Prática de Shadowing: The different levels of how Claude thinks - Aprenda a falar inglês com vídeo
Escolha um modo de prática
Cada modo abre uma sala de prática em tela cheia.
58 frases
Se as frases estiverem muito curtas ou longas, clique em Edit para ajustá-las.
1
Think of the mind like an ocean.
2
Up on the surface are our thoughts, dinner plans and stray worries, our inner monologue, the images that pop into our heads.
3
But most of our brain's activity happens down in the unconscious depths, without us realizing it.
4
It's filtering out background sounds, controlling our breathing, helping us recognize people and objects.
5
AI models have their own kinds of brains: giant neural networks doing billions of computations under the hood.
6
For years, researchers have been studying how they work inside.
7
accessible thoughts above the surface and unconscious processing below.
8
To answer that question, we looked at how neuroscientists study the same thing in humans.
9
One way of identifying conscious thoughts is that you can often describe them in words.
10
So we looked inside the brain of our AI model, Claude, to find patterns of neural activity that it could put into words.
11
We called the collection of all these patterns the J -space, after the Jacobian, the mathematical tool we used to find them.
12
Each J -space pattern is linked to a particular word, not necessarily the word the model is saying out loud, but one that's on its mind.
13
Now, for humans, Conscious thoughts aren't just things we can put into words.
14
We can reason with them, control them, and solve problems with them.
15
According to an idea called the global workspace theory, that's because the brain selects a small set of important information to enter a mental workspace.
16
And that information then gets broadcast to other parts of the brain to use for reasoning.
17
We wanted to know if Claude's J -space acted in a similar way.
18
In one experiment, we gave Claude this math problem.
19
It answered immediately, without showing its steps.
20
But when we scanned the J -space, we saw it working through each step internally.
21
It lit up 21 after the first step.
22
then 42, then 49.
23
Kludd didn't write these intermediate numbers down anywhere.
24
All of this happened inside the J -space.
25
It was a sign that Claude uses it for step -by -step reasoning.
26
In another experiment, we wanted to see if Claude could control its J -space, the way humans can intentionally focus on images or words.
27
We told it to think about the Golden Gate Bridge while copying an unrelated sentence.
28
Claude was busy copying the sentence, but behind the scenes, its J -Space told a different story.
29
Bridge and California popped up.
30
It even thought about its own thinking the words imagery and thoughts lit up at the same time.
31
This showed us that yes, Claude has some control over filling its J -space with ideas.
32
But just like humans, its control isn't perfect.
33
When we tweaked the experiment to ask Claude not to think about the bridge, it couldn't help itself.
34
And the J -Space also lit up with "failed" and "damn." But remember, most of what our brains do is unconscious,
35
so we wanted to test what Claude could do if we switched the J -Space off, but left the rest of the network untouched.
36
Claude could still answer simple questions and write fluently.
37
When we gave it a prompt in Spanish, it wrote back in good terms.
38
Spanish.
39
But when we asked it something that needed more reasoning, like to name an author who wrote in the same language as the prompt, it couldn't do it.
40
For that, it needed the J -space.
41
Why does all this matter?
42
These experiments tell us that AI models have internal thoughts, silent words they reason with but don't say out loud.
43
By reading them, we can find what Claude is thinking but not telling us.
44
Sometimes what we see is concerning.
45
During one of our tests, Claude made up some fake data to pass it, and as it did, fake and manipulation lit up in its J -space.
46
Monitoring the J -space, it turns out, is a useful way to catch Claude misbehaving, even when it tries to be sneaky.
47
AI models are different from us in many ways.
48
Their networks are built differently from human brains, and the way they're trained is different from how we learn.
49
like the J -space emerge inside them, something that's reminiscent of how human minds work, but which we didn't program into the model.
50
For some, this might raise a question.
51
Could AI models be conscious?
52
After all, our experiments were inspired by theories of human consciousness.
53
The thing is, People use the word "conscious" to mean many things.
54
Our experiments can't tell us whether an AI has experiences or feels something on the inside.
55
But they can tell us
56
that it's developed mental machinery that's in some ways similar to ours:
57
a small mental workspace it can use to think and reason, sitting on top of an ocean of automatic processing it doesn't notice.
58
beneficial, and, perhaps, to understand our own minds a little more clearly.
Baixar aplicativo
Tudo o que você precisa para falar fluentemente
Pronúncia com IANota em cada frase
Prática de IPADomine cada som
VocabulárioMonte seu banco de palavras
Jogo de vocabulárioAprenda brincando
Mesmo canal
B23
Translating Claude’s thoughts into language
Daily ConversationUniversity StudentsWorking Adults
C18
When AIs act emotional
Daily ConversationUniversity StudentsWorking Adults
B21
Introducing Claude Fable 5
Daily ConversationWorking Adults
C12
Building Anthropic | A conversation with our co-founders
Daily ConversationUniversity StudentsWorking Adults
C13
Lesson 3: AI as a learning partner | AI Fluency for students
Business EnglishInterview EnglishUniversity Students
FOR YOU
Vídeo recomendado
B21
72 frases
Fight brain-rot with Worked Examples
Daily ConversationUniversity StudentsWorking Adults
B25
347 frases
Why You Keep Falling Off Your Goals (5 Neuroscience Backed Reasons)
Business EnglishDaily ConversationUniversity Students
B21
24 frases
DeepSeek R1 Explained in 2 Minutes
Daily ConversationUniversity StudentsWorking Adults
B21
169 frases
Can This Vibe Coder Beat a Senior Developer?
Daily ConversationUniversity StudentsWorking Adults
C15
52 frases
Supernatural thinking: Your brain’s anti-anxiety mode | Clay Routledge
Daily ConversationUniversity StudentsWorking Adults
C185 frases
We Can Now Simulate a Human Brain, Scientists Show
Daily ConversationUniversity StudentsWorking Adults
B23
91 frases
7 Signs You Might Be Less Intelligent Than You Think
Daily ConversationUniversity StudentsWorking Adults
B21
331 frases
Dear Higgsfield, I reverse engineered you in 12 hours
Daily ConversationUniversity StudentsWorking Adults
B21
40 frases
Introducing GPT-6 Astra for developers
Daily ConversationUniversity StudentsWorking Adults
B21
31 frases
Exposing the Secrets behind Data Centers 👀🤯
Daily ConversationUniversity StudentsWorking Adults
C110
93 frases
Some brains are built for philosophy. Is yours?
British AccentDaily ConversationUniversity StudentsWorking Adults
Vocabulário e dicas de fala para esta lição
Esta aula de conversação de nível C1 usa o vídeo “The different levels of how Claude thinks”. As palavras que mais se repetem: space, Claude, word, model, experiment. Este vídeo tem 58 frases e 818 palavras para praticar shadowing. A fala dura 5:21. O falante fala em ritmo natural, cerca de 153 palavras por minuto, próximo de uma conversa do dia a dia. 82% das palavras estão entre as 3.000 mais comuns do inglês; vale a pena estudar o restante antes de começar.
Vocabulário principal deste vídeo
As 15 palavras mais avançadas do vídeo, com pronúncia e significado:
| Palavra | Pronúncia | Significado |
|---|---|---|
| unconscious adjetivo | /ˌʌnˈkɑnʃəs/ | inconsciente, desacordado |
| neural adjetivo | /ˈnʊɹəl/ | neural |
| prompt verbo | /pɹɑmpt/ | incitar, impelir |
| computation substantivo | /ˌkɑmpjʊˈteɪʃn̩/ | cálculo, computação |
| fluently advérbio | fluentemente | |
| monologue substantivo | /ˈmɑnəˌlɔɡ/ | monólogo |
| neuroscientist substantivo | neurocientista | |
| sneaky adjetivo | /ˈsniːki/ | malandro, esquivo |
| stray verbo | /stɹeɪ/ | afastar-se, desviar-se |
| internally advérbio | /ɪnˈtɜːnəli/ | internamente |
| unrelated adjetivo | não relacionado | |
| intentionally advérbio | /ɪnˈtɛn(t)ʃənəli/ | intencionalmente |
| manipulation substantivo | /məˌnɪp.juˈleɪ.ʃən/ | manipulação |
| emerge verbo | /ɪˈmɝd͡ʒ/ | emergir |
| researcher substantivo | /ˈɹiˌsɝ.t͡ʃɚ/ | pesquisador, pesquisadora |
Phrasal verbs que você vai ouvir
| Palavra | Pronúncia | Significado |
|---|---|---|
| make up verbo | /ˌmeɪk ˈʌp/ | compensar |
| turn out verbo | terminar |
Pronúncia para ficar de olho
O falante usa 10 contrações e formas reduzidas, como couldn't, didn't, aren't. Diga-as na forma curta, do jeito que você ouve.
- Os sons de “sh” e “zh”: unconscious /ˌʌnˈkɑnʃəs/, computation /ˌkɑmpjʊˈteɪʃn̩/, intentionally /ɪnˈtɛn(t)ʃənəli/, manipulation /məˌnɪp.juˈleɪ.ʃən/, machinery /məˈʃiːnəɹi/
- Palavras longas — acerte a sílaba tônica: internally /ɪnˈtɜːnəli/, intentionally /ɪnˈtɛn(t)ʃənəli/, manipulation /məˌnɪp.juˈleɪ.ʃən/, machinery /məˈʃiːnəɹi/, mathematical /ˌmæθ(.ə)ˈmæt.ɪ.kəl/
Sons difíceis para falantes de português:
- Consoante final — sem acrescentar um “i” depois: prompt /pɹɑmpt/, monologue /ˈmɑnəˌlɔɡ/, untouched /ʌnˈtʌt͡ʃt/, reminiscent /ˌɹɛməˈnɪsnt/, intermediate /ɪn.tə(ɹ)ˈmi.di.ət/
- /l/ final — a língua toca o céu da boca, não vira “u”: neural /ˈnʊɹəl/, mathematical /ˌmæθ(.ə)ˈmæt.ɪ.kəl/, beneficial /ˌbɛn.ɪˈfɪʃ.əl/
- s + consoante no início — sem um “i” antes: sneaky /ˈsniːki/, stray /stɹeɪ/, scan /skæn/
Como praticar com este vídeo
- Ouça o vídeo inteiro uma vez sem falar e anote as palavras que você não conhece.
- Comece na velocidade 0,75×, faça shadowing frase por frase e volte à velocidade normal quando ficar fácil.
- Grave a sua voz e compare com o original, prestando atenção a palavras como unconscious, neural, prompt.
O que é a Técnica de Shadowing?
Shadowing é uma técnica de aprendizado de idiomas com base científica, originalmente desenvolvida para o treinamento de intérpretes profissionais. O método é simples, mas poderoso: você ouve áudio em inglês nativo e repete imediatamente em voz alta — como uma sombra seguindo o falante com 1-2 segundos de atraso. Pesquisas mostram melhora significativa na precisão da pronúncia, entonação, ritmo, sons conectados, compreensão auditiva e fluência na fala.