Noah Stephens-Davidowitz

Nerd, computer scientist ( Cornell CS prof. I spend a lot of time thinking about lattices and sometimes other things.

[Noah Stephens-Davidowitz @NoahSD]

@ml_fest_at Yeah. I comment on that in the post. It makes sense technically, but
1) more-or-less any mathematical notation is formally equivalent to some kind of set membership; that doesn't mean we should write it as such; and
2) this hasn't caught on because it doesn't match our intuition.

[Noah Stephens-Davidowitz @NoahSD]

I wrote up a little blog post proposing a slightly different way to write asymptotic notation. https://t.co/Hhf49sT4nd In short, I think asymptotic notation should often be written with an INequality. E.g., f(n) <= O(n^2), f(n) < o(log n), f(n) > 2^{-o(n)}, f(n) >= n^{-O(1)}.

[Noah Stephens-Davidowitz @NoahSD]

I know that this is not an original idea. It seems that many authors use this notation already, at least in some contexts. I have not seen anyone advocate for its widespread adoption, so I thought I would.

[Noah Stephens-Davidowitz @NoahSD]

And, using these conventions means that authors must think a bit about what they actually mean when they use asymptotic notation, which is a good thing.

[Kyle Cheney @kyledcheney]

Left: Interim U.S. Attorney Ed Martin dismisses Jan. 6 case of Jose Padilla on Jan. 21, 2025

Right: Interim US Attorney Ed Martin seeks to withdraw as Padilla’s counsel of record on Feb. 5, 2025

He signed a dismissal for a client he was still recorded as representing.

[Noah Stephens-Davidowitz @NoahSD]

Almost three years later, I found a proof that I really like and wrote it up formally: https://t.co/bHwyVjG8Pz .

[Noah Stephens-Davidowitz @NoahSD]

Dumb question that's bugging me: What's a really nice clean proof that 2^n grows faster than any polynomial?

(Clean is obviously ill-defined here, but I mean short, intuitive, and using as little fancy math as possible.)

[Interintellect 🧭 @interintellect_]

A look back at The Future of Publishing: A festival in New York City. @SethS_D discusses writing a book in 30 days with the help of AI.

[Gautam Kamath @thegautamkamath]

Congratulations to @Eshan_chattopad and David Zuckerman on winning the Held Prize! For their groundbreaking work on randomness extraction.

[Huck Bennett @huckbennett]

Day 1 of professing at CU Boulder.

[Timothy Gowers @wtgowers @wtgowers]

A great preprint appeared on arXiv this morning by Marcelo Campos, Marcus Michelen, Julian Sahasrabudhe and Matthew Jensen -- the first improvement by more than a constant factor to the lower bound for sphere packing in large dimensions since 1947. 1/10

https://t.co/uKvM0rfaWA

[Seth Stephens-Davidowitz @SethS_D]

First review of Who Makes the NBA, from one of my heroes @tylercowen, who says "It's quite good!" and "I really liked this book." https://t.co/CHPOp5v0RO

[Seth Stephens-Davidowitz @SethS_D]

My new book, Who Makes the NBA, is out today!!!! This is one of my favorite charts from the book: the advantage fathers pass to their sons in various fields. I expanded it based on twitter comments. https://t.co/9LFhaxpqDV

[Seth Stephens-Davidowitz @SethS_D]

My 3rd favorite chart from Who Makes the NBA (out today!!!) https://t.co/h1SMWpdCzV

[Seth Stephens-Davidowitz @SethS_D]

DAY 3 OF MY CHALLENGE TO WRITE A (GOOD) BOOK ON THE NBA IN 30 DAYS, THANKS TO AI's DATA ANALYSIS TOOLS. Have you ever wondered: Who would be the best basketball player of all time if everybody were the same height? 🧵

[Hanlin Ren @dfc7027894e168c]

Amazing: Zeyong Li showed that the complexity class S_2 E requires near-maximum circuit size on 𝐚𝐥𝐦𝐨𝐬𝐭 𝐞𝐯𝐞𝐫𝐲 input length, as well as a "genuine" (no infty often caveats) pseudodeterministic ZPP^NP algorithm for range avoidance!
https://t.co/DuVOK6ht6P
(1/4)

[Noah Stephens-Davidowitz @NoahSD]

@CornellCIS @Cornell_CS I wish I could do something half as cool as this awesome work by @Jimantha .