More cherry blossoms in Hyde Park! Last week was mostly pink, this week they're white. Also featured: the (almost complete) Obama center.
Posts by
Cherry blossoms in Hyde Park! Probably the peak will be next weekend.
Definitely the worst person whose opinion keeps getting shoved in my face on a regular basis every time I read the NYT (with the possible exception of Ross Douthat).
Anyway, there's a standard reference for this discussion, which is Bill Thurston's famous article "On Proof and Progress in Mathematics", see e.g. here: flm-journal.org/Articles/1CC...
A lot of this goes to the question of what we actually want from the process. Fundamentally, mathematicians don't "want" proofs, they want *understanding*. Some (a lot) of understanding comes from proofs, which is why they're desirable. But uncoupling proofs from understanding doesn't make sense.
There are at least a couple of ways to try to get around this. 1: have the LLM output the proof in lean code, so that it can be formally verified as correct. 2: try to figure out how to have the LLM generate good (informal) mathematical ideas that are humanly readable/understandable.
Now suppose you're deluged by thousands of papers of this kind. It's almost impossible to get any "value" out of them, even if some of them are right - which ones?
It depends how hard the solution is to check. If I referee a paper and it's wrong, it's usually easy to tell it's wrong because something about it doesn't make sense. If an LLM has been optimized to chain together statements that sound very much like they make sense, very subtle errors can creep in.
This paper was linked to by one of the commentators, and seemed interesting to me at a glance. One of the authors is Jordan Peterson; nevertheless it's still logically possible that the paper has some scientific value.
www.yorku.ca/mar/Mar%20et...
Unfortunately Abebooks is owned by Amazon, so I boycott them too. They obviously go out of their way not to advertise this on their website, so it's up to us to let people know.
Unfortunately AbeBooks is a subsidiary of Amazon. I boycott them too (for the same reason).
Thanks so much for this! Just bought 4 books on thriftbooks!
Idk who owns it, but been very happy w/thriftbooks. I guess it could be benefiting someone awful on the back end but afaik, I’m just participating in used book/secondhand economy like on eBay: keeping stuff out of landfills & money in small hands. They *always* have a copy of whatever obscure thing.
I applaud the sentiment, and I have been boycotting Amazon for 14 months now. But I find bookshop.org very hard to use, especially for any rare or technical books. Personally I have had more success with Powells.com and moesbooks.com and even occasionally ebay.com. Any other suggestions?
Origami curve on the sphere - it's possible to fold the sphere over itself in such a way that this curve folds up over itself at the same time!
Mathematician here. There is no issue of falsification. The papers were simply incorrect, ie there were logical errors. This is unfortunate, and four prominent examples speaks to carelessness, but there is absolutely no reason to think it represents anything unethical.
This is one of the more evil proposals I have ever read. Where is the outrage on this? Where are my senators? Dick Durbin? Hello??
www.nytimes.com/2026/03/16/h...
Voted early in Chicago for Juliana Stratton (Senate) and Robert Peters (IL-02). Appalled to see all the crypto dark money propping up their opponents in the Democratic primary.
"But late Friday .. OpenAI said it had reached its own agreement with the Pentagon to provide its A.I. for classified systems. OpenAI said it had found a way to put safeguards into its technologies that would somehow prevent the systems from being used in ways that it does not want them to be." 2/2
"Sam Altman, the chief executive of OpenAI, said in a memo to employees this week that “we have long believed that A.I. should not be used for mass surveillance or autonomous lethal weapons,” which is the same stance as Anthropic’s." 1/2
You know if you fold a piece of paper and make a cut and unfold it you get a cool shape? What if you do the same fold 10 times before you make the cut?
This is a great idea! I’m a big fan of Powells when I’m in Portland but I always forget about their online store!
All I can say is: youtu.be/v4if--NgcQg
Deleted my Facebook and Instagram accounts, canceled my Amazon prime account, boycotted Amazon and Whole Foods since the inauguration. Was never on X and never will be. Never subscribed to any paid AI service. But I can and should do more, and so should you.
www.resistandunsubscribe.com
Article about the speed with which China is rushing past the US in every aspect of their economy, including (especially) tech. Tracks with my own experiences in Beijing in 2025 seeing the unbelievable depth of talent in the math department at Peking University.
www.nytimes.com/2026/02/10/o...
Ironically, this pro-crypto website is a great resource to identify politicians who want to properly regulate crypto so that you can give them your support. www.standwithcrypto.org/politicians
(2/2)
www.nytimes.com/2026/01/15/t...
Crypto - the currency of choice for criminals and nitwits. The good news is that Republicans and Democrats are united in their regulatory response to this lawless behemoth. The bad news is that they both embrace it.
(1/2)
Credit: David Smead; excerpted from www.jstor.org/stable/25678...