Mark Carney on Elon Musk supporting Quebec sovereignty: "The future of Canada will be decided by Canadians and the priorities of Canadians, the priorities of the people of Quebec."

One of the three withdrawn manuscripts was “Algebraicity of Kuga–Satake Correspondences for K3 Surfaces” which I had started trying to understand; it’s a question I quite like. The ideas seemed to be exciting, though of course this is no longer clear. In cases like this with a human-authored manuscript, one can often still extract something interesting from the wrong argument (which is often still based on some perhaps-correct intuitions of the author). It’s less clear if that is possible here, and my experience is that AI is not as helpful for this kind of fuzzy task as one might hope.

“In a three-month randomized controlled trial with practicing patent attorneys, we found that, while routinely using AI lifts work quality across the board, its effects on on-the-job learning depend on seniority: senior lawyers who used AI for 90 days demonstrate stronger judgment than those who didn’t, whereas junior lawyers show no average skill gain, their scores splitting into more strong scores and more low scores.”

amidst the hoopla, the social constructivist view of mathematics completely slipped my mind math as an invention rather than a discovery; an abstract formal language, created differently in different parts of the world (ethnomathematics is a fascinating field!), with which people seek to describe patterns and make sense of the world in that view, AI-generated proofs are useful (primarily to mathematicians), albeit limited in their utility, and also completely miss the point when used as anything more than fancy solvers

So, it seems OpenAI is about to release something like 400 AI-generated proofs on a public server. To the best of my understanding, people are still trying to figure out how the forced Navier–Stokes example actually works. I know of several experts who are studying it, and it is definitely not an easy task. Until mathematicians digest that example, however, its impact on mathematics as a living subject seems necessarily limited. What questions does it open? What new ideas does it suggest? What does it teach us about fluids, beyond the yes/no information contained in the theorem itself? Mathematics does not grow simply by accumulating correct statements. Results have to be understood, connected, explained, challenged, reused. Until that happens, they risk remaining lettera morta. Now imagine adding 400 more AI-generated proofs. For each of them, until some human mathematician picks it up, studies it, and connects it to the existing mathematical culture, isn't it in essentially the same position? So my question is genuinely: what is the intended mathematical value of releasing hundreds of proofs at once? What happens if there simply aren't enough people willing or able to read them? To be clear, I think mathematicians should discover mathematics by any means necessary, including scavenging through artificial, alien-looking mathematical material, since there may be extraordinary things to find there. But humans are not machines, as attention, understanding, taste, and mathematical culture are scarce resources. What happens if the rate at which mathematics is generated becomes much greater than the rate at which mathematicians can absorb it?

I need to retweet this again for every round of AI math discourse where people start castigating ungrateful mathematicians' ungrateful mourning. It captures aptly what they are mourning. (Note that mourning a lost past does not mean stopping the future.)

I believe in empowering people to fulfill their potential and do what they love. That necessarily means the volume of output of meaningful work will increase, even in the case where people choose to take more time to do higher-quality work. I also believe in giving people tools that make them more effective in their work, to improve focus, reduce toil, and extend human capabilities. That, too, increases the rate of production, again, even if people on an individual level take more time to learn, experiment, and produce at a high level of quality. These changes necessarily create new problems, up to and including resource pressure and overflow downstream that must be managed, as a byproduct of more people having a better quality of life and greater access to opportunity. This position is not accelerationist. The core objective is not to produce more, faster, as a good in-and-of-itself. Accelerationists would not give a flying fuck about managing and mitigating the impacts, and would claim the problems would solve themselves as long as you continue increasing the rate of change. This position does recognize that a likely second-order effect of mass empowerment (the primary goal) is an increase in the production of information that needs to be addressed elsewhere in the system. That is responsible. Handwaving at ~the environment~ as a reason why the status quo is good enough additionally does not track or resonate with me as someone who worked directly on climate change mitigation and adaptation strategies. The goal is not for eight billion people to live in dirt hovels cooking off coals. It is for more people to live longer, fuller, healthier, safer lives doing what is most fulfilling to them in ways that also make most efficient use of the resources we have available.

I think there’s a lot of truth to this. I welcome these results but if we want humanity to understand them it’s also important to support and reward human expertise in mathematics, and we (the mathematical community and society at large) need to find ways to do so.

it’s hard to not see much of the hypebros’ excitement about AI results like this as a sort of giddy desire to see expertise torn down, and experts torn a new one. there’s a kind of promise of anyone being able to do anything without putting in any hard work or effort that the marketing taps into, a sort of ‘get rich quick’ notion but for knowledge work, where all you need to do is ask the right question and the machine does it all for you. the very suggestion that it still doesn’t work that way and you have to do some work and know what you’re talking about to interpret the output really pisses those folks off.

scientific publishing was already experiencing an unprecedented overflow and validation crisis; dropping 722 papers with bold claims as an engineering-as-marketing project makes this an order of magnitude worse. if your idea of science is just a list of splashy headlines about Major Breakthroughs!, you’ve completely missed the point.

To Ed’s point, there is woefully little these days in terms of either material analysis of AI or a Walter Banjamin-esque assessment of how digital reproduction will change the act of human production, creative or otherwise

I can only assume OpenAI's executive staff had a meeting whose goal was to locate novel ways of getting the public to hate them to add to the already teetering pile and landed on 'let's have our main policy guy announce he wants to take away Social Security'

There’s a new kind of AI-enabled Dunning-Kruger where a little bit of knowledge and basic competence, paired with the robots’ abilities, can convince a layperson they are equivalent in skill to an expert. We had a bit of this with engineers and Google-fu, but it’ll get worse.