シャドーイング練習: Did Anthropic Kill Git? - 動画で英語スピーキングを学ぶ

練習モードを選ぶ

各モードは全画面の練習ルームで開きます。

65 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
Did Anthropic kill Git?
2
This week, a lot of people on X decided it did.
3
One post with over 600 ,000 views says Anthropic has replaced Git internally and the titans are falling.
4
Another says this is much bigger than it looks.
5
It all traces back to one post from Anthropic's own account, seen more than 6 .5 million times.
6
Part of the story is true.
7
Anthropic did stop using Git, but only for its own code, inside its own company.
8
Git itself is fine.
9
Anthropic keeps almost all of its code in one giant repository, a monorepo.
10
In May, about 2 ,500 changes were merged into it every day.
11
By September, it was over 30 ,000 a day.
12
On September 9th alone, 30 ,401 changes went in.
13
Their chart even has a dip in late July labeled the quiet week, and right after it, the line goes almost straight up.
14
Claude writes a lot of that code, and Anthropic says every bottleneck they remove leads to even more of it.
15
At that rate, the repository got heavy.
16
The history of the main branch alone is almost 100 gigabytes.
17
Their CI system checks out the code about 25 million times a day before you count developers, agent sandboxes, and research jobs.
18
Anthropic did the maths.
19
If every one of those checkouts fetched the repo the normal way, that would be over 600 petabytes of Git traffic every day.
20
Git was designed for a different kind of user.
21
Every clone is a full copy of the repository with all of its history, and it can do almost everything offline.
22
That's great for a person on a laptop on a plane.
23
At Anthropic, almost every client is an agent or a CI job, and those are always online.
24
They tried the usual fixes first, like shallow clones and sparse checkouts.
25
At the peak, they were running several hundred internal Git mirrors just to keep up.
26
So they built their own.
27
It's called Scion, and it moves the work from your machine to the server.
28
Commits, merges, rebases, blame, and history all happen on the server so there's nothing to push.
29
The client keeps a small cache and only downloads the files it actually reads.
30
A CI job that needs a diff can ask the server for it and never check out the code at all.
31
Between the client and the server sit two layers of caches, one on every machine and one in every region.
32
And each layer answers over 99 % of requests by itself.
33
Anthropic says it borrowed ideas from Google's Piper and Meta's Sapling, which were built for huge human teams.
34
On a developer machine, Git status took almost four seconds, and Scion does it in 19 milliseconds.
35
For agents, setting up a fresh work tree went from seven minutes to five seconds.
36
The whole system uses about 200 times less CPU and memory than the old mirrors, and CI got a quarter of its CPU back.
37
These are Anthropic's own measurements, and nobody outside the company has checked them yet seven people built it.
38
In five months, they made Scion, a code review tool called Antique, and a few other systems,
39
and they moved the whole monorepo over while the merge rate went up 10 times.
40
Claude wrote a lot of that code too.
41
The team had handwritten a pipeline to import the old Git history,
42
and an early version of Claude Mythos rewrote it to use 20 times less memory and run four times faster.
43
Every engineer on the project now rates Claude as better at writing code than they are.
44
They also say a big part of their job became telling Claude no
45
when it took shortcuts or piled on checks nobody needed.
46
On Hacker News, the thread passed 800 points and 900 comments in a day.
47
The most replied comment says that every time the writer stops reviewing Claude's code, they later find janky crap.
48
Others said a central server you always have to reach sounds a lot like subversion or perforce.
49
The older systems get replaced.
50
One person asked why Anthropic didn't just split up the monorepo.
51
On X, Dex Horthy pointed at the trade -off most people will feel.
52
Today, if GitHub is down, you can still commit on your laptop.
53
With Scion, if the server is down, you can't.
54
So did Anthropic kill Git?
55
No. In their own words, Scion is by no means a suitable replacement for Git for general use.
56
They say it only makes sense for truly enormous repos, with tens of thousands of engineers or swarms of agents.
57
It isn't open source and they don't plan to release it.
58
Your repos on GitHub don't change, and Claude code still works with plain Git.
59
A small team has already posted an unofficial open source version, but Anthropic has nothing to do with it.
60
What Anthropic wants is for Git to pick up some of these ideas,
61
like servers that do more of the work and answer questions about history through an API.
62
They say they want to build that with GitHub, GitLab, and the open source community,
63
and they expect other companies using AI coding tools to hit the same wall within a year.
64
If that happens, the changes will most likely land on the server side,
65
and the Git on your laptop will keep working the way it does today.

アプリをダウンロード

流暢に話すために必要なすべて

AI発音採点文ごとに採点
IPA練習すべての音をマスター
単語学習自分の単語帳を作る
単語ゲーム遊びながら覚える

このレッスンの語彙とスピーキングのポイント

このスピーキングレッスンは、動画「Did Anthropic Kill Git?」を教材にしています。 繰り返し出てくる語は次のとおりです:Git, Anthropic, code, server, Claude。 この動画には、シャドーイング用の文が65文、単語が881語あります。 音声の長さは4:49です。 話す速さは速く、1分あたり約183語です。音のつながりや弱く発音される音が多くなります。 英語の頻出3,000語に含まれる単語は79%だけなので、語彙は難しめです。

この動画の重要語彙

動画に出てくる覚えておきたい単語15語を、発音と意味つきで紹介します。

単語発音意味
server 名詞/ˈsɝvɚ/サーバー
laptop 名詞/ˈlæp.tɑp/ノートパソコン, ラップトップ
merge 動詞/mɝd͡ʒ/合わせる, 合併する
mirror 名詞/ˈmiɹɚ/鏡, 姿見
layer 名詞/ˈleɪ̯ɚ/層, 膜
commit 動詞/kəˈmɪt/預ける, 任せる
clone 名詞/kloʊn/クローン
checkout 名詞/ˈt͡ʃɛkˌaʊt/レジ
cache 名詞/kæʃ/キャッシュ
chart 名詞/t͡ʃɑɹt/チャート, 図表
replacement 名詞/ɹɪˈpleɪsmənt/代替
thread 名詞/ˈθɹɛd/糸
label 名詞/ˈleɪ.bəl/荷札, 貼り札
plain 名詞/pleɪn/平地, 平野
suitable 形容詞[ˈsjuː.tə.bɫ̩]相応しい, 相応な

動画に出てくる句動詞

単語意味
check out 動詞試験する, 検査する
go up 動詞上がる, 登る
pick up 動詞拾う, 拾い上げる

この動画の文法

話し手がよく使っている文型を、動画の実際の表現とともに紹介します。

文型動画での表現
現在完了形 have/has + 過去分詞 — 過去の出来事が今も関係しているhas replaced · has checked · has already posted
受動態 be + 過去分詞 — 誰がするかより、何が起きるかに焦点を当てるwere merged · was designed · were built

注意したい発音

話し手はdon't, can't, didn'tなど、短縮形や弱形を5回使っています。聞こえたとおりの短い形で発音しましょう。

  • 「th」の音: thread /ˈθɹɛd/, mathematics /mæθ(.ə)ˈmæt.ɪks/
  • 「sh」と「zh」の音: cache /kæʃ/, measurement /ˈmɛʒ.ə.mənt/, shallow /ˈʃæloʊ/, unofficial /ʌn.əˈfɪʃ.əl/, shortcut /ˈʃɔːtkʌt/
  • 長い単語(アクセントの位置に注意): developer /dɪˈvɛləpɚ/, mathematics /mæθ(.ə)ˈmæt.ɪks/, internally /ɪnˈtɜːnəli/, unofficial /ʌn.əˈfɪʃ.əl/, repository /ɹɪˈpɒz.ɪ.t(ə.)ɹɪ/

日本語話者が苦手な音:

  • /v/ — /b/ にならないように、上の歯を下唇に当てる: server /ˈsɝvɚ/, developer /dɪˈvɛləpɚ/, subversion /sʌ̈bˈvɝ.ʒən/
  • /f/ — 「フ」ではなく、上の歯と下唇で出す: offline /ɔfˈlaɪn/, unofficial /ʌn.əˈfɪʃ.əl/, fetch /fɛt͡ʃ/, perforce /pɚˈfɔɹs/
  • 子音の連続 — 間に母音を入れない: clone /kloʊn/, chart /t͡ʃɑɹt/, replacement /ɹɪˈpleɪsmənt/, thread /ˈθɹɛd/, plain /pleɪn/

この動画での練習方法

  1. まず声を出さずに動画を最後まで聞き、知らない単語をメモします。
  2. まず0.75倍速で一文ずつシャドーイングし、慣れてきたら通常の速度に戻します。
  3. 自分の声を録音して元の音声と比べます。server, laptop, mergeなどの単語に特に注意しましょう。

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。

シャドーイングのやり方: ステップ別の完全ガイドを読む →