taking notes on PDE as a outmoded subject, thought criminal
Zembla
Joined July 2021
- Tweets1.5K
- Following149
- Followers105
- Likes88.8K
sloth12 retweeted
It's also notable that the coterie of mathematicians that OpenAI courted doesn't seem to be involved. Instead there is one relevant scientist (Jaffe) and Gross (?) Nekrasov (?) Douglas (???)
sloth12 retweeted
Bitter lesson 2: AI is better than humans at AI alignment research. (See attached AI-generated proof.)
sloth12 retweeted
sad part is finding out presidents of the united states care about their ego than actual progress in world peace. this is just straight coping. i have lost all respect for presidents of the united states 😿
sloth12 retweeted
In the Q&A session Buckmaster says Levent had access to internal models and compute of Anthropic during their work. As most inferred/suspected, of course. Though sometimes omitted in discussions.
sloth12 retweeted
Don't bullshit with your advisory group.
Just enforce the data protection and condidentiality and give universal access to researchers.
Your work is to build models and our work is to solve the problems. We are very good, far better than you, at providing readable and understandable solutions, so enhancing your product.
For marketing reasons you decided to solve the problems yourself. You are doing a poor work. Your solutions are BS. Your marketing impact will be higher if your customers can show enhanced the power of the tool you provide.
We’re working with an independent advisory group of mathematicians to help OpenAI responsibly share advances in AI and mathematics.
The group will advise on how we assess and communicate new mathematical results, uphold academic and professional standards, and build tools that support mathematical research and learning.
Through this work, we want mathematicians to be at the center of shaping how AI supports mathematical understanding and how its benefits reach the wider community.
openai.com/index/advisory-gr…
sloth12 retweeted
According to a new paper of Constantin-Ignatova-Vicol, the OpenAI Navier-Stokes blow-up construction cannot give you a forcing term which is analytic. In particular, it does not work with zero forcing - as in formulation (A) of the Clay problem.
sloth12 retweeted
interesting; physics suggests almost all problems have inelegant solutions (of the form "solve this PDE")
sloth12 retweeted
OpenAI: we will alienate mathematicians from mathematics. We will do a worse job and spend more money, and you will like it
Mathematicians: we are not fans of that
Englightened centrists: Let us teach the controversy
sloth12 retweeted
Extremely sensible. Peter Thiel's olympics of drugs right there. You can count on Yau to do the right thing.
Among the oldest Deligne signed but Yau and Serre didn't. I note that Yau is launching Journal of AI-Assisted Mathematics intlpress.com/journals/journ… and singatories Yu Deng and Kontsevich are on the editorial board.
sloth12 retweeted
This reference list for Navier Stokes blowup is outrageous.
No Chen-Hou, no Alpoge-Buckmaster, no Córdoba and Martínez-Zoroa, no Elgindi.
They are not properly attributing any previous work to the problem
sloth12 retweeted
Mathematics is a human activity, there is no contradiction. A significant chunk of us will simply not pay attention to proofs which do not yield insight. You can throw money at academia to bend to AI digestion, good luck with that though! We would choose Exit.
sloth12 retweeted
I could care less what Mr. Kalshi advertisement has to say about my life choices lmao
Timothée Chalamet says there's something wrong with you if you're a Nets fan in New York
But he means no disrespect
"My grandma was a huge Knicks fan, so that's what turned me on to them."
"She had a New York Giants teddy bear and a New York Jets teddy bear, but she only had a New York Knicks teddy bear."
"No disrespect to Nets fans... If you're a Nets fan, you should be on like a government list or something."
"There's something wrong with you."
