AI and the risk of catastrophe
I am reminded of an episode from history involving Bertrand Russell and the U.S. writer and editor Leo Rosten. In the early 1950s, Rosten had approached Russell in London to write an article on agnosticism…
artificial intelligence
Verifiability, access, and allocation in maths
When I was talking about this post with a friend, he asked a thought-provoking question. In the post, I’d argued that I believe it should be appropriate for mathematics to consider a solution to a…
Privileged access is bad for science
Despite some problems of its own, the ongoing Microsoft Majorana saga has been striking a pleasant contrast with the rush by artificial intelligence (AI) companies like Anthropic and OpenAI to show their proprietary models can…
Lesson learnt? None.
A particular line from the open statement mathematicians wrote against the design of the new Caltech Mathathon, due to happen in late October, has stuck with me: Mathematicians do not only prove theorems: they also…
The bridge AI-writing builds
From ‘Why not normalise use of AI in writing?’, by Amit Kumar, assistant professor in the Institute for Development and Communication, Chandigarh, in The Hindu, August 29, 2026: Whose authority is being protected by that…
The mathematician and the oracle
On August 24, Levent Alpöge (@__alpoge__) posted on X.com that they had solved a maths problem with Claude’s help in a way that was contrary to a human proof of it, albeit a reputedly unsatisfactory…
AI shouldn’t distract journalism into a new trap
The following is a review of Journalism in the Age of AI: From Acceleration to Reimagination by Seth C. Lewis, Rodrigo Zamith, and Tomás Dodds, available here. The authors write, and I agree, that journalism…
The Meta Ray-Ban glasses and Indian law
1. India is not likely to find success by regulating hardware because (i) such regulations will be hard to enforce in practice and in principle, (ii) the regulatory regime could set up a new license…
So you use AI to write…
You’re probably using AI to write. Both ChatGPT and Google AI Studio prefer to construct their sentences in specific and characteristic ways and anyone who’s been a commissioning editor for at least a few years…
AI slop clears peer-review
Here’s an image from a paper that was published by Nature Scientific Reports on November 19 and retracted on December 5: This paper made it through peer review at the journal. Let that sink in…
Using AI to fight misinformation
In his latest newsletter, Bruce Schneier springboarded off of articles in Washington Post and The Atlantic to write: There’s a balance between the cost of the thing, and the cost to destroy the thing, and…
Who funds quantum research?
An odd little detail in a Physics World piece on Microsoft’s claim to have made a working topological qubit: Regardless of the debate about the results and how they have been announced, researchers are supportive of the efforts…