Arvind Narayanan

cs.princeton.eduWhat will be left for us to work on?ICML 2026 invited keynote — slides and edited transcript, presented click-by-click as delivered. Arvind Narayanan, Princeton University.

I had the honor of giving a keynote at the International Conference on Machine Learning last week. I addressed the widespread anxiety about how we should adapt as AI capabilities increase. I was thrilled by the talk’s reception, so I have made my slides available www.cs.princeton.edu/~arvindn/tal...

Arvind Narayanan reposted

Jameel Jaffer

@jameeljaffer.bsky.social

This is a great conversation between computer scientist and former @knightcolumbia.org visiting scholar Arvind Narayanan (@randomwalker.bsky.social) and @ezraklein.bsky.social about Arvind's @knightcolumbia.org essay, "AI as Normal Technology." www.nytimes.com/2026/10/09/o...

nytimes.comOpinion | Intelligence Isn’t PowerThe computer scientist Arvind Narayanan explains that just because A.I. is intelligent doesn’t necessarily mean it’s powerful.

Arvind Narayanan reposted

Princeton University Press

@princetonupress.bsky.social

@randomwalker.bsky.social, co-author of AI Snake Oil, talks to Ezra Klein for @nytopinion.nytimes.com about why AI is something we've seen before (and something we might actually have a roadmap for dealing with): buff.ly/KkS1IBK

buff.lyIntelligence Isn’t PowerThe computer scientist Arvind Narayanan explains that just because A.I. is intelligent doesn’t necessarily mean it’s powerful.

Arvind Narayanan reposted

Jan Jęcz

@jeczjan.eurosky.social

20 Sept

Replying to @rebeccawilliams.info

From @sayash.bsky.social and @randomwalker.bsky.social book “AI Snake Oil”. Published in September 2024, but hasn’t aged even a bit

Arvind Narayanan reposted

Per Arne Godejord

@pagodejord.bsky.social

14 Sept

A valuable and worthwhile read from @randomwalker.bsky.social and @sayash.bsky.social www.normaltech.ai/p/the-ai-as-...

normaltech.aiThe AI-as-Normal-Technology view of loss of control incidentsA middle ground between the cybersecurity and AI safety communities

Arvind Narayanan reposted

Ethan Mollick

@emollick.bsky.social

24 Aug

From @randomwalker.bsky.social: state-level data center bans have very little impact overall on AI progress due to rising efficiency. If your concern is AI development, concentration of power, future use, or anything else other than "is a datacenter near me?" this is not a substitute for policy.

Arvind Narayanan reposted

Eamon Duede

@eduede.bsky.social

18 Aug

Excellent @randomwalker.bsky.social essay on how frontier labs might navigate various layers of the value stack to survive. open.substack.com/pub/aisnakeo...

open.substack.comUp the Stack: How AI’s Escape From the Commodity Trap Risks Enterprise Lock-inCritics and boosters are both looking in the wrong place

Arvind Narayanan reposted

Melanie Mitchell

@melaniemitchell.bsky.social

2 Aug

I'm just now reading @randomwalker.bsky.social's writeup of his ICML keynote. It is very insightful. Everyone thinking about the future of work in the age of AI should read this. www.normaltech.ai/p/what-will-...

normaltech.aiWhat will be left for us to work on?My keynote at ICML 2026

cs.princeton.eduWhat will be left for us to work on?ICML 2026 invited keynote — slides and edited transcript, presented click-by-click as delivered. Arvind Narayanan, Princeton University.

29 July

To understand and empathize with how workers in many or most fields outside software experience advances in AI capabilities, I propose a little thought experiment. substack.com/@aisnakeoil/...

Arvind Narayanan reposted

Rafael M Batista

@rafmbatista.bsky.social

18 July

This talk made its rounds earlier this week and with good reason. I finally made my way through it (in 6 min increments!) & found it wonderful. Many of the ideas felt new; including new frameworks to think about #AI + language to talk about it. Leaves you smarter, even if you disagree w/ it.

Arvind Narayanan@randomwalker.bsky.social

This is one of the most important questions about AI and one I've thought about. In countries like India, much of the economy is informal—heavily reliant on tacit knowledge that's not written down anywhere, processes never codified, "I know a guy who knows a guy" networks, and pervasive corruption.

Shriram Krishnamurthi@shriram.bsky.social

Reading Arvind Narayanan made me think: one of the great (and most undersung) achievements of the 20th century was rural electrification. If "AI is the new electricity", what is "rural AIification"? "Rural" here is obviously a metaphor, though it has some literal elements too.

newsletter.deenamousa.comAI could keep poor countries poorKicking over the development ladder

AI could keep poor countries poor newsletter.deenamousa.com/p/ai-could-k... by @deenamousa.com. Past waves of technology and globalization benefited poor countries because they got to export the products and services created through their cheap labor.

AI probably won't have this effect — it substitutes cheap labor even more cheaply. (Implication: this makes it even more important for them to find ways to use AI to improve their local economies.)

Shriram Krishnamurthi@shriram.bsky.social

arjunramani.comDispatches from IndiaWhat I learned covering the world's biggest country for The Economist

Two important and depressing readings on related points: Dispatches from India arjunramani.com/india-dispat... by @arjunr.bsky.social. There's a lot in there and I won't try to summarize it, but it helps explain why this project is not just a matter of building technology.

"Many of the basic facts needed to operate in Indian society are not recorded anywhere, but rather sit inside the heads of various information brokers who live off the rents from keeping that information private." In other words, codification will be actively resisted by rent seekers.

Shriram Krishnamurthi@shriram.bsky.social

This underground economy is trapped in a vicious cycle: low wages + extreme illegibility → no incentive for private actors to build AI for their use cases → everything stays informal + low productivity.

@thiagokrause.bsky.social

I don't know if governments and/or NGOs can help break this cycle, but if they can, its impact would easily exceed that of rural electrification. India has successfully built and deployed digital infrastructure for legibility in the past — Aadhaar and UPI — so there is hope.

Arvind Narayanan reposted

Thiago Krause

cs.princeton.eduWhat will be left for us to work on?ICML 2026 invited keynote — slides and edited transcript, presented click-by-click as delivered. Arvind Narayanan, Princeton University.

Very much worth reading. Too long and thoughtful to summarize here, but some of the main points are that LLMs will probably generate a lot of work, vaguely-defined “super intelligence” is unrealistic, and judgment will become ever more important.

Arvind Narayanan@randomwalker.bsky.social

cs.princeton.eduWhat will be left for us to work on?ICML 2026 invited keynote — slides and edited transcript, presented click-by-click as delivered. Arvind Narayanan, Princeton University.

If you find the premise of the question ridiculous I totally get it. But I assure you the question is on almost everyone's mind at AI conferences, and I believe in meeting people where they are. If we don't challenge the narrative it will continue to be embedded in the way the tech itself is built.

Arvind Narayanan@randomwalker.bsky.social

That's why I'm particularly delighted that the talk was so well-received, including by many people working at frontier AI labs. I have a lot more writing coming on the newsletter about AI's ability to augment and amplify human potential — if it is designed and deployed appropriately.

cs.princeton.eduWhat will be left for us to work on?ICML 2026 invited keynote — slides and edited transcript, presented click-by-click as delivered. Arvind Narayanan, Princeton University.

If you haven't read AI as Normal Technology, this talk is probably the easiest way to get a high-level overview. If you're already familiar with the core ideas, Part 1 of the talk is largely a summary of what I and @sayash.bsky.social have already written, while Parts 2 and 3 have new ideas.

normaltech.aiWhat will be left for us to work on?My keynote at ICML 2026

(There is also a static version of the slides + transcript on my newsletter in case the animated slides don't render correctly on your device: www.normaltech.ai/p/what-will-...)

cs.princeton.eduWhat will be left for us to work on?ICML 2026 invited keynote — slides and edited transcript, presented click-by-click as delivered. Arvind Narayanan, Princeton University.

Second, even though we should take recursive self-improvement seriously, there is no milestone that companies might achieve in the lab that will suddenly put us all out of work.

Third and finally, jobs of the future will be radically different, and a lot of adaptation will be needed. I shared my thinking about what this might look like and ended with a vision of human/AI “co-superintelligence”.

cs.princeton.eduWhat will be left for us to work on?ICML 2026 invited keynote — slides and edited transcript, presented click-by-click as delivered. Arvind Narayanan, Princeton University.