跟讀練習: Introducing the Agents API - 通過影片學習英語口語
選擇練習模式
每種模式都會開啟全屏練習室。
25 句
如果句子過短或過長，請點選 Edit 進行調整。
1
Getting a long-running agent into production takes a lot of work, even with a capable model.
即便有能力出眾的模型支撐，要讓一個長期執行的智慧體投入生產環境也需要大量工作。
2
You need to connect tools, track progress, manage context, and secure and maintain the infrastructure around it.
3
Today, we're launching the Agents API to handle that infrastructure for you.
4
The Agents API brings a hosted version of the Codex Harness to your applications.
5
OpenAI handles orchestration, sessions, and context management so that you can stay focused on building.
6
Let's look at an example.
7
Suppose we want to build an agent that helps investigate incidents in our production stack.
8
It would need access to observability data and recent code changes, along with our team's instructions for how to handle an outage.
9
We can connect all of our necessary tools through MCPs and give the agent our investigation runbook via a skill.
10
You control the agent's execution environment and the tools it has access to.
11
That includes connecting a sandbox, whether it's through through OpenAI, a third party provider, or using your own infrastructure.
12
In a situation like this, we'd also be working through a ton of logs, likely more than we could fit in the model's context window.
13
With programmatic tool calling, the agent can process those logs and filter the results in code,
14
meaning fewer tokens spent passing around raw data and more spent on the information the agent needs.
15
For the largest tasks, independent work can be delegated via multi-agent orchestration.
16
In this case, we might want one sub-agent to examine recent changes, another to check telemetry, and then have the lead agent unify their findings.
17
But even in a single agent session, long-running context windows still work really well with compaction.
18
It gives the model a summary of all prior work completed so the agent can continue its investigation.
19
Once the work is complete, the findings should be served as a report that the on-call team can review, a likely root cause, the supporting evidence,
20
and suggested next steps bundled into a single shareable file.
21
And there we have it.
22
With the agents API, we were able to drive an entire workflow without building or maintaining any of our own agent infrastructure.
23
Infrastructure that will continue to get better alongside new models and new capabilities.
24
We're excited to bring you an ever-improving harness behind one API.
25
Happy building.
同頻道
B24
Introducing GPT-Image-2.5 in the API
Daily ConversationWorking Adults
B21
Introducing GPT-6 Astra for developers
Daily ConversationUniversity StudentsWorking Adults
B21
Introducing GPT-6 Astra: the most intelligent and aligned model in the world.
Business EnglishUniversity StudentsWorking Adults
B14
Introducing dots, always-on agents built to handle everything.
Business EnglishWorking Adults
C13
Reimagining IT with ChatGPT
Business EnglishUniversity StudentsWorking Adults
B2Using Voice in ChatGPT Work
Business EnglishWorking Adults
FOR YOU
推薦影片
C18
24 句
Introducing Automations: always-on coding agents
Business EnglishUniversity StudentsWorking Adults
C11
143 句
MCP vs API: The protocol every developer needs to know
Business EnglishUniversity StudentsWorking Adults
C15
222 句
The Multi-Agent Architecture That Actually Ships — Luke Alvoeiro, Factory
Business EnglishUniversity StudentsWorking Adults
C11
309 句
Claude Code: The Advanced Guide (99% of Devs Skip These Features)
Business EnglishUniversity StudentsWorking Adults
C16
10 句
What's new in Google Cloud in less than 1 minute
Business EnglishUniversity StudentsWorking Adults
C11
141 句
How AI is Changing Data Scientist's Workflow
Business EnglishUniversity StudentsWorking Adults
C111
98 句
What are AI Agents?
Business EnglishUniversity StudentsWorking Adults
C12
338 句
I Wrapped an API Into a Real SaaS (full build / step-by-step)
Business EnglishUniversity StudentsWorking Adults
C12
144 句
WWDC26: Run local agentic AI on the Mac using MLX | Apple
Business EnglishUniversity StudentsWorking Adults
C12
68 句
Prompt engineering - Introduction
Business EnglishUniversity StudentsWorking Adults
C12
540 句
Incremental Cooking in UE 5.7: A Dive Into the UE Cook Pipeline | Unreal Fest Chicago 2026
Business EnglishUniversity StudentsWorking Adults
本課的詞彙與口語要點
這節 C1 級別的口語課以影片“Introducing the Agents API”為素材。 影片中反覆出現的詞有：agent, infrastructure, Api, tool, context。 這段影片共有 25 個句子、392 個單詞可供跟讀。 講話部分時長為 2:17。 說話人語速自然，每分鐘約 172 個詞，接近日常對話。 只有 81% 的單詞屬於英語最常用的 3,000 詞，詞彙難度較高。
影片中的重點詞彙
影片中最難的 8 個單詞，附發音和釋義：
| 單詞 | 發音 | 釋義 |
|---|---|---|
| orchestration 名詞 | 配器 | |
| compaction 名詞 | /kəmˈpækʃən/ | 壓實 /壓實 |
| sandbox 名詞 | /ˈsæn(d).bɑks/ | 砂箱 |
| workflow 名詞 | /ˈwɝkfloʊ/ | 工作流 |
| delegate 名詞 | /ˈdɛlɪɡət/ | 代表, 使節 /使節 |
| token 名詞 | /ˈtoʊkən/ | 代幣 /代幣, 籌碼 /籌碼 |
| bundle 名詞 | /ˈbʌn.dl̩/ | 捆, 包 |
| stack 名詞 | /stæk/ | 堆 |
需要注意的發音
說話人用了 3 次縮略和弱讀形式，例如 we're, we'd。請按聽到的簡短形式來說。
- “sh” 和 “zh” 音: compaction /kəmˈpækʃən/, shareable /ˈʃɛəɹ.ə.bəl/
- 長單詞——注意重音位置: programmatic /ˌpɹəʊ.ɡɹəˈmæt.ɪk/, telemetry /tɪˈlɛm.ɪ.tɹi/, capability /ˌkeɪ.pəˈbɪl.ɪ.ti/
中文母語者容易讀錯的音：
- 詞尾子音 — 要讀清楚，後面不要加母音: harness /ˈhɑɹ.nəs/, outage /ˈaʊ.tɪd͡ʒ/, programmatic /ˌpɹəʊ.ɡɹəˈmæt.ɪk/, sandbox /ˈsæn(d).bɑks/, delegate /ˈdɛlɪɡət/
- /r/ 和 /l/ 的區別: shareable /ˈʃɛəɹ.ə.bəl/, telemetry /tɪˈlɛm.ɪ.tɹi/
- /æ/ — 嘴張大，不要讀成 /e/: compaction /kəmˈpækʃən/, programmatic /ˌpɹəʊ.ɡɹəˈmæt.ɪk/, sandbox /ˈsæn(d).bɑks/, stack /stæk/
如何用這段影片練習
- 先完整聽一遍影片，不要開口，記下不認識的單詞。
- 先用 0.75 倍速逐句跟讀，熟練之後再回到正常速度。
- 錄下自己的聲音並與原聲對比，特別注意 orchestration, compaction, sandbox 這類單詞。
什麼是跟讀法？
跟讀法 (Shadowing) 是一種有科學依據的語言學習技巧，最初開發用於專業口譯員的培訓，並由多語言者Alexander Arguelles博士普及。這個方法簡單而強大：您在聽英語母語原聲的同時立即大聲重複——就像是一個延遲1-2秒緊跟說話者的影子。與被動聽力或語法練習不同，跟讀法強迫您的大腦和口腔肌肉同時處理並模模擬實的講話模式。研究表明它能顯著提高發音準確性，語調，節奏，連讀，聽力理解和口語流利度——使其成為雅思口語備考和真實英語交流最有效的方法之一。