保留正文与补采附件。源页面当前未重新核验；具体限制见交付清单。
SubscribeSign in Home Notes Get the book Book exercises About us At a glance
LatestTopDiscussions
A big-tent or small-tent AI safety movement?
The unstated disagreement that underpins safety debates
4 hrs ago • Arvind Narayanan and Sayash Kapoor
101
5
12
AI existential risk probabilities are (still) too unreliable to inform policy
How speculation gets laundered through pseudo-quantification
Sep 28 • Arvind Narayanan and Sayash Kapoor
152
21
25
The AI-as-Normal-Technology view of loss-of-control incidents
A middle ground between the cybersecurity and AI safety communities
Sep 14 • Sayash Kapoor and Arvind Narayanan
274
54
49
AI agents can't yet do open-ended AI research
Early evidence from two case studies
Aug 5 • Sayash Kapoor and Arvind Narayanan
194
27
25
What will be left for us to work on?
My keynote at ICML 2026
Jul 13 • Arvind Narayanan
349
30
61
Up the Stack: How AI’s Escape From the Commodity Trap Risks Enterprise Lock-in
Critics and boosters are both looking in the wrong place
Jul 9 • Arvind Narayanan
173
24
31
Why AI hasn’t replaced software engineers, and won’t
Coding agents as normal technology
Jun 11 • Arvind Narayanan and Sayash Kapoor
600
89
110
Did Google’s AI agents really build an operating system for $916?
The importance of independent evaluation
May 22 • Sayash Kapoor and Arvind Narayanan
144
25
11
Do AI Risks Require Extraordinary Government Intervention?
Let’s not skip the hard work of AI governance
May 21 • Sayash Kapoor and Arvind Narayanan
85
9
19
Open-world evaluations for measuring frontier AI capabilities
Introducing CRUX, a new project for evaluating AI on long, messy tasks
Apr 16 • Sayash Kapoor and Arvind Narayanan
99
7
14
New Paper: Towards a science of AI agent reliability
Quantifying the capability-reliability gap
Feb 24 • Sayash Kapoor and Arvind Narayanan
328
36
71
AI Won’t Automatically Make Legal Services Cheaper
Applying the AI as Normal Technology framework to legal services
Feb 12 • Justin Curl, Sayash Kapoor, and Arvind Narayanan
86
13
15
Fact checking Moravec's paradox
The famous aphorism is neither true nor useful
Jan 29 • Arvind Narayanan
174
10
25
A guide to understanding AI as normal technology
And a big change for this newsletter
Sep 9, 2025 • Arvind Narayanan and Sayash Kapoor
234
29
42
Could AI slow science?
Confronting the production-progress paradox
Jul 16, 2025 • Sayash Kapoor and Arvind Narayanan
325
37
70
AGI is not a milestone
There is no capability threshold that will lead to sudden impacts
May 1, 2025 • Sayash Kapoor and Arvind Narayanan
199
37
40
AI as Normal Technology
A new paper that we will expand into our next book
Apr 15, 2025 • Arvind Narayanan and Sayash Kapoor
636
40
132
Is AI progress slowing down?
Making sense of recent technology trends and claims
Dec 18, 2024 • Arvind Narayanan and Sayash Kapoor
171
15
29
We Looked at 78 Election Deepfakes. Political Misinformation is not an AI Problem.
Technology Isn’t the Problem—or the Solution.
Dec 13, 2024 • Sayash Kapoor and Arvind Narayanan
119
6
23
Does the UK’s liver transplant matching algorithm systematically exclude younger patients?
Seemingly minor technical decisions can have life-or-death effects
Nov 11, 2024 • Arvind Narayanan and Sayash Kapoor
103
11
7
FAQ about the book and our writing process
What's in the book and how we wrote it
Oct 4, 2024 • Arvind Narayanan and Sayash Kapoor
90
13
6
Can AI automate computational reproducibility?
A new benchmark to measure the impact of AI on improving science
Sep 18, 2024 • Sayash Kapoor and Arvind Narayanan
48
2
2
Start reading the AI Snake Oil book online
The book was published on September 24, 2024
Sep 10, 2024 • Arvind Narayanan and Sayash Kapoor
199
8
10
AI companies are pivoting from creating gods to building products. Good.
Turning models into products runs into five challenges
Aug 19, 2024 • Arvind Narayanan and Sayash Kapoor
230
24
27
© 2026 Sayash Kapoor and Arvind Narayanan · Privacy ∙ Terms ∙ Collection notice
Start your SubstackGet the app
Substack is the home for great culture