シャドーイング練習: Updated Essential AI Skills For 2026 - YouTubeで英語スピーキングを学ぶ
C1
So it's only been five months since I gave you my 2026 Essential AI Skills Roadmap.
295 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
So it's only been five months since I gave you my 2026 Essential AI Skills Roadmap.
2
But oh my goodness, things have changed so much since the beginning of this year.
3
So I feel like I gotta give you an updated version.
4
So this video is it,
5
the updated Essential AI Skills Roadmap.
6
And I'm going to structure like a progression from the most basic skills to intermediate to advanced.
7
And pay attention because at the end of this video,
8
there will be a little quiz to make sure that you retain all of this information.
9
Now without further ado, let's go.
10
A portion of this video is sponsored by BALT.
11
Level one, securing the basics.
12
I think these are skills that every single modern person that like participates in society should have about AI.
13
And the first of these might actually surprise you because this is a tech AI learning channel.
14
But the first skill I'm going to mention is investing.
15
Yes, because in my opinion,
16
in 2026, if you are investing your money,
17
I think you really need to have an AI thesis.
18
You see the age old wisdom of just investing into the S&P 500
19
or whatever index it is that you're investing in does not count for the fact
20
that you are having massive exposure to AI,
21
whether you know that or not.
22
Because many of those companies in these funds are either AI companies or heavily integrated and invested in AI.
23
So I think it's really important for you to make a
24
decision of how much AI exposure you want to have with your investments in relation to your current like career
25
and your life in general.
26
For me, for example, I am massively exposed to AI through my career.
27
Plus I have like a very unstable career because I'm a YouTuber and entrepreneur.
28
So my investment philosophy reflects this.
29
I try to use my knowledge about AI to invest in things that I believe in,
30
but I also hedge against AI to an appropriate amount.
31
Anyways, I'm not going to go into more detail about this
32
because this is not a personal finance investment channel and this is hashtag not financial advice.
33
I just think it's really important for you,
34
for everybody these days, to think about AI when you are investing.
35
I'll also leave a little prompt in description.
36
At a very meta level,
37
you can put it into an AI chatbot to help you
38
figure out what your investment thesis should be based upon your specific life situation.
39
Now let's move on to the next AI skill that I think everybody should know in 2026,
40
which is prompting.
41
Prompting is the way that you interact with AI,
42
like any type of AI.
43
So knowing how to prompt well lays the foundation for everything that you do with AI.
44
It is so important.
45
I've talked about prompting a lot on this channel,
46
so I'm going to put on screen now two basic frameworks that you should definitely know if you don't already.
47
And you can check out this video over here where I go through the basics of prompting in much more detail.
48
Now let's move on to the last skill of the basics category
49
that I think everybody should know in 2026 and that is mastering a core set of tools.
50
You see in the AI world there's like tools popping up all over the place.
51
There's like 10 releases per day,
52
probably more at this point.
53
This can get really overwhelming if you're just trying to chase the hype of tools.
54
Now what I find to be much more productive
55
and sustainable is to master a core set of tools and just master them really really well.
56
If you want to be an AI minimalist
57
and you only want to have like one AI tool in your arsenal then I would recommend
58
that you pick from the category of a general AI chatbot.
59
This is like your Claude,
60
Chachapiti, Gemini, Rasa, Poe, etc,
61
etc. I'm going to put on screen examples of chatbots that you can choose from.
62
Honestly, even these like general chatbots are so powerful these days.
63
So you can ask questions like plan on my itinerary to go to France in two weeks,
64
generate images, audios, videos, animations,
65
even like prototype entire apps.
66
So if you've mastered a skill of prompting,
67
just using one of these chatbots well will get you very far.
68
Now, if you do want to be a little bit more fancy,
69
you can also check out a little bit more specialized AI tools.
70
Like most people will probably benefit from having a specialized news and research AI tool,
71
like perplexity, for example.
72
You can look at the news that you're interested in and be able to do more extensive research.
73
And for learning new things,
74
which I think everybody should be learning new things,
75
you can also check out a product like Notebook LM.
76
Now, if you want to harness the power of AI even more,
77
then I would also recommend checking out one or two AI tools that is also specific to your job.
78
For example, if you're a software engineer,
79
then you probably also want to master an AI coding tool.
80
And if you're a marketing person,
81
maybe you can also check out a tool that's specialized for,
82
say, content and SEO, like Jasper or Surfer SEO, for example.
83
I'm going to put a prompt in a description,
84
which you can copy paste to your favorite AI chatbot
85
and it will ask you some questions
86
and it will help you figure out what are some tools
87
that you can check out in for your specific job if you wish.
88
All right
89
so those are the basic skills I think everybody needs to know in 2026 such as like survive
90
and keep up with society but
91
if you're someone who's interested in harnessing the power of AI more
92
and integrating that into your own life
93
and workflows more then you want to move on to the intermediate skills
94
and this is the transition into thinking about what are called AI agents.
95
AI agents are software systems that use AI to pursue goals and complete tasks on behalf of users.
96
You see previously, you're just using AI tools for it to like do something very specific
97
or to give you suggestions for things.
98
Like, hey, like chat GPT,
99
can you like write me a paragraph about cars?
100
Or like, can you plan my itinerary?
101
This is like one request and you get one response back.
102
But with an AI agent,
103
you can give it like an overarching goal.
104
Like, hey, I got this new office.
105
I showed some pictures of your new office.
106
Help me decorate this office.
107
By the way, I literally did this with an AI agent called Manus.
108
It brainstormed different aesthetics with me,
109
ended up going with the light academia vibes.
110
Then it was able to take our exact office and decorate it
111
so it has this aesthetic and even link those specific products like the rug,
112
the lamp, the curtains.
113
You see, that's the power of agents, agentic workflows.
114
You can give it an overarching goal and it can break down the steps
115
that it needs to do and then go with and execute these steps.
116
I hope you can see how this is much more powerful than just like chatting with a tool.
117
So Manus is an example of the agent,
118
a web-based agent that can do stuff for you.
119
And it's a really great entry point into learning about how to prompt and work with agents.
120
But what puts you at the next level and is very popular these days is what are called local AI agents.
121
A local AI agent is firstly an AI agent that takes action
122
and completes tasks on its own and is also local
123
which means it's an agent that lives and runs on your specific machine like your computer.
124
And this is very very powerful.
125
It allows you to build out custom automations and workflows just for you.
126
For example, I have an agentic workflow that posts for my calendars,
127
my gmail, my investments, slack,
128
and notion to put together a personal daily digest to me straight to my Apple Notes.
129
I also have an AI workflow that updates AI news,
130
tracks AI topics, deep dives into them,
131
and helps me draft up scripts for my YouTube videos.
132
Of course, I still spend a lot of time researching more into these scripts and perfecting them,
133
and I still have to like film them myself.
134
Hello, I'm Tina.
135
I'm still real.
136
But overall, it saved me so much time,
137
and I'm able to increase the quality of my videos so much by this.
138
Another example, just to make it three,
139
I also have an investment dashboard that pulls in like custom news and curated suggestions for my specific investments too.
140
There are of course so many other workflows that you can do.
141
I'm going to put on screen now some other examples.
142
All you have to do is learn how to use a local AI agent like Cowork,
143
OpenClaw, Hermes for example.
144
I'm sure as the weeks
145
and the months go on there's going to be a lot more products in this local AI agent category as well.
146
But all you got to do is start off by choosing one of them depending on two factors.
147
The first one is how technical you are.
148
Some products are geared towards people who are non-technical
149
and don't know how to code and others are more towards people who do know how to code.
150
And the second factor is if you want to go with open source
151
or closed source ai claude chachi pt gemini are considered closed
152
source ai wall models like kimmy minimax llama are considered open
153
source i am not going to go into too much more details about what open source is
154
because I made an entire video about this,
155
which you can check out over here.
156
But basically, it boils down to capability, cost, and privacy.
157
Closed source models tend to be more capable and more powerful,
158
although that gap is closing.
159
While open source models are much lower cost,
160
we're free to run, and private if you choose to run them yourself.
161
So you don't have to be like sending your entire bank account information to Anthropoc or OpenAI, for example.
162
Anyways, I'm going to put a prompt in the description as well to help you figure out
163
which local AI agent that you should start with given your specific skill set and your needs.
164
You can just stick that into your favorite AI chatbot and it should help you out.
165
And that leads us to the next level, the advanced level.
166
Bolt is my go-to for creating really fast AI-powered functional web apps.
167
You can go from idea to a deployed production-ready app just by describing what you want in plain English.
168
And now it has a connector layer that lets your project pull in capabilities like shaders,
169
Notion, GitHub, and more through MCPs.
170
I'm building a new landing page for my Lonely Octopus AI Agents Bootcamp and I want it to look actually good,
171
not just functional like it is now.
172
So I use Bald to build out the full landing page and it pulls in shaders directly.
173
Here's an animated purple hero background and there's no manual setup.
174
No going to shaders and copying embed codes.
175
It's just there.
176
I can tweak it conversationally if I want the animation slower or a different shade of purple.
177
I just describe it and Bald can adjust it.
178
That prompt tweak loop is what makes this so powerful compared to just grabbing a static shader somewhere else.
179
You're building and refining in one place.
180
To set it up you just go to personal settings find a connectors panel
181
and toggle shaders on for your project that's it you can
182
now build in bolt for free the link is in the description thank you
183
so much ball for sponsoring this portion of the video now back to the video the difference between intermediate
184
and advanced skills is
185
that for intermediate you're kind of like doing things on your
186
own right like for your personal workflows maybe for some business workflows
187
but they're still more or less for yourself
188
but for people who are interested in AI from a commercial perspective or even as a career,
189
then it's important to learn the advanced skills.
190
And I gotta say, these advanced AI skills unlock so many opportunities.
191
These are skills that are in such high demand and knowing these gives you such an unfair advantage.
192
And that is when you might transition to the next skill,
193
which is actually building your own AI agents.
194
So if you're considering building commercial pipelines where other people are going to be using it,
195
potentially you might even have clients who are going to be using it,
196
you need these to be stable, reliable, and lowering costs.
197
For example, us as a company,
198
Lonely Octopus, we also have B2B clients.
199
And for them, we build up different types of agentic workflows and products.
200
For example, for a private equity company,
201
we built a custom reporting pipeline.
202
It pulls data from their tools like CRM,
203
analytics, spreadsheets, and also from their custom database on a schedule and then assembles it into a formatted report.
204
Private equity companies seem to have to do a lot of reporting,
205
so this has saved them so much time.
206
They're able just to send off these automated reports with their KPIs,
207
client updates, and board summaries.
208
We've also helped companies build onboarding agents where they can onboard new hires
209
and clients that go through a series of steps that is custom for their company.
210
It tracks their progress and they can customize it to however they like.
211
These are just a couple examples.
212
We have so many requests from people to help them build these different AI agents.
213
So if you're someone interested in like freelancing or even like working in the AI field,
214
just know that learning how to build your own AI agents is a very,
215
very in demand skill set kind of related to this building MCPs,
216
which allow AI agents to plug into third party apps and different data sources is also really,
217
really in demand as a skill.
218
By the way, I just want to say I called this like months ago when I made my MCP video.
219
So, you know, anyways, if you learn how to do this,
220
this is also a skill set that's very in high demand.
221
Not going to go into way too much detail about this.
222
I have so many different videos covering how to build AI agents about MCP,
223
and we actually run an entire four-week bootcamp where we teach people how to build AI agents.
224
So I'm going to link all of these resources into the description of the video,
225
which you can check out if you want to dive deeper into it.
226
And just to end off this section,
227
I'm going to show some different tools,
228
both non-technical and technical coding tools that you can learn in order to build AI agents.
229
Moving on to the final AI skill,
230
and that is AI coding.
231
AI coding, otherwise known as agentic engineering,
232
is using an AI agent in order to write code and build software.
233
Some of the examples of things that me and my team have built include this internal tool that helps us create slides,
234
documents, resources using our style and our branding.
235
And we're also creating a learning app which I hope I can share more details with you about soon.
236
This manga generator and accounting software.
237
Now there are pros and cons to AI coding.
238
The pro is that if you learn AI coding this is like boss level unlock.
239
Not only are you able to build like big complex production
240
grade products you'll be able to do this at like a fraction of the cost and time.
241
Like we're talking about like more than 10 times cost and time savings.
242
I honestly cannot even explain like how big of a deal this is.
243
If there is one domain that AI has already like completely transformed then
244
that would be coding like the difference between someone who knows how to do like AI coding
245
and doesn't know AI coding it is just so fast
246
if you want to truly unlock AI
247
and the power of AI you really got to know how
248
to do AI coding now the con of this is
249
that you low-key do already need to know how to code
250
I will use this by explaining it's kind of like putting lipstick on a pig you know like
251
if your coding sucks
252
and you don't know how to code it's like putting lipstick
253
on a pig it doesn't actually matter how good the AI coding tool it is
254
that you're using and like how fancy your setup is,
255
it's just not, it's not,
256
it's not going to work.
257
You're not going to get far.
258
Like you're going to run to a lot of issues along the way.
259
And coding is one of those things that does take some effort in order to learn.
260
Like all the other skills that we talked about earlier,
261
these are things that you can master in the span of days to weeks.
262
But to actually properly learn coding and then AI coding,
263
you do need to be spending at least like a couple months,
264
maybe like two to three months to properly master things.
265
I actually made a video about how to learn to code in the age of AI
266
if you want to check it out,
267
which I'll link over here.
268
But yes, it does take some effort
269
and commitment to actually learn how to code and be able to harness the power of AI coding.
270
Now, I don't think everybody has to do this.
271
I think you can get like really far along in AI without learning how to code and doing like AI coding, right?
272
That's why this is like the most advanced last skill.
273
You would be totally fine like building simpler workflows,
274
prototypes, and things like that without knowing how to code.
275
I would only recommend making a commitment to learn how to code if you don't know how to already,
276
and then learning AI coding if you're someone who's genuinely interested in really digging deep into AI and harassing its powers.
277
Plus you can also reduce your own costs a lot more
278
because you can stop paying for like subscriptions for different products
279
and you can just like AI code your own clones of them which is what I do.
280
Stuff like no taking accounting or budgeting softwares for example.
281
So I have not yet made a full video on AI coding.
282
It is in the pipeline.
283
I have heard your request.
284
I promise it is coming.
285
Give me a little bit more time and when it is out I will link over here so you can check out.
286
And that's it.
287
We're at the end of this video.
288
Wow.
289
These are the essential AI skills from the most basic, intermediate to advanced.
290
Now, as promised, I will show on screen a little quiz.
291
Please answer these questions
292
and put them into the comments to help you retain the information we have just covered
293
because science says this is the best way for learning.
294
And thank you so much.
295
Have fun learning AI and I will see you guys in the next video or live stream.
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
このレッスンについて
このレッスンでは、AIスキルの基礎について学びながら、英語スピーキング練習を行います。ビデオから抽出された文を使って、特に投資、プロンプティング、ツールの習得に関する重要な概念を理解し、実際に英語を話す力を強化します。このレッスンは、特に2026年に必要とされるAIに関連する技能を学ぶことに焦点を当てています。
重要な語彙とフレーズ
- 投資 (investment) - 資金を特定のプロジェクトやビジネスに投入すること。
- プロンプティング (prompting) - AIとやりとりするための方法。
- AI (人工知能) - 人間の知能を模倣するテクノロジー。
- ツール (tools) - 使用するためのソフトウェアや技術。
- 基礎 (basics) - まず最初に学ぶべき基本事項。
- 哲学 (thesis) - 投資や考え方における基本的な信念。
- 露出 (exposure) - 特定の状況に対する影響や関与。
- ヘッジ (hedge) - リスクを軽減するための戦略。
練習のコツ
このビデオの内容に沿って、英語スピーキング練習を効果的に行うためには、以下のポイントを意識してください。まず、自分のペースでビデオを視聴し、音声を聞き取ることから始めてください。言葉のスピードが早いと感じる場合は、再生速度を調整してみましょう。理解できるフレーズを見つけたら、そのフレーズを繰り返すことで、shadow speechの練習を行います。
特に、shadowing siteを活用すると、リスニングと発音の両方を強化できます。ビデオの最後にはクイズがあるため、情報をしっかり保持するために、その内容についても復習を行ってください。英語を話す力を伸ばすために、YouTubeで英語学習を効率よく活用して、IELTSスピーキング対策にも役立てましょう。
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。