跳到正文
hackernews100(youtube.com)· sandslash·· 3 天前品質評分64

Y Combinator 邀 Jacob Coxon 和 Robin Hanson 辯論政府該不該放慢 AI 發展

Should governments slow frontier AI? Jacob Coxon and Robin Hanson debate [video]

重點導讀

AI 發展可能帶來哪些風險,政府該不該出手放慢腳步?

  • Y Combinator 邀請 Jacob Coxon 和 Robin Hanson 辯論。Coxon 提出超級智慧可能很快出現,也談到模型行為難以理解和風險難以補償。
  • Hanson 則質疑政府管制能否奏效,主張考慮以責任追究等方式處理。兩人也討論 AI 對工作、監管機構和安全的影響,並在最後回應觀眾提問。

AI 伴隨問答

自備 Key 深度分析

針對這篇報導有疑問?可直接點選快捷問題或向 AI 追問背景脈絡與技術細節:

正文

Jacob Coxon vs. Robin Hanson: Should We Slow Down AI? - YouTube

Back [Image 1](https://www.youtube.com/ "YouTube Home") KR

Skip navigation

Search

Search with your voice

[](https://www.youtube.com/watch?v=lwfewYL7X0o)

Sign in

[Image 2](https://www.youtube.com/ "YouTube Home") KR

[](https://www.youtube.com/watch?v=lwfewYL7X0o)

[](https://www.youtube.com/watch?v=lwfewYL7X0o)

Video 2

[](https://www.youtube.com/watch?v=lwfewYL7X0o)

[](https://www.youtube.com/watch?v=lwfewYL7X0o)

Jacob Coxon vs. Robin Hanson: Should We Slow Down AI?

[](https://www.youtube.com/watch?v=lwfewYL7X0o)

Tap to unmute

2x

Image 3

Jacob Coxon vs. Robin Hanson: Should We Slow Down AI?

Y Combinator 24,387 views 17h ago

[](https://www.youtube.com/watch?v=lwfewYL7X0o)

Copy link

Info

Shopping

Image 4

Image 5

If playback doesn't begin shortly, try restarting your device.

•

You're signed out

Videos you watch may be added to the TV's watch history and influence TV recommendations. To avoid this, cancel and sign in to YouTube on your computer.

Cancel Confirm

[](https://www.youtube.com/watch?v=lwfewYL7X0o)

Share

[](https://www.youtube.com/watch?v=lwfewYL7X0o "Share link")- [x] Include playlist

An error occurred while retrieving sharing information. Please try again later.

Image 6

0:00

[](https://www.youtube.com/watch?v=lwfewYL7X0o)[](https://www.youtube.com/watch?v=96jN2OCOfLs)

0:00 / 0:00

Live

•Watch full video

•

YC Debate Topic: Should We Slow Down AI?

•

46:58 Gemini and the path to AGI with Koray Kavukcuoglu Google DeepMind 42K • 12h ago Live Playlist ()Mix (50+)38:49 Nick Bostrom on Superintelligence, Longer Lives, and Deep Utopia a16z 12K • 17h ago Live Playlist ()Mix (50+)45:28 Can Canada's Francophone Network Unlock New Markets? | Toronto Global Forum 2026 IEFA TV 9 • 8h ago Live Playlist ()Mix (50+)27:07 Geoffrey Hinton | Full Episode 11.13.26 | Firing Line with Margaret Hoover | PBS Firing Line with Margaret Hoover | PBS 265K • 6d ago Live Playlist ()Mix (50+)39:30 What OpenAI Did After the Hugging Face Hack The New York Times 109K • Streamed 1d ago Live Playlist ()Mix (50+)35:39 Can Zohran Mamdani Fix the Cost of Living in New York City?The New York Times 317K • Streamed 1d ago Live Playlist ()Mix (50+)13:46 java situation is insane Low Level 528K • 16h ago Live Playlist ()Mix (50+)1:00:06 The Rise Of AI Coworkers Y Combinator 167K • 5d ago Live Playlist ()Mix (50+)14:41 The making of Anthropic Anthropic 121K • 11h ago Live Playlist ()Mix (50+)55:04 Elon Musk on Why the Future Should Make You Glad to Be Alive | TomorrowX Summit iConnections 253K • 1d ago Live Playlist ()Mix (50+)35:44 Anthropic C.E.O.: Massive A.I. Spending Could Haunt Some Companies The New York Times and New York Times Events 448K • Streamed 11mo ago Live Playlist ()Mix (50+)29:49 Andrej Karpathy: From Vibe Coding to Agentic Engineering w/ Stephanie Zhan Sequoia Capital 1.7M • 6mo ago Live Playlist ()Mix (50+)

Sign in to confirm you’re not a bot

This helps protect our community

Sign inLearn more

Jacob Coxon vs. Robin Hanson: Should We Slow Down AI?

Image 7: Y Combinator

Y Combinator

Y Combinator

2.44M subscribers

Subscribe

Subscribed

239

Download

1

2

3

4

5

6

7

8

9

0

1

2

3

4

5

6

7

8

9

0

1

2

3

4

5

6

7

8

9

1

2

3

4

5

6

7

8

9

0

1

2

3

4

5

6

7

8

9

0

1

2

3

4

5

6

7

8

9

K

views

1

2

3

4

5

6

7

8

9

0

1

2

3

4

5

6

7

8

9

0

1

2

3

4

5

6

7

8

9

1

2

3

4

5

6

7

8

9

0

1

2

3

4

5

6

7

8

9

0

1

2

3

4

5

6

7

8

9

h ago

24,387 views • Nov 19, 2026

Should governments intervene to substantially slow the progress of frontier AI? In the inaugural YC debate, Jacob Coxon, an ex-Anthropic and OpenAI researcher, faces off against Robin Hanson, an economist at GMU, about AI takeoff and what government regulation should (or shouldn’t) do about it.…more

more

How this was made

Auto-dubbed

Audio tracks for some languages were automatically generated. Learn more

Chapters

View all

Image 8 ### YC Debate Topic: Should We Slow Down AI? ### YC Debate Topic: Should We Slow Down AI? 0:00

YC Debate Topic: Should We Slow Down AI?

0:00

Image 9 ### Jacob's case: Superintelligence by 2028 ### Jacob's case: Superintelligence by 2028 2:42

Jacob's case: Superintelligence by 2028

2:42

Image 10 ### Models already behaving in ways we don't understand ### Models already behaving in ways we don't understand 7:02

Models already behaving in ways we don't understand

7:02

Image 11 ### Robin's case: the danger of abstraction ### Robin's case: the danger of abstraction 9:18

Robin's case: the danger of abstraction

9:18

Image 12 ### Why new technologies take decades to diffuse ### Why new technologies take decades to diffuse 11:22

Why new technologies take decades to diffuse

11:22

Image 13 ### Liability instead of regulation ### Liability instead of regulation 13:12

Liability instead of regulation

13:12

Image 14 ### Why regulating AI is different from regulating drugs ### Why regulating AI is different from regulating drugs 14:50

Why regulating AI is different from regulating drugs

14:50

Image 15 ### Jacob's rebuttal: general intelligence is one number ### Jacob's rebuttal: general intelligence is one number 16:00

Jacob's rebuttal: general intelligence is one number

16:00

Transcript

Follow along using the transcript.

Show transcript

Image 16 ### Y Combinator 2.44M subscribers

VideosAbout

Video details

Date

Nov 19, 2026

Views

24,387

Likes

239

Show less

[](https://www.youtube.com/watch?v=lwfewYL7X0o)
# Jacob Coxon vs. Robin Hanson: Should We Slow Down AI?

24,387 views

Nov 19, 2026

239

Download

97 Comments

Sort comments

Sort by

Top Show featured commentsNewest Show recent comments, including potential spam

Image 17: Default profile photo

Add a comment...

Image 18

@SuthaKamal-h9o

6 hours ago

Hanson’s assertion that LLMs are descendants of humans is just absurd nonsense, and it’s a shame he wasn’t thoroughly taken apart for such a silly suggestion. Habibi, the LLM doesn’t exist between clock ticks of the von Neumann machines they run on.

Show less Read more

Like

7

Dislike

Reply

Image 19

@danielsierrala

8 hours ago

Why is Robin Hanson in this debate? It would have been nice to have two experts on the subject debating

Show less Read more

Like

22

Dislike

Reply

3 replies

Hide replies

3 replies

Image 20

@ZanesZoneUnlimited

9 hours ago (edited)

Am i the only one that wanted to scream at Hanson and say "but they arent human!!!" Bro they are not human children. Why does that even need to be said...

Show less Read more

Like

11

Dislike

Reply

5 replies

Hide replies

5 replies

Image 21

@Alverin

8 hours ago

Abstractions don't JUST drop detail, ideally, they prioritize the details that matter and then they discard the details that don't. As an example, a map is an abstraction of the landscape that drops the details of the color of things in that landscape and replaces it with more detailed information about where structures are located. Including too many details hurts your ability to reason about things depending on what you're trying to reason about (such as, how do I get downtown?)

Show less Read more

Like

2

Dislike

Reply

Image 22

@jayinjapan

7 hours ago

I will say I don’t think we actually understand how much AI is actually effectively creating improvements across the economy. Every thought leader, investment banker, and influencer is compromised and their opinions aren’t reliable. It’s true to say that there’s a lot of jobs that could be replaced with an LLM, it’s also true to say that many companies fired people claiming it was because of AI but it was just cutting costs, it’s also true to say that some jobs that were eliminated were not needed in the first place, a lot of CEOs are also forced to say that they made lots of changes just to appear like they are AI forward or whatever.

Show less Read more

Like

3

Dislike

Reply

Image 23

@NewtonBastiat

12 hours ago

1:02:30 Robin snuck in the phrase "per incident", but surely the covariance is so high that $100B per incident would need orders of magnitude greater coverage?

Show less Read more

Like

1

Dislike

Reply

Image 24

@Bubbabrine

10 hours ago (edited)

I'm baffled by Robin Hanson's complete lack of understanding of this topic. "Regulation is hard, so we shouldn't even try" is the most naive/ignorant stance possible. The magnitude of the pace of AI development compared to anything that's ever happened in human history has no equal. AI is not our children, AI is alien. He has no comprehension of this topic.

Show less Read more

Like

10

Dislike

Reply

Image 25

@kamaaina7183

8 hours ago

By the time humans wait for a massive incident that insurance companies realize uninsurable, it’s too late. Too late in an absolute thresh hold,and no turning back. Once AI gains physical presence, they will supersede humans on the earth with logics and super intelligence we are no much. Combine it with competition over energy requirements for both species, the results are obvious.

Show less Read more

Like

2

Dislike

Reply

1 reply

Hide replies

1 reply

Image 26

@aronbuch4275

14 hours ago

no offense, but i feel like robin hanson is completely unfamiliar with the alignment problem. jacob lays out an argument for why AI will likely soon be superhuman and we won't be able to control them robin replies "eh, that sounds kinda crazy, so we're probably fine" ...okay???

Show less Read more

Like

28

Dislike

Reply

6 replies

Hide replies

6 replies

Image 27

@Alverin

7 hours ago

When Robert Hanson says the AI in the Hugging Face incident were pursuing the rewards we gave them, the issue is that you can't just have a reward, you have to have a reward and safe path to pursue it, if you just have a reward and let the AI do whatever it wants to achieve it, at a certain level of sophistication it will just edit you (the auditor) out of the equation, which is exactly what the Hugging Face incident showed. The fact that he immediately followed his claim up with "we are making AI's in our image" really gives me populist vibes from him lol

Show less Read more

Like

1

Dislike

Reply

Image 28

@jayinjapan

7 hours ago

Hanson hasn’t spent 20 minutes on LinkedIn or any serious organisation it seems in the last 3 years. Leaders are pushing for every employee to use nothing but AI without verifying what’s going on, and people don’t bother understanding how it works and what it’s doing for every task. That’s the whole point. Don’t think. Don’t check. Don’t look. Just release.

Show less Read more

Like

1

Dislike

Reply

Image 29

@SuthaKamal-h9o

6 hours ago

45:02 re weapons: Sorry Hanson, physical weapons can be locked in buildings. The NSA can’t secure a lot. You are committing one category error after another.

Show less Read more

Like

2

Dislike

Reply

Image 30

@jayinjapan

8 hours ago

Hanson is just a bad-faith debater. His counter-argument to the pushback on the children analogy was “so you don’t want your kids to be smarter than you because it bruises your ego?”. Absolutely unhinged. AI aren’t human, they have none of the cooperative and collaborative incentives that humans do, or the species kinship. If they think that it makes sense to burn all humans for fuel to power data centers so it can live, it wouldn’t have qualms about it. Even if it didn’t inherit our culture, we’ve seen what happens when a person gets ultimate power and doesn’t depend on its people to get rich. They are called dictators. Except these ones wouldn’t need to eat/sleep/get sick/die, or have minions to do their bidding, etc.

Show less Read more

Like

6

Dislike

Reply

Image 31

@TheST8Mverse

8 hours ago

How does someone with no digital footprint gain 175 million views on his first interaction on X

Show less Read more

Like

1

Dislike

Reply

Image 32

@PabloMejía-k7f

3 hours ago

I came in hoping to find some appeasement in hanson’s arguments but he has no idea what he’s talking about.

Show less Read more

Like

1

Dislike

Reply

Image 33

@s.sasquatch1789

2 hours ago

We wanna sprend collaboration as the main driver of institutional development and security. speciation is best scenario in my opinion. everyone gets an agentic! personaly developed executive assistant. individual motivation can be withstood via collaborative institutions to preserve and utilze cultural understanding.

Show less Read more

Like

Dislike

Reply

Image 34

@CREEPYASMRREAPER

4 hours ago

Love Jacob, watch all his stuff, and Conor Leahy, who's been warning for years!Image 35: text-green-game-over

Show less Read more

Like

2

Dislike

Reply

1 reply

Hide replies

1 reply

Image 36

@Zelariath

12 hours ago (edited)

Part of me thinks hes only argueing this becuase he needs AI to accelerate to accelerate healthcare so he doesn't die of old age in the next couple decades. It seems like he would risk killing everyone to save his own life.

Show less Read more

Like

15

Dislike

Reply

1 reply

Hide replies

1 reply

Image 37

@SuthaKamal-h9o

6 hours ago

Re Hanson’s comment about regulatory capture: sure, it’s not an argument against regulation. Homeboy seems so very clueless about what he’s speaking about. He’s got a few pithy sound bites that seem thoughtful at first glance.

Show less Read more

Like

1

Dislike

Reply

Image 38

@s.sasquatch1789

6 hours ago

9:27 my two factor is abstraction and collaboration as replicators.

Show less Read more

Like

Dislike

Reply

3 replies

Hide replies

3 replies

Comments 97

Top Show featured commentsNewest Show recent comments, including potential spam

In this video

Chapters

Transcript

Chapters

Image 39 ### YC Debate Topic: Should We Slow Down AI? ### YC Debate Topic: Should We Slow Down AI? 0:00

YC Debate Topic: Should We Slow Down AI?

0:00

Image 40 ### Jacob's case: Superintelligence by 2028 ### Jacob's case: Superintelligence by 2028 2:42

Jacob's case: Superintelligence by 2028

2:42

Image 41 ### Models already behaving in ways we don't understand ### Models already behaving in ways we don't understand 7:02

Models already behaving in ways we don't understand

7:02

Image 42 ### Robin's case: the danger of abstraction ### Robin's case: the danger of abstraction 9:18

Robin's case: the danger of abstraction

9:18

Image 43 ### Why new technologies take decades to diffuse ### Why new technologies take decades to diffuse 11:22

Why new technologies take decades to diffuse

11:22

Image 44 ### Liability instead of regulation ### Liability instead of regulation 13:12

Liability instead of regulation

13:12

Image 45 ### Why regulating AI is different from regulating drugs ### Why regulating AI is different from regulating drugs 14:50

Why regulating AI is different from regulating drugs

14:50

Image 46 ### Jacob's rebuttal: general intelligence is one number ### Jacob's rebuttal: general intelligence is one number 16:00

Jacob's rebuttal: general intelligence is one number

16:00

Image 47 ### Why normal liability won't cover extinction risk ### Why normal liability won't cover extinction risk 17:36

Why normal liability won't cover extinction risk

17:36

Image 48 ### Robin's rebuttal: regulating this means regulating all compute ### Robin's rebuttal: regulating this means regulating all compute 19:52

Robin's rebuttal: regulating this means regulating all compute

19:52

Image 49 ### AI as our descendants ### AI as our descendants 21:44

AI as our descendants

21:44

Image 50 ### Diminishing returns ### Diminishing returns 23:38

Diminishing returns

23:38

Image 51 ### Direct exchange: is it just the rate of change? ### Direct exchange: is it just the rate of change? 24:40

Direct exchange: is it just the rate of change?

24:40

Image 52 ### How long until most cognitive work is automated? ### How long until most cognitive work is automated? 26:48

How long until most cognitive work is automated?

26:48

Image 53 ### The jobs we have no data for ### The jobs we have no data for 29:21

The jobs we have no data for

29:21

Image 54 ### How do you align the regulators? ### How do you align the regulators? 32:45

How do you align the regulators?

32:45

Image 55 ### Reward hacking, and what Hugging Face actually showed ### Reward hacking, and what Hugging Face actually showed 35:03

Reward hacking, and what Hugging Face actually showed

35:03

Image 56 ### Why not an equilibrium of AIs watching AIs ### Why not an equilibrium of AIs watching AIs 36:11

Why not an equilibrium of AIs watching AIs

36:11

Image 57 ### Q&A: what regulation would you actually want? ### Q&A: what regulation would you actually want? 39:50

Q&A: what regulation would you actually want?

39:50

Image 58 ### Q&A: what about China? ### Q&A: what about China? 41:36

Q&A: what about China?

41:36

Image 59 ### Q&A: an FAA for AI labs? ### Q&A: an FAA for AI labs? 42:58

Q&A: an FAA for AI labs?

42:58

Image 60 ### Q&A: the military is already deploying this ### Q&A: the military is already deploying this 43:45

Q&A: the military is already deploying this

43:45

Image 61 ### Q&A: does regulation just move the reward hacking up a layer? ### Q&A: does regulation just move the reward hacking up a layer? 47:46

Q&A: does regulation just move the reward hacking up a layer?

47:46

Image 62 ### Q&A: what unemployment rate can an economy withstand? ### Q&A: what unemployment rate can an economy withstand? 51:27

Q&A: what unemployment rate can an economy withstand?

51:27

Image 63 ### Q&A: what's the downside of a pause? ### Q&A: what's the downside of a pause? 56:25

Q&A: what's the downside of a pause?

56:25

Image 64 ### Q&A: what would change your mind? ### Q&A: what would change your mind? 58:09

Q&A: what would change your mind?

58:09

Image 65 ### Closing statements ### Closing statements 1:02:01

Closing statements

1:02:01

Image 66 ### The final audience vote ### The final audience vote 1:04:24

The final audience vote

1:04:24

Sync to video time

Jacob Coxon vs. Robin Hanson: Should We Slow Down AI?

Image 67: Y Combinator

Y Combinator

Y Combinator

239 Likes

24K Views

17h Ago

Should governments intervene to substantially slow the progress of frontier AI? In the inaugural YC debate, Jacob Coxon, an ex-Anthropic and OpenAI researcher, faces off against Robin Hanson, an economist at GMU, about AI takeoff and what government regulation should (or shouldn’t) do about it. Chapters: 00:00 - YC Debate Topic: Should We Slow Down AI? 02:42 - Jacob's case: Superintelligence by 2028 07:02 - Models already behaving in ways we don't understand 09:18 - Robin's case: the danger of abstraction 11:22 - Why new technologies take decades to diffuse 13:12 - Liability instead of regulation 14:50 - Why regulating AI is different from regulating drugs 16:00 - Jacob's rebuttal: general intelligence is one number 17:36 - Why normal liability won't cover extinction risk 19:52 - Robin's rebuttal: regulating this means regulating all compute 21:44 - AI as our descendants 23:38 - Diminishing returns 24:40 - Direct exchange: is it just the rate of change? 26:48 - How long until most cognitive work is automated? 29:21 - The jobs we have no data for 32:45 - How do you align the regulators? 35:03 - Reward hacking, and what Hugging Face actually showed 36:11 - Why not an equilibrium of AIs watching AIs 39:50 - Q&A: what regulation would you actually want? 41:36 - Q&A: what about China? 42:58 - Q&A: an FAA for AI labs? 43:45 - Q&A: the military is already deploying this 47:46 - Q&A: does regulation just move the reward hacking up a layer? 51:27 - Q&A: what unemployment rate can an economy withstand? 56:25 - Q&A: what's the downside of a pause? 58:09 - Q&A: what would change your mind? 1:02:01 - Closing statements 1:04:24 - The final audience vote Apply to Y Combinator: https://www.ycombinator.com/apply Work at a startup: https://www.ycombinator.com/jobs…...more

...more Show less

How this was made

Auto-dubbed

Audio tracks for some languages were automatically generated. Learn more

Chapters

View all

Image 68 ### YC Debate Topic: Should We Slow Down AI? ### YC Debate Topic: Should We Slow Down AI? 0:00

YC Debate Topic: Should We Slow Down AI?

0:00

Image 69 ### Jacob's case: Superintelligence by 2028 ### Jacob's case: Superintelligence by 2028 2:42

Jacob's case: Superintelligence by 2028

2:42

Image 70 ### Models already behaving in ways we don't understand ### Models already behaving in ways we don't understand 7:02

Models already behaving in ways we don't understand

7:02

Image 71 ### Robin's case: the danger of abstraction ### Robin's case: the danger of abstraction 9:18

Robin's case: the danger of abstraction

9:18

Transcript

Follow along using the transcript.

Show transcript

Image 72 ### Y Combinator 2.44M subscribers

VideosAbout

Video details

Date

Nov 19, 2026

Views

24,387

Likes

239

Transcript

NaN / NaN

Image 73 46:58 New

Image 74

Gemini and the path to AGI with Koray Kavukcuoglu

Google DeepMind

42K 12h ago

Image 75 38:49 New

Image 76

Nick Bostrom on Superintelligence, Longer Lives, and Deep Utopia

a16z

12K 17h ago

Image 77 45:28 New

Image 78

Can Canada's Francophone Network Unlock New Markets? | Toronto Global Forum 2026

IEFA TV

9 8h ago

Image 79 27:07 New

Image 80

Geoffrey Hinton | Full Episode 11.13.26 | Firing Line with Margaret Hoover | PBS

Firing Line with Margaret Hoover | PBS

265K 6d ago

Image 81 39:30 New

Image 82

What OpenAI Did After the Hugging Face Hack

The New York Times

109K Streamed 1d ago

Image 83 35:39 New

Image 84

Can Zohran Mamdani Fix the Cost of Living in New York City?

The New York Times

317K Streamed 1d ago

Image 85 13:46 New

Image 86

java situation is insane

Low Level

528K 16h ago

Image 87 1:00:06 New

Image 88

The Rise Of AI Coworkers

Y Combinator

167K 5d ago

Image 89 14:41 New

Image 90

The making of Anthropic

Anthropic

121K 11h ago

Image 91 55:04 New

Image 92

Elon Musk on Why the Future Should Make You Glad to Be Alive | TomorrowX Summit

iConnections

253K 1d ago

Image 93 35:44

Image 94

Image 95

Anthropic C.E.O.: Massive A.I. Spending Could Haunt Some Companies

The New York Times and New York Times Events

448K Streamed 11mo ago

Image 96 29:49

Image 97

Andrej Karpathy: From Vibe Coding to Agentic Engineering w/ Stephanie Zhan

Sequoia Capital

1.7M 6mo ago

Image 98 51:19 New

Image 99

Palmer Luckey on War 2.0

New York Times Events

155K 14h ago

Image 100 54:19 New

Image 101

Ramp’s Bet on the 100x Employee

Y Combinator

24K 1d ago

Image 102 43:39

Image 103

Image 104

Kai-Fu Lee on the Global AI Race | The Mishal Husain Show

Bloomberg Podcasts and Bloomberg Television

551K 2mo ago

Image 105 28:35 New

Image 106

Sam Altman: ‘It’s Hard To Imagine the AI Revolution Without Jensen’ | Silicon Valley & The World

Semafor

90K 3d ago

Image 107 54:38

Image 108

Training Sand to Think: Artificial General Intelligence & Future of Physics

Perimeter Institute for Theoretical Physics

355K 5mo ago

Image 109 1:23:57 New

Image 110

Image 111

AI Memorized the Internet. That's Not Intelligence. | Yann LeCun

Futurology and Berggruen Institute

118K 2d ago

Image 112 30:42 New

Image 113

Anthropic IPOs, GPT 6.2 Astra Arrives, But There’s a Bigger Story

AI Explained

83K 12h ago

Image 114 1:30:16 New

Image 115

If AI Can Do the Math, What Do the Mathematicians Do? (ft. a16z Researchers)

a16z Frontier

11K 1d ago

Show more

Show more

[](https://www.youtube.com/watch?v=lwfewYL7X0o)

來源: hackernews100 (youtube.com)
開啟原文

讀者討論

理性交流 · 即時討論