📢 Excited to share my latest work: "A Perspective on the Regularity Theory of Degenerate Elliptic Equations"
https://t.co/P7aAUJrTRk
🧵Here's a thread unpacking the main ideas! 👇 #PDE#MathResearch@cimatoficial
@Frances01896069@ScienceOrMyth Frencesco. What’s your take on the referee problem? I don’t think it has been discussed enough. If papers are (by default) AI-generated, what incentive do we have to referee them? The author’s role starts to look very close to the referee’s, but with a very different reward.
¡Te invitamos a la séptima sesión del Seminario de Matemáticas Aplicadas! 😉
Este miércoles de septiembre, elDr. Javier Gómez Serrano impartirá la charla: "Modern Mathematics in the Age of Artificial Intelligence".
👉Enlace:
https://t.co/X0IdzTr0BL
Autoformalization is now very easy.
You can formalize long PDE or probability papers (including many prerequisities, none of which are probably in mathlib) in 24-48 hours. Harder projects can take more time, of course, but not much more.
I wrote a blog post explaining how to do it.
I'm also making all my lean skill files public.
Great interview!
Brady clearly did his homework. He brought up a bunch of questions and follow-ups I was genuinely curious about and had been hoping someone would ask.
NEW INTERVIEW
Tristan Buckmaster finds himself at the eye of massive storm over (denied) claims that AI stole his work on its way to solving Navier-Stokes
Oh, and he has a two-month-old son at home!
The man is not getting much sleep.
https://t.co/1w1GzLR31g
NEW INTERVIEW
Tristan Buckmaster finds himself at the eye of massive storm over (denied) claims that AI stole his work on its way to solving Navier-Stokes
Oh, and he has a two-month-old son at home!
The man is not getting much sleep.
https://t.co/1w1GzLR31g
¿Haces ciencia, tecnología o cultura y necesitas un impulso para tu proyecto? Fundación #BBVA lanza las Becas Leonardo: 50 apoyos individuales de hasta 50,000 USD para impulsar tu trabajo.
➡️Revisa los detalles del lanzamiento y los requisitos aquí: https://t.co/crwTVnC1YV
@mathandcobb One step further: solve all the problems. No need to disclose the LLM or the author. Revive Bourbaki!
Actually, no need to solve them either: “solve and verify” should be enough for a couple of months. Once everyone figures out the trick, repeat.
In the mid-19th century, a brain teaser about map colors quickly turned into an obsession. Today, the hunt is still on for a simpler solution to the deceptively tricky problem. Recently, a team of mathematicians made progress. https://t.co/be4wXAIY8l
Happy, those able to know the causes of things.Really insightful post by Nestor on how to frame the current progress in mathematics by LLMs and by AI companies.
https://t.co/yxwSY0Efcl
We are seeing a new trend in submissions to @arxiv (and presumably to conferences and journals): Authors submitting papers whose contents they likely do not understand. 1/
Assuming we have the money, what we need now is the creativity and courage to propose a better future. We got tons of that, the money will follow once we show what's worth. Dream big!
Do we defend math as it has always been done, or use this opportunity to reimagine it?
Thanks, Saraí!
I like the direction you’re pointing in, and I’d dare to add: If we have the opportunity to start from scratch, how we do things better?
2/ the change we want to happen, because we do have a say. 🔥
I encourage everyone to create spaces in institutes and departments for discussing the future of math in our context. The conversation is not the same in Mexico as it is in Canada.
We face many problems (inequalities, evaluation metrics, etc) whose solutions have been constrained by money and the momentum of tradition. That momentum is shaken, possibly broken.
2/ the change we want to happen, because we do have a say. 🔥
I encourage everyone to create spaces in institutes and departments for discussing the future of math in our context. The conversation is not the same in Mexico as it is in Canada.
Statements of AMS https://t.co/hHdl71llVa, LMS https://t.co/LjQxn9LsrF and EMS https://t.co/9fupdKFdez for NS. IMU does not write a statement so far, basically because I am slow……