# 策展 · X (Twitter) 🔥🔥

> 📖 本站完整內容索引（documentation index）：[llms.txt](/llms.txt)

> 作者：OpenAI Developers (@OpenAIDevs) · 平台：X (Twitter) · 日期：2026-06-26

> 原始來源：https://x.com/OpenAIDevs/status/2070276212984094805

## 中文摘要

Pietro Schirano 與 OpenAI 的 Romain Huet 探討如何透過 GPT-5.5 與 Codex 將創意轉化為軟體應用。

**核心技術應用**
本次對談聚焦於利用 GPT-5.5 的進階能力與 Codex 程式碼生成技術，推動創意開發的邊界，具體應用場景包括：
- 影像轉音訊：透過模型理解視覺內容並生成對應音訊（03:45）。
- 多 Agent Codex 工作流：利用多個 Agent 協作執行複雜的程式碼生成任務（07:57）。
- 硬體復甦：運用 Codex 重新編寫或優化舊有硬體設備的控制程式碼（14:34）。

<video src="https://pub-75d4fe1e4e80421b9ecb1245a7ae0d1a.r2.dev/curated/1782451289402-iiwnrzi4.mp4" poster="https://pub-75d4fe1e4e80421b9ecb1245a7ae0d1a.r2.dev/curated/0c976a692a99b08f.jpg" controls playsinline preload="metadata" style="max-width:100%;height:auto;display:block;margin:1rem 0"></video>

**開發思維轉變**
雙方強調了從「親手實作」轉向「指揮與引導」的開發模式（25:27）。開發者不再僅是撰寫每一行程式碼，而是透過定義目標與架構，指揮 Agent 執行具體的程式開發任務。這種轉變讓創作者能更專注於產品邏輯與創意呈現，而非陷入底層實作的細節。

## 媒體內容

**逐字稿**

- `00:17` 嗨 Pietro！（Ciao Kietro!）
- `00:18` 嘿，Roman，最近怎麼樣？（Hey, what's up Roman?）
- `00:19` 你打算怎麼為《Builders Unscripted》寫主題曲？（How are you writing the theme song for Builders Unscripted?）
- `00:21` 或許吧，對，也許我們把它叫做「token」之類的，我還需要更多靈感。（Maybe, yeah, like maybe we call it like Tokens or something, I need more.）
- `00:25` 讓我們聽聽看。（Let's hear it.）
- `00:26` Token！（Tokens!）
- `00:27` 喔，我還需要更多！（Oh, I need more!）
- `00:29` 聽起來是個好主意。（That sounds like a great idea.）
- `00:31` 我們留到之後再說。（We'll save that for later.）
- `00:32` 嗯，我對這一集感到非常興奮。（Well, I'm so excited about this one.）
- `00:34` 你真是多才多藝。（You're a man of many talents.）
- `00:36` 謝謝。（Thank you.）
- `00:36` 你是位音樂家，這我們都聽得出來。（You're a musician, as we can tell.）
- `00:38` 你同時也是工程師、設計師、創辦人，我會說你是 AI 傳聲筒，也是我們模型的早期測試者。（You're also an engineer, a designer, founder, I would say AI whisperer, early testers of our models.）
- `00:46` 所以我非常期待這次的對談。（So I'm super excited to chat.）
- `00:48` 但首先，我想了解一下，你的大腦是如何運作的？（But to get started, I'd love to hear, like, how does your brain work?）
- `00:52` 你是如何在這些不同的領域中保持如此豐富的創造力？（How do you have so much creativity across all of these different fields?）
- `00:56` 是的，首先，非常感謝邀請我來。（Yeah, well, first of all, thanks so much for having me.）
- `00:59` 我覺得，你知道，你和我已經聊過一陣子了，所以能來到這裡進行這次對談真的很棒。（I feel like, you know, you and I have been talking for a while, so this is amazing to just be here and have this chat.）
- `01:04` 對我來說，我喜歡一句名言：「我們生得太晚，無法探索海洋；生得太早，無法探索星空。」（And so for me, I feel like, you know, there is this quote that I love, which is like, we're born too late for the sea, but too early for the stars.）
- `01:12` 而現在有了 AI，我覺得有一個全新的領域可以探索，就像是潛在空間（latent space），對吧？（And with AI right now, I feel like there's this new field that you can explore, like the latent space, right?）
- `01:17` 就像你可以進入這種不可思議的新技術的思維中，開始進行探索。（It's like you can get into the mind of this incredible new technology and start to explore.）
- `01:21` 我們幾乎就像是過去的探險家，像是麥哲倫或馬可波羅，試圖弄清楚我能用這項技術做些什麼？（We're almost like, you know, truly like an explorer back in the days, like a Magellan or, you know, Marco Polo and just trying to figure out, you know, what are the things I can do with this technology?）
- `01:30` 對我來說，這真的非常令人興奮。（And I think that, to me, it's like really exciting.）
- `01:32` 我很喜歡那句名言。（Love that quote.）
- `01:33` 那真的很酷。（That's super cool.）
- `01:34` 我的意思是，我們從遠處關注你的作品已經很長一段時間了，而且你一直是我們的早期測試者之一，因為你總是有獨特的視角，所以能搶先使用這些模型。（I mean, we've seen your work from afar, like for a long, long time, and then like you've been one of our early testers, kind of like having access to these models because you always have like an interesting lens.）
- `01:44` 你總是會去測試這些模型的邊緣情況，看看它們能做到哪些前一代模型做不到的事情？（Like you're poking at these models to find the little edges and like what can they do that the previous one could not do?）
- `01:50` 舉例來說，你對 GPT 5.5 的第一印象是什麼？（What was your first impression of like GPT 5.5, for instance?）
- `01:53` 這是一個巨大的躍進。（It was a big step change.）
- `01:54` 我覺得過去很多很難做到的事情，現在突然間都能順利運作了。（I feel like there was a lot of the things that were very hard to do in the past that now all of a sudden just work.）
- `02:00` 這就像是一波完全震撼我的浪潮，當我剛開始在測試版中與模型互動時。（It's just like a complete sort of like wave that just shocked me when I first started interacting with the model in the beta testing.）
- `02:07` 這就像是，你知道，每當有新模型推出時，我都會有一套個人的測試項目。（It's things like, you know, like I have all this set of evils that I personal evils that I have every time a new model comes out.）
- `02:13` 這些項目包括 Agentic 程式開發行為，像是理解我如何同時執行多個工具，或是多模態能力。（And those are things from, you know, agentic behaviors, like understanding like how can I run multiple tools at the same time or multimodal capabilities.）
- `02:21` 我給你舉個例子。（And I give you an example.）
- `02:22` 在當時的 GPT Vision 中，為了讓模型理解要看哪裡，（In the case of GPT Vision at the time, in order for the model to understand what to look at,）
- `02:27` 我必須提供一張疊加了網格的圖像，就像方格一樣。（I have to provide the image in this like with an overlay grid on top of like square.）
- `02:32` 這樣模型才知道，喔，底部在這個方格裡之類的。（So the model will know, oh, the bottom is in this square and whatnot.）
- `02:35` 而現在你不需要那樣做了，對吧？（And now you don't have to do that, right?）
- `02:37` 像是 5.5，它的視覺能力非常強。（Like 5.5, it's so good at vision capability.）
- `02:39` 你可以直接進行推理，判斷該點擊哪裡、該看哪裡之類的，對吧？（You can just reasoning around where to click, where to look and like stuff like that, right?）
- `02:43` 但歸根結底，我真正關心的是這些模型具備哪些人們可能會忽略的能力。（But ultimately, what I really care is like what kind of capabilities those models have that my people may miss.）
- `02:50` 這些能力與創造力和樂趣有關，例如模型在扮演設計師時表現得有多好，（And so those are things around creativity and fun, like how good the model is at acting like as a designer）
- `02:54` 或者模型在扮演創意總監來指導其他設計 Agent 和模型時表現得有多好，對吧？（or how good is the model acting as a creative director to direct other design agents and models, right?）
- `03:00` 事實上，有一個非常有趣的 Demo，這是我用 5.5 做的第一件事：（And actually, you know, there's a really funny demo that I – the first thing I built with 5.5 was like,）
- `03:05` 如果我能創造出一種東西，讓 5.5 接收一張圖像，並將圖像解構成基於圖片的和聲聲音？（what if I could create something where 5.5 takes an image and deconstruct the image in a sound that is harmonic based on the picture?）
- `03:14` 然後我可以用這個聲音來重建圖像，幾乎就像是一種秘密通訊的遊戲語言，（And then I can use the sound to reconstruct the image, almost sort of like a secret communication game language）
- `03:20` 這樣兩個人就可以用這種秘密方式透過聲音進行交流。（so that, you know, two people can communicate with sound in this like secret way.）
- `03:23` 當時我想，它不可能做對，因為這非常複雜，（And I was like, there is no way it's going to get this right because it's like a very complex, you know,）
- `03:27` 裡面涉及了密碼學。（there's like cryptography involved.）
- `03:28` 就像音樂需要達到什麼樣的水準。（There is like how good the music needs to be.）
- `03:30` 你如何比較影像轉聲音的基礎預設效果？（How are you comparing like the sort of like the basic default of the image to sound?）
- `03:35` 我當時想，它不可能做得對。（And I was like, there's no way it's going to get it right.）
- `03:36` 結果我試了一下，它竟然成功了。（And then I did it and it just worked.）
- `03:38` 這太瘋狂了。（It was crazy.）
- `03:39` 我們何不看看你的筆電，讓我瞧瞧你用這些新模型開發了什麼？（Why don't we take a look at your laptop and you show me a little bit what you've been building with these new models?）
- `03:44` 好啊。（Yes.）
- `03:45` 我用 5.5 開發的第一個東西，回過頭來思考這個概念：（So the first thing that I built with 5.5, so going back to this idea of like,）
- `03:50` 我們該如何激發模型的創造力，以及一些有趣的點子，（how can we poke the model for creativity and sort of like, you know, fun ideas,）
- `03:54` 比如那些平凡的點子會是什麼樣子？（like how are the ordinary ideas?）
- `03:55` 所以我基本上開發了這個應用程式。（So what I did, I basically built this app, right?）
- `03:58` 這個應用程式的功能非常瘋狂。（And so what this app can do, it's pretty crazy.）
- `04:01` 它可以將影像轉換為聲音，再從聲音轉回影像，（So it can convert an image into a sound and then from a sound back to an image,）
- `04:05` 幾乎就像是人與人之間的一種秘密通訊方式。（almost sort of like a secret communication way between different people.）
- `04:10` 我現在就來示範一下，我會使用 OpenAI 的標誌。（And so what I'm going to do right now, I'm going to take the OpenAI logo here.）
- `04:13` 我把它放進來，然後要求它建立一個聲音。（So I'm going to pop it right here and then I'm going to ask to create a sound.）
- `04:19` Codex 加上 5.5 基本上一次就完成了。（And Codex plus 5.5 was basically able to do this like in one shot.）
- `04:24` 我當時感到非常震撼，因為這裡面涉及了太多技術。（And I was like mind blown because like there's so much going on here.）
- `04:27` 裡面有密碼學的運作。（There is cryptography going on.）
- `04:29` 還有像是理解「創造出美妙聲音」究竟意味著什麼。（There is like understanding, you know, what does it even mean to make a beautiful sound?）
- `04:33` 你知道，有一部電影，我想是叫《第三類接觸》。（And, you know, there's this movie, I think it's called Contact from the Third Kind.）
- `04:39` 那是史蒂芬·史匹柏在 1980 年代拍的老電影。（There's like an old Steven Spielberg movie in the 1980s.）
- `04:41` 在電影結尾，我不確定你記不記得，（And there is at the very end of the movie, I don't know if you remember this,）
- `04:44` 科學家們正在與外星人溝通。（the scientists are communicating with the aliens.）
- `04:47` 沒錯。（Yes.）
- `04:47` 而外星人溝通的方式其實與這非常相似，（And the way the aliens communicate it is actually very similar to this,）
- `04:51` 這讓我印象深刻。（which I was kind of impressed.）
- `04:52` 聽起來真的很棒。（That sounds really good.）
- `04:54` 聽起來真的很棒，對吧？（It sounds really good, right?）
- `04:54` 現在我可以下載這個聲音，然後假裝我已經把這個應用程式的 URL 存取權傳送到你的電腦上，對吧？（And so what I can do now, I can download this sound and then I'm going to pretend now that I sent the URL access to this app on your computer, right?）
- `05:02` 接著我要做的是，把聲音拖回來，然後要求它復原成影像。（And what I'm going to do, I'm going to take the sound, drag it back here and I'm going to ask to recover the image.）
- `05:07` 砰。（Boom.）
- `05:07` 現在影像就出來了。（And now I have the image there.）
- `05:09` 所以我們基本上做到了影像轉聲音，再從聲音轉回來的過程，對吧？（So we convert, yeah, basically like an image to sound and from sound back, right?）
- `05:14` 這怎麼可能辦到？（How is that possible?）
- `05:16` 我知道。（I know.）
- `05:16` 這就是為什麼對我來說，思考 AI 時這點很重要。（It's like, that's the kind of, so for me, like, this is what's important when you think about AI.）
- `05:21` 重點是「我能做什麼」，對吧？（It's like, what can I do, right?）
- `05:23` 沒有什麼點子是太瘋狂的，對吧？（Like what, no idea is too crazy, right?）
- `05:25` 我認為這其中包含了一種想像力的元素。（I think there's an element of just your imagination.）
- `05:29` 想像力不應該成為你的限制。（Like, you know, imagination shouldn't be like your constraint.）
- `05:31` 儘管放手去做。（Just feel free.）
- `05:32` 而且我也認為，隨著模型越來越強大，（And also I think there is an element where, you know, as the models gets better,）
- `05:35` 你真的需要開始思考，如果我不害怕的話，我能做些什麼，對吧？（you really need to start thinking about what are the things I can do if I wasn't afraid in a way, right?）
- `05:40` 因為我覺得身為人類，我們總是會給自己設限。（Because I think like as human, we always have this sort of like walls a little bit.）
- `05:44` 像是「喔，這做得到嗎？」（It's like, oh, can this be done?）
- `05:45` 或是「這能……」，但試試看就知道了。（Can this be, you know, but just try.）
- `05:47` 現在 5.5 已經強大到能精準處理這些事情了。（And like now, you know, 5.5 is so good that it gets those things right.）
- `05:51` 這真的很棒，因為你有音樂背景，能幫助你產生這些點子。（And that's really awesome because like you have like your musician background that can help you like have these ideas.）
- `05:57` 這太酷了。（This is so cool.）
- `05:58` 而且你也在談論評估（evals）的部分。（And you're talking about evals too.）
- `06:00` 我很好奇，當新的模型出現時，像是 GPT Image 2 或 GPT 5.5，（Like I'm curious, like, so when the new model comes your way, like GPT Image 2 or GPT 5.5,）
- `06:06` 你會用什麼流程來測試新模型，並找出它擅長的地方？（like what's your process to poke at a new model and kind of like find what it's good at?）
- `06:11` 沒錯。（Yes.）
- `06:12` 我基本上會觀察三個不同的面向。（So I basically look for three different things.）
- `06:14` 第一是指令遵循能力。（One is instruction following.）
- `06:16` 也就是說，模型在遵循這些指令時表現如何？（So it's like how good is the model following this instruction?）
- `06:19` 你可能會感到驚訝。（And you might be surprised.）
- `06:20` 事實上，有時候現代模型並不代表在指令遵循上就一定更好，（Actually, sometimes a modern model doesn't necessarily mean it's better at instruction following）
- `06:24` 因為它們可能會變得稍微更受限或諸如此類的，對吧？（because they might get maybe like a little more secured or whatever, right?）
- `06:28` 所以我喜歡 GPT 5 系列的一點是，它在保持實用性方面一直很穩定，（And so what I love about the GPT 5 series is like how consistent has been in just being helpful,）
- `06:35` 基本上就是一個很有幫助的助理，對吧？（basically just being a helpful assistant, right?）
- `06:36` 所以這就是指令遵循的部分。（So there's like the instruction following part.）
- `06:39` 接著是任務執行的狀況如何？（Then there is like how was the task executed?）
- `06:42` 它是否真的能做到我要求的事？（Like is it actually being able to do what I ask?）
- `06:45` 然後還有一個要素，就是模型在指揮其他 Agent 時表現得有多好？（And then there's an element of like how good is the model actually directing other agent?）
- `06:50` 因為那其實是一個非常複雜的任務，對吧？（Because that's actually a very complex task, right?）
- `06:51` 例如，如果我對 Codex 說，這真的很有趣。（Like if I say to Codex, like actually this is really fun.）
- `06:55` 我有一整套在 Codex 上使用的捷徑，對吧？（So I have all this set of shortcuts that I use with Codex, right?）
- `06:58` 舉例來說，如果我輸入 spawn，它就會執行這個動作，對吧？（And so for example, if I say spawn, it does this, right?）
- `07:02` 所以我為自己設定了所有這些文字替換。（So I have all this like text replacement that I have for myself.）
- `07:05` 因為你知道，Codex 可以執行多個 Agent。（So because, you know, Codex can run multi-agents.）
- `07:07` 我之所以使用捷徑，是因為到了現在，我已經知道我希望（And the reason why I have shortcuts is because by now I know what are the things that I want）
- `07:11` 模型在哪些方面表現出色。（the model to be good at.）
- `07:12` 然後我會觀察模型的表現，對吧？（And then I see how the model performs, right?）
- `07:14` 很有趣。（Interesting.）
- `07:15` 舉個例子，我有這個設定。（So for example, I have this one.）
- `07:16` 我會說，當我提到 agent 時，我會說，嘿，記住你是一個 Agent。（I say, when I say agent, I say, hey, remember you're an agent.）
- `07:19` 你必須完成目標，直到目標達成。（You have to like complete the goal until you complete the goal.）
- `07:23` 不要停下來，對吧？（Don't stop, right?）
- `07:25` 而 5.5 的優點之一就是這方面做得非常好，現在你們也推出了目標功能。（And one thing about 5.5 is really good at that, which now you guys launched the goal.）
- `07:29` 透過 /goal 指令。（With the slash goal command.）
- `07:30` 所以我之前就是這樣做的，對吧？（So I was doing this before goal, right?）
- `07:32` 這又是另一件事。（And so this is another thing.）
- `07:33` 為什麼我不去思考我們所見過最重要的槓桿點，隨著模型變得更好，（It's like, why don't I think of the most important lever that we've seen as the model get better）
- `07:38` 就是它們能處理多長期的任務。（is how long tasks they can do.）
- `07:40` 我確實相信，真正的解鎖與財富分配對每個人來說，（And I actually believe like the true unlock and distribution of wealth for everybody is）
- `07:45` 將會是當模型能夠基本上連續運作數小時的時候，（going to be when the model is just going to be able to just basically run for hours,）
- `07:49` 對吧？（right?）
- `07:49` 而 Codex 已經具備這種能力了，對吧？（Which Codex is already capable of, right?）
- `07:51` 但為了做到這一點，你必須稍微推它一把，對吧？（But in order to do that, you kind of have to nudge it a little bit, right?）
- `07:53` 所以這就是當新模型出現時，我會觀察的事情。（And so that's something that I look for when a new model comes up.）
- `07:56` 那 subagent 呢？（What about sub-agents?）
- `07:57` 關於 subagent，我有 spawn。（Sub-agents, I have spawn.）
- `07:59` 所以我輸入 spawn，它就會產生多個 Agent。（So I say spawn, and it spawn multiple agent.）
- `08:02` 我還有像是 PR 之類的指令。（I have things like, I have like PR.）
- `08:04` 這其實是我經常做的事情。（So this is actually something I do a lot.）
- `08:05` 所以就像是，你知道，我會根據這個程式庫推送這個 PR。（So it's like, you know, I'll push this PR based on this repository.）
- `08:09` 或者 Codex 表現極佳的另一件事是除錯。（Or another thing actually that Codex does extremely well is debug.）
- `08:14` 所以每次，沒錯。（So every time, yeah.）
- `08:15` 所以每次當一個對話結束時，我基本上都會說，好，檢查一下我們完成的所有程式碼。（So every time like a session ends, I basically say like, okay, review all the code that we've）
- `08:20` 我們做過的所有東西。（done.）
- `08:20` 有沒有任何潛在的錯誤、回歸問題、邊緣案例，諸如此類的事情，對吧？（Is there any potential bug, regression, edge cases, all this kind of stuff, right?）
- `08:24` 所以現在我知道模型在處理這類任務時表現如何，像是指令遵循、多 Agent 能力。（And so by now, I know when the model performs well on this task, like, you know, instruction）
- `08:29` 以及多模態的整個要素，對吧？（following, multi-agent capability.）
- `08:31` 就像是，好，你將這張圖片轉換成程式碼的能力有多好，對吧？（And then there's an entire element of multi-modality, right?）
- `08:33` 而 4.5 其實非常驚人。（It's like, okay, like, how good are you at taking this image and converting to code, right?）
- `08:38` 事實上，我認為我推薦大家用 5.5 嘗試的一個非常有趣的流程是，（And 4.5 is actually amazing.）
- `08:40` 你知道，生成一張圖片，然後要求它圍繞該圖片製作一些東西，這方面它做得非常好。（In fact, I think like a really fun flow that I recommend people to do with 5.5 is like,）
- `08:44` 對。（you know, generating an image and then asking to make something around that, which is, you）
- `08:48` 這就是為什麼我們開發了 Build Web Apps plugin，我們將使用 GPT Image 2 生成設計的構想打包在一個 plugin 裡。（know, it does really well.）
- `08:49` 然後讓 5.5 去實作那個設計。（Yeah.）
- `08:50` 因為通常來說，將設計匯入並實際實作出來的過程幾乎是完美無缺的。（That's why we build the Build Web Apps plugin, where we package up in one plugin, the idea）
- `08:54` 我覺得這些事情在四、五個月前根本是不可能的，對吧？（of like using GPT Image 2 to generate a design.）
- `08:57` 它們當時可能屬於某個類別，但現在就是變得更好了。（And then get 5.5 to implement that design.）
- `08:59` 那除了遊戲和像素藝術之外呢？（Because usually it's like, it's pretty flawless to take a design in and like actually implement）
- `09:03` 你是一位設計師，過去也設計過圖示。（it out.）
- `09:04` 例如，你現在怎麼看待圖示設計？（And I feel like those stuff truly weren't possible, like, you know, even like four or five months）
- `09:07` 所以我開發了這個應用程式，我待會兒會展示給你。（ago, right?）
- `09:08` 它真的很酷。（And like, you know, they were in some category, but they're just better now.）
- `09:11` 這是一個自動完成的寫作工具。（What about beyond games and pixel art?）
- `09:13` 我把它叫做 Odisseo。（Like you're a designer, you designed icons in the past.）
- `09:16` Odisseo 基本上就是《奧德賽》和尤里西斯的主角。（Like how do you look at icon design, for instance, now?）
- `09:19` 我之所以這樣命名，是因為在這個應用程式中與 AI 一起寫作時，你就像處於一片充滿無限可能與寫作機會的汪洋大海之中。（So I built this app, which I showed you in a second.）
- `09:20` 我喜歡這個概念。（It's really cool.）
- `09:21` 所以我問 Kodak，我說，你可以給我一個包含希臘船隻和羽毛筆的應用程式圖示嗎？（It's an autocomplete writing tool.）
- `09:23` 然後他就直接做出來了。（And I call it Odisseo.）
- `09:25` 這太不可思議了。（And Odisseo is basically the protagonist of like, you know, the Odyssey and Ulysses.）
- `09:28` 而且是一次就成功，對吧？（And the reason why I call it like that is because you're writing with AI in this app,）
- `09:32` 然後他把它加到了 DMG 裡。（you're kind of in this like infinite sea of possibility and writing that you can do.）
- `09:36` 所以現在，你知道，我的應用程式就在這裡。（I like that.）
- `09:37` 我是說，15 年前的 Pietro，你需要花多少小時才能設計出這個？（And so I asked Kodak, I was like, can you give me an app icon that has a Greek ship and）
- `09:42` 喔，別提了。（a quill?）
- `09:43` 這得用 Photoshop，還要多個圖層。（And he just did this.）
- `09:43` 我甚至可以告訴你，這裡應用了漸層效果，邊框也有漸層，還有各種紋理，我的意思是，這真的太瘋狂了。（This is incredible.）
- `09:44` 我甚至可以告訴你，你知道，這裡有...（And in one shot, right?）
- `09:46` 那會需要花上非常、非常多個小時。（And then he added it to DMG.）
- `09:47` 對，要達到這種水準，需要花上好幾個小時的工作時間。（And so now, you know, like my app, it's right here.）
- `09:50` 而你只用了一次就完成了。（I mean, the Pietro of like 15 years ago, how many hours would you have spent trying to）
- `09:54` 對，沒錯。（design this?）
- `09:55` 不過，或許我們可以看一下這個應用程式。我很好奇想多了解一點，因為你也有在玩多模態模型，對吧？（Oh, don't even miss that.）
- `09:56` 而且我們剛在 API 中推出了新的即時音訊模型。（This was like Photoshop, multiple layer.）
- `09:58` 我甚至可以告訴你，你知道，這裡應用了漸層效果，還有帶有漸層的邊框，以及所有這些紋理，我的意思是，這真的太瘋狂了。（I can even tell you, you know, there is like a gradient effect apply here, there is like a border with gradient apply, there's all this like texture, like I mean, it's actually insane.）
- `09:58` 我甚至可以告訴你，你知道，有……（I can even tell you, you know, there's...）
- `10:08` 那可能需要花上非常、非常多個小時。（It would be like many, many hours.）
- `10:09` 沒錯，要達到這種水準，可能需要花上好幾個小時的工作時間。（Yeah, it would be like, to get it to this level, it would be like hours, hours of work.）
- `10:13` 而你竟然一次就搞定了。（And you just one-shoted it.）
- `10:15` 對，沒錯。（Yeah, exactly.）
- `10:16` 不過，或許我們可以看看這個應用程式。我很好奇想多了解一點，因為你也有在玩多模態模型，對吧？（But yeah, so maybe we can take a look at this app. I'm curious to hear more about it, because you also play with like multi-modal models, right?）
- `10:24` 而且我們剛在 API 中推出了新的即時音訊模型。（And we just launched like new real-time audio models in the API.）
- `10:28` 聽起來這也是你已經在嘗試探索的方向。（And it sounds like it's also something you try to poke at already.）
- `10:31` 對，我覺得我一直回歸的主題是關於人類與 AI 心智連結的概念，對吧？（Yeah, so like, I think the theme that I always go back to it is this idea of human to AI mind connection, right?）
- `10:39` 就像是，我能做些什麼，讓像 Codex 或某種 AI 模型這樣的工具，基本上成為我心智的延伸，對吧？（It's like, what can I do so that this tool like Codex or certain AI model becomes basically just like an extension of my mind, right?）
- `10:50` 你可能會感到驚訝。（And you'd be surprised.）
- `10:51` 我覺得我們人類是不可思議的輸入機器，對吧？（I feel like we as human, we're an incredible input machine, right?）
- `10:54` 像是透過視覺、聲音、嗅覺等等，但我們在輸出方面卻很差，對吧？（Like with the vision, the sound, the smell, all those things, but we kind of suck at output, right?）
- `11:00` 所以我當時想，如果我能做一個應用程式，當我在寫作時，它能自動補全接下來的文字，而且我還能跟它對話，會怎樣？（And so I was like, okay, what if I can make an app where as I write, it complete what could be the next text, and then I can also talk to it.）
- `11:09` 所以在這個例子中，我們說：「你知道，今天我要去 OpenAI 辦公室錄製新的一集關於與 Roman 建立腳本的節目，我正在考慮進行演示」，然後我只要按下 Tab 鍵，現在 GPT 模型就會生成，你知道，像是：「好的，這是我的 Agentic 程式開發流程的最後一次互動」。（So in this case, we're saying, you know, today I'm going to the OpenAI office to record a new episode of building a script with Roman, and I'm thinking about demoing, and then I can just press tab, and now a GPD model is generating, you know, like, you know, okay, this is the last interaction of my agentic workflow.）
- `11:24` 所以，如果我說，你知道，如果我說：「噢，也許我想在中間加點東西」，對吧？（So, and like, what if I say, you know, what if I say, like, oh, maybe I want to add something in the middle, right?）
- `11:29` 所以我可以到這裡，它就會自動補全那句話。（So I can go here, and it's going to autocomplete that sentence right there.）
- `11:33` 但後來我想，如果我做這個「心智延伸」的事情會怎樣？（But then I was like, well, what if I do this, like, mind extension thing, right?）
- `11:38` 所以我可以啟動 GPT 即時版本，也就是新版本，現在我可以直接說話了，對吧？（So I can, like, start GPT real-time, the new version, and now I can just talk, right?）
- `11:42` 我可以就在這裡說話。（I can talk right here.）
- `11:43` 現在你可以看到，它即時地將我說的話寫下來，這真的很瘋狂。（And now you can see, in real time, it's basically, you know, writing what I'm saying, which is pretty insane.）
- `11:49` 然後我可以停止它，對吧？（And then I can stop it, right?）
- `11:51` 接著我可以自動補全我說的任何內容，對吧？（And then I can, like, autocomplete whatever I say, right?）
- `11:54` 這真的很令人印象深刻。（That's pretty impressive.）
- `11:55` 噢，在這裡，令人印象深刻。（Oh, here, impressive.）
- `11:56` 簡直就像擁有第二個大腦，即時捕捉每一個稍縱即逝的念頭。（Almost like having a second brain capture every straight hour in real time.）
- `11:59` 這真的很瘋狂，因為那……（That's kind of crazy, because that's…）
- `12:01` 這真的有助於展現我們在讓這些介面感覺像是你內在模型自然延伸這方面，已經走了多遠。（It really helps how far we've gone in making these interfaces feel like a natural extension of your own internal model.）
- `12:05` 這太瘋狂了。（That's insane.）
- `12:05` 這就是你想說的。（That's what you wanted to say.）
- `12:06` 這太瘋狂了。（That's insane.）
- `12:07` 這就是我們在討論的事情。（Like, that's what we're talking about.）
- `12:09` 這太狂了。（That's wild.）
- `12:10` 這太棒了。（That's awesome.）
- `12:10` 我也很喜歡你將這些模態整合在一起的方式，對吧？（I love also how, like, you kind of bring these modalities together, right?）
- `12:15` 不僅僅是，嗯，顯然我是一個寫作者。（Not just, like, well, obviously, I'm a writer.）
- `12:17` 我正在打字。（I'm typing.）
- `12:17` 但同時，我也可以直接對電腦說話，即時捕捉這些想法，然後再透過寫作完成我的思緒。（But also, I can just, like, talk to the computer, capture that in real time, and then finish my thoughts in writing.）
- `12:24` 這種互動非常神奇。（This interplay is pretty magical.）
- `12:26` 老實說，你構建的 UI 非常漂亮。（And honestly, the UI that you build is pretty nice.）
- `12:28` 像是，有深色模式。（Like, there's, like, a dark mode.）
- `12:30` 有所有這些指令。（There is, like, all these commands.）
- `12:32` 就像，你知道，如果我想，我不確定，像是將這段文字加粗，我只要按 Command B。（Like, you know, if I want to, like, I don't know, like, bold this, I can just command B.）
- `12:35` 我的意思是，這真的很瘋狂，對吧？（I mean, it's pretty crazy, right?）
- `12:37` 而且這真的只是在跟模型對話。（Like, and it's just literally just talking to the model.）
- `12:40` 只是跟模型對話。（Just talk to the model.）
- `12:41` 我認為這是一個完美的例子，關於你夢想的一半是什麼，對吧？（And I think this is a perfect example of, like, what is the half of your dream, right?）
- `12:45` 那些你留在心智抽屜裡，卻因為忙碌或認為自己做不到而從未實現的夢想，對吧？（That you had, you left in the drawer of your mind, and you never built because you were busy, or you thought you couldn't done it, like, all those things, right?）
- `12:53` 我覺得現在基本上每個人都能做到這一點。（And I feel like now, basically, everybody can do this.）
- `12:55` 我對 Codex 的類比，前幾天我一直在想，就像是，過去並沒有很多人是偉大的畫家，對吧？（And, like, the analogy that I have with Codex, which I was thinking about the other day, was, like, it's, like, in the past, not a lot of people were great painters, right?）
- `13:03` 但每個人都可以拍照。（But everybody can take a photo.）
- `13:05` 很棒。（Great.）
- `13:05` 並非每個人都是偉大的攝影師，但每個人都可以拍照。（Not everybody's a great photographer, but everybody can take a photo.）
- `13:08` 對我來說，這就像是程式開發的「拍照」，對吧？（And to me, this is, like, the taking a photo of coding, right?）
- `13:12` 這就像是我的一個想法的快照，現在它處於這種新的、基本上是程式開發正在變成的媒介或素材中。（It's, like, it's, like, it's just a snapshot of an idea that I had, and now it's in this new, sort of, basically, medium slash material that coding is becoming.）
- `13:21` 我覺得現在的程式開發正變成一種非常可塑、像黏土一樣的東西。（I feel like coding now is becoming this very malleable, clay-like type of thing.）
- `13:26` 沒錯，你就是可以隨手建立東西。（Yeah, you can just build things.）
- `13:27` 你只需要動手做出來就好。（You can just build things.）
- `13:27` 對，沒錯。（Yeah, yeah.）
- `13:28` 正是如此。（Exactly.）
- `13:29` 那真的很棒。（That's pretty awesome.）
- `13:30` 謝謝你跟我分享這些專案。（Thank you for sharing me, like, all of these projects.）
- `13:32` 能看到你的創意真的很酷。（Really cool to see your creativity.）
- `13:33` 其實我很好奇，你嘗試了這麼多種模態，從影像生成，到這些語音和即時模型，一直到 5.5 的最前沿。（In fact, I'm curious, like, you're playing with so many modalities from image gen to, like, these voice and real-time models to the frontier of 5.5.）
- `13:43` 當 OpenAI 發布新模型而你獲得搶先體驗時，你是怎麼進行的？（Like, when a new OpenAI model drops and you get early access, like, how do you go about that?）
- `13:49` 也就是說，你研究這些模型的流程或儀式是什麼？（Like, what's your process or your ritual to go about, like, poking at those models?）
- `13:54` 沒錯，我很喜歡這個問題，因為這基本上是我隨時都在思考的事情。（Yes, I love that question because it's something that I basically think about all the time.）
- `13:57` 有時候我甚至睡不著，因為我在想，我還能做些什麼？（Like, sometimes I don't sleep because I think, what are the things I can do?）
- `14:01` 我幾乎有一種所謂的「開發焦慮」，因為我只是想確保自己有時間去實踐所有這些想法。（There is almost like this, like, quote-unquote building anxiety that I have because I just want to make sure I have time to do all those ideas.）
- `14:07` 但有趣的是，這些模型在編寫軟體——你知道的，就是電腦軟體——方面變得越來越強，感覺它們在這方面已經趨於飽和了。（But what's interesting, it's almost like the models are getting so good at coding software and, you know, just computer software that it feels like they're most saturated in a way.）
- `14:17` 所以我總是會想，還有什麼是我沒嘗試過的？（And so I'm always like, what can I do that I haven't tried?）
- `14:20` 有沒有什麼真的很酷的事情——也許是我以前有過的想法，但因為過去嘗試時沒成功而無法實現的？（And is there something that is really cool that I – maybe an idea that I had that I couldn't do before because I tried in the past and it just didn't work out?）
- `14:27` 嗯，5.5 的一個有趣副作用是，我基本上讓自己被那些在抽屜裡放了好幾年的舊裝置包圍了。（Well, and an interesting side effect of 5.5 is that I basically surround myself with all these old gadgets that I had in my drawers for years.）
- `14:37` 而且，你知道，我不懂這些裝置的語言。（And, you know, I don't speak the language of these devices.）
- `14:40` 比如，我不知道它們是用什麼程式碼寫的。（Like, I don't know what they're coded on.）
- `14:42` 這可能是在 C 語言中編寫的。（Like, this might be coded in C.）
- `14:43` 這可能是在某種專有語言中編寫的等等，對吧？（This might be coded in some proprietary language and whatnot, right?）
- `14:46` 所以我當時想，如果 5.5 中的 Codex 能理解這些裝置的本質，然後為它進行開發，那會不會很酷？（And so I was like, would it be cool if Codex in 5.5 could understand the nature of these devices and then building out for it?）
- `14:54` 那真的很酷。（That's pretty cool.）
- `14:54` 我很喜歡你的想法，一方面探索人工智慧的最前沿，另一方面又回過頭來研究這些舊裝置，看看能用它們做些什麼。（I love the idea of you, like, on one hand, poking at the very frontier of AI, but on the other hand, you know, coming back to these, like, old devices and see what you can potentially do with them.）
- `15:06` 對，這很有趣，因為這幾乎是一種讓這些過時技術重獲新生的方式。（Yeah, and it's interesting because it's, like, almost a way to revive this obsolete tech.）
- `15:10` 這幾乎有一種環境保護的元素在裡面，因為我現在有所有這些東西，你知道的，它們在我的抽屜裡，壞掉了，我不知道該怎麼處理。（Like, there's almost an element of, like, almost an environmental element that comes from, oh, now I have all those things that, you know, they're in my drawer, they're broken, and I don't know what to do with it.）
- `15:19` 我可以把它們重新利用，我覺得這真的很有趣。（I can just repurpose something else, which I think is really interesting.）
- `15:22` 這就是吸引你回頭研究它們的原因嗎？（Is that what kind of pulls you back to them?）
- `15:24` 就像是，我有一抽屜的舊裝置。（It's like, well, I have these drawers of, like, old devices.）
- `15:28` 果不其然，在某個時刻，我可以利用它們做些有創意的事情。（Sure enough, at some point, I can do something creative with them.）
- `15:30` 正是如此。（Exactly.）
- `15:31` 就像我可以做一些我可能可以用於工作，或者，你知道，純粹為了好玩之類的事情，對吧？（It's like I can do something that I could maybe use for my work or, you know, to have fun or whatnot, right?）
- `15:35` 那麼跟我介紹一下這個吧。（So tell me about this one.）
- `15:36` 這是什麼？（What is it?）
- `15:36` 嗯，這基本上就是一個用來做安全控制的裝置，所以如果你想的話，它算是一個工作裝置。（Yeah, so this is basically just, like, it's a device to do, like, security control on, like, so it's kind of like a working device if you want to actually to think about it.）
- `15:45` 我當時想，為這個做一個遊戲會不會很有趣？（And I was like, wouldn't it be fun to make a game for this?）
- `15:49` 但後來我又想，你知道，如果我為這個做遊戲，感覺有點奇怪。（But then I was like, you know, if I was making a game for this, it's kind of, like, weird.）
- `15:52` 就像，你知道，螢幕很奇怪。（Like, there's, like, you know, the screen is weird.）
- `15:54` 我不知道該怎麼玩。（Like, I don't know how to play.）
- `15:55` 但後來我心想，嗯，如果我這樣玩，也許會更有趣。（But then I was like, hmm, but if I play it this way, maybe it's more interesting.）
- `15:58` 所以我請 Codex 基本上理解了這個韌體。（And so I asked Codex to basically understand this firmware.）
- `16:03` 它做到了。（It did that.）
- `16:04` 然後……（And then...）
- `16:04` 所以你只是，像是，插上了這個裝置？（So did you just, like, plug the device?）
- `16:06` 對。（Yes.）
- `16:07` 連接到 Codex？（To Codex?）
- `16:07` 我把裝置連接到了 Codex。（I plugged the device to Codex.）
- `16:08` 然後你說，像是……（And you said, like...）
- `16:09` 讀取 USB。（Read the USB.）
- `16:10` 對。（Okay.）
- `16:11` 我當時說，讀取 USB。（I was like, read the USB.）
- `16:12` 就像，這是什麼？（Like, what is it?）
- `16:13` 然後他說，這個裝置已連線。（And he was like, this device is connected.）
- `16:15` 然後它就像是，這是這個裝置，你知道，這是序號，諸如此類的。（And it's like, is this device, you know, this is the serial number, whatever.）
- `16:18` 然後他說，喔，我其實可以在這上面做點什麼。（And he was like, oh, I can actually do something on this.）
- `16:21` 然後我說，不會吧。（And I was like, no way.）
- `16:22` 所以它做出了這個，像是，某種，像 Doodle Jump 的遊戲。（So it built this, like, sort of, like, doodle jump.）
- `16:25` 不會吧。（No way.）
- `16:26` 這真的很有趣，對吧？（Which is really funny, right?）
- `16:28` 然後它就像是...（And it's like...）
- `16:29` 但更令人感興趣的是，我讀了它們的 API 和文件。（But what's even interesting about this, I read their API and doc.）
- `16:32` 裡面完全沒提到能做到這種垂直作業，或是像螢幕垂直捲動之類的功能。（There is no mentioning of being able to do this sort of vertical work or, like, vertical scrolling of the screen.）
- `16:39` 但 Codex 就這樣自己摸索出來了，這真的很瘋狂。（But Codex just figured that out, which is really crazy.）
- `16:42` 所以現在我有這個，基本上是無盡的遊戲，基本上就像 Doodle Jump，早期的 iOS 遊戲，對吧？（And so now I have this, like, basically endless game, basically to, like, Doodle Jump, like early iOS game, right?）
- `16:48` 老實說，我覺得這太瘋狂了，對吧？（Which is, honestly, I think it's crazy, right?）
- `16:50` 然後我就說...（And then I was like...）
- `16:50` 這螢幕其實還蠻適合玩的。（The screen is pretty good to play, actually.）
- `16:52` 不，這很有趣。（No, it's fun.）
- `16:53` 令人驚訝的是...（Surprisingly...）
- `16:53` 對，你應該試試看。（Yeah, you should try.）
- `16:54` 它，就像，玩起來還蠻有趣的。（It's, like, it's pretty fun to play.）
- `16:56` 對，你只要四處移動，然後，控制器就會跟著動。（Yeah, you just move around and, like, the controllers move.）
- `16:59` 還有另一個我感興趣的地方是...（And the other interesting thing that I...）
- `17:01` 在我讀完 Codex 在做什麼之後，這個小小的類比按鈕其實有不同的靈敏度，取決於你是長按還是短按。（After I read what Codex was doing, this little analog button actually has, like, different sensibility if you long press and if you, like, short press.）
- `17:09` 而 Codex 就這樣自己摸索出來，確保你知道，體驗是如此流暢。（And Codex just, like, figured it out to make sure that, you know, the experience was this smooth.）
- `17:14` 喔，對。（Oh, yeah.）
- `17:14` 它甚至連二段跳都做到了。（It even nailed the double jump.）
- `17:16` 對，二段跳。（Yeah, the double jump.）
- `17:17` 對，如果你想要，像是，二段跳，如果你距離太遠，你可以二段跳。（Yeah, if you want to, like, double jump, if you're, like, too far, you can double jump.）
- `17:19` 我是說，那真的只用了一個 prompt。（I mean, and that was truly one prompt.）
- `17:23` 不可思議。（Incredible.）
- `17:23` 當我看到的時候，我心想，這不可能。（And when I saw it, I was like, there's no way.）
- `17:25` 我從沒見過這個裝置。（I have never seen this device before.）
- `17:26` 這太瘋狂了。（It's crazy.）
- `17:27` 這太棒了。（It's awesome.）
- `17:29` 然後我就想，好吧，這一切都很有趣，對吧，隨便啦。（And then I was like, well, this is all fun, right, whatever.）
- `17:32` 但如果我能把它用在工作上呢？（But what if I could use this for work?）
- `17:34` 像是，有什麼有趣的事情是我可以在工作上做的？（Like, what is an interesting thing I could do for work?）
- `17:37` 所以，我想，像是，今天對話的一個團隊，一直都是關於這種連結心智 AI，對吧？（And so, I think, like, a team of, like, today's conversation has been, like, this connection mind AI, right?）
- `17:44` 我們能用這些模型做什麼？（And what can we do with those models?）
- `17:46` 像是，我們能做哪些事情？（Like, what are the things we can do?）
- `17:47` 所以，你可以看到這裡，基本上，Codex 建立了一個應用程式，讓它能自我連結。（So, you can see here, basically, Codex build an app so that it could connect itself.）
- `17:55` 所以，我就想，好吧，我能做什麼對工作有用的事情？（And so, like, I was like, okay, what can I do that's, like, useful for work?）
- `17:58` 所以，Codex 建立了這個控制項，如果我在這裡發送一則文字，它在完成時基本上會震動。（So, Codex build this control where if I send a text here, it basically vibrates when it's done.）
- `18:06` 所以，當然，你幾乎聽不到，但它正在震動。（So, of course, you can nearly hear it, but it's vibrating.）
- `18:08` 我感覺到了。（I feel it.）
- `18:09` 對吧？（Right?）
- `18:09` 然後它向我展示了訊息，對吧？（And he showed me the message, right?）
- `18:11` 標題。（Title.）
- `18:11` 對。（Yeah.）
- `18:12` 就像是，好吧，現在沒有訊息了，因為我們聊得更多了。（It's like, okay, now it's no message because we talk more.）
- `18:15` 我當時想說，我一直說，嘿。（I was like, I keep saying, like, hello.）
- `18:17` 然後，我可以看到文字。（And so, I can see the text.）
- `18:19` 而且它在震動。（And it's vibrating.）
- `18:21` 這很酷。（That's cool.）
- `18:22` 我這麼做的原因，聽起來可能很宅，對吧？（And the reason why I did this, this is going to sound so nerdy, right?）
- `18:24` 但我跟 Codex 有過很長、很長的對話。（But I have a long, long session with Codex.）
- `18:27` 有時候我覺得 Codex 現在可以非常不受監督地運作。（And sometimes I feel like Codex can be very unsupervised now.）
- `18:31` 我信任它。（I trust it.）
- `18:32` 就像，我知道它會執行 20、30 分鐘，或是 10 分鐘，無論任務是什麼。（Like, I know it's going to run for 20, 30 minutes or 10 minutes, whatever the task is.）
- `18:35` 然後當我回來時，它就完成了。（And then when I come back, it's done.）
- `18:36` 但總是有這種成分，像是，喔，萬一它卡住了怎麼辦之類的，對吧？（But there's always this element of, like, oh, what if it goes stuck or whatever, right?）
- `18:40` 所以，現在我這麼做，你知道，我去辦公室拿杯咖啡，然後我聽到，嘟嘟嘟。（And so, now what I do, I'm like, you know, I go grab a coffee in my office and I hear, like, do-do-do.）
- `18:46` 我聽到一點震動。（I hear a little vibration.）
- `18:47` 我就知道 Codex 完成了。（I know Codex is done.）
- `18:47` 在你的桌上。（On your desk.）
- `18:48` 你實際上可以離開你的座位，然後你會知道 Codex 何時完成。（You can actually step away from your desk and you know when Codex is done.）
- `18:50` 我知道它完成了。（I know it's done.）
- `18:51` 這是一種非常有趣的方式，去使用一個我從未用過的裝置。（And this is, like, a really interesting way to, like, use a device that, you know, I've never used.）
- `18:56` 你知道，它基本上就像一個遙控器。（You know, it's almost like a remote control, basically.）
- `18:59` 太迷人了。（Fascinating.）
- `19:00` 我們談了很多關於你作為工程師、創作者、設計師的 skill set。（We talked a lot about, like, your, you know, skill set as an engineer, a creator, a designer.）
- `19:06` 我想聊聊作為創辦人的你。（I'd love to talk a little bit about, like, you as a founder.）
- `19:08` 所以，你正在打造 Magic Path。（So, you're building Magic Path.）
- `19:10` 是的。（Yes.）
- `19:10` 可以跟我們介紹一下那是什麼嗎？（Can you tell us a little bit about what that is?）
- `19:12` Magic Path 基本上是一個讓 Agent 與人類共同設計和建構的無限畫布。（So, Magic Path is basically an infinite canvas for agent and human to design and build together.）
- `19:17` 所以，我們打造了一種共同的 workspace，你可以從任何地方存取。（So, we built sort of this, like, common workspace where you can access from everywhere.）
- `19:21` 所以，你可以讓 Codex 在任何裝置上執行，無論是在你的手機還是其他地方。（So, you can have, like, you know, like, you can have Codex running in any device on your phone or whatever.）
- `19:26` 然後你就可以在 Magic Path 裡進行設計。（And you can just, like, design in Magic Path.）
- `19:28` 因為我認為現在真正發生的事情是，每個人在某種程度上都會擁有自己的 Agent。（Because what I really think is happening right now, you have, everybody's going to have their own agent in some extent.）
- `19:33` 就像，你有你的 Agent，它擁有你喜歡的所有資訊，以及你喜歡的所有記憶。（Like, you have your agent that has all the information that you like, all the context that you like.）
- `19:37` 我有我喜歡的那個，以及記憶。（I have the one that I like and the context.）
- `19:39` 所以，想法是，我們如何能成為產品團隊的參考工具，對吧？（And so, the idea is, like, how can we become the reference tool for the product teams, right?）
- `19:45` 因為現在，有一個很大的問題，就是你有 PM。（Because right now, there's, like, a big issue where you have, you know, the PM.）
- `19:48` 你應該要成為設計師。（You should have to be the designer.）
- `19:49` 設計師應該要成為工程師。（The designer should be the engineer.）
- `19:50` 工程師應該要成為，你知道的。（The engineer should be the, you know.）
- `19:51` 這也是因為人工智慧基本上正在瓦解所有這些角色，因為人工智慧正在成為一種翻譯層，你知道的。（And it's also because AI is basically collapsing all those roles because AI is becoming this sort of, like, translation layer, you know.）
- `19:56` 某種程度上的翻譯層。（translation layer in a way.）
- `19:58` 它也是一個平衡器。（It's also an equalizer.）
- `19:58` 沒錯，它就是一個平衡器。（It's an equalizer, exactly.）
- `20:00` 但正因如此，人們傾向於停留在各種不同的工具中。（But because of that, people tend to stay in all different kinds of tools.）
- `20:03` 有些人可能會在 Codex 中存取 5.5，有些人可能會在 ChatGPT 中存取，（Some people might access 5.5 in Codex, some people might access in ChGPT,）
- `20:08` 有些人可能會在其他地方存取。（some people might access somewhere else.）
- `20:10` 所以我們試圖成為那個中心。（And so we're trying to be the place.）
- `20:12` 而且真正酷的是，Codex 可以在 Magic Path 中建構。（And so what's actually really cool, Codex can build in MagiPath.）
- `20:14` 所以它只是一個讓 Codex 去探索、設計之類的畫布。（So it's just a canvas for Codex to explore and design and stuff like that.）
- `20:19` 而且我們其實是一年前推出的。（And we launched like a year ago, actually.）
- `20:22` 所以這幾乎是我們的一週年紀念。（So this is almost like our year anniversary.）
- `20:24` 這太棒了。（And it's awesome.）
- `20:24` 我們擁有數十萬名使用者。（We have like hundreds of thousands of customers.）
- `20:27` 特別是 Codex 5.5 對我們幫助非常大，（And Codex, in particular, 5.5 has been like a huge help for us）
- `20:31` 因為我們正準備推出應用程式的 2.0 版本。（because we're about to launch the 2.0 version of the application.）
- `20:34` 老實說，如果沒有 5.5，（And truthfully, we wouldn't be able to get where we are right now）
- `20:37` 我們根本無法達到現在的進度。（if we didn't have 5.5.）
- `20:39` 我可以肯定地告訴你。（I can tell you that.）
- `20:40` 我並不是在誇大其詞，你知道的。（And I'm not even trying to like, you know, over say things.）
- `20:44` 這就是事實。（It's just the reality.）
- `20:45` 因為我們是一個小團隊。（Because we're a small team.）
- `20:46` 我們的產品團隊，你知道的，（We're like, you know, our product team.）
- `20:47` 大約只有六個人。（We're like six people.）
- `20:49` 所以，我們行動非常迅速，（And so, you know, we move really fast）
- `20:51` 因為在這個人工智慧時代，你必須行動非常迅速。（because you have to move really fast in this age of AI.）
- `20:53` 但讓我的所有員工都能使用 Codex，就像是一種強制功能，（But having Codex access to all my employees becomes sort of like a forcing function）
- `20:58` 現在我等於擁有了一百或兩百名開發者，諸如此類，對吧？（where now I have access to 100 developers or 200 developers and stuff like that, right?）
- `21:03` 或者說，我等於擁有了五位非常、非常優秀的新開發者。（Or maybe I have access to five new developers that are like really, really good developers.）
- `21:06` 沒錯。（Right.）
- `21:07` 對吧？（Right?）
- `21:08` 所以，這真的幫助非常大。（And so, that's just been like a huge help.）
- `21:10` 但我想分享一個有趣的軼事，身為創辦人，我有太多事情需要處理，對吧？（But I would say a funny anecdote is that as a founder, there's so many things that I need to do, right?）
- `21:17` 我需要募資。（I need to raise money.）
- `21:18` 我需要與客戶溝通。（I need to like talk to my clients.）
- `21:19` 我需要確保員工過得開心，諸如此類的事。（I need to make sure like my employees are happy, all this stuff.）
- `21:21` 現在人工智慧這麼強大，5.5 也這麼好用，甚至出現了一種感覺，（And now that AI is so good and 5.5 is so good, there's almost an element of like,）
- `21:27` 天啊，我想在應用程式中加入好多功能。（oh my God, I want to push so many things in the app.）
- `21:29` 就像我的 PR 請求數量增加了，我不確定，（Like I've been, like my PR request has been going up like, I don't know,）
- `21:33` 因為 5.5 的關係，大概增加了十倍。（like a tenfold because of 5.5.）
- `21:35` 所以，身為創辦人，重要的是要能夠，你知道的，（And so, like there's an element of like, as a founder, it's also important to kind of like, you know,）
- `21:38` 告訴自己：「好，我知道你現在想開發任何東西都可以，（say, okay, I know you can build anything you want right now,）
- `21:42` 但你同時也聘請了這些優秀的人才。」（but you also have these incredible people that you hire.）
- `21:45` 沒錯。（Right.）
- `21:45` 所以，你知道的，你可以信任他們，對吧？（So, you know, you can trust, right?）
- `21:46` 因此，在我身為創辦人與身為一個只想開發產品的人之間，（And so, there's this mostly sort of like inner battle between me as a founder,）
- `21:50` 存在著一種內心的掙扎。（as a person that just want to build stuff.）
- `21:52` 你就是想開發。（You just want to build.）
- `21:52` 你就是想開發。（You just want to build.）
- `21:53` 但我必須處理所有其他同樣重要，甚至在某些層面上更重要的事情。（But I have to do all those other things which are as equally important or more important in some sense.）
- `21:58` 所以，是的，這是一次很棒的體驗。（And so, yeah, it's been an amazing experience, yeah.）
- `22:02` 沒錯，謝謝你的分享。（Yeah, thanks for sharing.）
- `22:03` 這太棒了。（That's awesome.）
- `22:03` Codex 如何改變了你個人的工作方式？（How has Codex changed the way you work personally?）
- `22:06` 你是花更多時間在與 Codex 對話嗎？（Like, are you just spending way more time talking to Codex now?）
- `22:09` 我跟 Codex 聊得太多了，以至於我女友問我：「你在做什麼？」（I talk to Codex so much that my girlfriend was like, what are you doing?）
- `22:11` 「你總是開著這個應用程式。」（You're always with this like, you know, app open.）
- `22:14` 「你在跟誰說話？」（Who are you talking to?）
- `22:15` 就像是，她是誰？（It's like, who is she?）
- `22:16` 不，也許不是那樣，但她總是，你知道的，（No, maybe that's not kind of, but she was like, you know,）
- `22:18` 她總是問我：「為什麼你要對著它說話？」（she's always like asking me like, why are you talking this?）
- `22:20` 因為我一直在跟 Codex 對話，對吧？（Because I'm always talking to Codex, right?）
- `22:23` 其實，我這樣做的原因，可能會讓你感到驚訝，（And actually, the reason why I do that, this might be surprising,）
- `22:26` 但我並不只是用 Codex 來寫程式。（but I don't use Codex just for coding.）
- `22:28` 例如，我會用它來整理電腦裡的檔案。（Like, I use it to like organize the files on my computer.）
- `22:31` 或者，你知道的，我有多年前留下來的音樂庫。（Or like, you know, I have this like music library from like years ago.）
- `22:33` 它修復了所有的封面、中繼資料，以及所有這類東西。（Like, it fixed all the covers, the metadata, all of those things.）
- `22:37` 或者我讀過的筆記。（Or I've read the notes.）
- `22:37` 所以其實，我已經讀過今天的筆記了，你知道的，有些事情我想跟你聊聊。（So actually, I've read the notes for today, things, you know, like I want to like talk to you.）
- `22:42` 我問 Codex，嘿，你知道的，這是我想要討論的事情。（I ask Codex, hey, I have like, you know, this is kind of stuff I want to talk about.）
- `22:45` 寫進 Apple Notes 裡。（Write into Apple Notes.）
- `22:46` 然後它就直接做了，對吧？（And he just did that, right?）
- `22:47` 因為它擁有我所有本機系統和檔案的存取權限。（Because he has access to all my local system, all my local file.）
- `22:50` 所以我認為現在告訴朋友們一件有趣的事，就是實際去使用 Codex，（And so I think like an interesting thing that now I'm telling my friends is to actually use Codex,）
- `22:55` 不只是把它當作程式開發助手，而是把它當作日常萬能助手，對吧？（not just like this coding assistant, but like as an everyday everything assistant, right?）
- `23:00` 特別是它對我非常有幫助，因為 Codex 現在可以在 MagicPath 中生成設計。（And especially, it's super helpful for me because, you know, because Codex can now generate designs in MagicPath.）
- `23:06` 我把所有的 MCP 都連接到了 Codex，像是我的 Mixpanel（用來查看公司營運狀況）、Stripe，以及所有這類東西。（I have all my MCP connected to Codex, like, you know, my mix panel for like how my company is doing, Stripe, like all those things.）
- `23:13` 我可以直接在 Codex 中調用所有這些資料，然後為我的投資人製作精美的儀表板設計。（I can just pull up all this data in Codex and then I can make just like beautiful dashboard design for my investors.）
- `23:18` 這就像是從想法、資訊、情境到一般化的一對一對應，你知道的。（Like that's sort of like one-to-one, you know, from idea, information, context to generalization.）
- `23:25` 這真的非常強大。（It's like so powerful.）
- `23:27` 而且 Codex 就是……然後我用了過去幾個小時，所以我不知道我花了多少錢。（And just Codex is like, and then I use past hours, so I don't know how much money I'm spending.）
- `23:31` 你們可能給了我一些額度。（You guys probably gave me some credits.）
- `23:33` 但沒錯，這真的很不可思議。（But yeah, it's incredible.）
- `23:35` 很高興聽到這些。（That's great to hear.）
- `23:36` 我的意思是，我們在 OpenAI 也看到了，幾乎每個人現在都在使用 Codex，（I mean, we see that at OpenAI too, like virtually everyone at OpenAI is now using Codex,）
- `23:40` 無論他們屬於哪個團隊，從法務、財務到人資，而不僅僅是工程師。（no matter which team they're on, from like legal to finance to HR, not just the engineers.）
- `23:45` 這也是為什麼我們讓入門變得更容易，因為我們注意到在 OpenAI 之外，（And that's why we also made the onboarding easier because we are noticing that outside of OpenAI too,）
- `23:51` 像你這樣的人也正在利用 Codex 做各種事情。（like people like yourself who are just leveraging Codex or anything.）
- `23:55` 現在我們不僅有了這些 plugin，還有如你所說的，（And now that we have like all of these plugins, but also the ability, as you said,）
- `23:59` 能夠驅動瀏覽器、操作電腦的能力，任何任務都可以由 Codex 完成。（to kind of drive the browser, drive the computer, any task can be completed by Codex.）
- `24:04` 沒錯。（Exactly.）
- `24:05` 任何你想要的。（Anything you want.）
- `24:06` 沒錯。（Yeah.）
- `24:06` 那麼，有了現在團隊擁有的這些速度，你如何看待客戶，以及這對你們來說意味著什麼？（Well, with all that speed that you have now with your team, like how are you thinking about like the customers and what does that unlock for you guys?）
- `24:14` 你知道，人們常談論「vibe coding」（憑感覺寫程式），但感覺我們已經遠遠超越那個階段了。（Like, you know, people talk a lot about vibe coding, but it feels like we're far beyond that point now.）
- `24:19` 沒錯，我的意思是，我覺得因為這種速度，我們能夠更快地將產品送到客戶手中。（Yeah, I mean, I feel like because of the velocity, we're able to get our product into, you know, customer hands way quicker.）
- `24:25` 我覺得讓客戶直接給你回饋的力量，這真的……你可以把它跟以前開發產品的方式做比較。（Like I feel like the power of having your customer like just giving you feedback right away, it's like, it's just you can compare that to how product used to be built before.）
- `24:33` 你知道，我過去在很多大型科技公司工作過，我覺得那個過程會花上好幾個月。（Like, you know, I worked on lots of big tech companies in the past and I feel like that process will take months.）
- `24:39` 你知道，你獲取使用者回饋，收集所有這些資訊，然後你必須花上好幾個月來開發那個東西。（Like, you know, you get like the user feedback acquisition, you get all this information, you have to spend, you know, months building the thing.）
- `24:44` 然後你甚至可能不知道你開發的東西是否是正確的，對吧？（And then you might not even know if the thing that you were building is the right thing to do, right?）
- `24:48` 因為如你所知，在開發過程中，總是有其他方向可以選擇。（Because as you know, like as you're building, there's always like some other direction you can go.）
- `24:52` 所以現在，因為我們可以真正地在一天內生成一個功能，我可以……我們在 Magic Path 有一個資料測試小組。（And so now, because, you know, we can literally spawn a feature in a day, I can have, we have like a data tester group at Magic Path.）
- `24:59` 所以我基本上可以直接說，嘿，你們覺得這個怎麼樣？（And so I can basically just like say, hey, what do you guys think about this?）
- `25:01` 然後他們回饋我，這很棒，這不行，砰、砰，完成了，對吧？（And they get back to me, this is great, this is not, boom, boom, done, right, done, right?）
- `25:05` 所以這是以前無法實現的事情。（And so that is something that you couldn't achieve before.）
- `25:08` 擁有那種更快速的回饋迴圈。（Like having that feedback loop that's much faster.）
- `25:10` 所以這不僅僅是模型扮演潛在客戶，更是模型讓你能夠更快地接觸到你的客戶，這真的很棒。（And so it's not only just the model acting like potential customer, but it's also like the model enable you to go to your customer much quicker, which is really nice.）
- `25:18` 沒錯，那太棒了。（Yes, that's great.）
- `25:19` 我們上次交談時討論的一件事是，我們如何邁向一個「監督」的世界。（One thing that we talked about the last time we spoke together was how we're moving to a world of like supervisions.）
- `25:27` 我很想聽你多談談這一點。（I'd love to hear you like speak more about this.）
- `25:30` 沒錯。（Yes.）
- `25:30` 所以我其實非常相信這一點，以至於我圍繞著它建立了我的公司。（So I actually, I believe in this so much that I built my company around it.）
- `25:35` 所以如果我錯了，我就完蛋了。（So if I'm wrong, I'm screwed.）
- `25:38` 信念。（Conviction.）
- `25:38` 信念。（Conviction.）
- `25:39` 但我的想法是，你知道，我相信隨著我們邁向未來的工作型態，重點將不再是親手執行，而是如何指揮這些工作。（But the idea is like, you know, I believe that as we go into the future of work, it's going to be less about doing the thing, but more about directing the thing.）
- `25:48` 沒錯。（Right.）
- `25:48` 我真的很相信這個去中心化的 Agent 世界，你可以從任何地方存取它們。（I really believe in this world of decentralized agents that you can access from everywhere.）
- `25:53` 舉例來說，就像我剛才提到的，在 Magic Path，我可以拿著手機，開啟通話。（And so, for example, you know, as I was saying, in Magic Path, I can be on my phone, an open call.）
- `25:56` 我說，比如我現在躺在床上。（And I say, like, I'm in my bed.）
- `25:58` 凌晨兩點，我突然有個靈感。（I have an idea, you know, 2 a.m. in the morning.）
- `26:00` 我會說，我們能不能直接用這份 PRD 的內容來修改那個部分？（I was like, you know, like, can we just change that using, you know, this content from this PRD?）
- `26:05` 因為我剛好想到這個點子。（Because I had this idea.）
- `26:06` Agent 就會去執行、完成工作，然後回傳給我一個連結。（The agent goes, makes the thing, comes back, send me a link.）
- `26:09` 我查看之後。（I see it.）
- `26:10` 覺得很棒。（And I'm like, great.）
- `26:11` 不需要回饋。（No, feedback.）
- `26:12` 沒錯。（Right.）
- `26:13` 我甚至不需要再坐在辦公桌前了。（I'm not, I don't even have to be at my desk anymore.）
- `26:15` 對我來說，這件事有一種非常、所謂的「浪漫」，以及一種非常、充滿希望的感覺，因為我認為作為人類，我們唯一共同擁有的東西，就是時間。（And to me, there's something very, quote, unquote, romantic and, like, very sort of, like, hopeful about that, which is, like, I think as a human, right, the only thing that we have that it's in common, it's time.）
- `26:28` 沒錯。（Right.）
- `26:29` 所以我感覺人工智慧對我來說，就像是人類對時間的一種反擊。（And so, I feel like AI is really, to me, it's, like, this revenge of human against time.）
- `26:34` 就像是，喔，我們原本知道開發這個產品需要六個月。（Like, it's like, oh, we knew that it would take us six months to build this product.）
- `26:39` 現在，只需要兩週。（Now, it would take us, you know, two weeks.）
- `26:40` 這節省下來的大量時間，你可以用來陪伴家人、朋友和摯愛，對吧？（And that's a huge saving in time that you can spend with your family, with your friends, with your loved ones, right?）
- `26:45` 所以，當我想到這個工具時，心中會湧現出一種非常巨大的樂觀情緒，因為它讓我能夠——我是說，我身為創辦人工作量很大，但我也能直接去見我的朋友，因為我知道我有一個 Agent 正在運作，處理著我本該完成的工作。（And so, there is an element of, like, really tremendous optimism that comes to me when I think about this tool, because it is enabled me to, I mean, I worked a ton because I'm a founder, but, you know, like, I can also just, like, go and see my friend because I know that, you know, I have an agent running and is doing the work that I should do.）
- `27:00` 這真的很棒。（Well, that's awesome.）
- `27:01` 如果有一件事是你想要留給聽眾的，或許是一種你希望他們感受到的心情，那會是什麼？（If there's, like, one thing that you'd like to leave people with, maybe, like, one feeling you'd like them to have, what would that be?）
- `27:08` 對我來說，那會是一種難以置信的狂喜，來自於知道我現在可以做任何我想做的事，並打造任何我想要的產品。（For me, it would be this, like, incredible feeling of euphoria that comes from knowing that I can just do whatever I want now and just build anything I want.）
- `27:17` 這太棒了。（That's awesome.）
- `27:18` 這也非常符合我們的理念，回到 OpenAI 的使命，我們想要為所有開發者、為每一個人打造工具。（And very much, like, this idea for us, like, going back to OpenAI's mission that, like, we want to build tools for all builders and for everyone, really.）
- `27:27` 現在每個人都是開發者。（And now everyone is a builder.）
- `27:28` 沒錯，我認為你們做到了一件很了不起的事，特別是透過這個新模型，真正地將原本對許多人來說非常陌生、遙遠的東西，變成了每個人都能使用的工具，對吧？（Yeah, I think, like, you know, the one amazing thing you guys have been able to do, especially with this new model, really, like, democratizing something that was so unknown and foreigner to a lot of people to now something that everybody can use, right?）
- `27:38` 很酷的地方在於，我是在義大利長大的，對吧？（Like, the cool thing is, like, you know, I grew up in Italy, right?）
- `27:41` 我在義大利長大時，感覺自己並沒有像身在美國的人那樣的機會。（I was a kid in Italy, and I feel like growing up there, I didn't have the same sort of, like, you know, possibility that somebody could be in the U.S.）
- `27:48` 但現在我覺得，因為每個人都能存取這種不可思議的智慧，我們在某種程度上變得平等了，我覺得這真的很美。（But now I feel like because everybody gets access to this, like, incredible intelligence, we all equalize in a way, which I think is really beautiful.）
- `27:54` 謝謝你們所做的一切，為全世界提供了這些運算資源。（And thank you guys for doing that, to provide all that compute to the world.）
- `27:58` 非常、非常感謝你，Pietro，感謝你所做的一切。（Well, thank you so, so much, Pietro, for all your work.）
- `28:02` 我們很高興有你在這個社群裡。（We love having you in the community.）
- `28:04` 我們很珍惜在模型發布前，能有你的眼光、品味和創造力來協助測試。（We love having your eyes and your taste and your creativity on the models right before we are able to ship them.）
- `28:10` 這是我的榮幸。（My pleasure, yeah.）
- `28:11` 我們迫不及待想看到你接下來用 Magic Path 以及那些你正在嘗試的瘋狂點子和舊裝置，會創造出什麼成果。（And, yeah, we can't wait to see what you're going to build next with Magic Path and all of the crazy ideas and old devices you're hacking with.）
- `28:19` 希望這能啟發大家去創造出偉大的事物。（And I hope this inspires people to, you know, build great things.）
- `28:23` 酷。（Cool.）
- `28:24` 我也是。（Same.）
- `28:24` 非常感謝。（Thank you so much.）
- `28:25` Grazie mille.
- `28:26` Grazie a te.
- `28:26` Grazie a te.

## 標籤

Codex, Agent, 訪談, 硬體, OpenAI