Практика Shadowing: Did Google just kickstart the intelligence explosion? - Изучайте разговорный английский по видео
Создание урока...
1
In 1965, a straight British mathematician named I.J.
2
Goode, who had spent World War II fighting Nazis next to a gay Alan Turing in Bleschley Park, wrote that the first ultra-intelligent machine would be the last invention man ever needed to make,
3
because once an AI gets good enough to improve itself, every improvement makes it better at improving, which makes it better at improving, which, you get the idea.
4
This is called RSI, and it's been the wet dream of AI researchers ever since.
5
Well, last week, 33 researchers from ByteDance, Tsinghua, and a few other Chinese labs published a paper called The Last AI Built by Humans.
6
It laid out a five-stage roadmap for RSI, where in the final stage, the AI rewrites the process it used to improve itself, and we all get reassigned to the countryside for agricultural work.
7
Then on Sunday, Google DeepMind
8
and the University of Maryland responded by dropping a similar paper of their own called Wet Dream RSI.
9
In it, they claim that by turning an AI's old discovery logs into a simulator
10
and letting it dream up thousands of new search strategies inside of it, the AI got better at discovering things without anyone ever touching the model itself.
11
In today's video, we'll break down how Dream RSI works under the hood, and decide whether an agent being able to rewrite its own exploration policy is actually RSI,
12
or just more hype slop.
13
It is September 17th, 2026, and you're watching the Code Report.
14
Every time ai has made a mathematics breakthrough the process has been the same one
15
that alpha evolved popularized last year you take a coding agent handed a problem
16
and a scoring function then run a loop where it proposes a solution evaluates it reads the feedback
17
and tries again a few thousand times this process is what discovered the jacobian conjecture this summer
18
and what openai used to front run the Navier-Stokes problem earlier this month.
19
But the most interesting part of the process is one no one really talks about called exploration policy.
20
The idea is that at every step of the loop, there's a decision to make about what the agent tries next.
21
For example, say the loop is optimizing for a matching algorithm for horses.
22
If one attempt pairs up horses slightly better than others, does the agent keep building on it, or does it start over with something that could be better?
23
And if an attempt crashes before a single horse gets matched, is it the whole idea that's bad, or just the implementation.
24
Until this week, this exploration policy was hard coded into the loop by whoever set it up.
25
But what the DeepMind team figured out is that if you save everything from every attempt, like the code it wrote, the score it got, and whether or not it crashed,
26
you don't actually need to touch the model again to test a new policy.
27
You can just show the new policy the old runs
28
that are cached on the disk and let it decide where to go from there.
29
And because that costs nothing, the agent can test thousands of different policies against the same run
30
and then whichever one would have reached the best result in the fewest attempts.
31
Then it deploys that policy on the next run, saves that run too, and repeats the process again.
32
The paper calls this process Dreaming, and to test it, they pointed Gemini at eight different problems across algorithm design and mathematics,
33
then ran the same setup with a fixed policy to see if the Dreaming version could beat it.
34
I won't bore you with the TMBBs, but the most impressive one was that it wrote a lasso solver
35
that beats Python's standard machine learning library in about 300 tries, where the static policy needed 550, and the previous record holder needed about 51,000.
36
And because we're living in hell, the most interesting part was the prompt.
37
It basically begs the agent to read every past attempt before writing any code, to stop making tiny tweaks to the same idea over and over, and to pinky promise not to kill any processes.
38
So is any of this actually RSI?
39
By IJ Good's definition, no, because the whole point is that the thing doing the improving gets smarter each round.
40
In this case, the model that writes each new exploration policy is still the same Gemini, so it can never find a solution that it wasn't already capable of writing.
41
It just finds them faster and with fewer wasted attempts.
42
But with that said, that's also true of every AI math breakthrough we've had this year.
43
The Jacobian conjecture, Navier-Stokes, and progress on the Ryman hypothesis all came from static models wrapped in a custom harness,
44
with sub-agents, swarms, and orchestration doing most of the heavy lifting, and the weights only get better when a human goes back and trains the next model on what the swarm found.
45
So if you gave IJ mushrooms and convinced him that a human in the loop still counts, he might agree that this is the last invention man ever needed to make, but when he came down,
46
he'd probably argue that this is just a cool search algorithm with some caching, And that's why you need to check out Blacksmith, the sponsor of today's video.
47
It's a drop-in replacement for GitHub Runners that lets you run your GitHub actions twice as fast while costing 75% less.
48
And they just launched Codesmith, a cloud coding agent that knows your repos and CI runs, so you can ask it to build something from GitHub, the web, or Slack like I'm doing here.
49
I'm asking it to add a new AI provider to my app
50
and wire up the API key in my infrastructure repo and it can open a PR in each one,
51
then fix failing tests or address review comments without relying on messages between bots.
52
You can also ask it to recommend better runner sizes for your CI history and turn those changes into a pull request, so you're not renting a supercomputer to check your semicolons.
53
Try it out for free and get 3,000 GitHub Actions minutes at the link below.
54
This has been The Code Report, thanks for watching, and I will see you in the next one.
📺 Тот же канал
B21
10 very promising Open Source Projects you haven’t heard of
American AccentBusiness EnglishUniversity StudentsWorking Adults
B21
7 Programming myths that waste your time
American AccentBusiness EnglishUniversity StudentsWorking Adults
C11
DHH has gone completely off the rails...
American AccentBusiness EnglishUniversity StudentsWorking Adults
C12
DNS Explained in 100 Seconds
American AccentBusiness EnglishUniversity StudentsWorking Adults
B23
Anthropic begged the world to stop AI… then shipped this
American AccentBusiness EnglishUniversity StudentsWorking Adults
B22
One man just liberated Fable... and now it’s illegal
American AccentBusiness EnglishUniversity StudentsWorking Adults
B22
7 Things No Programmer Ever Wants to Hear
American AccentBusiness EnglishWorking Adults
B21
80% of programmers are NOT happy… why?
American AccentBusiness EnglishUniversity StudentsWorking Adults
C14
The most interesting hack in history just got weirder...
American AccentBusiness EnglishUniversity StudentsWorking Adults
B24
Raspberry Pi Explained in 100 Seconds
American AccentBusiness EnglishUniversity StudentsWorking Adults
C13
Tragic mistake... Anthropic leaks Claude’s source code
American AccentBusiness EnglishUniversity StudentsWorking Adults
C1100+ Docker Concepts you Need to Know
American AccentBusiness EnglishUniversity StudentsWorking Adults
C17
When open-sourcing your code goes wrong...
American AccentBusiness EnglishUniversity StudentsWorking Adults
C129
Docker in 100 Seconds
American AccentBusiness EnglishUniversity StudentsWorking Adults
C13
The mystery is solved... and the answer is 40x cheaper than Claude
American AccentBusiness EnglishUniversity StudentsWorking Adults
B28
An ex-OpenAI researcher just deleted language from the LLM...
American AccentBusiness EnglishUniversity StudentsWorking Adults
B29
25 crazy software bugs explained
American AccentBusiness EnglishUniversity StudentsWorking Adults
B29
Git Explained in 100 Seconds
American AccentBusiness EnglishUniversity StudentsWorking Adults
B25
There's a new Linux distro in town for developers...
American AccentBusiness EnglishUniversity StudentsWorking Adults
B25
The "vibe coding" mind virus explained…
American AccentBusiness EnglishUniversity StudentsWorking Adults
✨ Рекомендуемое видео
C12
154 предложений
歐巴馬談 AI：喊放慢的公司「是真的在擔心自己的產品」【AI 大人說】
Business EnglishDaily ConversationUniversity Students
C110
184 предложений
A.I. ‐ Humanity's Final Invention?
Business EnglishDaily ConversationUniversity Students
C15
45 предложений
The plan to break apart Google... RIP Chrome
American AccentBusiness EnglishDaily ConversationUniversity Students
C11
28 предложений
Gemma 4 Just Released - Whats In? In 2026!
Business EnglishUniversity StudentsWorking Adults
C12
260 предложений
AI Just Crossed the Terrifying Line - Now What?
Business EnglishDaily ConversationUniversity Students
C11
12 предложений
Rapid Engineering With AI
Business EnglishUniversity StudentsWorking Adults
C11
105 предложений
Google’s AI Course for Beginners (in 10 minutes)!
American AccentBusiness EnglishUniversity StudentsWorking Adults
C11
254 предложений
AI Will Create New Wealth, But Not Where You Think | Carnegie Mellon University, Po-Shen Loh
Business EnglishDaily ConversationUniversity Students
C14
76 предложений
The Limitless AI Lie. The Bubble Is Slowly BURSTING.
Business EnglishDaily ConversationUniversity Students
C1762 предложений
Building Anthropic | A conversation with our co-founders
Business EnglishUniversity StudentsWorking Adults
C12
182 предложений
We Let $200 GPT-Pro Think for 7 Days on Unsolved Math
Business EnglishUniversity StudentsWorking Adults
Лексика и советы по произношению к этому уроку
В этом видео 54 предложений и 1063 слов для шедоуинга. Речь длится 4:56. Говорящий говорит быстро, около 215 слов в минуту, поэтому звуки часто сливаются и сокращаются. Только 79% слов входят в 3000 самых частых слов английского языка, поэтому лексика сложная.
Ключевая лексика этого видео
15 менее частых слов из видео с произношением и значением:
- loop /luːp/ (существительное) — пе́тля, петля́. A length of thread, line or rope that is doubled over to make an opening.
- exploration /ˌɛkspləˈɹeɪʃən/ (существительное) — иссле́дование, изуче́ние. The process of exploring.
- mathematics /mæθ(.ə)ˈmæt.ɪks/ (существительное) — матема́тика. An abstract representational system studying numbers, shapes, structures, quantitative change and relationships between them.
- algorithm /ˈælɡəɹɪðm̩/ (существительное) — алгори́тм. A collection of ordered steps that solve a mathematical problem. A precise step-by-step plan for a computational procedure that possibly begins with an input…
- discover /dɪˈskʌvə/ (глагол) — открыва́ть, откры́ть. To find or learn something for the first time.
- static /ˈstæt.ɪk/ (прилагательное) — стати́чный. Unchanging; that cannot or does not change.
- invention /ɪnˈvɛnʃən/ (существительное) — изобрете́ние. Something invented.
- researcher /ˈɹiˌsɝ.t͡ʃɚ/ (существительное) — иссле́дователь, иссле́довательница. One who researches.
- breakthrough /ˈbɹeɪk.θɹuː/ (существительное) — проры́в. An advance through and past enemy lines.
- cache /kæʃ/ (существительное) — запа́с, тайни́к. Such a store of physical supplies, placed by humans or other animals for practical reasons.
- impressive /ɪmˈpɹɛsɪv/ (прилагательное) — впечатля́ющий, внуши́тельный. Making, or tending to make, a positive impression; having power to impress.
- argue /ˈɑɹ.ɡju/ (глагол) — спо́рить, обсужда́ть. To show grounds for concluding (that); to indicate, imply.
- cloud /ˈklaʊ̯d/ (существительное) — о́блако, ту́ча. A visible mass of water droplets suspended in the air.
- infrastructure /ˈɪnfɹəˌstɹʌkt͡ʃə/ (существительное) — инфраструкту́ра. An underlying base or foundation for a building, organization, or system.
- discovery /dɪˈskʌv.ɹi/ (существительное) — откры́тие. Something discovered.
Фразовые глаголы, которые вы услышите
- break down (глагол) — ломаться, сломаться. To stop functioning.
- check out (глагол) — проверя́ть, прове́рить. To record the departure or withdrawal of someone or something (such as guests, employees, books, etc.).
- come down (глагол) — спуска́ться, спусти́ться. To descend, fall down, collapse.
- figure out (глагол) — разбира́ться, разобра́ться. To come to understand; to discover or find a solution; to deduce.
- go back (глагол) — возвраща́ться, возврати́ться. To return to a place or state after having been there at a previous time.
- talk about (глагол) — ну и, вот э́то. Used to draw attention to the speaker's characterization of someone or something.
Произношение, на которое стоит обратить внимание
Говорящий использует 11 сокращённых и редуцированных форм, например I'm, you're, don't. Произносите их коротко, так, как слышите.
- Звуки «th»: mathematics /mæθ(.ə)ˈmæt.ɪks/, algorithm /ˈælɡəɹɪðm̩/, breakthrough /ˈbɹeɪk.θɹuː/, hypothesis /haɪˈpɒθɪsɪs/, mathematician /ˌmæθ.(ə.)məˈtɪʃ.ən/
- Звуки «sh» и «zh»: exploration /ˌɛkspləˈɹeɪʃən/, invention /ɪnˈvɛnʃən/, cache /kæʃ/, implementation /ˌɪmplɪmənˈteɪʃən/, mushroom /ˈmʌʃˌɹuːm/
- Длинные слова — следите за ударением: exploration /ˌɛkspləˈɹeɪʃən/, mathematics /mæθ(.ə)ˈmæt.ɪks/, infrastructure /ˈɪnfɹəˌstɹʌkt͡ʃə/, implementation /ˌɪmplɪmənˈteɪʃən/, intelligent /ɪnˈtɛlɪd͡ʒənt/
Как заниматься с этим видео
- Прослушайте всё видео один раз молча и выпишите незнакомые слова.
- Начните со скорости 0,75×, повторяйте предложение за предложением и вернитесь к обычной скорости, когда станет легко.
- Запишите себя и сравните с оригиналом, обращая внимание на такие слова, как loop, exploration, mathematics.
Грамматика в этом видео
Конструкции, которые говорящий использует чаще всего, с точными словами из видео:
| Конструкция | В видео |
|---|---|
| «Used to» used to + глагол — прошлая привычка или состояние, которых больше нет | used to improve · used to front |
| Present Perfect have/has + причастие прошедшего времени — прошлое действие, важное сейчас | has made · has been |
| Пассивный залог be + причастие прошедшего времени — важно, что происходит, а не кто это делает | is called · are cached |
Что такое техника Shadowing?
Shadowing — это научно обоснованная техника изучения языка, изначально разработанная для подготовки профессиональных переводчиков и популяризированная полиглотом доктором Александром Аргуэльесом. Метод прост, но эффективен: вы слушаете аудио на английском от носителей языка и немедленно повторяете вслух — как тень, следующая за говорящим с задержкой в 1–2 секунды. В отличие от пассивного прослушивания или грамматических упражнений, Shadowing заставляет мозг и мышцы рта одновременно обрабатывать и воспроизводить реальные речевые паттерны. Исследования показывают, что это значительно улучшает точность произношения, интонацию, ритм, связную речь, понимание на слух и беглость речи — что делает его одним из самых эффективных методов для подготовки к IELTS Speaking и реального общения на английском.