跟读练习: Most devs don't understand what agents are - 通过YouTube学习英语口语
B2
OpenAI just introduced AgentKit, a complete set of tools for developers and enterprises to build, deploy and optimize agents.
106 句
如果句子过短或过长，请点击 Edit 进行调整。
1
OpenAI just introduced AgentKit, a complete set of tools for developers and enterprises to build, deploy and optimize agents.
2
This is cool, fairly exciting announcement, developers can now design workflows visually, embed agentic UIs faster.
3
Now I don't really care about this product, I'm not sure I'll end up using it really.
4
But what I am kind of interested in is how they're talking about agents.
5
Because if we look at this, this is not an agent to me.
6
We have here a set of deterministic steps.
7
We start, then we enter a jailbreak guardrail, which by the way just filters for malicious inputs I assume.
8
Then we have another LLM call here which just routes the input to one of three separate agents.
9
This to me is not an agent builder, this is a workflow.
10
I thought we as a kind of AI engineering typescripty community had landed on some definitions for what agents were
11
and what workflows were.
12
But it turns out no we didn't because OpenAI seems to have a different definition from the one that we've been using.
13
I want to talk through this debate so that you understand what agents are, what workflows are and why the distinction even matters.
14
And the best place to start is with Anthropic's famous article building effective agents.
15
This came out in December last year and it basically codified what an agent was and what a workflow was.
16
This is how Anthropic defines an agent, it's essentially a loop.
17
We'll talk more about what this loop is in a minute and what makes this agentic. But
18
if we zoom up to a workflow example here we can see a very similar example to what OpenAI just put out.
19
Instead of a loop here we have a directional flow, we have predetermined code paths.
20
And this is what Anthropic calls a workflow.
21
And it's kind of funny too that this famous article really talking about, you know, titled building effective agents actually talks through like six different kinds of workflows.
22
Let's go to TLDraw where we've got a bit of dark mode where we can actually dive into some of these concepts.
23
An agent is a loop where the LLM decides when to stop.
24
That loop is essentially multiple LLM calls one after the other.
25
Now if you call an LLM multiple times with the same information it's not going to do anything useful.
26
And so to make this an agent you kind of need to give it new information each time.
27
The way that works is the LLM calls tools.
28
It basically says execute this piece of code for me and then tell me what happened when that piece of code ran.
29
Just to dive into this for a minute it kind of looks like this.
30
Let's imagine our system has access to a tool called write file where it can write files to the file system.
31
The user can say to the agent, write a new file called gitignore.
32
Then the assistant comes back with a message here saying, okay, call this tool with this content and this path.
33
On our local machine then we execute the tool and we send the result back to the LLM.
34
And so this flow becomes a loop where the LLM is gaining more information each time.
35
This beautiful loop is what drives things like clawed code, coding agents, all the stuff that you're kind of used to using.
36
The key thing then is that the agent then decides when it's had enough.
37
So the agent can either continue to call tools or it can say stop.
38
At which point it will emit a special token
39
that just says stop and we can catch that in the frontend and no longer call the LLM again and again.
40
Workflows are of course much easier to define.
41
There's no loop here, it's just predetermined steps one after another.
42
You take one LLM call, you pass its result to another LLM call and you pass that result to another LLM call.
43
You might have some deterministic logic in these steps, like If the LLM call returns one thing, do one thing.
44
If it returns another, do another.
45
But all of those code paths are known ahead of time and written in code.
46
Workflows are neat, by the way, because you get opportunities to optimise the system.
47
For instance, you can have parallel workflows where you have multiple LLM calls at the same time.
48
We might take in or produce a chunk of text, split it into two parts, get the LLM to summarise each part of it, and then pass the results of those to another LLM call where we summarise the summaries.
49
Because the path to the solution is known up front, we can optimise it in all sorts of ways, which make workflows really, really powerful.
50
And by the way, if I had to pick between agents and workflows, like one that I could take to a desert island, I would probably pick workflows.
51
But that's just me because I'm a natural contrarian.
52
So let's sum up then.
53
Agent and workflow.
54
What are the differences?
55
What are they good at?
56
Well, the first thing to say that to qualify in this category, you need multiple LLM calls.
57
Like a single LLM call all by itself doesn't really qualify as either an agent or a workflow.
58
It's just a frickin' API call.
59
We don't need an extra definition for that.
60
To me, the key difference is who decides when to stop the program.
61
With an agent, as we saw, it is the LLM really.
62
The LLM can say, OK, I've done the work, let's now stop.
63
Whereas in a workflow it is predetermined steps that are known up front.
64
Now the reason that this entire distinction matters is that agents and workflows are good for different things.
65
An agent is really good when the path to the solution is unclear
66
or when you need to be able to generalise it to lots and lots of different tasks.
67
Coding Coding assistants are a really, really, really good example of this.
68
Because the coding assistant in Clawcode or Cursor doesn't know what kind of codebase it's going to go into, it doesn't know what kind of bug you're going to throw at it, and so it needs to be able to adapt on the fly.
69
In other words, agents are really, really good at improvising.
70
But workflows are much better when the path to the solution is known up front.
71
When you need to do the same thing a thousand times, you always want a workflow.
72
Because as we saw with the parallelizable steps you can basically optimise it in all sorts of different ways.
73
Whereas an agent you really leave the optimisation up to the agent itself.
74
Agent is like jazz, you know, it's all improvisation, all feel.
75
And workflows are like classical music where you can spend ages optimising the upfront set up
76
so that the final output is as good as it can be.
77
The next thing to say though is that agents and workflows are a spectrum, not a hard definition.
78
Most systems out there you will see will be somewhere on this gradient between agent and workflow.
79
For instance, a pure agent where the LLM is solely in charge of deciding when to stop, well, I don't want to deploy that because that thing is going to eventually run forever.
80
And so most agents have a max steps counter, in other words a deterministic stop in the code to prevent the agent running infinitely.
81
This is so common that tools like the AISDK actually have a max steps parameter to their agents.
82
Going further down we have agents that contain workflows.
83
Many agents are able to call workflows from within tools.
84
Which by the way allows you to build really really smart systems
85
because you get the generalizability of the agent and then you're able to optimise the tools that that agent has.
86
Finally, of course, you can have workflows that contain loops.
87
This might be that you produce some text and you evaluate it multiple times to refine the output continuously.
88
The difference here, of course, is does the LLM itself have the ability to break the loop early?
89
For me that's a sign that it's an agent rather than a workflow.
90
But these terms are on a spectrum
91
and most systems out there will use some combination of each or have agents within workflows or workflows within agents.
92
And so the definitions are useful because they allow you to think about problems in terms of patterns.
93
And so it kind of hurts me a little bit when I see as agent workflows grow more complex.
94
Ah, what did we do to deserve this?
95
This is just so confusing.
96
Now of course I'm annoyed I suppose because I'm interested in agents versus workflows as like a pedagogical tool, as a teaching tool.
97
Because I do find the definitions useful for communicating what you're trying to build and the trade-offs between them.
98
But also there's a sense that everyone's using the word without there necessarily being a good definition behind it.
99
I only hope that this definition will spread
100
that the anthropic definition of just two calls in a loop will be what people land on.
101
Now if you're digging what I'm putting out then you will love AIHero.dev.
102
I'm going to be releasing something soon which is going to mash together AI
103
and TypeScript and give you the ability to ship really powerful AI applications with the language
104
that you know and you know that I love.
105
Thanks so much for joining along folks.
106
I will see you very soon.
下载应用
Everything you need to speak fluently
AI PronunciationScore every sentence
IPA PracticeMaster every sound
VocabularyBuild your word bank
Vocab GameLearn while playing
上下文背景
在這段視頻中，講者討論了OpenAI新推出的AgentKit工具，這是一個幫助開發者和企業構建、部署和優化代理的完整工具集。雖然這個產品引起興趣，但更核心的議題是代理及工作流程之間的區隔。講者認為，許多開發者對代理的理解可能存在誤區，因此他希望澄清這些概念，以增進對人工智能工程的理解。
日常交流的五個關鍵短語
- 什麼是代理？ - 用於詢問關於代理的基本概念。
- 工作流程是什麼？ - 用來了解工作流程與代理的區別。
- 這種方法的好處是什麼？ - 探討代理或工作流程的優勢。
- 可以給我個例子嗎？ - 請求具體示範以便理解。
- 如何最佳化這個過程？ - 尋求提升效率的建議。
逐步跟讀指導
要掌握這段視頻中的內容，以下是一些實用的逐步跟讀技巧：
- 主動聆聽：首遍觀看視頻時，專注於聆聽講者的每一句話，試著理解整體意義，而不必過於關注細節。
- 重複跟讀：將視頻暫停，在每個重要段落後跟讀，特別是講者提到的短語。這樣可以提高您的英语发音和語調。
- 寫下關鍵詞：記錄下視頻中出現的關鍵詞和短語，幫助您在重複聆聽時更好地理解上下文。
- 分析語言結構：注意講者使用的句型和語法結構，並在視頻結束後，嘗試用相同的結構自己造句。
- 反覆練習：在幾天內重複觀看視頻，利用看YouTube学英语提升您的聽力和口語技能。您也可以使用shadowspeaks或其他工具來進行有針對性的練習。
透過這種方式，您將能更好地把握代理和工作流程的概念，也能提高自己的聽說能力，讓學習過程變得更高效。有助於您在實際應用中靈活運用這些知識。
什么是跟读法？
跟读法 (Shadowing) 是一种有科学依据的语言学习技巧，最初开发用于专业口译员的培训，并由多语言者Alexander Arguelles博士普及。这个方法简单而强大：您在听英语母语原声的同时立即大声重复——就像是一个延迟1-2秒紧跟说话者的影子。与被动听力或语法练习不同，跟读法强迫您的大脑和口腔肌肉同时处理并模仿真实的讲话模式。研究表明它能显着提高发音准确性，语调，节奏，连读，听力理解和口语流利度——使其成为雅思口语备考和真实英语交流最有效的方法之一。