쉐도잉 연습: 12 Important Concepts In the Age of AI Software Development - 영상으로 영어 말하기 배우기

연습 모드 선택

각 모드는 전체 화면 연습실로 열립니다.

164 문장

문장이 너무 짧거나 길면 Edit를 눌러 조정하세요.

Hey guys, so obviously things are changing.

They have changed a lot over the past couple years, especially the past year, and it's only speeding up.

And we've gone from kind of a standard workflow of writing

code to using different AI tools with more attention on things like planning, reading, reviewing, and testing code.

Now with that said, just because you won't be writing as much code anymore, or maybe not even writing any code,

it doesn't mean that you shouldn't understand what's going on and what the AI is generating.

I think being able to read code is now one of the most important skills.

So I think the focus is shifting a bit.

Fundamentals like functions, loops, conditionals, data types, that's all still a must.

You have to learn those.

Without those, you're pretty much going to be lost.

But it's less important to memorize every function or method in a language or a library or framework.

Most developers were looking that stuff up before anyway.

I think what really matters now is why it belongs there, the data that it's working with,

and just how it fits into the rest of the project.

So in this video, I want to go over the concepts

that I think are most important to understand if you're going to be using AI as a developer.

And this is not a full lesson on each one.

It's more of a checklist of things to pay attention to when AI gives you code.

So let's get into it.

So before we dive in, this video is sponsored by CodeRabbit.

You know that feeling when a PR comes in and there's 47 files changed, GitHub just throws them at you alphabetically.

There's no sense of what depends on what, there's no logical order.

And now that AI is writing more of our code, these PRs are only getting bigger.

So CodeRabbit review was built specifically for this.

It reorganizes the diff into groups of related changes with ordered layers

so that Foundational stuff like data models come before the code

that depends on them and every layer has an AI summary

and inline diagrams You can also drop inline comments You can navigate with keyboard shortcuts

and you can submit your github review without ever leaving the tab It's also opt -in per reviewer

so your teammates who prefer the default github experience can just ignore it entirely

And it's free to try right now during early access comment.

You can read more about it in the link in the description below.

All right, guys, so we're going to cover a bunch of very important concepts, and I'm not going to just define what that concept is, but I'm going to relate it to agentic coding.

And of course, these are important regardless, even, you know, long before AI, but I'm going to, like I said, relate them to coding with AI.

Now, the first three are very closely related, and they're about the flow of your program.

So number one is control flow.

And this is the order that your code runs in.

So what happens first, what happens next, what happens if something is true, and so on.

And this includes things like if statements, loops, try catch blocks.

And this matters because AI can generate code that looks clean, but the path through that code could do things like return too early,

or maybe it skips validation, or maybe the error handling is in the wrong place.

So if you can't follow the flow, then you can't really debug the program.

Now, number two, as I said, is related to number one, it's data flow.

So data flow is how much, sorry, not how much, but how data moves through your application, and where does it come from, where does it end up,

and what changes along the way.

So for instance, a very common data flow pattern that we see in is where it starts from a user, right?

Usually within a form, within a web page, it gets sent to an API, it gets stored in the database, comes back as JSON, and then it shows up in the UI.

Now, this matters because AI might generate multiple files for one feature.

And if you can trace the data from that initial user action to the database and back to the screen,

that whole feature starts And this is what happens when something goes wrong.

So a request fails, a value is missing, a user enters bad data, a database query throws an error.

There's a lot of different things obviously that can cause errors and a lot of beginners

they only think about the happy path, right?

But real software has, it has to handle failure.

And AI can generate code that works when everything's perfect, but it all falls apart when something unexpected happens.

And I see this over and over again.

So you really need to ask, where can this fail?

And what happens if it does?

And does the user or the developer get a useful message to work with?

All right.

So you got control flow, data flow, and error flow, So the next one is scope.

And scope is where things like variables and functions and other values, it's where they're available in your code.

So if a variable is created inside of a specific function, can you use it outside of that function?

And the answer is usually no. That's scope.

And this matters because AI might create a variable in one place and try to use it somewhere else.

Or you might move code around and suddenly something breaks because that variable doesn't exist in that part of the program.

this value created where can it be accessed and who's allowed to change it Number five is input and output.

I would say most useful pieces of code, they usually take something in and give something back.

So if you think about a function, for example, it might take in a user ID and it might return the entire user.

An API endpoint might take form data and return a success message.

A component might take props and render something on the screen.

you can check if the function expects the right arguments, if it returns the right data, and if it fits in with the rest of the app.

Number six is state.

So state is data that changes over time.

And this could be something like a logged in user, a shopping cart, a form input, a loading status, or just data stored in a database.

And a lot of bugs happen when the program thinks that the state is one thing, but it's actually something else.

So AI might put state in the wrong place, it might it.

It might update one thing but forget to update something else.

So you want to pay attention to where state lives, who can change it, and what should happen when it changes.

Now the next three concepts are closely related to each other

and probably the most important in my opinion

because they all relate to just how files are created and put together in an efficient way.

So number seven, the first one of these three is abstraction.

So abstraction is hiding complexity behind something simpler to use.

So for example, a function called create user, it might validate input, it might hash a password, save a record to the database, and return the new user.

But from the outside, you just see a function called create user.

So AI will create abstractions constantly, helpers, services, hooks, middleware, utilities, all that.

And good abstraction makes code easier to use and files just to figure out what's happening.

So there is a fine line between, you know, too little abstraction and too much.

Now, modularity number eight is closely related to abstraction.

So once you start hiding logic behind functions and components and services,

the next question is where those pieces should live and what each one should be responsible for.

So instead of having Routes, controllers, models, components, utilities.

I could go on and on.

Config files.

The idea is simple, though.

It's each piece should have a job, right?

So you have routes.

100

They handle requests.

101

Controllers or handlers, they decide what happens.

102

Models deal with data.

103

Components deal with UI.

104

Utilities, they handle reusable helper logic.

105

So good modularity makes the code base much easier to navigate,

106

much easier to read, much easier to have multiple team members work on it and so on.

107

Now, number nine, architecture is the overall structure of the application.

108

So where does the front end live?

109

Where does the back end live?

110

The database?

111

How do they communicate?

112

What files are responsible for what?

113

And you don't need to be a software architect, but you should understand the basic shape of what you're building.

114

And this matters because with AI, you need to know where to point it.

115

You know, if you want to add an endpoint, where should that go?

116

If you want to change how data is displayed, is that the front end, is that the back end, or is it both?

117

You don't need every detail memorized, but you should understand the map.

118

So just to kind of recap those three concepts, you have abstraction, which is mostly about simplifying pieces of logic,

119

modularity, which is about organizing those pieces, as a whole and how everything fits together.

120

Now, the reason that this stuff is more important than ever, at least in my opinion, is because AI can generate a lot of code very quickly,

121

but it doesn't always understand the long -term shape of your project, no matter how much context you give it.

122

It might solve the immediate problem, but put the logic in the wrong place or create an unnecessary abstraction, duplicate something that already exists.

123

And that's the difference between agentic coding and vibe coding.

124

is basically saying build this feature and you just accept whatever comes back

125

and just hope it works and doesn't wreck something else.

126

Where a more serious AI workflow is being able to say, you know, this logic belongs in a service, not in a component, or this helper is unnecessary.

127

That's why when I use AI, I use very specific workflows along with skills and agents, sub -agents that let me make all those decisions.

128

I'm going to be using in my next video to really show you what, what I think a good AI workflow looks like.

129

So next we have side effects and side effects are when code changes something outside of itself.

130

So returning a value from a function is one thing, but five other things may happen as a result of that.

131

So things like updating a database, writing to a file, sending an email, modifying state, those can all be side effects.

132

So a function that has no side effects and just simply takes in and give something back, that's called a pure function.

133

And this matters because side effects are where a lot of bugs happen, right?

134

So they make code harder to predict because the function isn't just giving something back, it's changing the world around it in some way.

135

And AI will often mix pure logic and side effects together.

136

So it helps to be able to look at the code and ask, is this just calculating a value or is it changing something somewhere else that I need to know about?

137

Next we have the request response cycle and this one is really geared more toward obviously web development.

138

So in a browser a user will click a button, the browser sends a request, the server receives it, runs some code, maybe talks to a database and then sends a response back.

139

So if you understand that cycle then frameworks like Express,

140

Laravel, Django, Rails or whatever else they start to make a lot more sense and different frameworks have different syntax

141

But a lot of them are solving the same basic problem and doing the same thing, which is receiving a request, doing some work and returning a response.

142

And I would also lump in just like APIs or REST APIs in this as well.

143

So knowing HTTP status codes, knowing the different methods and so on.

144

So next is concurrency.

145

So concurrency is when multiple things are happening during the same period of time.

146

So maybe you have multiple API requests running, users editing data at the same time, background jobs processing, things like that.

147

And this can get tricky because bugs can happen based on timing, right?

148

Something works fine once but breaks when two things happen at nearly the same time.

149

of video conversions, so basically videos to articles, and one job might be processing, another might be waiting, and another might fail,

150

and then another might finish right as the UI is trying to update.

151

So I had a Big issue with that that I worked on for a week straight.

152

So you don't need to be a concurrency expert right away, but you should understand that software doesn't always just run in a perfect neat line.

153

And AI can handle the simple case, but you still need to think about what happens when that timing gets messy.

154

All right.

155

So I know that that was a lot, but the point of this is not to have you master all of this before touching AI.

156

Nobody masters all this overnight.

157

I certainly haven't even come close.

158

The point of this is that there are concepts

159

that you want to keep learning as you build

160

and as you learn because AI can generate code but you still need to understand the shape of what it's generating.

161

And the better you understand these concepts,

162

the better you can just getting AI to write code,

163

but knowing whether that code actually makes sense.

164

All right, so hope you guys enjoyed it, and I will see you next time.

앱 다운로드

유창하게 말하는 데 필요한 모든 것

AI 발음 평가문장마다 점수 확인

IPA 발음 연습모든 소리 마스터

단어 학습나만의 단어장 만들기

단어 게임놀면서 외우기

QR 스캔

Download on theApp Store Get it onGoogle Play

같은 채널

FOR YOU

이 레슨의 어휘와 말하기 포인트

이 B2 수준 말하기 레슨은 영상 “12 Important Concepts In the Age of AI Software Development”을(를) 바탕으로 합니다. 가장 자주 반복되는 단어는 다음과 같습니다: code, function, happen, user, flow. 이 영상에는 섀도잉할 문장 164개와 단어 2352개가 있습니다. 말하는 구간의 길이는 14:00입니다. 화자는 분당 약 168단어로, 일상 대화에 가까운 자연스러운 속도로 말합니다. 단어의 84%가 영어에서 가장 많이 쓰이는 3,000단어에 속합니다. 나머지는 연습 전에 미리 확인해 두세요.

이 영상의 핵심 어휘

영상에서 가장 어려운 단어 15개를 발음, 뜻과 함께 정리했습니다.

단어	발음	뜻
workflow 명사	/ˈwɝkfloʊ/	워크플로
helper 명사	/ˈhɛlpɚ/	조수, 도우미
memorize 동사	/ˈmɛm.əˌɹaɪ̯z/	기억하다, 암기하다
unnecessary 형용사	/ʌnˈnɛs.əˌsɛɹ.i/	불필요하다
depend 동사	/dɪˈpɛnd/	의존하다, ...에 달려 있다
alphabetically 부사		알파벳순의
shortcut 명사	/ˈʃɔːtkʌt/	바로가기
beginner 명사	/bəˈɡɪnɚ/	초심자, 초보자
syntax 명사	/ˈsɪn.tæks/	통어론, 구문론
duplicate 동사	/ˈd(j)uː.plɪˌkeɪt/	복사하다
hash 명사	/ˈhæʃ/	해시
query 명사	/ˈkwɪɹ.i/	질문
render 동사	/ˈɹɛn.dɚ/	일어나게 하다, 하게하다
modify 동사	/ˈmɑ.dɪˌfaɪ/	수정하다
teammate 명사	/ˈtiːmˌmeɪt/	팀메이트

영상에 나오는 구동사

단어	뜻
come back 동사	돌아오다

이 영상의 문법

화자가 가장 많이 쓰는 문형을 영상 속 실제 표현과 함께 정리했습니다.

문형	영상 속 표현
수동태 be + 과거분사 — 누가 하는지보다 무슨 일이 일어나는지에 초점	be lost · is sponsored · was built
현재완료 have/has + 과거분사 — 과거의 일이 지금도 관련이 있을 때	have changed · we've gone · haven't even come
관계절 who / which + 절 — 사람이나 사물에 대한 추가 정보	abstraction, which is · modularity, which is · thing, which is

주의할 발음

화자는 doesn't, I'm, don't 같은 축약형과 약화된 형태를 25번 사용합니다. 들리는 대로 짧게 발음하세요.

“sh”와 “zh” 소리: abstraction /æbˈstɹæk.ʃn̩/, shortcut /ˈʃɔːtkʌt/, conditional /kənˈdɪʃ.ə.nəl/, hash /ˈhæʃ/, validation /ˌvæl.əˈdeɪ.ʃən/
긴 단어 — 강세 위치에 주의: concurrency /kəŋˈkʌɹənsi/, developer /dɪˈvɛləpɚ/, unnecessary /ʌnˈnɛs.əˌsɛɹ.i/, conditional /kənˈdɪʃ.ə.nəl/, validation /ˌvæl.əˈdeɪ.ʃən/

한국어 화자가 어려워하는 소리:

/f/ — ㅍ(/p/)으로 바꾸지 말고 윗니를 아랫입술에 대기: workflow /ˈwɝkfloʊ/, config /ˌkənˈfɪɡ/, simplify /ˈsɪmplɪfaɪ/, modify /ˈmɑ.dɪˌfaɪ/
/v/ — ㅂ(/b/)과 구별하기: variable [ˈvæɹ.i.ə.bl̩], developer /dɪˈvɛləpɚ/, navigate /ˈnæv.ɪ.ɡeɪt/, validate /ˈvæl.ɪ.deɪt/, validation /ˌvæl.əˈdeɪ.ʃən/
/z/ — ㅈ이 아니라 성대를 울리는 /s/: memorize /ˈmɛm.əˌɹaɪ̯z/, browser /ˈbɹaʊ.zəɹ/, organize /ˈɔɹɡənaɪz/

이 영상으로 연습하는 방법

먼저 말하지 않고 영상을 끝까지 듣고 모르는 단어를 적어 둡니다.
0.75배속으로 한 문장씩 섀도잉을 시작하고, 익숙해지면 보통 속도로 돌아갑니다.
자신의 목소리를 녹음해 원본과 비교하고, workflow, helper, memorize 같은 단어에 특히 주의합니다.

쉐도잉이란? 영어 실력을 빠르게 키우는 과학적 방법

쉐도잉(Shadowing)은 원래 전문 통역사 훈련을 위해 개발된 언어 학습 기법으로, 다언어 학자인 Dr. Alexander Arguelles에 의해 대중화된 방법입니다. 핵심 원리는 간단하지만 매우 강력합니다: 원어민의 영어를 들으면서 1~2초의 짧은 지연으로 즉시 소리 내어 따라 말하는 것——마치 '그림자(shadow)'처럼 화자를 따라가는 것입니다. 문법 공부나 수동적인 청취와 달리, 쉐도잉은 뇌와 입 근육이 동시에 실시간으로 영어를 처리하고 재현하도록 훈련합니다. 연구에 따르면 이 방법은 발음 정확도, 억양, 리듬, 연음, 청취력, 말하기 유창성을 크게 향상시킵니다. IELTS 스피킹 준비와 자연스러운 영어 소통을 원하는 분들에게 특히 효과적입니다.

섀도잉 방법: 단계별 전체 가이드 읽기 →

쉐도잉 연습: 12 Important Concepts In the Age of AI Software Development - 영상으로 영어 말하기 배우기

연습 모드 선택

앱 다운로드

같은 채널

Does Anyone Use JavaScript Anymore?

I've Changed My Opinion On DEVELOPERS Vibe Coding

Nobody is hiring vibe coders. Sorry

추천 영상

Is AI Doom Real... or Just Good Marketing?

AI's Impact on Education: Pros, Cons, and Coping

What is AI and ML? [Pt 2] | Generative AI with .NET for Beginners

The 7 phases of AI-driven development

What is clean code? - Uncle Bob

Cloud Providers are changing RAPIDLY [REUPLOAD]

How to Build Unfair Advantage in your 20s

How I'd Learn AI From Scratch in 2026 (skip the useless 80%)

The Art Of Winning In Tech

Cần cập nhật gì về AI - Ali Abdaal #AI #Update #ielts

7 Programming myths that waste your time

Has AI killed Tailwind?