Luyện nói tiếng Anh bằng Shadowing qua video: 12 Important Concepts In the Age of AI Software Development

Chọn chế độ luyện tập

Mỗi chế độ mở một phòng luyện tập toàn màn hình.

164 câu

Nếu các câu quá ngắn hoặc quá dài, hãy bấm Edit để chỉnh sửa.

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.

Tải Ứng Dụng

Mọi thứ bạn cần để nói trôi chảy

Chấm điểm AIChấm từng câu nói

Luyện IPAThành thạo mọi âm

Từ vựngMở rộng vốn từ

Game Từ vựngHọc qua chơi

Quét QR

Download on theApp Store Get it onGoogle Play

Cùng kênh

FOR YOU

Video gợi ý

180 câu

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

Business EnglishUniversity StudentsWorking Adults

25 câu

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

Business EnglishUniversity StudentsWorking Adults

52 câu

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

Business EnglishUniversity StudentsWorking Adults

104 câu

The 7 phases of AI-driven development

Business EnglishUniversity StudentsWorking Adults

80 câu

What is clean code? - Uncle Bob

Business EnglishUniversity StudentsWorking Adults

365 câu

Top 1% Career Strategy: How to Get Hired in the AI Era

Business EnglishUniversity StudentsWorking Adults

55 câu

Cloud Providers are changing RAPIDLY [REUPLOAD]

Business EnglishUniversity StudentsWorking Adults

165 câu

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

Business EnglishUniversity StudentsWorking Adults

89 câu

The Art Of Winning In Tech

Business EnglishInterview EnglishUniversity Students

11 câu

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

IELTS PrepBusiness EnglishUniversity Students

69 câu

7 Programming myths that waste your time

Business EnglishUniversity StudentsWorking Adults

261 câu

Has AI killed Tailwind?

Business EnglishUniversity StudentsWorking Adults

Từ vựng và ghi chú luyện nói cho bài học này

Bài luyện nói trình độ B2 này dựa trên video “12 Important Concepts In the Age of AI Software Development”. Những từ được nhắc lại nhiều nhất trong bài: code, function, happen, user, flow. Video này có 164 câu và 2352 từ để luyện shadowing. Phần lời nói dài 14:00. Người nói nói với tốc độ tự nhiên, khoảng 168 từ mỗi phút, gần với hội thoại hằng ngày. 84% số từ nằm trong 3.000 từ tiếng Anh thông dụng nhất; phần còn lại nên xem trước khi luyện.

Từ vựng quan trọng trong video

15 từ khó nhất trong video, kèm phiên âm và nghĩa:

Từ	Phiên âm	Nghĩa
helper danh từ	/ˈhɛlpɚ/	phó thủ
developer danh từ	/dɪˈvɛləpɚ/	nhà phát triển
endpoint danh từ		đầu mút, mút
memorize động từ	/ˈmɛm.əˌɹaɪ̯z/	ghi nhớ, ghi chép
depend động từ	/dɪˈpɛnd/	phụ thuộc
alphabetically trạng từ		theo thứ tự bảng chữ cái
reviewer danh từ		nhà phê bình
shortcut danh từ	/ˈʃɔːtkʌt/	lối tắt, đường tắt
simplify động từ	/ˈsɪmplɪfaɪ/	đơn giản hóa
beginner danh từ	/bəˈɡɪnɚ/	người bắt đầu, người mới học
syntax danh từ	/ˈsɪn.tæks/	cú pháp
duplicate động từ	/ˈd(j)uː.plɪˌkeɪt/	nhân bản, sao chép
conditional danh từ	/kənˈdɪʃ.ə.nəl/	thức điều kiện
hash danh từ	/ˈhæʃ/	thịt băm, thịt bằm
lump danh từ	/lʌmp/	đống

Cụm động từ bạn sẽ nghe

Từ	Nghĩa
come back động từ	trở lại, trở về

Ngữ pháp trong video

Những cấu trúc người nói dùng nhiều nhất, kèm đúng cụm từ trong video:

Cấu trúc	Trong video
Câu bị động be + quá khứ phân từ — nhấn vào việc xảy ra, không phải người làm	be lost · is sponsored · was built
Thì hiện tại hoàn thành have/has + quá khứ phân từ — việc đã xảy ra nhưng còn liên quan đến hiện tại	have changed · we've gone · haven't even come
Mệnh đề quan hệ who / which + mệnh đề — thêm thông tin về người hoặc vật	abstraction, which is · modularity, which is · thing, which is

Phát âm cần chú ý

Người nói dùng 25 dạng rút gọn, ví dụ doesn't, I'm, don't. Hãy nói theo dạng ngắn đúng như bạn nghe.

Âm “sh” và “zh”: abstraction /æbˈstɹæk.ʃn̩/, shortcut /ˈʃɔːtkʌt/, conditional /kənˈdɪʃ.ə.nəl/, hash /ˈhæʃ/, validation /ˌvæl.əˈdeɪ.ʃən/
Từ dài — đặt trọng âm cho đúng: concurrency /kəŋˈkʌɹənsi/, developer /dɪˈvɛləpɚ/, unnecessary /ʌnˈnɛs.əˌsɛɹ.i/, conditional /kənˈdɪʃ.ə.nəl/, validation /ˌvæl.əˈdeɪ.ʃən/

Những âm người Việt thường gặp khó:

Cụm phụ âm cuối — đọc đủ từng âm, đừng nuốt âm cuối: depend /dɪˈpɛnd/, checklist /ˈt͡ʃɛklɪst/, syntax /ˈsɪn.tæks/, lump /lʌmp/, cart /kɑɹt/
Âm /tʃ/, /dʒ/ và /ʒ/ — tiếng Việt không có, đừng đọc thành “ch” hay “gi”: agentic /eɪˈd͡ʒɛn.tɪk/, checklist /ˈt͡ʃɛklɪst/, logical /ˈlɑd͡ʒɪkəl/
Âm /æ/ — mở miệng rộng hơn âm “e”: abstraction /æbˈstɹæk.ʃn̩/, variable [ˈvæɹ.i.ə.bl̩], navigate /ˈnæv.ɪ.ɡeɪt/, validate /ˈvæl.ɪ.deɪt/, handler /ˈhænd.lə/

Cách luyện với video này

Nghe hết video một lần, chưa cần nói, và ghi lại những từ bạn chưa biết.
Bắt đầu ở tốc độ 0,75×, nói đuổi từng câu, rồi quay lại tốc độ bình thường khi đã quen.
Ghi âm giọng mình rồi so với bản gốc, chú ý các từ như helper, developer, endpoint.

Phương Pháp Shadowing Là Gì?

Shadowing là kỹ thuật học ngôn ngữ có cơ sở khoa học, ban đầu được phát triển cho chương trình đào tạo phiên dịch viên chuyên nghiệp và được phổ biến rộng rãi bởi nhà đa ngôn ngữ học Dr. Alexander Arguelles. Nguyên lý cốt lõi đơn giản nhưng cực kỳ hiệu quả: bạn nghe tiếng Anh của người bản xứ và lặp lại to ngay lập tức — như một "cái bóng" (shadow) đuổi theo người nói với độ trễ chỉ 1–2 giây. Khác với luyện ngữ pháp hay học từ vựng bị động, Shadowing buộc não bộ và cơ miệng phải đồng thời xử lý và tái tạo ngôn ngữ thực tế. Các nghiên cứu khoa học xác nhận phương pháp này cải thiện đáng kể phát âm, ngữ điệu, nhịp điệu, nối âm, kỹ năng nghe và độ lưu loát khi nói — đặc biệt hiệu quả cho người luyện IELTS Speaking và muốn giao tiếp tiếng Anh tự nhiên như người bản ngữ.

Phương pháp shadowing: đọc hướng dẫn từng bước đầy đủ →

Luyện nói tiếng Anh bằng Shadowing qua video: 12 Important Concepts In the Age of AI Software Development

Chọn chế độ luyện tập

Tải Ứng Dụng

Claude Code Mods Explained: Customize Your Setup

Does Anyone Use JavaScript Anymore?

I've Changed My Opinion On DEVELOPERS Vibe Coding

Nobody is hiring vibe coders. Sorry

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

Top 1% Career Strategy: How to Get Hired in the AI Era

Cloud Providers are changing RAPIDLY [REUPLOAD]

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?

Từ vựng và ghi chú luyện nói cho bài học này

Từ vựng quan trọng trong video

Cụm động từ bạn sẽ nghe

Ngữ pháp trong video

Phát âm cần chú ý

Cách luyện với video này

Phương Pháp Shadowing Là Gì?