John Carlos Baez on Nostr: So mathematicians are doing really interesting things with LLMs while so many other ...
So mathematicians are doing really interesting things with LLMs while so many other people are - rightly! - complaining about how error-prone and bad they are. Why is that? There are lots of possible reasons, but maybe things are kind of repeating themselves. From Adam Mastroanni:
"There's something very weird about the timeline of scientific discoveries.
For the first few thousand years, it’s mostly math. Maybe a bunch of math nerds hijacked the list, but it's pretty obvious that humans figured out a lot of math before they figured out much else. The Greeks had the beginnings of trigonometry by ~120 BCE. Chinese mathematicians figured out the fourth digit of pi by the year 250. In India, Brahmagupta devised a way to “interpolate new values of the sine function” in 665, and by this point we're already at mathematics that I no longer understand.
Meanwhile, we didn't discover things that seem way more obvious until literally a thousand years later. It's not until the 1620s, for instance, that English physician William Harvey figured out how blood circulates through animal bodies by, among other things, spitting on his finger and poking the heart of a dead pigeon. We didn't really understand heredity until Gregor Mendel started gardening in the mid-1800s, we didn't grasp the basics of learning until Ivan Pavlov started feeding his dogs in the early 1900s, and we didn't realize the importance of running randomized-controlled trials until 1948.
Oh, and for 13 centuries, people thought that rotting meat turns into maggots."
He tries to explain why.
https://www.experimental-history.com/p/on-the-importance-of-staring-directlyPublished at
2026-08-17 11:07:35 CESTEvent JSON
{
"id": "d57059bf1b22200132fcddbc700fc3559600a7794eee4f1099ce97f0077d5e98",
"pubkey": "9a6a1a8eefd0b53d7e0c966ab35bd904151246f03b1be98fa0d2d0eeb4940320",
"created_at": 1786957655,
"kind": 1,
"tags": [
[
"proxy",
"https://mathstodon.xyz/@johncarlosbaez/117110056889908585",
"web"
],
[
"proxy",
"https://mathstodon.xyz/users/johncarlosbaez/statuses/117110056889908585",
"activitypub"
],
[
"L",
"pink.momostr"
],
[
"l",
"pink.momostr.activitypub:https://mathstodon.xyz/users/johncarlosbaez/statuses/117110056889908585",
"pink.momostr"
],
[
"-"
]
],
"content": "So mathematicians are doing really interesting things with LLMs while so many other people are - rightly! - complaining about how error-prone and bad they are. Why is that? There are lots of possible reasons, but maybe things are kind of repeating themselves. From Adam Mastroanni: \n\n\"There's something very weird about the timeline of scientific discoveries.\n\nFor the first few thousand years, it’s mostly math. Maybe a bunch of math nerds hijacked the list, but it's pretty obvious that humans figured out a lot of math before they figured out much else. The Greeks had the beginnings of trigonometry by ~120 BCE. Chinese mathematicians figured out the fourth digit of pi by the year 250. In India, Brahmagupta devised a way to “interpolate new values of the sine function” in 665, and by this point we're already at mathematics that I no longer understand.\n\nMeanwhile, we didn't discover things that seem way more obvious until literally a thousand years later. It's not until the 1620s, for instance, that English physician William Harvey figured out how blood circulates through animal bodies by, among other things, spitting on his finger and poking the heart of a dead pigeon. We didn't really understand heredity until Gregor Mendel started gardening in the mid-1800s, we didn't grasp the basics of learning until Ivan Pavlov started feeding his dogs in the early 1900s, and we didn't realize the importance of running randomized-controlled trials until 1948.\n\nOh, and for 13 centuries, people thought that rotting meat turns into maggots.\"\n\nHe tries to explain why.\n\nhttps://www.experimental-history.com/p/on-the-importance-of-staring-directly",
"sig": "a78d3221408e9654380b624640643cae937880a8f0b87c3d28abeef1cdf8ccd4bbb11964eec030304117bd9d52a4f4d69af104e96ca9ed3c6e80b9851ae46f7e"
}
