跟读练习: building an app is easy, actually - 通过视频学习英语口语

正在创建课程...
1
Usually when people talk about building apps on here, and by here I mean like on YouTube, they always advertise it as a method to make money.
2
Which is a little ironic because YouTube itself wasn't even a profitable app until around 2014.
3
In fact, most apps nowadays aren't profitable unless they're B2B or business-to-business, meaning you're selling to people who themselves are selling to people.
4
So I find it kind of funny when people claim
5
that you can drop out of school or quit your job and just build apps for a living.
6
because yeah you can build apps for a living
7
but you'll be building them for a tech company
8
or a bank at a job or you could just quit your job
9
and build apps and not make a living
10
or you could quit your job and make a living
11
but you won't be building apps I'm not saying it's impossible
12
to make a living I'm just saying that's not possible and even
13
if you do get to do all three at the same time be careful what you wish for
14
so let's just make an app that actually does something useful for fun
15
because we want to because that's the best way to to learn anyway.
16
I assume somewhere on your laptop right now is a PDF you haven't read.
17
Maybe it's a research paper you bookmarked and forgot, but me no want open PDF and read word.
18
We want stimulating activity.
19
So how do you actually just sit down and read a PDF fast?
20
The solution is a technique you've probably seen before, called RSVP, Rapid Serial Visual Presentation, where instead of moving your eyes left to right to read something,
21
you just have the words flash on screen really fast one
22
or two words at a time even faster than you could say them out loud.
23
Normally the fastest a human can read is around 250 words per minute but this technique can get you to read 350,
24
400 or even 500 words per minute without any previous training.
25
Now let me just say there are plenty of free apps out there where you can paste in some text
26
and it will flash each word on the screen one at a time for you
27
but so far I haven't found anything that's both free and works on PDFs.
28
So how about we build one?
29
And let's call it Read Faster, because I was going to make a creative name, but then I realized half of the most downloaded apps on Google Play don't have creative names.
30
So why bother?
31
Now that you have the idea for the app and the name, you might feel a little underprepared.
32
But many tech startups, in order to become truly billion-dollar companies, go through a long period of aggressively spending money from outside investors to grow.
33
During the losing money stage, these startups are called pre-profit.
34
Some tech startups don't even charge money yet, which makes them pre-revenue.
35
Some are still working on the actual app and haven't released it yet, which makes them pre-launch.
36
Some haven't even started working on the app and just have a PowerPoint presentation on what the app would do.
37
This is known as pre-product.
38
Lastly, some startups don't even know what the app would be about, just that it's going to use AI.
39
This is known as pre-idea and it's about where 99% of startups are, so the fact that you have an idea already puts you way ahead.
40
Now even though you might want to get to blindly vibe coding the app straight away, the components of an app are actually pretty easy to understand, so I think you might as well learn them.
41
First, you need a UI or user interface, also known as a front end.
42
This is all of the buttons and screens that the user will see and interact with on the app.
43
The UI for your app will be slightly different when it's on the web versus iOS versus Android.
44
Next, you may need a server-side framework, also known as a backend.
45
This is what will manage all the data for a user
46
so that they can make an account or use the same app across multiple devices.
47
For example, let's say you make an app like Tinder
48
that requires people to make an account and five people download your app.
49
In order for them to sign in, you need some type of server that, somewhere, keeps track of all their passwords and data and who they matched with and what their current Tinder ELO is.
50
Every time someone logs in, they have to talk to the server first.
51
If you want this to happen, you need to code a backend.
52
But if it's just a game like Flappy Bird, where no one has an account, and there's no leaderboard, and the only data it tracks is the highest score each person has on their phone, then you don't need to make a backend.
53
Flappy Bird would be considered an entirely front-end app.
54
The last thing is packaging and publishing.
55
This just means combining all the code into a format that a phone or web browser can understand.
56
and then you just publish to the App Store or Google Play, or it could even just run on a website.
57
Now that we at least understand what's happening, we can vibe code the app.
58
Well, maybe not you, maybe you're an actual coder, but I will vibe code using Bolt, which is just a free option that used to be popular, but may or may not be completely obsolete in two months.
59
Never mind, it's already obsolete.
60
Okay, so instead we're going to use Google AI Studio to build the app.
61
You can write a prompt here for the app, which obviously I'm going to ask Claude to write.
62
I'm going to give it the app requirements, and then the framework I want it to use, and then the single design requirement.
63
A framework is just a pre-written set of tools, libraries, and other stuff that just makes your app logic standardized, easier to code, more secure, and a lot more reliable.
64
The technology we'll use is the free Kotlin multi-platform made by JetBrains.
65
You may have heard the Kotlin programming language before, but this is a technology that's multi-platform, meaning it allows you to write code once and deploy it to Android,
66
iOS, desktop, to a server, and the web just from one code base.
67
In a multi-platform architecture, your app's divided into two pieces.
68
The first is the shared logic.
69
For us, that's the logic for the program to read the PDF, to sequence the words, to calculate the words per minute and the timing of each word.
70
Basically, everything that is done behind the scenes to flash the words on the screen.
71
That code is written once in Kotlin and runs on every platform.
72
So you only have to maintain one thing, just one code base.
73
The second piece is the UI.
74
With Compose Multi-Platform, JetBrains' declarative UI framework, which works across Android, iOS, desktop, and web, you write the UI once and it renders natively everywhere.
75
Meaning everywhere, the app can run smoothly and without any lag.
76
The recommended way to set this up is to download IntelliJ IDEA, which is a coding workspace.
77
This is where you'll be writing all your code.
78
Then you just go to plugins and download the one called Kotlin Multiplatform.
79
After this, the template and starter files will already be there for you.
80
So you already have the skeleton.
81
Now you just have to code.
82
And if you've never written Kotlin before, imagine Java, but about 40% less code.
83
You can code an app just once and you could ship it out to run cleanly on any platform.
84
If you're really getting into app development, you probably want to try this out, so check the description.
85
By now, Google AI Studio should have fully generated the app
86
and now you could preview it and you could also look at all the code.
87
Once we've previewed it and tried it and we see that it works, the next step is to thoroughly test the app across various cases,
88
various inputs, and all kinds of button presses to make sure that there are no bugs.
89
After completely ignoring that step, we move on to deployment.
90
Well, okay, you're not supposed to completely ignore testing, but personally, I feel like playing around with the preview you get is enough testing.
91
Now it's time to deploy, which means put the app out on the internet for anyone to use slash install.
92
In our case, we'll deploy it to the web
93
because it costs money to get your app on the App Store or Google Play, and I know you don't want to do that.
94
Now, you could deploy straight to Google Cloud.
95
That's what Google wants you to do.
96
That's why it's the only option when you click the big publish button.
97
And technically, Google does give out $300 of free Google Cloud credit to everyone.
98
So you technically don't have to pay, but still requires you to put in your credit card information.
99
And I know you don't want to do that.
100
Anyway, we're not going to deploy with Google Cloud.
101
We're going to deploy using Vercel, which is free.
102
Vercel basically hosts the app on the web so
103
that Someone across the world can type in the link to your app and access it online and use it.
104
Before that, one thing you might want to do first is go to the code tab here, and then on the right, you can download all the code as a zip file in
105
case you want to work on it later inside an actual code editor and not Google AI Studio.
106
But if you don't feel like downloading the code, you don't have to.
107
Now to deploy with Vercel, we first need to use GitHub as a middleman.
108
Just hit the settings button in the top right and keep scrolling right until you get to GitHub.
109
Now, just sync your code to GitHub.
110
Make a GitHub account if you need to.
111
It probably takes under a minute, like I made a new account just for this.
112
And then you follow the steps to transfer your code from Google AI Studio over to GitHub, which should be quick.
113
Once all that is finished, you now have a GitHub repository that is storing all your code for the app.
114
Now search up Vercel, make an account, and then click Start Deploying.
115
Copy your link to your GitHub repository and paste it here.
116
And then follow all the steps after that until eventually you can click the final button and deploy your app.
117
Pretty soon it'll finish deploying and now you plus anyone else can access and use the app on the web.
118
This app doesn't actually have a backend.
119
So it doesn't store any user data.
120
It doesn't need you to log in.
121
And it would have done it using Firebase, which is just a Google-owned service that'll take care of everything you need for databases,
122
authentication, cloud functions, sending notifications, stuff like that.
123
All you have to do is click yes when it asks you if you want to use Firebase.
124
And now you have a backend.
125
Anyway, there's no steps left, besides maybe improving your app if you want to.
126
But other than that, it's done.
127
Use it once and forget about it.
128
Tell your friends who don't care about it.
129
Add it to your resume.
130
They'll get auto rejected by the AI screening system anyway, whatever you want to do with it.
131
You can even LARP as an actual software developer.
132
I know I will.

這個視頻適合誰?

無論你是剛開始練習英语口语练习的新手,還是準備雅思口语练习的考生,這段視頻都很適合。它用親切的口語化表達講述開發APP的趣事,語速適中,內容貼近生活,非常適合用來練習shadow speak(影子跟讀)。跟讀時既能積累實用詞彙,又能模仿自然的語調,輕鬆提升口語流利度。

值得積累的詞彙和短語

  • B2B:即business-to-business,指“企業對企業”的商業模式。
  • pre-profit:“盈利前”,指公司尚未開始賺錢的階段。
  • blindly vibe coding:形象表達“盲目地隨意編碼”,類似於我們常說的“亂敲代碼”。
  • be careful what you wish for:諺語,意思是“當心你許的願”,提醒別人不要輕易奢望。

如何模仿口音更地道?

這位講者的口音清晰自然,沒有過多重口音,很適合練習shadow speech。注意以下兩點:一是連讀,比如“make a living”會讀成“mak-a-living”,跟讀時要跟上節奏;二是語氣起伏,講者在講笑話或強調觀點時會加重語氣,比如“but you won't be building apps”,模仿時要體現出這種情緒變化。每天花10分鐘跟讀,堅持一周就能明顯感覺口語更流暢!

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。

影子跟读法: 阅读完整分步指南 →