跟读练习: You’re Not Behind (Yet): Learn AI Agents in 13 Minutes - 通过YouTube学习英语口语
C1
Most people think they're using AI well when they get a decent answer from ChatGPT.
241 句
如果句子过短或过长，请点击 Edit 进行调整。
1
Most people think they're using AI well when they get a decent answer from ChatGPT.
2
That was enough six months ago.
3
It's not enough anymore.
4
The next shift is AI agents and the gap between people who understand them and people who don't,
5
it's about to get very expensive.
6
I've spent years in the boardrooms of billion-dollar companies
7
and the good news here is that agents are much simpler than most people think.
8
So in this video, I'll show you exactly how they work,
9
when to use them, and how to start before everyone else catches up.
10
An AI prompt and an AI agent are completely different.
11
But most of us are still stuck with old habits.
12
Let me give you the simplest way to think about this before we go any further.
13
This is our first framework right off the bat.
14
I call it ARR.
15
If a task is autonomous,
16
recurring, and reviewable, it's a strong candidate for an agent.
17
If it needs live judgment,
18
or it only happens once,
19
or can't be reviewed clearly, then use a prompt.
20
That one distinction alone will put you miles ahead of most people using AI today.
21
The internet gave us search, so we started googling.
22
AI gave us LLMs, or large language models.
23
But most people today still think of AI as a glorified search.
24
Now agents are here and we're still making the same mistake again.
25
We think of agents as just more capable chatbots.
26
They're not.
27
A chatbot waits for your next prompt.
28
An agent figures out its next move.
29
Prompting is like sitting next to a student driver.
30
You still have to guide them,
31
correct them, and stay very alert.
32
An agent, on the other hand is a hired driver you set the destination hand over the keys
33
and just sit in the back seat it handles the route the traffic
34
and all the step-by-step decisions that's the mental shift we have to make
35
so here's a prompt write me a linkedin post
36
and here's an agent watch my industry every monday find the
37
three most relevant stories study my previous posts draft the new post based on those stories in my voice,
38
revise against my style, and schedule it for Tuesday morning.
39
That's the power of AI agents.
40
And to wield that power,
41
you need to know what's actually running under the hood.
42
Everyone's talking about AI agents.
43
Almost nobody can tell you what's actually happening inside.
44
A chatbot predicts the next word.
45
An agent decides the next action.
46
Here's how a chatbot actually works.
47
It's a large language model.
48
When you type a question,
49
it's going to break that question into small units of words called tokens,
50
and it converts them into numbers.
51
And then it just finishes the sentence.
52
So if you said, Jack fell down and broke his crown,
53
now you know that, but LLM does not know that rhyme.
54
It will know words like bones and heart and crown,
55
and all of those words could make sense.
56
But based on its training,
57
it predicts that the most likely next word given that line is crown.
58
It's based on probabilities.
59
The agent has the same language model in the center,
60
but now there are four workers around it.
61
Analyst, planner, operator, auditor.
62
One finds the pattern, one decides the plan,
63
one does the work, one checks the results.
64
Let's make this real.
65
You can give an agent some instruction like every Monday at 7 a.m.,
66
review the past week's customer support tickets,
67
sales notes, and product feedback,
68
identify three biggest recurring issues,
69
summarize what changed, and email my leadership team a one page weekly brief.
70
That's it.
71
I mean, those are lots of steps,
72
but the agent will read the tickets,
73
notes, and feedback, and find the pattern.
74
Analyst.
75
It will then decide what matters most and what belongs in the brief.
76
Planner.
77
It will write and send the update.
78
Operator.
79
And then it'll check for weak logic,
80
missing context, or sloppy conclusions, and it'll refine it.
81
Auditor.
82
Now, by Monday morning, the brief is in your team's inbox.
83
You did not write the report.
84
You did not analyze it.
85
You just assigned the job of four people to one agent.
86
So here's your move.
87
Tonight, open ChatGPT agent mode and give it one recurring task,
88
but then watch what it does.
89
You'll see all four workers show up
90
in real time that's the anatomy of an ai agent now
91
that you understand the parts let's look at the entire loop the best thing about agents is
92
that they can adapt when things go wrong this is what makes agents genuinely different from everything
93
that came before it in the 1970s there was an air force colonel john boyd
94
and he studied a very intriguing puzzle from the korean war American pilots in their F-86 kept beating technically superior Soviet MiG.
95
Now the MiG was faster and it could climb higher.
96
It should have won, but it didn't.
97
And Boyd eventually found the difference.
98
The American pilots could see more from their cockpits and they could adapt faster.
99
So they got inside the enemy's decision cycle before the enemy could respond.
100
He called that loop, the OODA loop.
101
Observe, orient, decide, act.
102
And in the world of agents, it's the same thing.
103
That is the real test of an agent.
104
When the obvious path fails,
105
can it choose a better one?
106
Can it go through its own OODA loop?
107
So let me give you a concrete example.
108
You can build an automated workflow every Friday,
109
check this week's grocery prices,
110
build my shopping list, and place the order.
111
It works every Friday until one week your usual item is out of stock
112
and you have six friends coming for dinner on Saturday.
113
So that automated workflow is going to break.
114
Not because it's dumb, but because it's designed to be obedient.
115
It's designed to not think on its own.
116
An agent, on the other hand, does something very unique.
117
It sees the usual list.
118
It sees that that is not working, it finds substitutes.
119
It adjusts quantities for six people.
120
It checks your calendar, sees the dinner,
121
and rebuilds the entire order around it.
122
A workflow can follow the process.
123
An agent can reroute it completely.
124
That's the difference.
125
So when someone says they built an agent, ask one question.
126
When the first path breaks,
127
does the agent keep following the script or can it find a better path?
128
Can it find another way?
129
That's the agent adapting in real time.
130
So this begs the question, right?
131
If agents have autonomy, why do they still fail so often in real life?
132
That's next.
133
The most dangerous thing about AI agents is
134
that they will do wrong things faster and with more confidence than you ever could.
135
An agent is not magic, it's a multiplier.
136
I was working with a board and leadership team of a large consumer company.
137
I'm still working with them.
138
And they're profitable, well-run, great CEO.
139
And when I asked what was stopping them from using AI to drive customer acquisition,
140
for example, the CMO responded,
141
we have all the data,
142
but we'll still need to build a clean process so we can turn that into something useful, something insightful.
143
And I asked where the real challenge was.
144
And she said, you know,
145
we need the right people in the seats first.
146
That's the story everywhere.
147
Most AI problems are human problems in disguise.
148
An agent is just a mirror.
149
It reflects the quality of your thinking back at you.
150
It just amplifies it.
151
Give an agent vague goals,
152
sloppy directions, and no way to get feedback,
153
and it will drive the car straight into the tree,
154
faster and with more confidence than you ever could.
155
Here's the dangerous part.
156
An agent doesn't fix bad thinking, it formalizes it.
157
Usually, the agent fails because the human was vague,
158
not because the underlying model was bad or anything.
159
So, before you automate anything, run a GPS check.
160
Goal, proof, steps.
161
Goal.
162
Can I define the goal in one sentence very clearly?
163
Proof.
164
Can I tell what good looks like?
165
And how will I know if the agent got it right?
166
And steps.
167
Can I describe each and every step very clearly without a lot of hand-waving?
168
Unless you can do those three things very well,
169
your agent is not gonna make any difference.
170
I'll give you an example.
171
Here are two instructions for your agent.
172
First one, summarize my emails every morning.
173
It's good.
174
And here's the second one.
175
Every morning at 7 a.m.,
176
read my unread emails, categorize them by urgency,
177
draft replies to routine messages,
178
and flag anything from my top five customers.
179
So you see, there is a difference between those two instructions,
180
and that gap is exactly where the mess lives.
181
The winners who can wield the power of AI agents aren't just gonna be engineers.
182
They'll be the people who understand their work deeply enough to define it precisely.
183
Most companies want AI everywhere.
184
The ones actually winning are obsessively narrow.
185
If clarity is the bottleneck,
186
then the opportunity is not broad intelligence, it's narrow ownership.
187
I was visiting the customer conference of a construction software company
188
that I work with
189
and the product lead was on stage showing a demo of a single agent
190
that was focused on a very specific problem collecting field data
191
for a specific type of customers in a specific type of situation
192
and it was a beta launch the demo worked mostly with a few minor glitches here
193
and there but
194
when he showed the qr code at the end every hand
195
in the conference room went up with their phones everyone took a picture
196
because it solved a very specific
197
but very real pain they all had been living with for decades
198
that is where the real opportunity is in your career
199
or in your company narrow focus here's the test find a highly specific task people hate doing
200
but they have to do it repeatedly that's where the money is you know
201
we're entering an age where we'll have more agents than human beings on this planet.
202
On the business side, for every software company that exists today,
203
there will be an agent company trying to dethrone it.
204
The winners won't build the broadest agents first.
205
They'll build the one that understands one workflow,
206
one market, and one kind of user pain better than everyone else.
207
By the way, if you If you want to keep this conversation going,
208
I write a short newsletter once a week,
209
just useful ideas, tools, honest reflections.
210
You can subscribe below.
211
It's free.
212
AI will reshape almost every role,
213
but it won't replace what makes you irreplaceable.
214
You know, AI is a giant decoupling machine.
215
For most of our modern history,
216
your income was tied to your hours.
217
at the top, you're always trading time for decisions.
218
Agents are breaking that link for the first time.
219
Now they do the work and you scale your judgment in areas where it matters most.
220
And that changes what's valuable.
221
We're entering an era of infinite output,
222
content, code, and analysis all becoming super cheap.
223
When intelligence becomes that cheap,
224
judgment becomes even more expensive.
225
When output becomes infinite, taste becomes scarce.
226
Every time you define a task clearly enough for an agent to run it,
227
you're not just training the system,
228
you're clarifying your own standards.
229
You're learning what good actually looks like.
230
Sure, some roles will be reshaped.
231
The paralegal, the junior analyst, the coordinator.
232
But every disruption has always created new roles that never existed before.
233
Before the internet, nobody imagined the role of an online community manager.
234
The question is not whether the shift happens.
235
It is whether you shape it or get shaped by it.
236
The most valuable person is no longer the one who can think the fastest.
237
It's the one who can define good work,
238
spot bad work, and know when to trust an agent and when to trust a human.
239
That is where your value is going to move.
240
Ironically, AI will make human life less robotic.
241
Thank you and I love you.
下载应用
AI 为你说出的每个句子打分
TRENDING
热门
背景与语境
在当今快速发展的人工智能领域，许多人对AI的理解仍然停留在表面，尤其是在如何有效使用AI代理方面。有鉴于此，我们的演讲者具有丰富的经验，他在多家亿万美元公司董事会工作多年。他分享了AI代理的工作原理以及如何快速上手的见解，强调了与仅仅依靠聊天机器人的旧习惯形成鲜明对比的必要性。
日常交流的五个关键短语
- AI代理的工作方式 - 了解AI代理如何运作是关键。
- 任务的自主性 - 具体任务是否可以独立执行？
- 递归与评估 - 任务是否会重复且可以进行评估？
- 区分聊天机器人与AI代理 - 聊天机器人只预测下一个词，而AI代理制定下一步行动。
- 有效的任务管理 - AI代理如何处理任务，如内容创作与社交媒体发布。
逐步跟读指南
为了有效提高你的英语口语能力，尤其在学习有关AI代理的内容时，可以采用以下步骤进行影子跟读（shadowspeaks）练习：
- 选择合适的材料 - 确保选择与AI相关的视频或音频，以便于提升专业词汇。
- 认真聆听并理解内容 - 在开始影子跟读之前，先了解视频的主题和主要观点。
- 分段进行跟读 - 将视频分为小段，逐段进行影子跟读（英语影子跟读），确保每个部分的流利度。
- 模仿发音和语音语调 - 注意演讲者的发音和语调，努力模仿，提升自己的语音表达能力。
- 复习与反思 - 完成跟读后，复习自己的录音，与原视频对比，不断修正自己的发音。
通过这种方法，学习者能够在专业领域内自信地表达自己的观点，并掌握更复杂的语言结构，提升自己的shadow speech（影子演讲）技能，进一步推动个人在英语口语沟通中的提升。
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。