A 13-year-old in Hangzhou snaps a photo of a dull vocabulary list, asks an AI chatbot to turn it into a story, and thirty seconds later the list has become something he actually wants to read. The video of him doing it reached about 20 million views. In this episode of Echo Stories English, Emma and Ryan break the method down step by step, check it against decades of memory research, and fix its one big weakness.

This article is a detailed summary of the episode (level B1–C1, about 35 minutes), with copy-ready prompts and a 15-minute daily routine you can start tonight.

▶ Watch the full video on YouTube: 20M Views: How “China’s Little Elon Musk” Learns English with AI — use the chapter times below to jump to any section.

Video chapters

TimeChapter
00:00Intro: the 13-year-old who broke the internet
02:22Who is Yang Xizhe? (“China’s Little Elon Musk”)
06:39The 5-step AI method
13:58The science of memory: why rote learning fails
22:13Prompts, the upgrade and the “1 to 100” rule
29:36Wrap-up and final takeaway

1. Who is Yang Xizhe?

Yang Xizhe attends Fuchun Middle School in Hangzhou, in eastern China. His vocabulary video was filmed over Chinese New Year 2025, when he was still in sixth grade. He used Doubao, a popular Chinese AI chatbot made by ByteDance (the company behind TikTok), to learn English words — and the video went viral.

The “future Elon Musk” nickname came later, from his hackathon projects:

  • An AI vision system that helps visually impaired people put on make-up — first place at the Moonshot 48 hackathon.
  • A smart lunchbox with a camera and sensors so parents can see what their children really eat at school — Best Pitch at HackMIT China.
  • At a hackathon run by RedNote, a company offered 100,000 yuan (about US$14,000) for one of his tools.

The most useful part of his story is how ordinary it is. According to one interview he ranks around 50th in his grade, spends roughly three hours a day on coding and AI, and simply keeps his grades above average. His parents have no tech background, learned about AI to keep up with him, and deliberately downplay the “future Elon Musk” label.

That is exactly why the method is worth copying: it is not a genius trick, it is a smarter way for a normal student to do boring homework.

2. What goes wrong with the usual way

Most of us learn words like this: word on the left, meaning on the right, cover one side and repeat. That is rote learning. Before a test we cram fifty words in one night — and by Monday they are gone.

The episode sums it up with a simple picture: your brain doesn’t keep what you repeat, it keeps what you use. Words you only repeat are visitors. Words you use in a story, pull out of your memory and meet again later become residents — they move in.

3. The five-finger method: Snap · Story · Say · Test · Meet it again

Yang’s routine has four steps; Echo Stories adds a fifth to fix its weakness. Hold up one hand:

  1. Snap it (thumb). Photograph the vocabulary list in your textbook and upload it to the chatbot. No typing, two seconds.
  2. Story it (pointer). Ask the AI for a short, fun story that uses every word on the list. Twenty separate words become one small world. In the episode, eight words — exhausted, unpredictable, a delay, take something for granted, in the long run, used to, get used to, hit the nail on the head — are woven into a story about a tired freelancer stuck at an airport. You don’t try to remember “exhausted”; you remember the tired man at the airport, and the word comes along for the ride.
  3. Say it (middle). The step most people skip: the AI reads sentences aloud, you write them down, and the AI corrects your mistakes. Reading lets your brain be lazy; writing from sound leaves nowhere to hide.
  4. Test it (ring). Ask the AI for a personalised quiz on the same words: fill in the blank, choose the right word, write your own sentence. The moment you have to stop, search your memory and choose (“get used to wake up or waking up?”) is where the learning happens.
  5. Meet it again (little finger). The upgrade — see section 5.

4. Why it works: four pieces of memory research

The AI is new; the ideas behind it are not.

IdeaResearchKey finding
Stories connect wordsBower & Clark, Stanford, 1969Students who linked ten words in a story recalled about seven times as many words
Effort makes words stickLaufer & Hulstijn, 2001 — the Involvement Load HypothesisNeed, search and evaluation; writing your own sentence is usually strongest (see also the 1978 generation-effect study)
Testing beats re-readingKarpicke & Roediger, Science, 2008After a week, self-testers recalled about 80% of Swahili–English pairs; re-studiers only about a third
Spacing beats crammingEbbinghaus, 1885 — the forgetting curve; Webb; Uchihara, Webb & Yanagisawa, 2019Memory drops fast in the first days; learners often need around ten meetings with a word in context

The two ideas to remember are retrieval practice (pulling information out of your memory instead of reading it again) and spaced repetition (reviewing after a day, three days, a week — right before you forget).

The episode is honest about the limits: nobody has studied Yang’s own results. What exists is strong evidence for each piece — stories, effort, testing and spacing — and his routine bundles all four into something easy enough for a twelve-year-old to do every day.

5. The weakness, and the “Story Matrix” upgrade

One list, one story, then on to the next list: the old words never come back, so the forgetting curve still wins. The fix is the Story Matrix — plan things so each word appears two or three times, in different stories, on different days.

With forty words a month, instead of four stories and done, write about twelve short stories over three weeks, each mixing new words with returning ones. On Monday “exhausted” is the freelancer at the airport; on Thursday it is a marathon runner; next week a nurse after a night shift. Three memories, three paths back to the same word.

6. Copy-ready prompts

Change “B1” to your level (not sure? take the free placement test).

  1. Story: “Here is my vocabulary list. Write a short, interesting story of about two hundred words at B1 level that uses every word on the list. Make the words bold. Then give me a Vietnamese translation underneath.” (Swap Vietnamese for your language.)
  2. Dictation: “Now read five sentences from the story to me one by one. Wait for me to type each sentence. Then correct my spelling and grammar, and explain my mistakes simply.” Voice mode makes this even better.
  3. Quiz: “Make a quiz from these words. Five fill-in-the-blank questions, three multiple-choice questions, and two questions where I must write my own sentence. Don’t show the answers until I finish.” That last line is everything — if you see the answers, there is no retrieval.
  4. Meet it again (three days later): “Here are the words from my last two lists. Write a new story with a completely different setting, using ten of these old words and five new ones. Then quiz me only on the old words.”

AI is a double-edged sword. Three rules: check strange words in a learner’s dictionary (Cambridge, Oxford); ask the AI whether a sentence sounds natural to a native speaker; and take whatever it says with a grain of salt.

7. The “1 to 100” rule

The best line of the episode comes from Yang himself, roughly: AI’s ability is not your ability. You need a “one” first — then AI can amplify it into a hundred. If you are zero, whatever you multiply by, you still get zero.

So the AI writes the story, but you read it aloud. The AI reads the dictation, but you write it. The AI makes the quiz, but you answer without peeking — and then write one sentence about your own life. That is your “one”.

8. A 15-minute daily routine

MinuteTask
1Snap your list, or pick ten words you met this week
2–6Ask for the story and read it aloud once
6–10Dictation
10–14Quiz — no peeking
15Write one sentence about your life with your favourite new word

Every three or four days, add a “meet it again” story with old words. Keep it up for a month and each word will have met you several times, in several stories.

Key vocabulary from the episode

Word / phraseMeaningExample
to go viralto spread very quickly onlineHis vocabulary video went viral and reached twenty million views.
a prodigya young person with unusual talentPeople online call him a prodigy, but he says he simply practices every day.
to downplayto make something seem less importantHis parents downplay the “future Elon Musk” label.
rote learninglearning by repetition without real understandingRote learning helped me pass the test, but I forgot the words a week later.
to cramto study a lot just before an examI crammed fifty words the night before the test.
to weave something into somethingto include something smoothly in a bigger wholeThe AI wove all eight words into one short story.
retrieval practicelearning by recalling from memoryRetrieval practice works better than reading your notes again.
the forgetting curvehow fast we forget without reviewWithout review, the forgetting curve drops fast in the first few days.
spaced repetitionreviewing at growing intervalsSpaced repetition helps new words stay in your long-term memory.
a double-edged swordsomething with both good and bad effectsAI is a double-edged sword: it saves time, but it can make mistakes.
to take something with a grain of saltnot to believe something completelyTake the AI’s grammar notes with a grain of salt.
to amplifyto make bigger or strongerAI can amplify your effort, but it cannot replace it.

Keep learning with Echo Stories

This article is educational commentary and analysis based on an Echo Stories English video; information about Yang Xizhe comes from the public sources referenced in the episode.