This page collects the most important essays we’ve written on this newsletter. All essays are by Arvind and Sayash unless otherwise specified.
Foundational essays
The original essay together with the other essays in this section contain our overarching framework for understanding AI progress and impacts.
AI as Normal Technology
The original essay that we are expanding into our next book
Apr 2025
A guide to understanding AI as normal technology
And a big change for this newsletter
Sep 2025
What will be left for us to work on?
My keynote at ICML 2026
Jul 2026 • Arvind Narayanan
Do AI Risks Require Extraordinary Government Intervention?
Let’s not skip the hard work of AI governance
May 2026
AGI is not a milestone
There is no capability threshold that will lead to sudden impacts
May 2025
AI existential risk probabilities are too unreliable to inform policy
How speculation gets laundered through pseudo-quantification
Jul 2024
AI safety is not a model property
Trying to make an AI model that can’t be misused is like trying to make a computer that can’t be used for bad things
Mar 2024
On the Societal Impact of Open Foundation Models
Adding precision to the debate on openness in AI
Feb 2024 • Sayash Kapoor, Rishi Bommasani, Daniel E. Ho, Percy Liang, and Arvind Narayanan
Applications of the framework
We apply the AI-as-normal-technology framework to analyze various industries, occupations, and institutions.
Why AI hasn’t replaced software engineers, and won’t
Coding agents as normal technology
Jun 2026
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 2026 • Arvind Narayanan and Akash Kapur
AI Won’t Automatically Make Legal Services Cheaper
Applying the AI as Normal Technology framework to legal services
Feb 2026 • Justin Curl, Sayash Kapoor, and Arvind Narayanan
Fact checking Moravec’s paradox
This famous aphorism is neither true nor useful
Jan 2026 • Arvind Narayanan
Could AI slow science?
Confronting the production-progress paradox
Jul 2025
We Looked at 78 Election Deepfakes. Political Misinformation is not an AI Problem.
Technology Isn’t the Problem—or the Solution.
Dec 2024
AI companies are pivoting from creating gods to building products. Good.
Turning models into products runs into five challenges
Aug 2024
Technical research and explainers
We lead a team working on the science of AI agent evaluation, and regularly share our research and technical commentary here.
AI agents can’t yet do open-ended AI research
Early evidence from two case studies
Aug 2026
Did Google’s AI agents really build an operating system for $916?
The importance of independent evaluation
May 2026 • Stephan Rabanser, Sayash Kapoor, Rishi Bommasani, Andrew Schwartz, and Arvind Narayanan
Open-world evaluations for measuring frontier AI capabilities
Introducing CRUX, a new project for evaluating AI on long, messy tasks
Apr 2026 • [Many authors]
New Paper: Towards a science of AI agent reliability
Quantifying the capability-reliability gap
Feb 2026 • Stephan Rabanser, Sayash Kapoor, and Arvind Narayanan
Is AI progress slowing down?
Making sense of recent technology trends and claims
Dec 2024 • Arvind Narayanan, Benedikt Ströbl, and Sayash Kapoor
Can AI automate computational reproducibility?
A new benchmark to measure the impact of AI on improving science
Sep 2024
New paper: AI agents that matter
Rethinking AI agent benchmarking and evaluation
Jul 2024
AI scaling myths
Scaling will run out. The question is when.
Jun 2024
AI Snake Oil
This newsletter was previously called AI Snake Oil, also the title of our first book together. That project was about calling out dubious and harmful AI applications deployed here and now, whereas AI as Normal Technology is more forward looking. Here are selected essays from that earlier version of this newsletter.
Start reading the AI Snake Oil book online
The book was published September 2024
Sep 2024
FAQ about the book and our writing process
What’s in the book and how we wrote it
Oct 2024
Does the UK’s liver transplant matching algorithm systematically exclude younger patients?
Seemingly minor technical decisions can have life-or-death effects
Nov 2024 • Arvind Narayanan, Angelina Wang, Sayash Kapoor, and Solon Barocas
Scientists should use AI as a tool, not an oracle
How AI hype leads to flawed research that fuels more hype
Jun 2024
Introducing the REFORMS checklist for ML-based science
ML-based science is in trouble. Clear reporting standards for researchers could help.
Aug 2023
ML is useful for many things, but not for predicting scientific replicability
How the veneer of AI is used to legitimize awful ideas
Aug 2023
The bait and switch behind AI risk prediction tools
Toronto recently used an AI tool to predict when a public beach will be safe. It went horribly awry.
Nov 2022