シャドーイング練習: They want mediocre developers... - YouTubeで英語スピーキングを学ぶ
C1
I miss the good old days when technical debt
133 文
文が短すぎたり長すぎる場合は、Editをタップして調整してください。
1
I miss the good old days when technical debt
2
and student debt were the only debts we had to worry about in the software world.
3
Looking back at it, technical debt was actually quite fun.
4
We would take shortcuts to move faster,
5
ignore best practices to meet a deadline,
6
or delay updating a library because we were in no mood to work through the side effects.
7
The worst that could've happened was
8
that you would end up paying for it later through increased complexity in systems that nobody wanted to touch.
9
But ultimately, this increased complexity would increase your job security because
10
if If you were the only person who understood a messy part of the system,
11
you would suddenly become very hard to replace.
12
This is why for the past 20 years,
13
software engineering has been a pretty nice gig,
14
with a lot of perks and more than decent paychecks.
15
And I'm not making this video to tell you that all these are going to end,
16
but I want to discuss a new emerging kind of debt which will actually make our lives much more difficult.
17
Adios Money recently posted about a new concept called Comprehension Debt
18
and in this Monday morning review will look at the importance of being able to think
19
and solve problems in the generated code era.
20
Comprehension debt is the growing gap between how much code exists in your system
21
and how much of it anyone on the team can genuinely explain.
22
At first, this doesn't look like a problem,
23
since most AI-generated code looks really clean at first and it might look like all the best practices are being followed.
24
But underneath that, something subtle starts to shift because the understanding of the system begins to slowly degrade.
25
In a recent study, engineers were asked to learn a new library.
26
Some of them used AI to generate solutions,
27
others used it more as a guide to explore concepts and ask questions.
28
Both groups completed the task in roughly the same time,
29
so from a productivity perspective, everything looked identical.
30
But when they were tested afterwards,
31
the group that relied on generation performed significantly worse,
32
especially when it came to debugging.
33
And debugging is where things become very real very quickly.
34
If you've been involved in software for a while,
35
you probably know that reading and understanding code at a deep level is one of the most difficult parts of the job.
36
I'm not talking here about going through some random syntax filled with for loops and if statements.
37
I'm talking about complex implementations,
38
which have interactions, implicit assumptions,
39
and behavior that only makes sense once you follow the entire execution path end-to-end.
40
This part is so difficult that you'll go through code you wrote yourself six months ago
41
and you'd still be confused trying to reconstruct what you were thinking at the time.
42
When everything works, speed feels like the most important metric.
43
The moment something breaks, the only thing that matters is whether someone can explain what the system is doing and why.
44
And, if I have issues understanding my own code from six months ago
45
and I can proudly call myself an average developer with a weird accent and a copied YouTube channel,
46
I expect
47
that most of us will have a lot of issues understanding thousands of lines of code generated in a couple of minutes.
48
For a long time, writing code and reviewing it acted as a natural bottleneck in development.
49
It slowed teams down, but it also forced engineers to read code written by others,
50
to question decisions, and to build a shared understanding of how the system behaved.
51
The process was one of the main ways knowledge spread across a team.
52
With AI-generated code, the volume changes the nature of that interaction.
53
Code is produced faster than it can be meaningfully reviewed,
54
and because it is usually clean and well-structured,
55
it gives off all the signals that historically indicated quality.
56
And the AI labs are telling you to ignore the problem.
57
You spend your tokens generating code,
58
and then they provide you the tools to review that generated code for the convenient price of $15-25 for each review.
59
In other words, the token sellers just want you to buy more tokens,
60
ideally hundreds of thousands of dollars worth of tokens.
61
I keep trying to avoid highlighting the absurdity of some of these CEOs' comments,
62
but they are making it very tough for me to ignore them.
63
screams trying to inflate the bubble more than focusing on quantity over quality.
64
Sure, you could easily spin up Todo apps in an endless loop to burn through $250,000 in tokens.
65
That would be completely pointless and will consume a few more million gallons of water,
66
but at least the billionaire Jensen Huang with his geeky leather
67
jackets will be happy to see his Nvidia stock keep going through the roof.
68
But I digress.
69
Back to the topic of comprehension debt,
70
there's also a more subtle failure mode that starts to show up in AI-heavy workflows.
71
When an implementation changes and the associated tests are updated to match,
72
the conversation quietly shifts.
73
Instead of asking whether the code behaves correctly,
74
you now have to ask whether the behavior itself is correct.
75
The test passing no longer answered the most important question.
76
As we discussed in last Monday's episode,
77
another common reaction is to rely more on specifications.
78
You define the behavior in detail,
79
review the specification carefully, and let AI handle the implementation.
80
On paper, this creates a clean separation between intent and execution.
81
In practice, translating a specification into working software involves a large number of implicit decisions that rarely get captured in full.
82
Data structures, edge cases, error handling,
83
and performance trade-offs all shape the final behavior in ways that are difficult to fully describe up front.
84
This year's DORA report also highlights this new emerging comprehension crisis.
85
In teams where engineering practices are already solid,
86
where workflows are clear, and where people have a strong mental model of the system they are working on,
87
AI tends to make everything move faster in a way that actually compounds quality.
88
Developers spend less time on repetitive tasks,
89
decisions are made with more context,
90
and the system evolves in a relatively controlled way.
91
But in teams where things are already messy,
92
where architecture has grown organically without much discipline,
93
and where understanding is fragmented across individuals,
94
AI scales all of those issues at the same time.
95
The same acceleration that helps good systems becomes a multiplier for bad ones,
96
and the difference between the two becomes much more visible.
97
What's interesting is that the report states
98
that around 90% of developers are already using AI in their work and more than 80% believe it improves their productivity.
99
But at the same time,
100
a significant portion of them openly admit that they do not fully trust the code that AI produces.
101
So you end up in a situation where developers rely on a tool daily,
102
move faster because of it,
103
and still carry a constant layer of uncertainty about what is actually being generated.
104
The DORA data also highlights something that reinforces this dynamic even further.
105
AI adoption correlates with higher throughput,
106
meaning teams are shipping more changes and moving faster through their backlog.
107
At the same time, delivery instability increases,
108
which means those changes introduce more issues into the system.
109
So basically, you move faster,
110
create more issues, you fix those issues faster,
111
and create even more issues.
112
Your KPIs go through the roof,
113
management has nice graphics to look at,
114
clients are happy to see results,
115
and token sellers are making a lot of money.
116
Everybody is happy and Jensen Huang gets richer.
117
If I'm being honest, I spent the past couple of months worrying about the future of software engineering.
118
AI code generation is far from perfect,
119
it has improved a lot in the past couple of months.
120
However, something feels off.
121
The AI labs are pushing these coding agents hard and we have to pay very little to use their models.
122
However, their computing costs are huge and they are very far from running a profitable business.
123
This is clearly the customer acquisition phase,
124
where their pricing is so low you can't ignore them
125
and their agents are slowly becoming an integral part of the software workflow.
126
And I believe that this comprehension debt is exactly what they need to make sure you can't really walk away later.
127
This is not going to happen overnight,
128
but imagine a world 10 years from now where most of the software engineers have lost most of their critical thinking,
129
problem solving and debugging skills because they are used to delegating all of it.
130
At that point, Anthropic could increase the cost of their agents to thousands of dollars per month
131
and there would be little we could do because most of the skills are lost,
132
we have no idea what our code actually does
133
and the jobs pipeline was destroyed when we decided AI can do a better job than juniors.
アプリをダウンロード
話したすべての文をAIが採点
TRENDING
人気動画
このビデオで話す練習をする理由は？
ソフトウェア開発の世界では、技術的な問題や学生の借金を除けば、昔はそれほど心配することがありませんでした。このビデオでは、コードを生成するAIの登場がもたらす新たな「理解の負債」について触れています。これを通じて、実際の問題解決能力や深く考える力が試される場面が多いです。この話題をシャドーイングすることで、専門的なトピックに対する語彙力が向上し、同時に発音やリズム感も磨かれます。こうしたスキルは、ソフトウェア開発者にとって必要不可欠なのです。
文法と表現のコンテキスト
- 「At first, this doesn't look like a problem」 - 「最初は、これは問題のようには見えない」という表現から始まり、状況の説明や問題提起の方法を学ぶことができます。
- 「if I have issues understanding my own code」 - 「自分のコードを理解するのに問題がある場合」という条件文は、英語の仮定形を実践する良い例です。
- 「reading and understanding code at a deep level」 - 「深いレベルでコードを読み理解する」というフレーズは、技術的な語彙を強化し、実際の文脈での使用法を学ぶ手助けになります。
- 「something starts to shift」 - 「何かが変わり始める」という表現は、変化を表す言い回しとして役立ちます。
一般的な発音の罠
このビデオの中で、特に注意が必要な単語やフレーズには、「comprehension debt」や「complexity」といった発音があります。これらは話し手のアクセントやイントネーションによっても変わりやすいため、反復練習が重要です。シャドーイングでは、これらの発音を実際のコンテキストで練習することで、自然に身に付けることができます。また、shadow speechや
shadowspeakを意識することで、より滑らかな英語のスピーキング能力を向上させることができるでしょう。
シャドーイングとは？英語上達に効果的な理由
シャドーイング（Shadowing）は、もともとプロの通訳者養成プログラムで開発された言語学習法で、多言語習得者として知られるDr. Alexander Arguelles によって広く普及されました。方法はシンプルですが非常に効果的：ネイティブスピーカーの英語を聞きながら、1〜2秒の遅延で声に出してすぐに繰り返す——まるで「影（shadow）」のように話者を追いかけます。文法ドリルや受動的なリスニングと異なり、シャドーイングは脳と口の筋肉が同時にリアルタイムで英語を処理・再現することを強制します。研究により、発音精度、抑揚、リズム、連音、リスニング力、そして会話の流暢さが大幅に向上することが確認されています。IELTSスピーキング対策や自然な英語コミュニケーションを目指す方に特におすすめです。