🌏 中文版
Read the full bilingual transcript and play any sentence
WATCH & LEARN
Video and full subtitles
Before asking AI to do a task, what does the speaker say you need to know? Name one skill you still need.
Play the video, switch subtitle modes, or choose a clip to locate its practice range.
Practise this clip
Identify which parts of a task can be automated and which decisions remain yours.
Practise describing a change with “gets automated” and a remaining responsibility with “I still need to…”.
Name one repetitive task and one judgment you will keep.
AI can resize the images, but I still need to decide which image tells the story.
Practise this clip
Listen for the knowledge a person needs before delegating a task to AI.
Turn a vague request such as “make it good” into criteria another person can check.
Give two criteria for an AI-generated page, email or report.
A good report names the decision and shows one example. I will check whether the conclusion follows from the evidence.
Practise this clip
Distinguish producing an initial version from deciding that the work is finished.
Use “The first draft…, but…” to name a concrete gap and the revision it needs.
Describe a recent first draft and one change you would make before sharing it.
The first draft lists our features, but it does not explain the benefit. I will add a customer example.
Full transcript · 413 segments
Sentence-grouped source captions with Traditional Chinese machine translations. Source timestamps retained; translations have not been reviewed sentence by sentence.
Text is hidden. Use the timestamps or sentence navigation to practise listening; switch modes to reveal it.
Click any sentence to play it in the video above. Loop and speed apply to your selection.
Related expressions · tap for meaning and an authored example.
what good looks like
The qualities that make a result fit its purpose.
Before we ask AI to write the page, let’s agree on what good looks like.
first draft
An initial version that still needs checking and revision.
The first draft lists the features, but it does not explain who needs them.
Practice your understanding and speaking (optional)
02 Write what you understood
Compare with the reference answer
Know what you want and what a good result looks like. Marketing, product, sales and distribution are examples of skills you still need. Explaining a task and judging its result are separate from producing a first draft.
Compare the main idea, not the exact wording.
03 Make it your own
Describe one task you want AI to help with. State two observable quality criteria and one decision you will make yourself.
See an authored example
I want AI to draft a landing page. A good result names the target reader and includes one concrete example. I will decide which benefit matters most.
Questions and examples are authored practice material, not a verbatim transcript. Your text stays on this page and is not submitted or scored.
This is for two kinds of readers: people with scattered interests who keep being told to "pick a lane," and people unsure whether to hand their work to AI. After reading, you can sort your skills into three layers and run one exercise: list your daily work and ask AI which parts can become a repeatable process.
What this is
Dan Koe published If You Have Multiple Interests, Please Take Advantage of the Next 10 Years on 2026-09-03. It runs about 36 minutes with a single speaker.
Scope first. This guide is based on two machine transcripts: the full caption text without timestamps (probably auto-generated) and a timestamped Whisper transcript. I read both in full. I did not listen through the audio myself or check the on-screen visuals; quotes are machine-transcribed text. Everything below is one person's view, not research findings.
Disclosure: he sells things in the video
Partway through he promotes his creator bootcamp, and later mentions Eden, a writing tool his team built. So a conclusion like "a personal brand is the highest-leverage vessel" comes from someone with a stake in it. I take his framework and practical advice only, and I'm not endorsing any paid product.
His opening: three kinds of people
| Type | His description | His verdict |
|---|---|---|
| Purists | Reject AI, retreat to simple living | Reject the external world (work, economy) |
| Automatons | Outsource everything to tech | Reject the inner world (thinking, judgment) |
| Digital Renaissance person | Develops many domains, uses AI as leverage | The path he recommends |
He borrows the four quadrants from Ken Wilber's integral theory (interior/exterior × individual/collective; see Wikipedia) to argue that each extreme drops part of reality. His version is simplified. I only confirmed the quadrants' origin, not that the whole theory suits career planning.
Why he says specialist paths are narrowing
Three claims, all the speaker's own:
- Proof of work replaces the resume, an audience replaces the employer, taste replaces credentials.
- The modern "job" is only about 250 years old (I did not verify this; treat it as his claim).
- AI is good at known information, not at the unknown you reach through experience, so human value moves to taste, direction, and judgment.
Stay skeptical: the trend claims come without data and the history comes without sources. The last point is the usable one, because it turns into action.
The core framework: a three-layer skill stack
| Layer | Content | His view | Ask yourself |
|---|---|---|---|
| 1 | Human nature and psychology: marketing, sales, writing, speaking | Everyone working for themselves needs it, since it decides what people will pay for | Do I know why people would care about this? |
| 2 | Domain obsession: fitness, philosophy, design, coding | The combination of interests creates uniqueness (his example: a fit philosopher) | Which two or three interests are rare together? |
| 3 | Tool skills: Photoshop, no-code, writing code | Most replaceable by AI; don't tie your identity to a tool | Is my identity a tool or a problem? |
His analogy: the scribe became the editor, the hand weaver became the machine operator. The mechanical part gets automated and people move up a layer. Whether the analogy holds is debatable, but "don't bet your whole identity on one tool" applies to almost everyone.
Self-education and the one-person business
He describes learning as apprenticeship: find a mentor (a book or a YouTube channel counts), use your vision and goals as a filter, and act before you study. In his words, you try first and only then is your mind primed to take the lesson.
Next he says to build a one-person business: content to attract an audience, code to build the product, domain knowledge for differentiation. He defers the details to future videos, so this one gives only an outline.
Three rules for using AI
This is the most directly usable part.
- Know what you want and what good looks like. If you don't know marketing, don't expect AI to do it well for you. He suggests starting with AI as a tailored learning tool.
- Set absurdly high standards. AI output feels like slop when you treat the first draft as the final one. A landing page starts rough and gets tuned until it works.
- Try to replace yourself. He thinks this is the fastest way to learn AI, because you quickly find what can and can't be automated.
Limits and trade-offs
- One speaker, one video, no opposing view, no data.
- He says one person can now do what a 10 to 50 person team used to do, depending on the task (around 8:43). That is his estimate, with no source.
- He uses Darwin's routine as an example of high output from few work hours. Secondary sources I found do mention the walks and short fixed work sessions, but the speaker's own hour counts are inconsistent, so I don't repeat them.
- "Build it and the money follows" skips distribution and luck. He says himself you still need marketing, product, sales, and distribution.
Three things to do this week
- Put your skills into the three layers on one sheet. Too many in layer 3 is the warning sign.
- Write down your day, one line per hour, give it to AI and ask which parts could become a repeatable process (a prompt, a skill, a tool). Treat the first result as a draft.
- Pick one thing you want to be known for (his "craft") and treat your other interests as fuel for it.
Three clips to listen and practise
Browse English video practice articles. Choose a clip, press Play, then use the listening goal and speaking prompt beside the video. Speed and loop controls apply to the selected clip; expand the original example after trying your own answer.
The three clips total 2 minutes 28 seconds. Each includes one selected line with a Chinese translation; these are not complete transcripts. Click the line to replay its source audio. Teaching notes and examples are written for this site. Turn on English CC for the remaining source captions. Selected-line timestamps have been checked against the original subtitle track; audio boundaries have not been independently reviewed.
Update history
-
2026-10-20: Rebuilt 413 timed semantic units before translation; source-word coverage retained. Selected passages now share sentence boundaries with the full transcript, which follows playback without requiring prior selection.
-
2026-10-20: Imported all 1066 supplied subtitle segments with Traditional Chinese machine translations and sentence replay. Original clip exercises retained; translations have not been reviewed sentence by sentence.
-
2026-10-19: Combined three source clips in one player with per-clip original listening goals, language notes and speaking tasks; retained the existing comprehension and phrase exercises. Removed duplicate players and untimed locating quotes. Added three selected bilingual lines with timestamps from the original subtitle track. These are not full transcripts; audio boundaries have not been independently reviewed.
-
2026-10-19: Added a guided clip exercise with revealable comprehension answers, contextual phrases and an original speaking task; added the English video practice tag.
-
2026-10-16: Added three shadowing clips with source-player captions (50, 53, and 45 seconds), timed from a Whisper machine transcript; the "10 to 50 person team" line now states it is the speaker's estimate.