所以我挺想先在这儿埋点种子，把这个念头带起来。

[9:44]

That task is running now. Comparing flights from SFO at a practical hotel for August 25th through September 2nd. I'll share the suggested itinerary when it's ready.

任务已经跑起来了。正在对比从 SFO 出发的航班，以及 8 月 25 日到 9 月 2 日期间比较实用的酒店。行程建议好了我就发给你。

[9:54] Nick Baumann

Cool. So, we've got this task already running and make sure it uses my existing Chrome browser for this one.

不错。这个任务已经在跑了——另外确保它这次用我现有的 Chrome 浏览器。

[10:01]

I'll check that. I've updated the travel task to use your existing Chrome session and keep it to preparing options, not booking.

好的，我确认一下。已经把这个差旅任务改成使用你现有的 Chrome session，并且只负责整理选项，不做预订。

[10:11] Nick Baumann

Awesome. So, right now it just it just sets a message back to that task. It's already running. But I think while we've got that going, I'm going to ask it to do another thing for me. While I've got you, I've got an email. I think there should be some Amazon receipt for a a mic I just bought recently. Um, it's for DX uses. I'm hoping you can help me expense it. Could you spin up another thread to handle taking care of that expense report for me?

很好。刚才其实就是往那个已经在跑的任务里回传了一条消息。不过趁它跑着，我想再让它帮我办件别的事。既然你在——我这儿有封邮件，应该是我最近买的一支麦克风的 Amazon 收据，是 DX 团队要用的，我想让你帮我报销。你能不能另外开一个 thread，专门把这份报销单搞定？

[10:35]

Sure, I'll take a look. I've started that expense task and it'll use your existing Chrome session. The expense task is running. It'll find the Amazon receipt and stage the Navon report, but not submit it until you confirm. I'll let you know when it's ready to review.

没问题，我看一下。报销任务已经启动，会用你现有的 Chrome session。任务正在跑，它会找到那张 Amazon 收据、把 Navan 上的报销单准备好，但在你确认之前不会提交。可以复核的时候我再通知你。

[10:52] Nick Baumann

Awesome. So, those are both running in the background. I've I've muted our voice so we can talk without being interrupted. Um, but we can already see that in Chrome, um, this task is using the browser. It's figuring out our flight options. Um, it'll figure out the hotel.

很好。这两个任务现在都在后台跑。我把语音静音了，这样我们说话不会被打断。不过在 Chrome 里已经能看到，这个任务正在用浏览器筛航班选项，接下来还会去找酒店。

[11:08] Nick Baumann

Frankly, this is something I just enjoy not having to do. Um, and it's nice to just have this handle in the background for me.

老实说，这类事我就是很享受「不用自己干」的感觉。有东西在后台替我处理掉，挺爽的。

[11:13] Claire Vo

What I like about voice is I do think sometimes people with AI really struggle still with like blank chat window syndrome, which is like they look at and they're like, "What am I going to ask? What am I going to ask this brilliant AI to do? And so what I love about first just like voice transcript to text which a lot of people have been using is it it is what our beloved multi-time guest um Hillary calls the yappers API which is like the best highest bandwidth way to communicate with an LLM is just to yap out loud and just contact them. But what I love about this like voicetovoice interaction is if anybody has been lucky enough to have an assistant like this is how it gets it gets done and it is just like such a nice delegation experience to be like hey can you take care of this hey can you fix this tell me what's going on blah blah blah blah blah and so I'm I'm guessing the hypothesis here with this voice kind of experience is there's going to be like better discovery of use cases and sort of like more sprawling use cases because people can sort of delegate in a more natural way as opposed to having to like sit there and think through how to instruct kind of in text the the models.

我喜欢语音的一点是：我确实觉得很多人用 AI 到现在还卡在所谓的「空白对话框综合症」——盯着输入框心想「我该问点啥？我该让这个聪明得要命的 AI 干点啥？」所以我特别喜欢——先说最基础的那层，语音转文字，很多人已经在用了——它其实就是我们那位多次做客的嘉宾 Hillary 说的 yappers API：跟 LLM 沟通带宽最高的方式，就是张嘴一顿输出。而这种语音对语音的交互我更喜欢，因为如果你有幸配过助理，就知道事情本来就是这么推进的——那种委派的体验特别舒服：「嘿，这个你帮我搞定一下」「这个你修一下」「跟我说说现在什么情况」，噼里啪啦一通说完就完了。所以我猜，语音这套体验背后的假设是：用户会发现更多用例、用例也会更发散，因为大家能用更自然的方式去委派，而不用坐在那儿琢磨怎么把指令用文字写清楚喂给模型。

[12:27] Claire Vo

Have you found that like there are specific things that you reach for voice with that you don't that you don't like type with your human human fingers? Are you a full voice? Like where are we?

你有没有发现，有些事你就是会用语音、而不会用手一个字一个字敲？你现在是彻底转语音了吗？处在什么状态？

[12:36] Nick Baumann

Yeah, a few thoughts there. One, I think it's like this iteration itself is like kind of a new primitive in that

有几点想法。第一，我觉得这一代交互本身就是一种新的 primitive，因为——

[12:43]

y

嗯。

[12:44] Nick Baumann

typically voice is great. you can like talk back and forth, but it's usually with a less strong model that is less capable.

通常语音体验是不错，你可以来回聊，但背后往往是一个能力更弱的模型。

[12:50] Claire Vo

Yep.

对。

[12:50] Nick Baumann

Whereas this, it's fully able to delegate and manage these, you know, fully, you know, five, six sole threads on its own, which is great because then you can have it do things on your behalf. Other thoughts there? Um, I find the like the thinking process and voice just a lot better when it comes to, you know, I could dictate for, you know, 2 minutes, send off a long paragraph, get back a couple long paragraphs and read those and it's really hard for me to get in flow. Um, I think of like planning in voice to be really really effective.

而这个不一样，它能完整地委派并管理五六个独立线程，这点很棒，因为你可以让它替你去干活。还有什么想法呢……我觉得「思考」这件事用语音顺畅太多了。比如我口述两分钟发出去一大段，再收回来几大段文字去读，这样我很难进入心流。而用语音做规划，我觉得效率特别高。

[13:26] Claire Vo

Hot hot take question.

问个暴论式的问题。

[13:29] Claire Vo

What do you think matters more on voice experience latency or intelligence? If you could only pick one, I'm gonna make you pick one.

你觉得语音体验里更重要的是延迟还是智能？只能选一个的话——我一定逼你选一个。

[13:35] Nick Baumann

I would say when I use chatbt, I put it on high intelligence for voice.

我自己用 ChatGPT 的时候，语音是开在「高智能」档的。

[13:40] Nick Baumann

I think there's there's like a middle ground. I think if there are good tools it can delegate to, then I care more about latency.

不过我觉得中间有个平衡点。如果它手上有足够好的工具可以委派出去，那我更在意延迟。

[13:47] Claire Vo

Yeah.

嗯。

[13:48] Nick Baumann

Um, so I don't know. I guess I I still say latency.

所以……说不好。我猜我还是选延迟吧。

[13:53] Claire Vo

Yeah. I think I think it's the I think it's what we're going to be talking and this is the Clairvo prediction. I think it's we're going to talk more and more about

对。我觉得——这是 Claire Vo 的预测——接下来我们会越来越多地讨论

[13:58] Claire Vo

latency in the in the second half of the year because I just think it's like almost the the thing to differentiate on right now is like how real time can these experiences really be? I see a lot of people abandon great great AI workflows because of a spinner or because of a delay. And so like the more you can close that gap, the more I think people can discover cool things and actually pull through and get some some stuff done.

延迟，尤其是今年下半年。因为我觉得现在最能拉开差距的点，就是这些体验到底能做到多实时。我见过太多人因为一个转圈、因为一点延迟，就放弃了原本很棒的 AI 工作流。所以这个差距压得越小，大家越能发现有意思的玩法，也越能真的走完流程、把事做成。

[14:22] Nick Baumann

Yeah. I feel like so many product assumptions in design and AI tooling is like built around this like there's going to be latency, there's going to be delay.

对。我感觉现在 AI 工具的产品设计里，好多假设都建立在「一定会有延迟、一定要等」上面。

[14:31] Claire Vo

Yep.

没错。

[14:31] Nick Baumann

And the more we shrink that we can kind of drop those assumptions and new products will, you know, will emerge.

延迟压得越小，我们就越能扔掉这些假设，新的产品形态自然就冒出来了。

[14:36] Claire Vo

Amazing. Okay. So, so what we've seen here just to to reiterate for people who are maybe not watching or listening is you can hotkey spin up voice. It can manage these like very intelligent five six threads. You can delegate tasks. It can be hooked up to plugins. It can use your browser. It can use appshots. It can use all these things. And it's just like a very high bandwidth, low friction way to talk to super intelligence that can get nice stuff done. You you mentioned that a little bit earlier that you're also loving like mobile and and this combination of like dictation and managing through mobile. Can you tell me how like being able to do some of this on your phone has changed your workflow? So maybe like give me an example of that.

太棒了。好，我给没看画面、只在听的朋友再复述一遍刚才发生了什么：你可以用快捷键唤起语音，它能管理五六个很聪明的 thread，你可以把任务委派出去，它能接各种插件、能用你的浏览器、能调用各种 app，这些全都做得到。所以这是一种带宽极高、摩擦极低的方式，去跟一个真能把事办成的超级智能对话。你刚才还提到你也很喜欢手机端——口述加上在手机上管理任务的这种组合。能不能讲讲，能在手机上干这些事之后，你的工作流发生了什么变化？给我举个例子？

[15:22] Nick Baumann

So I put together this site and so this is just a an aggregation of all these really cool tweets that you know both people in like the codeex and chatbt community and also people work at openai have shared with how they're using codeex chatbt. Um, this was done just with with chatbt, like the chatbt app where I had it use it own inapp browser, go through Twitter for like a couple hours and then embed these links and kind of add a little bit of, you know, description to them. And now what I do is I've got a thread on my work phone that is essentially managing this site and when I see something I like, I just drop in that link. I say add it and it does. We also have this prompts tab and I do the same thing there. If I have like an idea for how somebody could be maybe he's like you newer to like chatbt work I'll just like dictate my phone explain like hey there's this prompt for like having it manage your inbox and your finances can you add this to the site as well and in terms of actually editing the site um deploying it that can all happen from chatbt work on mobile

我做了这么一个站点，它其实就是把一堆特别酷的推文聚合起来——有 Codex 和 ChatGPT 社区里的人发的，也有 OpenAI 同事发的，讲他们各自怎么用 Codex 和 ChatGPT。这个站完全是用 ChatGPT 做出来的，就是 ChatGPT app，我让它用自带的 in-app 浏览器，在 Twitter 上刷了大概几个小时，然后把这些链接嵌进来，再给每条加一点描述。现在我的做法是：工作手机上留着一个 thread 专门管这个站点，看到喜欢的内容我就把链接丢进去，说一句「加上」，它就加好了。我们还有一个 prompts 页签，也是同样的玩法。比如我想到一个点子，可能是给刚上手 ChatGPT Work 的人用的，我就对着手机口述：「有这么一个 prompt，是让它帮你管理收件箱和财务的，你把它也加到站点上。」至于站点的实际编辑和 deploy，全都能在手机上的 ChatGPT Work 里完成。

[16:31] Claire Vo

and I think what people are not seeing here which again so many new features coming out that I think people aren't taking advant antage of is the domain here is chatgbt.site. So you can now deploy these like artifacts these sites live. So can you just walk us through a little bit how how that works or maybe we can do one just so people can like understand the power of these sites and and how you might use them or what their limitations are even.

我觉得大家可能没注意到——新功能实在太多了，很多人根本没用起来——这里的域名是 chatgpt.site。也就是说，你现在可以把这类 artifact、这类站点直接 deploy 成线上可访问的。你能不能带我们过一遍它是怎么跑起来的？或者干脆现场做一个，让大家理解 ChatGPT Sites 这套能力有多强、你会怎么用它，以及它的限制在哪儿。

[16:58] Nick Baumann

Yeah, I think like for like 10-second description for the technical audience,

行。我先给偏技术的观众来个 10 秒版的解释吧，

[17:03] Nick Baumann

it supports a SQL database. Um, it has S3 storage for files and it even has environmental environmental variables you can put there. For the non technical audience, it's essentially a website that you can store things on, which is great. And you can also control who sees it. You can you can filter by by email. So, it can be private, it can be fully public. That's up to you. But what we could do, if you're interested, is we could try building a site live.

它支持 SQL 数据库，有 S3 存储可以放文件，甚至还能配环境变量。对非技术的观众来说，它本质上就是一个能存东西的网站，这很棒。而且你还能控制谁看得到——可以按邮箱做过滤，所以它可以是完全私有的，也可以完全公开，全看你自己。要是你有兴趣，我们可以现场建一个站试试。

[17:29] Claire Vo

Let's do it.

来吧。

[17:30] Claire Vo

And see how to build live.

现场演示怎么建。

[17:31] Nick Baumann

Yeah. All right. Uh, so let's jump into it. So, I've got an idea for a site. I'm here right now with Claire Vo. She hosts the How I AI podcast, and we're hoping that we can build a site that takes all the best tips from the videos on her YouTube channel and makes it so you can like easily filter through them and have quick links to like different timestamps on her videos.

好，那我们开始。我这儿有个做网站的想法。（开始口述 prompt）我现在和 Claire Vo 在一起，她主持 How I AI 播客。我们想做一个网站，把她 YouTube 频道里所有视频中最好的那些技巧汇总起来，让人可以很方便地筛选，还能直接点链接跳到视频里对应的时间点。

[17:58] Claire Vo

Make it match the how I AA branding. We use good old AI blurp purple um and black and white and um let's have it if if possible categorized by tool. So like what AI tool is being used in the workflow and also function. So if it's for designers, it's if it's for engineers, it's for personal productivity, any of those things. Otherwise, I'm excited to see what you come up with.

让它符合 How I AI 的品牌风格。我们用的是那个经典的 AI 蓝紫色（blurple），再配黑白。另外如果可以的话，按工具分类——就是这个工作流里用的是哪个 AI 工具；再按职能分一层，比如是给设计师的、给工程师的，还是个人效率类的，诸如此类。其他的我就期待你能做出什么来了。

[18:26] Claire Vo

Awesome. That sounds great. and go ahead and deploy it privately and then we'll take a look at it you know when it's ready and then we can deploy it publicly from there if it if it looks good. While this is running, let me make you make you laugh because this um two human chat to chat GBTE I did recently with my mother and um she was like stuck on an administrative task and she was like getting all frazzled about she's like I'm going to have to look up all this stuff and blah blah blah blah blah blah blah and I was on the phone with her and I was like mom just hold on one second and I opened up chat GPT on my desktop and I turned on voice and then I was like can you just please tell what you need to do. And I turned the speaker phone on my phone and I put it up to the computer and had her like babble kind of like we did and I asked her a couple questions and then I was like, "Thanks." Pressed enter and I was like, "I did it for you. I did it for you." [laughter] So, I really do feel like people are underusing like twoerson voice chat. You can do it on your phone. You can use speaker phone. like it is this really nice again like high bandwidth way to get requirements into the system and it doesn't just have to be you alone.

（Nick）很好，听起来不错。先私有部署，等它好了我们看一眼，如果效果不错再从那儿公开发布。（Claire）趁它在跑，我讲个好笑的事儿。前阵子我跟我妈来了一次「两个人一起对着 ChatGPT 说话」。她当时卡在一件行政流程上，急得不行，一直念叨说我还得去查这个查那个、balabala 一大堆。我正跟她通着电话，就说：妈，你等我一秒。然后我在电脑上打开 ChatGPT，开了语音，跟她说：你把你要办的事说一遍就行。我把手机开免提举到电脑前，让她像我们刚才那样一通说，我又追问了她几个问题，然后说了句「谢谢」，一按回车——我说：好了，我替你办完了。（笑）所以我真觉得大家严重低估了「两个人一起语音聊」这种用法。手机上就能做，开个免提就行。这又是那种带宽极高的方式，能把需求直接灌进系统里，而且不一定非得你一个人对着它说。

[19:35] Nick Baumann

Yeah. I I've had times where

是啊。我也遇到过这种时候，

[19:37] Nick Baumann

you know I need context with somebody I work with.

就是我需要从某个同事那儿拿到上下文。

[19:39] Claire Vo

Yeah.

嗯。

[19:40] Nick Baumann

And I'm like please just like ramble dictate into Slack and send me like garbage. I don't care. And they're like no let's just like set up a meeting. And I'm like no don't don't do that. I don't need that. [laughter] I don't need this to be like pretty or anything. just just tell me that's fine.

我就跟他说：你随便对着 Slack 口述一段就行，发过来多乱都无所谓，我真不在乎。结果对方说：要不咱们约个会吧。我说：别别别，不用，我不需要开会。（笑）我不需要它多规整多漂亮，你说出来就行。

[19:58] Claire Vo

So, are we at this meeting could have been a voice note like level?

所以我们已经到了「这个会本来发条语音就够了」的阶段？

[20:02] Nick Baumann

Yeah, I think so. I mean, all it's it's more like give me information that unstructured however it is that my agent can can decipher and that's that's good enough for me.

对，我觉得是。更准确地说是：把信息给我就行，多不结构化都无所谓，只要我的 agent 能读懂，对我来说就够了。

[20:12] Claire Vo

This episode is brought to you by Hyper Agent, the platform for deploying always on agents that actually run your business. With Hyper Agent, you build agents in the cloud and deploy them where your work already happens, like Slack, Telegram, or email. [music] An agent will scan your inbox and draft replies to vendor follow-ups. Another monitors [music] competitors and spins up rich ad kits and landing pages. A third notices a deal going cold in Salesforce and writes the save email with full account context. These aren't chat bots waiting for a perfect prompt. They're proactive, learning your preferences, retaining your playbooks, and getting better with every run. One user built four agents to run an outbound sales pipeline, prospecting, outreach, follow-ups, CRM updates, all in a single afternoon. No local setup, no VPS bills, no fragile permissions on your laptop, just powerful agents with full control over skills, tools, and guard rails. [music] Hyper Agent was built by the team behind Air Table and How I AI listeners get $1,000 in free inference to start building. Claim yours at hyperagent.com/howi ai. Okay, so this is going to go spin up. We're going to let it kind of like whirl a little bit. Maybe we'll come back to what what we see. But there's one other use case that both you and I really like in codeex um

本期节目由 Hyper Agent 赞助——一个部署常驻 agent、真正帮你把业务跑起来的平台。用 Hyper Agent，你在云端搭好 agent，然后部署到你本来就在用的地方，比如 Slack、Telegram 或者邮箱。一个 agent 扫你的收件箱，替你起草回复供应商的跟进邮件；另一个盯着竞品动态，自动生成整套广告素材和落地页；第三个发现 Salesforce 里某笔单子快凉了，就带着完整的客户背景写一封挽回邮件。这些不是那种要等你写出完美 prompt 才动的聊天机器人，它们是主动的，会学你的偏好、记住你的打法，每跑一次都更好用。有个用户一个下午就搭了四个 agent，把外呼销售全流程跑通了——找线索、触达、跟进、更新 CRM。不用本地配置，不用付 VPS 账单，也不用在自己笔记本上折腾一堆脆弱的权限，就是一套强力 agent，技能、工具和护栏都由你完全掌控。Hyper Agent 出自 Airtable 原班团队，How I AI 的听众可以领 1000 美元免费推理额度来上手，去 hyperagent.com/howiai 领取。好，那这个建站任务就先跑起来了，让它转一会儿，待会儿我们再回来看看效果。不过还有另外一个 Codex 的用法，是你我都特别喜欢的——

[21:38] Claire Vo

that I thought you could show and particularly how how you prompt it which is codeex for video editing and for those that don't know it I do a lot of this for um for the podcast. I also do a lot of this just for like general work stuff. I'm cutting a lot of videos. So, um, tell me like how you came to this use case, why you feel it's so useful, and then how how easy is it to prompt kind of like video editing?

我觉得你可以演示一下，特别是你到底怎么写这个 prompt 的，就是用 Codex 剪视频。给不了解的人说一句，我自己就干了大量这种活儿——播客要剪，平时工作上也剪很多，我的视频剪辑量非常大。所以你说说，你是怎么走到这个用法上的，为什么觉得它这么有用，以及「用 prompt 指挥剪视频」这件事到底有多容易？

[22:04] Nick Baumann

You know, I I do a little bit of content as part of the DX team, and, you know, we're trying to reach people that might not be inside like the Twitter verse, which, you know, there's there's there's a lot of us there, but there's more of us not there. Um, and I think explaining, you know, how you can use these tools um, in really relatable ways where you're explaining to like a camera um, in like 45 seconds is great. And so I started just kind of like going to a park, filming myself, filming my camera, getting a, you know, a smattering of little clips. And so this is from like this is from yesterday. I went to the park. These are like most of these are takes that we're not going to use. Um, but what I'll do is I'm literally going to drag these into chatbt work. And then so what I've done and I'm going to share this, but I've created a plugin called UGC video. Actually, let me get the one that I use for I made a version that has like open eye branding, but you can actually see it here. And so this basically after having gone through like a few runs of this, what I've noticed is that just having like some guidance around like, hey, make sure that we're not entering like the safety zone with Instagram ads or let's do these formats. We want 9x6 also 4x5.

我在 DX 团队也做一点内容。我们想触达的是那些不在 Twitter 圈子里的人——我们这帮人在 Twitter 上确实不少，但不在上面的人更多。我觉得用特别接地气的方式，对着镜头 45 秒讲清楚这些工具怎么用，效果非常好。所以我就开始去公园，自己给自己拍，攒一堆零碎的小片段。这些就是昨天拍的，我去了趟公园。这里面大部分其实是我们不会用的废镜头。但我接下来要做的，就是直接把它们拖进 ChatGPT Work。我做了一个叫 UGC video 的 plugin，等下我会分享出来。等一下，我换成我平时常用的那个——我还做过一个带 OpenAI 品牌风格的版本，不过在这儿你就能看到。基本上，跑过几轮之后我发现，加一些指引特别管用，比如「注意别踩进 Instagram 广告的安全区」、「按这些画幅出片」——我们要 9:16，也要 4:5。

[23:24] Nick Baumann

That's great. But what I'll generally do and I'll dictate this. So we've got a bunch of clips here. I recorded a video for the record and replay feature. It starts with me just explaining that I'm apartment hunting and that's kind of a pain. And then I walk through and show the feature in real life. I show myself actually apartment hunting. And I show how Codex can do it for me. And then I have a closer about how I this skills that's been made that's been made for me can be put on a heartbeat and I can have Codeex just search for apartments in the background. Can you go through these clips first? like pull the transcripts, find the best takes, and then kind of piece this together into a UGC style video. So, that's running a few things that to kind of explain how this works. So, I've only submitted like 20 some odd like video clips. And so, what code is going to is going to do is going to start by processing those clips to get just the transcripts. This will help it understand like the story, what we're talking about, and even like my light description will help it understand that. And then what I'll do then, actually, it's ask me some questions. Sure, we'll do organic vertical. Let's do complete and clean.

这就很好了。不过我一般会这么做，而且是用口述的方式。（对 Codex 说）这儿有一堆片段。我录了一支关于 record and replay 功能的视频。开头是我讲自己在找房子、这事儿挺折磨人的。然后我一路演示，实拍展示这个功能，展示我真的在找房，再演示 Codex 怎么替我做。最后有个收尾，讲我把别人替我做好的这个 skill 挂到 heartbeat 上，Codex 就能在后台自动帮我搜房源。你能不能先把这些片段过一遍，把转写文本拉出来，挑出最好的几条，然后拼成一支 UGC 风格的视频？这里同时在跑好几件事，我解释一下它是怎么工作的。我只提交了二十来个视频片段。Codex 接下来会先处理这些片段，只取转写文本，这能帮它理解整个故事、我们在讲什么，加上我刚才那段粗略的描述，它就更清楚了。然后我要做的是——哦，它其实在反过来问我问题。行，我们选 organic vertical（原生竖版）。风格选 complete and clean。

[24:49] Nick Baumann

Sure. So, it'll first get the transcript and that's going to go through the takes and actually like process images from them to understand what I'm showing. And then it'll choose like the best takes. And when I'm recording, I'll say things like cut or that's a bad take or that's a good take. And it's able to like use that as guidance, you know, like you know what's actually useful from these. H

好。所以它会先拿到转写文本，然后把这些镜头过一遍，还会真的抽帧看画面，理解我在演示什么。接着挑出最好的那几条。而且我录的时候会说「cut」「这条废了」「这条不错」，它能把这些当成提示，知道这些素材里哪些是真正能用的。

[25:12] Claire Vo

you're going to give me so many ideas because I have like deeply abandoned my once great TikTok because I guess like UGC is hard. Like short for short form is hard. Um and I just have such a high bar that I'm constantly like recording these clips and being like that was garbage and then I get frustrated. I walk away. But you've given me this idea which is I could just like keep recording until like there's one that feels good and then just dump it all. I don't even have to remember which one is good. Dump it all, give some voice notes about how to edit and then and then kind of assemble it. And then I I love this process of just having the model transcribe, look at the video, come up with like good good cuts and put it together. I use it a lot for our trailers. So, anytime I do a very long talk, like a 30 to 60 minute talk, or I do a podcast that's like 60 minutes, I dump it in and I'm like, give me a 60-second hype video with the funniest parts of the talk and like give me a little bit of buffer for for clipping and then it it clips that really well. It actually has like pretty good funny taste. Um, which is nice. And so, I do that. And then I also do like I teach like a three-hour long workshop um every quarter on how to like transform your engineering organization to be kind of like AI engineers. It's very very long.

你这要给我一堆灵感了。因为我那个曾经还挺火的 TikTok 已经被我彻底荒废了，我猜就是因为 UGC 太难做了——短视频太难做了。而且我对自己要求特别高，老是录一堆片段然后觉得「这就是垃圾」，一沮丧就撂挑子走人。但你给了我一个思路：我可以一直录，录到有一条感觉对了为止，然后把所有素材一股脑全丢进去。我甚至不用记住哪条是好的，全丢进去，再用语音说几句该怎么剪，然后让它拼起来。我特别喜欢这个流程——让模型去转写、去看画面、想出好的剪辑点，再把它拼成片。我们做预告片就大量用这个。所以每次我讲完一场很长的演讲，比如三十到六十分钟的，或者录了一集六十分钟的播客，我就丢进去说：给我剪一支 60 秒的高能预告，把演讲里最好笑的部分挑出来，前后各留一点余量方便我再修。它剪得真的很好。它对「好笑」的品味其实相当不错，这点很难得。所以我经常这么干。另外我每个季度还会讲一次三小时的工作坊，教怎么把工程组织转型成 AI 时代的工程团队。特别特别长。

[26:31] Nick Baumann

It's like three-hour Zoom clip and I also do that to like clip chapters, clip shorts, do all that sort. It's just like all this tedium of of video editing. It's quite quite good at. And then just tell us like what does what does this plugin actually do? Yeah. So, the plugin is really just based on like all the steering I've done uh really this week of like making these videos, whether it's, you know, describing the kind of formats I need or, you know, that you don't want to have text in certain areas because it'll be cut off or natural kind of break points or how to do captions. Um, it's, you know, literally I just I had this massive thread where I did a few of these videos and there was so much essentially like bespoke knowledge and this process built up that I was like, "Hey, can you make a plugin out of this?" And there's actually a plug-in creator skill um in the chatbt app. So, it's pretty easy. And then I did a few tweaks and now I just more reliably get, I guess, the the happy path that I want.

（Claire）就是一段三小时的 Zoom 录像，我也会用它来切章节、切短视频，全都干。视频剪辑里所有这些琐碎的活儿，它做得相当好。那你说说，这个 plugin 具体做了什么？（Nick）好。这个 plugin 其实就是把我这一周做这些视频时给出的所有引导沉淀了下来——比如我需要哪些画幅，比如某些区域不能放文字、否则会被裁掉，比如自然的断点在哪儿、字幕该怎么处理。真的就是，我有一个超长的对话线程，在里面做了好几支视频，攒下了大量这种很定制化的经验和流程，于是我就说：你能不能把这些做成一个 plugin？而且 ChatGPT 应用里本来就有一个 plugin creator 的 skill，所以做起来挺容易。然后我又微调了几下，现在就能比较稳定地拿到我想要的那条「顺利路径」了。

[27:31] Claire Vo

And can you show us an out? I know you did one of these for yesterday. Can you show us like what what the kind of output is that you can get?

能给我们看看成品吗？我知道你昨天做了一支。能不能展示一下最后出来的东西大概是什么样？

[27:38] Nick Baumann

Yeah. So, here was my starting prompt. Again, I just dumped in a bunch of clips. And one thing I want to know, you talked about like the tedious stuff.

可以。这是我一开始的 prompt。还是老样子，我就把一堆片段丢了进去。还有一点我想说，你刚才提到那些琐碎的事——

[27:47] Claire Vo

Yeah.

嗯。

[27:48] Nick Baumann

Even like little things like going through and deleting bad takes or finding like the right segments. I I feel like I'm I'm so willing just to hand that over.

哪怕是很小的事，比如一条条过一遍删掉废镜头，或者找准要用的片段，我都特别乐意直接交出去。

[27:59] Claire Vo

Yep.

对。

[27:59] Nick Baumann

And and just get that off my plate. I don't want to think about it. So yeah, I I I I started with a pretty similar prompt. I just I kicked off with a little bit for before at one point and actually I was in an Uber. I said this on my phone. I remembered I was like, "Oh yeah, I need I need a 4x5 also." So I I sent that in as like a steer and it was like, "All right, got it. Updating the updating the delivery matrix." And then this is actually where, you know, I was like, "Why don't you ask me questions?" So I had to update the plugin um at the same time. And this is what the output looks like. So this is actually a few few different like you know back and forth steers. Um here we'll actually haven't like watched this full one but we'll see how it looks.

让这些事彻底从我手上拿走，我不想再花脑子想它。所以我一开始用的 prompt 跟刚才那条差不多。我先起了个头，中间有个时候我人在 Uber 上，用手机说了一句——我突然想起来：哦对，我还需要一个 4:5 的版本。我就把这句当成一次调整发过去，它回：好的，收到，正在更新交付清单。然后就是在这儿，我说了句「你干嘛不先问我几个问题呢」，所以我同时又去改了一下 plugin。这就是最后出来的成品。这中间其实来回调整了好几轮。这支我自己还没完整看过，我们一起看看效果。

[28:44] Nick Baumann

It's insane trying to find apartment in SF or New York City right now and it's impossible to get to all the listings before they're gone.

（视频播放）现在在旧金山或者纽约找房子简直离谱，房源根本抢不过来，等你联系上早就没了。

[28:50] Nick Baumann

So what's what's funny I don't if you can see that it's it even added blurs to these addresses here.

好玩的是，不知道你能不能看清——它甚至还自动把这些地址打了码。

[28:55] Claire Vo

Real smart. No I can see it.

太聪明了。能看见。

[28:57] Nick Baumann

Yeah. Um, I actually I did like a couple of these for like that were releasing like on Instagram next week

对。我其实还做过几支下周要在 Instagram 上发的，

[29:03] Nick Baumann

and there's like sensitive like release material information by Slack that it like bubbled up in my chatbt work

里面涉及一些还没发布的敏感物料信息，是从 Slack 里被它带到我的 ChatGPT Work 里来的。

[29:09] Nick Baumann

and it did this like insanely granular work of like blurring out

而且它打码这活儿做得细到离谱，

[29:14] Nick Baumann

literally lines and then following it while I'm scrolling and it does all this like verification where it it adds these blurs. It checks its own work and then it like does that over and over again until it's good and then it gives you like, you know, a good result.

真的是一行一行地糊，我往下滚动的时候码还能跟着走。整个过程它是有自检的：先打上码，再回头检查自己做得对不对，然后反复迭代，直到满意为止，最后给你一个相当不错的成品。

[29:29] Claire Vo

I probably six months ago coded a redactic for the How AI podcast because I had so many people being like, "Whoops, I left like my API key open in that like one piece of code that we were talking about." And so we actually like built one. Um, I need to try it again with codecs because it's probably I the one challenge we have is the blurs following kind of like screen shares. But

大概半年前我给 How I AI 播客也写过一个打码工具，因为老有人跟我说：「糟了，刚才讲的那段代码里我的 API key 露出来了。」所以我们真的做了一个。嗯，我得拿 Codex 再试一次，因为我们唯一的难点就是打的码怎么跟住屏幕共享的画面。不过——

[29:54] Nick Baumann

yeah, and actually that gets back to the like I don't my I don't even want to deal with the tedious work.

对，这其实又回到刚才那点——我压根就不想碰这些琐碎的体力活。

[30:00] Claire Vo

Like not only actually it's a non-starter that I'm going to make blurs

不只是不想——让我自己动手去打这些码，这事儿根本就不成立，

[30:05] Claire Vo

automatic like do that manually. That's just that's not going to happen.

手动一个个弄？那绝对不可能发生。

[30:08] Nick Baumann

No. What's even more annoying though is having to think about all right well how do I generate outputs that don't have confidential information

是啊。但更烦人的其实是，你还得动脑子想：我怎么才能让产出里不带机密信息，

[30:13] Nick Baumann

and how do I still make it seem genuine that's also a pain that I just I don't think about I just I leave it to leave it leave it to the chat GBT app

同时又让它看起来还是真实的、不假。这也是件麻烦事，我现在干脆不想了，全丢给 ChatGPT app 去处理。

[30:21] Claire Vo

I'm going to give one other benefit for the content creators out there for this like AI editing of content which is it's real embarrassing to edit your own content. It's like really embarrassing to look at yourself and watch all these videos of yourself and be like, "Oh yeah, I was like real cute and funny in that one and that one I am not like I make no sense as a human." So I like to offload like the critique of myself and the critique of my performance off to another model so I don't have to look at my face so much as well, especially as like a solo creator, solo founder, somebody who's like working on their own work. So, I do like having a third, you know, like a neutral third party do these cuts so I don't have to sit there and like listen to my own videos all day.

我再替内容创作者补一个用 AI 剪片的好处：自己剪自己的内容其实特别尴尬。你得一遍遍盯着自己看，回放自己的视频，然后心里想「哦这段我还挺可爱挺有梗的」「那段我完全不知道自己在说什么，根本不像个正常人」。所以我特别乐意把「评判自己、评判自己的表现」这件事外包给另一个模型，这样我就不用老盯着自己的脸。尤其当你是独立创作者、独立创业者、一个人干活的时候更是如此。所以我很喜欢让一个中立的第三方来定剪辑点，这样我就不用整天坐在那儿听自己讲话了。

[31:07] Nick Baumann

Yeah, I know. It's it's it's great. Um, back to like the

是啊，我懂。这真的很好。嗯，回到刚才说的——

[31:11] Nick Baumann

it like it definitely lowers like the bar like the barrier to entry for stuff like this

它确实把这类事情的门槛拉低了非常多。

[31:15] Nick Baumann

and if I just need to like go to a park, you know, I recorded like three videos. I recorded to three videos over the course of like an hour and a half couple days ago.

我只要去趟公园就行了。前两天我在一个半小时里录了三条视频。

[31:25] Nick Baumann

And then I uploaded like 50 or 60 clips.

然后我上传了大概五六十个片段。

[31:29] Claire Vo

Yep.

嗯。

[31:29] Nick Baumann

And I dictated like what the different narratives for those three videos were

接着我口述了这三条视频各自要讲的叙事线，

[31:33] Nick Baumann

and because I have a plugin and a task that already understands what I want.

因为我已经有一个 plugin、还有一个很懂我要什么的 task 在那儿了。

[31:37] Claire Vo

Yep.

嗯。

[31:37] Nick Baumann

You know, it took a while. I I went to bed and I came back to it, but I had I had the I had what I needed and we actually shipped one to our Instagram a couple days ago.

它确实跑了挺久，我去睡了一觉，回来再看，要的东西都在了。前两天我们还真把其中一条发到了 Instagram 上。

[31:45] Claire Vo

Okay. I have one more question on this and then maybe we'll see if the if the website is up. Do you need Soul Ultra for this? Because that's what I see below.

好，这块我最后再问一个问题，然后我们看看网站起来了没有。这个必须用 Sol Ultra 吗？因为我在下面看到的就是它。

[31:55] Nick Baumann

Yeah, I would. Um, in my experience, there's just so many, as much as anything, just for speed.

我会用。就我的经验来说，别的先不谈，光为了速度就值得。

[32:02] Nick Baumann

There's so many little like micro tasks involved in checking all these frames, in checking these transcripts, and there's a lot of parallel processing that can happen

这里面有太多细碎的小任务了——一帧一帧地检查、一段一段转写文本地核对，而且很多是可以并行跑的，

[32:14] Nick Baumann

and a lot of like parallel judgment, if you will. And what soul ultra is is it's basically this framework for multi- aent um for 56 soul. And so I find that it's just more efficient and I get better outputs when I use Soul Ultra.

也有很多判断可以并行做。Sol Ultra 本质上就是一套跑多 agent 的框架，配 5.6 Sol 用的。所以我发现用 Sol Ultra 效率更高，产出质量也更好。

[32:29] Claire Vo

Love it. Okay, cool. Let's see if the How AI website is is up and running. Look at that. Black and purple and white. Just like I said, [laughter] [gasps]

太喜欢了。好，我们来看看 How I AI 的网站是不是已经跑起来了。你看这个！黑色、紫色配白色，跟我说的一模一样。[笑][惊叹]

[32:45] Nick Baumann

we also have AI identify which thumbnails are going to do the best. This the by the way, this episode, just so you know, it's all look no hands. We had some problems.

我们还让 AI 来判断哪些封面图（thumbnail）效果会最好。顺便说一句，这一整期都是「你看，我没上手」——全程没碰键盘，中间我们确实也出过一些状况。

[32:54] Claire Vo

Oh, I see what it is. It's the use cases. So, each episode is showing up multiple times. That was a very good episode about codeex browser use where I let it shop for me for my Hawaiian vacation. Um, and it did quite quite a lovely

哦，我知道是怎么回事了。是 use cases 那一块——每期节目重复出现了好几次。那期讲 Codex 浏览器操作的其实特别好，我让它替我给夏威夷度假买东西。嗯，它干得相当、相当——

[33:08] Nick Baumann

Yeah. Yeah,

对，对，

[33:08] Claire Vo

lovely job.

——相当漂亮。

[33:09] Nick Baumann

If you if we want to get really fancy, we could probably have it have different thumbnails for each clip. We could even have like use image gen to make like a, you know, a singular thumbnail.

如果我们想再花哨一点，完全可以让每个片段配不同的封面图，甚至用 image gen 生成一张专属封面。

[33:19] Claire Vo

Love it.

太喜欢了。

[33:20] Claire Vo

Amazing. Okay, there's good old Alex over and over and over again. And then we're going to hit three of me. It just built it. Awesome.

太厉害了。好，这儿是老朋友 Alex，一遍又一遍地出现。再往下我们会连着看到三个我。它就这么把网站搭出来了。牛。

[33:28] Nick Baumann

Yeah.

对。

[33:28] Claire Vo

And then this can be published. Just like remind me how it how it gets published, who I can share it with, how that all works.

然后这个东西是可以发布出去的。你再跟我说一下，它具体怎么发布、我能分享给谁、整套机制是怎么运作的。

[33:34] Nick Baumann

Yeah. So let's say and let's say I wanted to like share this with you.

好。比如说——比如说我想把这个分享给你。

[33:39] Nick Baumann

Um I can say share this. And so now you can you can put in you know whoever you really want to. It could be totally publicly available. You could add

我就点「分享」。然后你可以把想给的人填进去，可以设成完全公开，也可以加

[33:48] Nick Baumann

um individual individual email addresses.

单个单个的邮箱地址。

[33:50] Claire Vo

Yep.

嗯。

[33:50] Nick Baumann

And it'll share with those folks.

它就会分享给这些人。

[33:52] Claire Vo

Amazing. And then how do you log in?

太棒了。那对方怎么登录呢？

[33:55] Nick Baumann

Uh it's login with Chat GBT.

用 ChatGPT 账号登录。

[33:57] Claire Vo

Great. Easy peasy. I do have a Chat GBT account underneath that email [laughter] address. So we are

好，那太简单了。我那个邮箱底下确实有 ChatGPT 账号（笑），所以我们

[34:02] Claire Vo

we are good to go. Um, just to recap, we did a couple use cases. So, we talked a little bit about voice, high bandwidth, Yappers API, um, just basically like talking to a computer empowered assistant that can just go do a bunch of your work for you, um, with access to browser and other things, including just being able to screenshot whatever's on the screen and have context. We did a um an a tragic bee site which I did not know it has its own computer and memory and browser and all that stuff. So that's fun fact for Claire. Um or you can have it actually go off and make a website which you can share publicly or share with specific email addresses which you just showed. And then again, my favorite use case, which is take a bunch of clips um for UGC, dump them in, make it the model's problem, including doing redaction and blurring, which is for any of my fellow YouTubers out there, a a huge and very annoying issue. And I'm very happy um personally to have seen this this workflow and get it solved. Um Nick, a couple questions uh lightning round questions for you. We will get you back to all your all your voice chatting. Question number one is on that topic. In this world where we're all loving our Chad GBT voice experience, are is everybody in the open AI office just like mumbling to themselves and to their computer? Like how are we going to intersect all this voice capability with the realities of like existing around humans that find all this chitter chatter annoying? Like how does it actually work? I think it's like a like an almost genuinely uncomfortable question because when I'm at home, I exclusively dictate and when I'm at my desk around other people, I will sometimes whisper, but I'm more likely to like use my keyboard. And you know, we have all these meeting rooms and these like little pods people hop into. And I see people like hopping in there for not calls more often these days because they're they're clearly

完全没问题。我快速回顾一下，我们today过了几个用例。一个是语音——高带宽的「话痨 API」（Yappers API），说白了就是对着一个有电脑能力的助手讲话，它就能替你把一堆活干了，还能用浏览器之类的工具，包括直接截屏、把屏幕上的东西变成上下文。然后我们看了 agent 那一块（我之前都不知道它自己就带一台电脑、有记忆、有浏览器这些东西，这算是给 Claire 长知识了）。或者你也可以让它直接去做一个网站，做完可以公开分享，也可以指定邮箱分享——你刚刚演示的就是这个。再就是我最喜欢的那个用例：把一堆素材片段（UGC）一股脑丢进去，让模型自己去头疼，连打码、模糊处理都一起做了——这对我们这些做 YouTube 的人来说，是个特别大又特别烦人的问题。我个人真的很高兴今天看到这个工作流，把它给解决了。Nick，接下来问你几个快问快答，问完就放你回去继续「语音聊天」。第一个问题正好接着刚才那个话题：现在大家都爱上了 ChatGPT 的语音体验，那 OpenAI 办公室里是不是每个人都在自言自语、对着电脑碎碎念？我们怎么把这套语音能力，跟「身边还有活人、而且人家觉得这些叽叽喳喳很烦」的现实调和？实际到底是怎么运作的？（Nick Baumann）我觉得这个问题问得……说实话有点让人不自在。因为我在家的时候是完全靠口述的；但我坐在工位上、周围都是人的时候，我有时候会压低声音说，但更多时候还是用键盘。而且我们公司有那么多会议室、那种一人电话亭，我最近越来越多看到有人钻进去——但不是去开会，他们明显是

[36:04] Nick Baumann

going in there to like use dictation or to use voice and codecs. And clearly people are adapting to this as like a better experience. But frankly, we you know, I don't know if the answer is we all walk around like with those masks around us that like shield our voice to anybody but the agent, but it's it's not ideal right now. But you can but we can tell that there there's a there's a better there's a good version and that's using voice but it's not ideal for our current you know

进去用听写、或者用语音操作 Codex 的。很明显大家在往这个方向适应，因为体验确实更好。但坦白讲，我不知道最终答案是不是我们每个人都戴一个罩子，把声音挡住只让 agent 听见——反正现在这个状态肯定不理想。不过我们已经能感觉到：有一个更好的版本存在，那就是用语音；只是它跟我们现在的

[36:32] Nick Baumann

office makeup

办公室格局、

[36:33] Claire Vo

infrastructure. Yeah I I imagine we like press a button and like a tube comes down

基础设施还不太匹配。（Claire Vo）是啊。我想象中应该是按个按钮，天上降下来一根管子罩住你。

[36:39] Claire Vo

that would be ideal. Yeah that's that's what I'm talking about press our button and the tube the tube goes up. Okay. We're going to we're going to figure this out um until we get the direct codeex to mind connection which I'm sure will be a 2027 release. Okay. Um, second question. If if people really wanted to like feel the AGI, what are what are the things that you really think are just huge step changes in the last 3 months? Maybe like voice aside, what is like the one thing that you would tell people to try or that you're really thinking is is the new interface because there's so many things that you've shown here

（Nick）那就完美了。（Claire）对，我说的就是这个——按个按钮，管子降下来、说完再升上去。行吧，这个问题我们慢慢解决，直到哪天 Codex 能直连大脑为止——我估计那是 2027 年的版本。好，第二个问题。如果有人真的想「感受到 AGI」，你觉得过去三个月里，哪些东西是真正的阶跃式变化？语音先放一边，你最想让大家去试的那一件事是什么？或者你心里认定的那个新交互界面是什么？因为你今天展示了太多东西了。

[37:16] Nick Baumann

in terms of like in terms of feeling the AGI. I think a lot of people like you and me have already felt the AGI. I think if we go back to you know 2022 2023 around the release of chatbt.

说到「感受到 AGI」——我觉得像你我这样的人，其实早就感受过了。回头看 2022、2023 年 ChatGPT 刚发布那会儿，

[37:28] Claire Vo

Yep.

对。

[37:28] Nick Baumann

I think a lot of people felt the AI then you know what what AI can be became

我觉得很多人在那时候就已经感受到了 AI 的冲击——「AI 到底能是什么样子」这件事，突然变得

[37:34] Nick Baumann

understandable like you understand this conception of a a chatbot that I say things to and it intelligently says say says things back to me. That's that's pretty cool. And then I think with the advent of Kodi agents and like the Kodi agent harness in particular, I think people in the developer community and like AI tinkerers, they felt this AGI moment where wow now now this AI is going out and like reading files and creating files like running commands and this is crazy how I tell it to do I I tell it one thing and then it figures out all these other things on its own and that was a huge AGI moment for people. But I do think there is a certain barrier like barrier to entry with that where if you're not an engineer or you're not a thinker, you know, there's this is limited to a smallish group of people. But now, you know, with something like chat GBT work in the web, I think we're taking a lot of the primitives um you know, like the coding agent harness, like having an always on agent and we're making them not only accessible to the web, but they're accessible to your phone. And so for people that are already using the chat GBC app, which is a lot, they can just switch over to the work tab, they can ask a question like, "Hey, can you check my Gmail and also my finances?"

可以理解了：你能想象出「聊天机器人」这个概念——我对它说话，它很聪明地回我话，挺酷的。再往后，随着 coding agent、尤其是 coding agent harness 的出现，开发者群体和那些爱折腾 AI 的人又经历了一次「AGI 时刻」：哇，现在这个 AI 会自己去读文件、写文件、跑命令了；我只交代一件事，它自己就把其他一堆事都推演出来了——这对很多人是巨大的震撼。但我也确实觉得，这里有个门槛：如果你不是工程师、不是那种爱折腾的人，能享受到的人群就很小。而现在，有了 ChatGPT Work 网页版这样的东西，我们等于是把很多底层能力——比如 coding agent harness、比如「常驻在线的 agent」——不仅搬到了网页上，还搬到了你手机上。所以对那些本来就在用 ChatGPT app 的人（数量非常大），他们只要切到 Work 那个标签页，就能问一句：「帮我看一下我的 Gmail，还有我的账单财务。」

[38:48] Nick Baumann

Um, which I know are two, you know, plugins that you can install really quickly and just monitor those and let me know if anything stands out or anything has gone arry. And now they suddenly have this agent that's, you know, maybe twice a day checking their email, checking their finances. And I think that like small thing, you know, when they first get a notification, you know, letting them know maybe a package has arrived or there's some weird charge. I think those are like the small moments that people might kind of like it might be like a foot in the door for a lot of people.

这两个插件都能很快装上，然后让它盯着，「有什么值得注意的、或者哪里出岔子了就告诉我」。于是他们突然就拥有了一个 agent，一天大概两次帮他们查邮件、查账单。我觉得就是这种小事——当他们第一次收到通知，比如告诉他们包裹到了，或者账上有一笔奇怪的扣款——这些小瞬间，可能就是很多人踏进门的第一步。

[39:18] Claire Vo

Okay. Amazing. And then I have one last question I ask every guest and because you speak to AI, I'm really curious here what what you say. when I I saw a little bit of your prompting strategy. It's a little bit of what I do, but when AI is not responding, when it is not doing what you want, when it is not using that plugin, what is your prompting strategy? Do you yell? Do you all caps?

好，太棒了。最后还有一个问题，我每位嘉宾都会问，而你又是「用嘴跟 AI 说话」的人，所以我特别好奇你的答案。我刚刚看到了一点你写 prompt 的策略，跟我的路数有点像。那当 AI 不听话的时候、它没按你想要的做、它就是不肯用那个插件的时候，你的 prompt 策略是什么？你会吼它吗？会全大写吗？

[39:42] Nick Baumann

Well, it depends if I'm sitting on my desk or if I'm in my living room. Those are [laughter] I guess the my treatment of of AI is different. [laughter] Um, it might be all caps in the office as I like bang on my keyboard.

这得看我是坐在工位上，还是在自家客厅里（笑）——我对 AI 的态度在这两种场合是不一样的（笑）。在办公室可能就是全大写，一边敲键盘一边砸。

[39:58] Claire Vo

Yeah,

对。

[39:58] Nick Baumann

at home I'll turn on dictation,

在家我就打开听写，

[40:01] Nick Baumann

maybe share a few expletives.

顺带骂几句脏话。

[40:03] Claire Vo

Yeah,

对。

[40:03] Nick Baumann

try to redirect things.

然后再想办法把它拽回正轨。

[40:05] Claire Vo

I'm just like constantly constantly being like, "This is trash. Why Why are you Why are you this way?" Um, I used to be so polite and now my bar is so high. I'm like, "You you can do better. I believe in you." I would say lately I feel like I have been less upset. Um things are just working better, you know.

我基本上是不停地在说：「这也太烂了。你为什么？你怎么会是这样？」我以前特别客气，现在门槛被拉得特别高，我会说：「你可以做得更好，我相信你。」（Nick）我倒是觉得最近我没那么容易上火了，因为东西确实跑得更顺了。

[40:26] Claire Vo

I've been more upset.

我反而更容易上火了。

[40:27] Nick Baumann

People People are Are you getting more upset?

大家……你是越来越容易上火了吗？

[40:29] Nick

Yeah, because my bar is higher. I'm like, I know you can do this. You're not a dumb dumb. Come on.

对，因为我的标准更高了。我会想，我知道你能做到这个，你又不是笨蛋，来吧。

[40:36] Claire

I understand. Yeah. All right. Yeah. Treat it treat as equal or superior. You're surprised when it when it doesn't come through.

我懂。是的。好的。对，把它当成平等甚至更强的对手来对待，所以它做不出来的时候你会觉得意外。

[40:42] Claire

Okay. Well, Nick, this has been super fun. Where can we find you? How how can we be helpful?

好的。Nick，这期聊得太有意思了。大家在哪儿能找到你？我们能帮上什么忙吗？

[40:47] Nick

Yeah. Find me on Twitter X um my name Nick Bowman_. Um if you have any questions about codeex chatbt work, anything we're releasing OpenAI, feel free to reach out to me. Uh DMs are open. Great to uh great to be on with you, Claire.

可以在 Twitter（X）上找我，我的账号是 @nbaumann_。如果你对 Codex、ChatGPT Work，或者 OpenAI 正在发布的任何东西有问题，欢迎来找我，私信是开放的。很高兴能和你一起聊，Claire。

[41:03] Claire

Yeah, thanks for being here. Thanks so much for watching. If you enjoyed this show, please like and subscribe here on YouTube or even better, leave us a comment with your thoughts. You can also find this podcast on Apple Podcasts, Spotify, or your favorite podcast app. [music] Please consider leaving us a rating and review, which will help others find the show. You can see all our episodes and learn more about the show at howiaipod.com. See you next time.

谢谢你来。也非常感谢各位收看。如果你喜欢这期节目，请在 YouTube 上点赞订阅，或者更好的是，留言告诉我们你的想法。你也可以在 Apple Podcasts、Spotify 或你常用的播客 App 上找到这档播客。[music] 也欢迎给我们打个分、写个评价，这能帮更多人发现这档节目。所有往期节目和节目介绍都在 howiaipod.com。下期见。