Hacker News: eastWestMath

New comment by eastWestMath in "A new kind of symmetry shakes up physics"

eastWestMath — Tue, 18 Apr 2023 18:02:59 +0000

Is this some flavour of groupoids symmetry?

New comment by eastWestMath in "Llama.cpp 30B runs with only 6GB of RAM now"

eastWestMath — Sat, 01 Apr 2023 07:51:05 +0000

I was in a PLT group in grad school going into robotics. I could spend all day ranting about how Python is just completely unsuitable for professional software development. Even something like F# would be an enormous improvement.

New comment by eastWestMath in "NYPD is refusing to comply with NYC’s new surveillance tech laws"

eastWestMath — Fri, 31 Mar 2023 15:41:32 +0000

As a police officer you also typically have a high school education, and your idea of whether or not something is constitutional carries little to no weight.

New comment by eastWestMath in "Taichi lang: High-performance parallel programming in Python"

eastWestMath — Fri, 10 Mar 2023 07:03:33 +0000

I used to be very optimistic about Julia, but my enthusiasm has really cooled off the last year. I find JAX+jax_dataclasses gives me 90% of what I wanted from Julia (a nice monad library that is compatible with jit’ed code and the ability to cache compiled code would make it a solidly better language).

New comment by eastWestMath in "Google stopping coding competitions"

eastWestMath — Thu, 23 Feb 2023 02:45:31 +0000

These are all, of course, things one should learn in an algorithms class.

New comment by eastWestMath in "Google stopping coding competitions"

eastWestMath — Wed, 22 Feb 2023 19:48:17 +0000

Yeah I’ve always been a bit leery of code competitions. I saw a lot of undergrads spend a lot of effort on these contests, and I was never convinced it was the best use of their time (especially because the university I did my graduate studies at had a professor who put a lot of effort into running the competitive programming club).

I think a group of 15-20 bright and enthusiastic CS students under the guidance of a tenured CS professor could accomplish a lot with the time/effort they were spending and learn just as much!

New comment by eastWestMath in "Symbolic Discovery of Optimization Algorithms"

eastWestMath — Sat, 18 Feb 2023 04:43:29 +0000

> Our method discovers a simple and effective optimization algorithm, Lion (EvoLved Sign Momentum).

Come on now

New comment by eastWestMath in "[dead]"

eastWestMath — Sat, 11 Feb 2023 15:56:58 +0000

I didn’t make it past that point, I got confused about how such meaningless dreck was at number 2(!) on my feed.

Mods may want to take a peek if the author used a few sock puppets to give this post a push...

New comment by eastWestMath in "David Sabatini lands millions from private donors to start new lab"

eastWestMath — Tue, 07 Feb 2023 21:14:26 +0000

Yeah, it’s spending a lot of money to start a lab that:

- will probably be stonewalled out of any collaboration with reputable universities (because his reputation is so toxic),

- will have trouble attracting talent (because his reputation is so toxic),

- people will be ready to believe any sort of accusation of impropriety coming out of that lab from sexual harassment to more general academic fraud (because his reputation is so toxic).

New comment by eastWestMath in "In a moment, mathematicians merge probability and number theory"

eastWestMath — Sun, 15 Jan 2023 22:04:15 +0000

The paper is absolutely a construction on categories.

New comment by eastWestMath in "In a moment, mathematicians merge probability and number theory"

eastWestMath — Sun, 15 Jan 2023 22:01:35 +0000

“I’ve discovered a new thing about Y in X’s! Oh, no I haven't actually looked at the X literature about Y.” This is hardly specific to category theory.

New comment by eastWestMath in "In a moment, mathematicians merge probability and number theory"

eastWestMath — Sun, 15 Jan 2023 21:51:34 +0000

My issue is that I can’t see a cursory literature review into the established category literature on these methods - how do they even know they have something new/novel if they haven't checked?

This is a bit like the recent news about the “ML discovered a new fast matrix multiplication” press push out of Google, where researchers in the field just sort of rolled their eyes.

New comment by eastWestMath in "In a moment, mathematicians merge probability and number theory"

eastWestMath — Sun, 15 Jan 2023 21:45:58 +0000

Oh, I’m talking about the paper they’re hyping up. I’m just not terribly impressed, and I think if was written by people at UCSD instead of Harvard nobody would have even noticed it.

New comment by eastWestMath in "In a moment, mathematicians merge probability and number theory"

eastWestMath — Sun, 15 Jan 2023 20:02:39 +0000

Coming from category theory, and having followed the work studying probability theory categorically (e.g. Fritz, Perrone, Lucyshyn-Wright’s stuff, Leinster’s notion of the magnitude of a category), is this actually interesting? The actual category theory seems pretty rudimentary, and I’m not sure if they’ve invented anything new on that side (they don’t reference any of the authors I’ve mentioned, so it doesn’t seem like there was much scholarship on that end).

It feels like Quanta just takes press releases from a few big departments and asks a writer to get a few quotes to flesh out an article.

New comment by eastWestMath in "PyTorch 2.0"

eastWestMath — Fri, 02 Dec 2022 19:28:59 +0000

It reminded me of the sort of lazy Wikipedia regurgitation that a lot of undergrads used to give when I was teaching. So it is a bit jarring to see a response like that in a non-compulsory setting.

New comment by eastWestMath in "The dangers of high status, low wage jobs"

eastWestMath — Wed, 09 Feb 2022 20:18:43 +0000

I feel like the NYTimes could take more papers under its subscription - e.g. give subscribers the option to bundle in a state and municipal paper, that the NYTimes essentially vets as a local affiliate (I’m guessing this would probably lead to substantial overhead on their end, though).

New comment by eastWestMath in "National Collegiate Athletic Association vs. Alston [pdf]"

eastWestMath — Mon, 21 Jun 2021 20:17:53 +0000

Just because someone doesn’t go to college at 18 doesn’t mean they’re “skipping” higher education. Spending a few years making 30k a year to play basketball would be awesome - my father had teammates in college who put off grad/med school to go play in a shitty European leagues or random minor leagues around North America, it never sounded like they regretted it.

If they followed Hockey’s example, where playing a year in the MJ gives you a year of tuition, it could work out great for young players.

New comment by eastWestMath in "Cheerleader’s Snapchat rant leads to ‘momentous’ Supreme Court case on speech"

eastWestMath — Mon, 26 Apr 2021 22:36:49 +0000

> Should she have been fired on the spot without even having any form of communication with her employer?

Too be honest, that depends entirely on the joke and her role, doesn’t it?

New comment by eastWestMath in "Public-ownership rental as a third option to renting or owning a house"

eastWestMath — Sun, 18 Apr 2021 23:24:54 +0000

Banking seems like the best way to phrase this to prevent people from screaming about socialism.

New comment by eastWestMath in "Public-ownership rental as a third option to renting or owning a house"

eastWestMath — Sun, 18 Apr 2021 23:20:42 +0000

Ugh.