Jacob Coxon 與 Robin Hanson 在 Y Combinator 辯論政府是否應放慢前沿 AI 發展
政府是否应该放缓前沿人工智能的发展?雅各布·考克森与罗宾·汉森就此展开辩论 [视频]
Y Combinator 舉行首場辯論,由曾任 Anthropic 和 OpenAI 研究員的 Jacob Coxon 與 GMU 經濟學家 Robin Hanson 討論政府是否應大幅放慢前沿 AI 發展。辯論涉及 AI 起飛、模型行為、技術擴散,以及政府監管、責任追究與算力管制等議題。
Back
Skip navigation
Search
Search with your voice
Jacob Coxon vs. Robin Hanson: Should We Slow Down AI?
Tap to unmute
2x
Jacob Coxon vs. Robin Hanson: Should We Slow Down AI?
Y Combinator 23,242 views 16h ago
Copy link
Info
Shopping
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
Share
- [x] Include playlist
An error occurred while retrieving sharing information. Please try again later.
0:00
0:00 / 0:00
Live
•Watch full video
•
YC Debate Topic: Should We Slow Down AI?
•
27:07 Geoffrey Hinton | Full Episode 11.13.26 | Firing Line with Margaret Hoover | PBS Firing Line with Margaret Hoover | PBS 264K • 6d ago Live Playlist ()Mix (50+)38:49 Nick Bostrom on Superintelligence, Longer Lives, and Deep Utopia a16z 11K • 15h ago Live Playlist ()Mix (50+)14:41 The making of Anthropic Anthropic 104K • 10h ago Live Playlist ()Mix (50+)31:46 Satya Nadella on the New Algorithm New York Times Events 28K • 12h ago Live Playlist ()Mix (50+)51:19 Palmer Luckey on War 2.0 New York Times Events 143K • 13h ago Live Playlist ()Mix (50+)37:25 This algorithm changed how I think about compression | Information theory, part 3 3Blue1Brown 1M • 2d ago Live Playlist ()Mix (50+)16:55 Mustafa Suleyman: AI Is Starting to Improve Itself, and That's Dangerous Yahoo Finance 16K • 2d ago Live Playlist ()Mix (50+)39:30 What OpenAI Did After the Hugging Face Hack The New York Times 107K • Streamed 1d ago Live Playlist ()Mix (50+)22:59 Bernie Sanders VS. 15 Gen Z On AI Use More Perfect Union and 2 more 1.7M • 2mo ago Live Playlist ()Mix (50+)31:12 FULL DISCUSSION: Google's Demis Hassabis, Anthropic's Dario Amodei Debate the World After AGI | AI1G DRM News 1.2M • 9mo ago Live Playlist ()Mix (50+)19:45 No More Heroes - Terence Tao Talks to Numberphile About AI Numberphile 1.1M • 1mo ago Live Playlist ()Mix (50+)18:27 Even AI Can't Beat This Man at Programming Turing and Turing Now - Science & Tech News 822K • 9d ago Live Playlist ()Mix (50+)
Sign in to confirm you’re not a bot
This helps protect our community
Jacob Coxon vs. Robin Hanson: Should We Slow Down AI?
Y Combinator
2.44M subscribers
Subscribe
Subscribed
227
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
23,242 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
### 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
### Jacob's case: Superintelligence by 2028 ### Jacob's case: Superintelligence by 2028 2:42
Jacob's case: Superintelligence by 2028
2:42
Models already behaving in ways we don't understand
7:02
### 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
Why new technologies take decades to diffuse
11:22
### Liability instead of regulation ### Liability instead of regulation 13:12
Liability instead of regulation
13:12
Why regulating AI is different from regulating drugs
14:50
Jacob's rebuttal: general intelligence is one number
16:00
Transcript
Follow along using the transcript.
Show transcript
### Y Combinator 2.44M subscribers
Video details
Date
Nov 19, 2026
Views
23,242
Likes
227
Show less
Jacob Coxon vs. Robin Hanson: Should We Slow Down AI?
23,242 views
Nov 19, 2026
227
Download
92 Comments
Sort comments
Sort by
Top Show featured commentsNewest Show recent comments, including potential spam
Add a comment...
@SuthaKamal-h9o
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
5
Dislike
Reply
@ZanesZoneUnlimited
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
10
Dislike
Reply
5 replies
Hide replies
5 replies
@danielsierrala
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
19
Dislike
Reply
2 replies
Hide replies
2 replies
@kamaaina7183
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
@Alverin
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
@CREEPYASMRREAPER
Love Jacob, watch all his stuff, and Conor Leahy, who's been warning for years!
Show less Read more
Like
2
Dislike
Reply
1 reply
Hide replies
1 reply
@NewtonBastiat
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
@s.sasquatch1789
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
@Alverin
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
@SuthaKamal-h9o
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
@s.sasquatch1789
28:00 figure one is collecting that data
Show less Read more
Like
Dislike
Reply
@007musk
I found nay arguments more convincing
Show less Read more
Like
2
Dislike
Reply
@IemonandIime
Robin Hanson: "I have no idea what I'm talking about lmao"
Show less Read more
Like
Dislike
Reply
@PabloMejía-k7f
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
@jayinjapan
We could you know just…..turn off the power to the data center? So long as you don’t give Claude arms it should be fine.
Show less Read more
Like
Dislike
Reply
@s.sasquatch1789
9:27 my two factor is abstraction and collaboration as replicators.
Show less Read more
Like
Dislike
Reply
3 replies
Hide replies
3 replies
@jayinjapan
“We are remarkably more coordinated than animals”…yeah and look what we’ve done to the animals.
Show less Read more
Like
Dislike
Reply
@umrashrf
Looks like the younger generation is really about future
Show less Read more
Like
Dislike
Reply
@l81z
Slowing down AI means abstaining from doing some things to/with AI. Obviously it’s a good idea sometimes; not as a blanket Government decree. I bet AI could generate a strong debate too. The military should develop Anti-AI tech.
Show less Read more
Like
Dislike
Reply
@jayinjapan
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
Comments 92
Top Show featured commentsNewest Show recent comments, including potential spam
In this video
Chapters
Transcript
Chapters
### 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
### Jacob's case: Superintelligence by 2028 ### Jacob's case: Superintelligence by 2028 2:42
Jacob's case: Superintelligence by 2028
2:42
Models already behaving in ways we don't understand
7:02
### 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
Why new technologies take decades to diffuse
11:22
### Liability instead of regulation ### Liability instead of regulation 13:12
Liability instead of regulation
13:12
Why regulating AI is different from regulating drugs
14:50
Jacob's rebuttal: general intelligence is one number
16:00
Why normal liability won't cover extinction risk
17:36
Robin's rebuttal: regulating this means regulating all compute
19:52
### AI as our descendants ### AI as our descendants 21:44
AI as our descendants
21:44
### Diminishing returns ### Diminishing returns 23:38
Diminishing returns
23:38
Direct exchange: is it just the rate of change?
24:40
How long until most cognitive work is automated?
26:48
### 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
### How do you align the regulators? ### How do you align the regulators? 32:45
How do you align the regulators?
32:45
Reward hacking, and what Hugging Face actually showed
35:03
### 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
Q&A: what regulation would you actually want?
39:50
### Q&A: what about China? ### Q&A: what about China? 41:36
Q&A: what about China?
41:36
### 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
Q&A: the military is already deploying this
43:45
Q&A: does regulation just move the reward hacking up a layer?
47:46
Q&A: what unemployment rate can an economy withstand?
51:27
### 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
### 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
### Closing statements ### Closing statements 1:02:01
Closing statements
1:02:01
### 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?
Y Combinator
Y Combinator
227 Likes
23K Views
16h 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
### 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
### Jacob's case: Superintelligence by 2028 ### Jacob's case: Superintelligence by 2028 2:42
Jacob's case: Superintelligence by 2028
2:42
Models already behaving in ways we don't understand
7:02
### 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
### Y Combinator 2.44M subscribers
Video details
Date
Nov 19, 2026
Views
23,242
Likes
227
Transcript
NaN / NaN
Geoffrey Hinton | Full Episode 11.13.26 | Firing Line with Margaret Hoover | PBS
Firing Line with Margaret Hoover | PBS
264K 6d ago
Nick Bostrom on Superintelligence, Longer Lives, and Deep Utopia
a16z
11K 15h ago
The making of Anthropic
Anthropic
104K 10h ago
Satya Nadella on the New Algorithm
New York Times Events
28K 12h ago
Palmer Luckey on War 2.0
New York Times Events
143K 13h ago
This algorithm changed how I think about compression | Information theory, part 3
3Blue1Brown
1M 2d ago
Mustafa Suleyman: AI Is Starting to Improve Itself, and That's Dangerous
Yahoo Finance
16K 2d ago
What OpenAI Did After the Hugging Face Hack
The New York Times
107K Streamed 1d ago
Bernie Sanders VS. 15 Gen Z On AI Use
More Perfect Union and 2 more
1.7M 2mo ago
Fundraiser
FULL DISCUSSION: Google's Demis Hassabis, Anthropic's Dario Amodei Debate the World After AGI | AI1G
DRM News
1.2M 9mo ago
No More Heroes - Terence Tao Talks to Numberphile About AI
Numberphile
1.1M 1mo ago
Even AI Can't Beat This Man at Programming
Turing and Turing Now - Science & Tech News
822K 9d ago
RAM Altman
Gamers Nexus
527K 10h ago
Inside Anthropic, the $965 Billion AI Juggernaut | The Circuit
Bloomberg Originals
3.7M 5mo ago
Why Most Companies Fail to See Bottom-Line Returns From Deploying AI Agents
Forbes
25K 6d ago
Elon Musk on Why the Future Should Make You Glad to Be Alive | TomorrowX Summit
iConnections
245K 1d ago
Saagar Enjeti on The Future of The New Right | Interesting Times
Interesting Times and 3 more
85K 15h ago
Jeff Bezos on Jensen Huang, Getting Into The AI Race, & Picking James Bond
Semafor
636K 2d ago
Politics Chat, November 19, 2026
Heather Cox Richardson
242K Streamed 7h ago
The Kind of Thinker Who Will Thrive in the AI Era | Head of Google Research (Yossi Matias)
Sinead Bovell
576K 4w ago
Elon Musk Admits “I Feel Pretty Foolish” (Full NYT Interview)
Matt Pocius on Tesla Stock & Money
548K 1d ago
AI, Consciousness and the Future of Humanity | Mustafa Suleyman & Yuval Noah Harari
Yuval Noah Harari
185K 1d ago
OpenAI’s Cofounder & Research Chief Explain Self-Improving AI
a16z
46K 2d ago
10 Habits PhDs Must Unlearn To Build a Startup | Startup School
Y Combinator
63K 2w ago
Show more
Show more
來源:Hacker News 熱門(buzzing.cc 中文翻譯) · youtube.com
27:07 New
38:49 New
14:41 New
31:46 New
51:19 New
37:25 New
16:55 New
39:30 New
22:59
31:12
19:45