跟讀練習: Claude Just Revealed AI's Biggest Problem - 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。
63 句
如果句子過短或過長，請點選 Edit 進行調整。
1
For thousands of years, humans have built tools to make hard things easier.
2
A hammer does not make you worse at building.
3
But is AI different?
4
Because AI helps your brain, not your hands.
5
And that raises an uncomfortable question.
6
What happens to the skills we outsource?
7
Is AI brain rot real?
8
What should we do?
9
Well, I see a big difference in the media headlines and actual real-world studies.
10
So, let's see an actual experiment with a group of 52 junior software engineers split into two groups,
11
one that used AI for coding and one that didn't.
12
I have two questions here and the results on both kind of surprised me.
13
1. Did AI make them faster?
14
Well, the AI group finished their tasks a good two minutes, or about 8% faster.
15
So, they had time to watch a 2 minute papers episode.
16
Which is never 2 minutes.
17
But then, wait a second, the result is not statistically significant.
18
I'll downgrade this a bit, so did AI make people code faster?
19
Maybe.
20
Now, second question, did the AI make them dumber?
21
Now, hold on to your papers, fellow scholars, because the answer is, kinda yes.
22
After having them take a quiz, we get 50% for the AI people, and 67% for the no AI people.
23
Whew, that is almost a too great difference.
24
Huge difference!
25
And this is statistically significant.
26
So this one would be much harder to dismiss as noise.
27
I gotta say, this is a fairly large effect for such a smaller study.
28
So where is the largest gap?
29
The AI group really underperforms, when exactly?
30
Well, in debugging.
31
Yes, you get an AI to write a piece of code, and at some point, there will be a problem.
32
And if you rely on the AI too much, sooner or later, you will lose the capability to fix something that goes wrong.
33
I would be careful with that.
34
And I find this study informative because it tells you how much worse it is and in what areas.
35
So, what should you do going forward?
36
Well, in my opinion, three things.
37
Dear Fellow Scholars, this is Two Minute Papers with Dr. Károly Zsolnai-Fehér.
38
1. If you use AI, use it mainly for automating and speeding up things you already understand.
39
It is excellent for that.
40
2. For things you don't know yet.
41
Ask questions.
42
Keep your mind sharp.
43
That is the way of the scholar.
44
3. When something breaks, don't just ask the AI to fix it.
45
No, first try it yourself.
46
Then, have it explain to you what you missed.
47
You will learn so much more.
48
We are scholars here, so let's be measured.
49
This is not an ultimate study.
50
No, this is not the final word on this question.
51
This is 52 mostly junior developers, one Python library, one short task, and one quiz.
52
Oh, and it has a chat-style assistant, not a full agentic coding system.
53
Which would probably make this kind of difference even larger.
54
So, plenty of limitations apply, but as a pointer, I feel it is quite valuable.
55
TLDR, if you use AI as a tutor, it will sharpen your mind.
56
I use Lambda to reproduce AI research papers, often in minutes.
57
It's also great to train your own models or fine-tune an existing one.
58
Run inference or text to image or video, easy peasy.
59
Running a DeepSeek chatbot or agent, super fast, super reliable.
60
Lambda gives you powerful Nvidia GPUs to run your own experiments.
61
I test ideas from the papers I cover, and moments later, results!
62
Love it!
63
Seriously, try it out now at lambda.ai.papers.
FOR YOU
推薦影片
B23
49 句
AI Won't Replace You — But This Will (Stress Physiologist Explains)
Business EnglishDaily ConversationUniversity Students
B21
154 句
The 7 Trillion AI Gamble Is Failing. Big Tech is TRAPPED Right NOW.
Business EnglishDaily ConversationUniversity Students
B21
169 句
Can This Vibe Coder Beat a Senior Developer?
Daily ConversationUniversity StudentsWorking Adults
B22
27 句
Sydney University staff strike over AI concerns
Business EnglishDaily ConversationUniversity Students
B21
85 句
AI Can Write Code. And That’s the Problem.
Daily ConversationUniversity StudentsWorking Adults
B24
68 句
What Happens If China Invades Taiwan? Elon Musk's AI Warning
Business EnglishDaily ConversationUniversity Students
B23
43 句
Translating Claude’s thoughts into language
Daily ConversationUniversity StudentsWorking Adults
B238
26 句
Advantages and Disadvantages of AI
Daily ConversationUniversity StudentsWorking Adults
B22
268 句
Trump’s Canada Tariffs Are Backfiring in Red States | The Week
Business EnglishDaily ConversationUniversity Students
B21
546 句
Do Elite Degrees Still Matter In The AI Era? | Jamie Beaton from Crimson Education | EP 410
Daily ConversationUniversity StudentsWorking Adults
B21
386 句
Communicating the Future: Defining Where We Want AI To Take Us
Business EnglishDaily ConversationUniversity Students
C15
54 句
Did Google just kickstart the intelligence explosion?
American AccentBusiness EnglishDaily ConversationUniversity Students
本課的詞彙與口語要點
這節 B2 級別的口語課以影片“Claude Just Revealed AI's Biggest Problem”為素材。 影片中反覆出現的詞有：papers, scholar, difference, lambda, faster。 這段影片共有 63 個句子、602 個單詞可供跟讀。 講話部分時長為 3:54。 說話人語速自然，每分鐘約 154 個詞，接近日常對話。 82% 的單詞屬於英語最常用的 3,000 詞，其餘的詞建議在練習前先學一下。 全部 63 個句子中有 10 個是問句，適合練習疑問語調和簡短回答。
影片中的重點詞彙
影片中最難的 15 個單詞，附發音和釋義：
| 單詞 | 發音 | 釋義 |
|---|---|---|
| scholar 名詞 | /ˈskɑlɚ/ | 學生 /學生 |
| lambda 名詞 | /ˈlæm.də/ | 拉姆達 /拉姆達, 蘭布達 /蘭布達 |
| quiz 名詞 | /kwɪz/ | 測驗 /測驗, 競猜 /競猜 |
| automate 動詞 | /ˈɔ.təˌmeɪt/ | 自動化 /自動化 |
| chatbot 名詞 | /ˈtʃætbɑt/ | 聊天機器人 /聊天機器人 |
| outsource 動詞 | /ˈaʊtˌsɔːs/ | 外包, 委外 |
| sharpen 動詞 | /ˈʃɑɹpən/ | 磨, 削尖 |
| underperform 動詞 | 表現不佳 /表現不佳 | |
| pointer 名詞 | /ˈpɔɪn.təɹ/ | 教鞭 |
| reproduce 動詞 | /ˌɹi.pɹəˈdus/ | 複製 /複製 |
| tutor 名詞 | /ˈtjuːtəː/ | 私人教師 /私人教師, 家庭教師 /家庭教師 |
| limitation 名詞 | /lɪmɪˈteɪʃən/ | 限制 |
| dismiss 動詞 | /dɪsˈmɪs/ | 解僱 /解僱 |
| headline 名詞 | /ˈhɛd.laɪn/ | 標題 /標題 |
| hammer 名詞 | /ˈhæmɚ/ | 摃錘仔 /𫼱錘仔, 錘子 /錘子 |
影片中出現的短語動詞
| 單詞 | 發音 | 釋義 |
|---|---|---|
| going forward 副詞 | 放眼未來 /放眼未來 | |
| speed up 動詞 | /spiːdˈʌp/ | 加速, 加快 |
值得反覆跟讀的句子
影片中簡短而完整的句子，可以直接用在日常對話裡：
- What happens to the skills we outsource?
- What should we do?
- For things you don't know yet.
- We are scholars here, so let's be measured.
需要注意的發音
說話人用了 6 次縮略和弱讀形式，例如 don't, didn't, gotta。請按聽到的簡短形式來說。
- “sh” 和 “zh” 音: sharpen /ˈʃɑɹpən/, limitation /lɪmɪˈteɪʃən/
- 長單詞——注意重音位置: statistically /stəˈtɪs.tɪ.kə.li/, informative /ɪnˈfɔːmətɪv/, limitation /lɪmɪˈteɪʃən/, capability /ˌkeɪ.pəˈbɪl.ɪ.ti/, uncomfortable /ʌnˈkʌm.fɚ.tə.bl̩/
中文母語者容易讀錯的音：
- /v/ — 不要讀成 /w/，上齒輕觸下唇: informative /ɪnˈfɔːmətɪv/, developer /dɪˈvɛləpɚ/
- 詞尾子音 — 要讀清楚，後面不要加母音: quiz /kwɪz/, agentic /eɪˈd͡ʒɛn.tɪk/, automate /ˈɔ.təˌmeɪt/, chatbot /ˈtʃætbɑt/, downgrade /ˌdaʊnˈɡɹeɪd/
- /æ/ — 嘴張大，不要讀成 /e/: lambda /ˈlæm.də/, chatbot /ˈtʃætbɑt/, hammer /ˈhæmɚ/
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 scholar, lambda, quiz 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。