serious-discussion

[blazing pawn]

I have social anxiety I think but I dont get anxious about like big or existential things

Like "the future" or whatever

[modest rune]

Oh I mean more like

How are you with loss of control

And feeling kinda out of it

Do you panic easily

[sharp mulch]

Kinky

[blazing pawn]

[modest rune]

Go away moth is a child

[blazing pawn]

Idk thats a good question

I do like being in control of my environment but thats not so much for anxiety reasons as it just sensory overload/hyperacusis/etc

[modest rune]

Yeah and also like

Even the chillest people can have bad experiences

Not me

But the rest

[blazing pawn]

Built different etc etc

[modest rune]

Yes

[cold needle]

Constructed unusually.

[blazing pawn]

Yeah idk I dont do well with sensory overload so I'm not sure psychedelics would be smart

[modest rune]

Yeah maybe not

[sacred geyser]

yeah you wouldnt like it then, i pretty much always have bad trips from sensory overload

[blazing pawn]

I am not opposed to drugs it is simply that a combination of coincidences have made them incompatible with my life and brain

[modest rune]

So what ur saying is ur lame

[inner finch]

the future scares me

[blazing pawn]

Yes

[devout nacelle]

I scare the future.

[sacred geyser]

good thing its not here yet

[inner finch]

tru

[cold needle]

I am the future.

[inner finch]

the future is, however, not real

[cold needle]

i am so scared of manan

[sacred geyser]

so ur just getting scared by nothing

[inner finch]

yes

[sacred geyser]

based

[inner finch]

i used to be so afraid of everything

i went to a haunted house as a kid and i was afraid af

[cold needle]

i used to be afraid of my teachers

[inner finch]

i ran into the operator guys room

and hid

[devout nacelle]

...said no one ever, even the little kids in my extended family bully me

[blazing pawn]

Idk what im afraid of

[inner finch]

As hey should

sahi karte hain

[inner finch]

ssocti

sociey

...

[sacred geyser]

im afraid of it being too warm out

[devout nacelle]

I'm very scared of conflicts.

[blazing pawn]

I am kinda afraid of teachers sometimes because of [REDACTED]

[sacred geyser]

i should have decided to be born earlier then

[blazing pawn]

but thats not like

real fear

Just distrust

[devout nacelle]

I try my best to avoid them in most situations

Even when I have to heavily compromise on a lot of things

[inner finch]

i dont think im vrey scared of anything

except typos

[cold needle]

sociey

[devout nacelle]

Morza

[inner finch]

[sacred geyser]

im afraid of experimental physicists

[inner finch]

sause

[blazing pawn]

Yeah I think im distrustful of a lot of things but Im not particularly afraid of them

[devout nacelle]

Oh, I'm also afraid of butterflies

[inner finch]

[sacred geyser]

same

[inner finch]

bro butterflies are afraid of me wdym

[devout nacelle]

For some reason insects which fly around rapidly scare the hell out of me

[sacred geyser]

butterflies too fast and large i dont like

[cold needle]

take an entomology course manan you might become less scared

[devout nacelle]

Right

[inner finch]

im generally not afraid of insects

[devout nacelle]

They look pretty on paper

[cold needle]

i like my insects now

[inner finch]

even the cockroach type things

[cold needle]

they r wigglys

[devout nacelle]

Unfortunately I have a physical form Ultra

[inner finch]

wigglys

[sacred geyser]

ultra is too busy gaslight gatekeep girlbossing to have fears

[devout nacelle]

Cockroaches are fine

[cold needle]

my cute wigglys turned into pupas

[inner finch]

😌

[cold needle]

they become butterfly soon

[devout nacelle]

😌

[inner finch]

botterfinger

[sharp mulch]

I see the stars in your eyes believe in your lies aurora

[devout nacelle]

[inner finch]

what was that

[devout nacelle]

I just thought Ange was shilling something from ORV

I should try it out

[inner finch]

what is orv

[sharp mulch]

I see you light up the sky a dance in the night aurora

ORV is Omniscient Reader's Viewpoint

[inner finch]

oh

[devout nacelle]

Also Ze Tian Ji

But so many chapters

[sharp mulch]

[devout nacelle]

Light mode

[sharp mulch]

Is me

[inner finch]

the sus moon

[sharp mulch]

Maybe I'll update my status sometime soon

[devout nacelle]

Oh yeah I should do that too

Been stuck up on Untamed ever since

[sharp mulch]

Hehe

[devout nacelle]

Good drama though

[blazing pawn]

Ange what are your thoughts on solo leveling

I read the first chapter

It was bad

[mortal oasis]

See

The real thing to realize is

[devout nacelle]

Too bad they didn't let it be on screen gay romance

[mortal oasis]

All anime is bad

[blazing pawn]

Its not an anime

[sharp mulch]

SL is pretty generic

The plot is like

Nonexistent

[blazing pawn]

the first chapter opened with an elaborate explanation of the characters personality and relationships

So boring

[sharp mulch]

It appeals to the male fantasy of being buff

SJW is hot though

And the art is good

But the plot is like

Negative

In comparison, ORV has hot chars (YJH and LHS), has good art, and has a good plot as well

[inner finch]

[sacred geyser]

are these webcomics

[mortal oasis]

Weebcomics*

[sharp mulch]

ORV was a webnovel

[bronze pelican]

weebcomics

[sharp mulch]

It is being adapted as a manhwa right now

Also it's Korean not Japanese ok

[modest rune]

Whats sjw

[sharp mulch]

Char

[modest rune]

Huh

[sharp mulch]

Character

[modest rune]

Oh the main character

Or a side character

I get you

[sharp mulch]

Sung Jin-Woo

[sacred geyser]

gotcha; idk why but i was thinking manhwa or webcomic or sumn

[sharp mulch]

Main character

[bronze pelican]

still weeb

[inner finch]

korean weeb

[sharp mulch]

Sorry it's Koreaboo

Honestly

[modest rune]

Isn’t that the mc from elected

[inner finch]

kweeb more like dweeb

[modest rune]

Eleceed*

[sharp mulch]

Get your creepy obsessions right

[pale orchid]

that's pretty good

[bronze pelican]

kweeb

[sharp mulch]

Seo Ji-Woo

Lol

[inner finch]

Short for Seocial Jiustice Wooriar

[cold needle]

Wooriar

[sharp mulch]

(edited)

[deep mango]

tfw social justice worrier

[sharp mulch]

justic

[cold needle]

worrier

[pale orchid]

manhwan

[limber perch]

social justice worriers are known to have high levels of anxiety

[inner finch]

[vivid halo]

[deep mango]

[neat lintel]

Good webcomic

[blazing pawn]

This doesnt sound right

[neat lintel]

étale

french

[blazing pawn]

The analogy with the topological case comes from the restriction of the cover to any open subset U giving an isomorphism of the topological fiber product Y x_X U with p^{-1}(U)

so p is trivial over U if Y x_X U -> U is trivial over U which makes sense

But it feels like this definition is missing the local part

[vivid halo]

hmm

[neat lintel]

mmh

[blazing pawn]

shouldnt we need a covering of S by the image of morphisms Y -> S such that ur fiber product is trivial over each Y

Hmm maybe im thinking about the definition of trivial in the scheme case wrong

[vivid halo]

this is Stacks 37.36 "finite free locally dominates etale" for the correct statement I think

[blazing pawn]

Does this mean that the underlying topological space of the scheme X is a finite disjoint union of the underlying topological space S

[vivid halo]

for a trivial cover yes

[blazing pawn]

then like

how is the trivial cover X x_S Y -> Y local in any sense

like how is this gonna be locally trivial

[vivid halo]

wait you just said it's a trivial cover?

that means it's locally trivial

[blazing pawn]

Er yeah but like

it feels too narrow

[vivid halo]

trivial reads as globally trivial

hence locally trivial

[blazing pawn]

yeah but it seems like this is saying that finite etale covers are not just locally but globally trivial

[vivid halo]

how so?

locally blah doesn't imply globally blah

[blazing pawn]

Like in the analogy w/ topological spaces if i take any open subset U of X and form the fiber bundle Y x_X U -> U

this is trivial iff the cover is trivial over U

so in the topological case we need a bunch of inclusion maps U -> X that cover X such that when you form the fiber product with Y over each U its trivial

like, covering by trivial maps

[vivid halo]

mhm, and covers are always locally trivial

[blazing pawn]

But here you only need a single map above X

instead of a bunch of them that cover it

[vivid halo]

do you?

I don't think that's what affine morphism means

[blazing pawn]

well its saying that you have a finite locally free surjective morphism Y -> S instead of a family of morphisms Y_i -> S whose image covers S

[vivid halo]

X and S can still be globally complicated things

[blazing pawn]

like in the topological case being a cover just means you can cover the base space with any number of trivial U

but here its saying that you cover it by a single trivial Y

cause surjectivity

like i dont see whats local about this claim, it seems like its saying something global about covers

uh if its not clear im sorta thinking of the map Y -> S as an inclusion map in the way you have a ton of inclusion maps U -> X in the topological case and cover = when you pull back to the fiber product its trivial over each U

you usually just have a cover by some arbitrary number of U, but here your cover consists of a single Y -> S

[vivid halo]

I'm not really clear where this interpretation of affine comes from

just because Y->S is affine doesn't mean you can like

just smoosh everything local into global

maybe I'm being totally stupid but the way I'm reading what you're saying is something of a misunderstanding of what it means for a morphism to be affine; it does NOT mean that local properties become global properties

[blazing pawn]

Uh idk what you mean by affine like

[vivid halo]

this is true e.g. if the morphism is affine and the source and target are affine

[blazing pawn]

What im saying is that in the topological case a cover means that you can lift along inclusion maps U -> X such that the fiber product maps trivially into U

its local cause like. its trivial over elements of a cover right

not the whole space

But here you life along an arbitrary surjective map Y -> X and i dont see whats local about that at all

Like why is it just one map Y -> X instead of a bunch of maps Y_i -> X that cover X?

[vivid halo]

let's back up a little

are we unhappy with the fact that finite etale covers are locally trivial?

as in the previous screenshot?

[blazing pawn]

I dont understand why the defn of locally trivial for schemes is what it is

it does seem like theres anything local about it

[vivid halo]

no, there absolutely is, unless as usual Szamuely is writing like a moron

locally trivial means you can cover by affines such that over each the thing is trivial

this does NOT mean globally trivial

[blazing pawn]

What

[vivid halo]

oh okay I see the confusion now sorry

hold on

[blazing pawn]

ur defn makes sense

certified szamuely moment

[vivid halo]

you're not going to like my answer to this

the answer is of the form "I think I chained together a bunch of statements from the stacks project that make this work?"

[blazing pawn]

[vivid halo]

so I think you can maybe get away with the fact that being etale is etale local in the base and f:X->S (assumed finite affine surjective) is etale iff there exists an fpqc morphism Y->S such that the pullback of f by Y->S is etale, then use surjective etale implies fpqc

idk my brain is fried

(I'm really fucking bad at these kinds of proofs by the way so any time you ask me for help here I might be failing you lol)

[blazing pawn]

[vivid halo]

I'm sorry

in an attempt to cover my ass but also give somewhat realistic advice about what doing AG really feels like

I asked a certain really well respected postdoc in AG/motivic stuff about this sort of issue in AG

namely how many people actually know the nuts and bolts like this off the top of their head

and how many people honestly just get by cobbling together proofs from stacks/EGA/SGA

most people I ask say there's honestly maybe like

only a dozen or so people that can just vomit out proofs like this instantly without looking at stacks or other references

[blazing pawn]

Hell field

@vivid halo ok so to recap locally trivial actually means that you have a covering by injective morphisms Y_i -> S, Y_i affine, such that when you lift to the projection on the fiber X x_S Y_i -> Y_i is trivial?

and this is ultimately equivalent to the claim that there is a finite locally free surjective Y -> S such that X x_S Y -> Y is trivial?

"But if you write this incomprehensibly no one will be able to understand what you're saying and it'll make it harder for everyone"
"Lol," said the geometer, "lmao"

[vivid halo]

yes

this is a bitch answer but honestly don't worry about this kind of detail right now

if you have a good picture in your mind of finite etale morphisms you're fine for now

definitely return to this kind of annoying detail on a second pass, e.g. when you're reading Hartshorne later or something

[blazing pawn]

Yeah makes sense

[vivid halo]

but you'll be surprised by how much you can get away with being an utter moron with the nuts and bolts of AG while still having LOADS of juicy research accessible

[blazing pawn]

I guess itll help me to have the big picture in my head when im doing that

[vivid halo]

not endorsing being an AG moron but also pointing out it's not as awful a deficit as you might think

e.g. my undergrad advisor as an actual algebraic geometer is literally profoundly disabled

but he knows Macaulay2 too well

and knows complex AG too well

so he's fine

[blazing pawn]

$\begin{tikzcd}
X \times_S Y \ar[rrr, "\text{etale}"] \ar[ddd] & & & Y \ar[ddd] \
& \operatorname{Spec}(B \otimes_A C) \ar[lu] \ar[r, "\text{etale}"] \ar[d] & \operatorname{Spec}(C) \ar[ru] \ar[d, "\text{free, fg}"'] \
& \operatorname{Spec}(B) \ar[ld] \ar[r] & \operatorname{Spec}(A) \ar[rd] \
X \ar[rrr] & & & S
\end{tikzcd}$

Lmao literally just kill me

ok

makes sense

Stop sulling me!

I am trying to work things out with a diagram

Oh

This is so non symmetrical... why

tikzcd moment

[jaunty gate]

That's impressive latex competency

[blazing pawn]

its really not

You need to know maybe 3 things generously

[jaunty gate]

The labels

[blazing pawn]

the labels arent hard you just do "label" in ur arrow box thing

[jaunty gate]

No

I mean that's why it's not symmetric lol

[blazing pawn]

Oh wait im illiterate

Yea

Sigh

Slightly better i guess but tikzcd is just generally kinda ugly

[fathom swallow]

Ultragaslight

[blazing pawn]

Ah whatever

its just for reference anyway

What the fuck is a prime divisor

Oh No

Oh my god

[neat lintel]

[blazing pawn]

What does this mean..... what does this mean.......

@vivid halo what is a prime divisor in this context am i going to have to torture myself

[narrow rock]

integral subscheme of codim 1

[blazing pawn]

Whats a prime divisor of a principal ideal tho

[vivid halo]

oh oh oh

[blazing pawn]

Oh god oh fuck

[vivid halo]

so you know what a divisor on a curve is?

[blazing pawn]

Uh

no

[vivid halo]

okay here's what's going on

well okay here's a small glimpse of the super general situation first

a divisor in say a variety X of dimension d is a subvariety of codimension 1, that is a subvariety of dimension d-1

this is like

way too fucking hard to say anything about in general

but when X is a curve something nice happens, divisors are all dimension 0

[blazing pawn]

Uh is dimension in the variety dimension sense

[vivid halo]

mhm

[blazing pawn]

so the function field is algebraic

over the base field

Oh so that means it would look like a base change sort of?

[vivid halo]

I mean you're fine thinking of this in terms of like manifolds or whatever

[blazing pawn]

Or idk

[vivid halo]

in terms of usual dimension

codimension 1 just means the thing is 1 dimension smaller than the whole thing

and in the case of curves this is very very simple since this just means dimension 0 subschemes, that is (possibly unreduced) points

[blazing pawn]

My defn of variety is just the transcendence degree of the function field

is there a reason that the function field being algebraic implies that the subvariety is discrete

[vivid halo]

think of (affine or projective) variety as something cut out by polynomials in (affine or projective) space

this might include non reduced stuff, this is what schemes are for

[blazing pawn]

Oh so its like

er hm

[vivid halo]

another way to say it is like

ignore everything fancy with schemes

these things are cut out by polynomials

[blazing pawn]

So i dont really have much intuition for why dimension in the variety sense should look anything like dimension in the usual sense

[vivid halo]

the point of schemes is e.g. to keep track of non-transverse intersections

I can tell you that in the simplest case where you've over an algebraically closed field, dimension behaves in the way you expect

[blazing pawn]

well that would imply K(X) | k is algebraic so K(X) = k right

[vivid halo]

e.g. if you start with projective space P^n over your field, cut out some shape by a single irreducible polynomial, this will be codimension 1

[blazing pawn]

and then A = k...?

[bronze pelican]

Isnt a "prime divisor" of a principal ideal just a prime ideal that contains your principal ideal?

[vivid halo]

by two polynomials in general position, this will be codimension 2

and so on

[sly thistle]

Oh what’s nGroupoid explaining today

[vivid halo]

well so in general a prime divisor of an integral locally Noetherian scheme (ignore the adjective except integral) is an integral closed subscheme of codimension 1

on a curve this has a simple description

[blazing pawn]

oh well it just looks like gluing so in the non affine case you get points

[vivid halo]

integral codimension 1 subschemes of curves are literally just like

[blazing pawn]

Uh okay wait back up

[vivid halo]

points with multiplicities

[blazing pawn]

so nG my definition of dimension of a variety is the transcendence degree of the function field over the base field

[vivid halo]

that's a reasonable definition

although

[blazing pawn]

So like

[vivid halo]

this only works when X is normal

the intuition is that normal varieties are those where everything can be stuffed into the function field

in general the correct definition is like

[blazing pawn]

Actually this makes way more sense thinking about the irreducible subsets defn of dimension

[vivid halo]

there's a notion of Krull dimension of a ring

cover the thing by affines, take Krull dimensions of each ring

[blazing pawn]

like dim 0 subset means that there are no non trivial irreducible subsets of your space and points are always such a subset thus your subvariety is just gonna look like points

[vivid halo]

yup exactly

[blazing pawn]

Yeah

[vivid halo]

the only thing you might have to worry about in dimension 0 is like

you might have a point but with non-reduced structure

if you assume reduced it's just points

[blazing pawn]

uh what does non reduced structure mean exactly

[vivid halo]

here's an example

Spec(k[x]/(x^2))

is non-reduced

the picture is like

well the underlying reduced scheme is a point

but the whole thing is like some quadratic "fuzz" around this point

the picture to keep in mind is like

you know how y=x^2 crosses the x-axis only once, but you really want to count this single root with multiplicity 2?

it's exactly this idea, but encoded in the scheme theoretic intersection of x=0 and y=x^2.

[blazing pawn]

I kinda get it like

[vivid halo]

it's kinda hard to make precise but this is like

one of the handful of real goals of upgrading from schemes to varieties

[blazing pawn]

its the same point as the variety Spec(k[x]/(x))

the reduced structure is there to like

[vivid halo]

varieties can't keep track of intersection multiplicities like this

schemes can

[blazing pawn]

let you know that you're not quite working in that

[vivid halo]

yea

[blazing pawn]

cause x is nilpotent in this

[vivid halo]

YES

the nilpotent elements are EXACTLY what is happening here

in varieties you aren't allowed nilpotent functions

when you upgrade, the nilpotent functions allow fuzz like this

I hope this makes sense, this is like

[blazing pawn]

Yes i think i get it

[vivid halo]

one of the central insights of schemes

I mean you can go DEEP into this stuff, e.g. look up Serre's intersection formula

[blazing pawn]

Can you think of the fuzz as being like

[vivid halo]

you might imagine how it might be a little subtle how to like, formulate intersection theory of curves in a surface in a way that takes care of these intersection multiplicites, but purely algebraically instead of topologically where you can perturb stuff a bit

[blazing pawn]

in the non reduced case your non prime ideals cant "get close" to your prime ideals

[vivid halo]

the way you might think about it is in the topological setting you can perturb a non-transverse intersection like this to a transverse intersection where the sum of the number of crossings is what you expect

AG is more rigid, you can't do these kinds of wiggles like you can in topology

so you have to settle for the non-reduced structure encoding this

[blazing pawn]

I think i understand the geometric intuition

You have your non prime ideals which are not detected by varieties

[vivid halo]

yea this stuff is disgustingly hard to make precise (if you want intersection theory to really work, you need DAG not just AG)

but yes this is the correct picture

[blazing pawn]

ok

[vivid halo]

if you're really curious there's something called Serre's intersection formula that does this in general

where it makes sense of these intersection multiplicities at least locally by like

computing fucking ext groups of rings

this is homological, hence why DAG comes up

anyways

[blazing pawn]

Oh thats another thing i pretended to know things about when i was lying to summer programs

[vivid halo]

btw this is why scheme theory gets so fun

example

Spec(k[x]/(x^2)) that I mentioned earlier

you want to imagine this as a "disembodied tangent vector"

there is an algebraic notion of tangent space T_X of X in AG

T_X is literally Hom(Spec(k[x]/(x^2)),X)

[blazing pawn]

Like

Always or

Whats X here

[vivid halo]

X is literally anything

[blazing pawn]

its always specifically Hom(Spec(k[x]/(x^2)), X)?

[vivid halo]

for a first order tangent vector when X is defined over k, yes

[blazing pawn]

Wack

[vivid halo]

you want to think of Spec(k[x]/(x^2)) as like

a point plus a tangent vector

what does it mean to map this to X?

[blazing pawn]

Oh

[vivid halo]

it means pick a point in X

and pick a tangent vector at that point

[blazing pawn]

is the idea here that reducedness means that its "really close to 0" in the same way tangent vectors/derivatives are "really small"

[vivid halo]

Yup!

[blazing pawn]

Epic

that makes a lot of sense

[vivid halo]

There's some general notion of "jets"

which are like "higher order tangent vectors"

where you replace k[x]/(x^2) with k[x]/(x^{n+1})

same idea

[blazing pawn]

I think i see

Well i dont have much of an intuition for it but it seems like a natural generalization

[vivid halo]

yea it's like tangent vectors only see up to first derivative

[blazing pawn]

so i buy it

[vivid halo]

n-jets see up to n-th derivative

hopefully believable

[blazing pawn]

kiiiiinda?

its cool for now

Ill think about it more later

[vivid halo]

yea I'm honestly just trying to expose you to a lot of stuff

[blazing pawn]

[vivid halo]

so like, what are schemes good for that varieties aren't?

there's three main things

1. non-reduced structures

[velvet dagger]

Credits to Knutson

[vivid halo]

2. non-algebraically closed structures

3. non-irreducible structures

(the last of these is a little annoying depending on how you define variety)

it's really the first two that are the richest

1 allows for like infinitesimal structure and all that

2 allows for arithmetic

[blazing pawn]

Oh

Ok so like back to this

how exactly is a prime divisor of a principal ideal defined?

is this just prime ideal containing the principal ideal

[vivid halo]

should be yea

[bronze pelican]

Yeah

An ideal divides another ideal if it contains that other ideal

[blazing pawn]

wouldnt that just be all prime ideals because if p is prime and x is in p then p contains (x)

[neat lintel]

If anyone knows basic number theory would rly appreciate help with a problem in the ENT channel

It’s a problem from last chapter of Stein shakarchi

[terse flax]

yes

[limber perch]

tomorrow

[terse flax]

yesterday

[errant kiln]

all my troubles seemed so far away

[bronze pelican]

Random general advice for students who need letters of recommendation for their applications for whatever: Provide your letter writers with all your application information such as application essays, cover letter, resume, etc. Give them plenty of time to look over your materials and write the letter, so several weeks ahead of the deadline.

[crystal stone]

Really?

None of my writers saw any of that

I just said "Hey you wanna write letters for me"

and they all told me yes

Then I sent them a list of schools w/ deadlines, then I email them to follow up on deadlines

That was it

[vivid halo]

[bronze pelican]

Not saying you have to do this but it makes it 10 times easier for the reccomender to write their best letter for you. Also preparing your application materials that early to be seen by another pair of eyes makes it so much easier for you to obtain and incorporate feedback to make your part of the application as strong as possible.

[crystal stone]

Don't sully me

[sharp mulch]

I agree with PTY

[crystal stone]

Well...I guess I didn't get in anywhere

So y'know

Don't do what I did

[mortal oasis]

why does cool stuff require such in depth math

i dont want to learn differential geometry that seems hard

[vivid halo]

PTY has the right idea

[cold needle]

this sounds like good advice ptyamin i will keep this in mind

actually this is going to be relevant really soon isn't it for me

for REU stuff

if i start working on applications as soon as the list comes out for next summer or

something

uh

then i have until february

so if i ask during winter for letters

?

maybe thats too soon actually

oh okay

ok so what things should i think about that aren't specific to a program that i could work on before

maybe first i should google these things lmfoa

"how to apply to reu"

oh i see

ok

[sharp mulch]

You should ask for letters in the autumn

[cold needle]

ok

i have two people i think might be good idea

[sharp mulch]

Give your letter writers ~6 weeks at least

[vivid halo]

^

6 just means they'll not do anything for 4

[sharp mulch]

Lol

[vivid halo]

still

4 is adequate

[sharp mulch]

In my experience, some profs might not agree to write letters with 4 weeks notice

[vivid halo]

aahahaha

okay fair

[sharp mulch]

For grad school, you want to notify your profs like

3 months in advance

[vivid halo]

oh yea for sure

[cold needle]

dam ok

[bold ferry]

what bout 3 yrs

[sharp mulch]

And then you should email when the deadline is 4 weeks away, 3 weeks away, 2 weeks away, 1 week away, 3 days away, and 1 day away

If it hasn't been submitted

[vivid halo]

ask professor first day of class if they will write you a letter

[sharp mulch]

I think I asked for my NSF letters mid August

[limber perch]

everyone's obsessed with publishing shit to journals

but how do you actually access these journals

are they all monetized

[narrow rock]

University subscription

[limber perch]

is the journal like an actual journal

or is it just a collection of math papers

yea ok

hey that's not happening to me any time soon

but good to know ig

[cold needle]

is it ok to ask a professor i had a couple semesters before

for a letter

i know they still remember but is that awkward or what

oh

ok

yeah i had him twice

once last fall and once over the summer

so not immediate ig

but i think maybe having him twice is good

the other one im thinking of i will have this fall and i had him before so i think thats good

[bold ferry]

u use scihub

[cold needle]

i just need to absolutely sauce his class

[deep mango]

Ultra can.

[cold needle]

Ultra.

[narrow rock]

letter.

[cold needle]

Ultraletter.

[deep mango]

little slimmy participated in class a whole lot, and showed a lot of excitement for the wonders of mathematics!

sincerely, RYC

[narrow rock]

[deep mango]

I wrote a blurb about a prof who won the distinguished undergrad teaching award at berkeley

😎

even ange liked it 😎

[cold needle]

wow

ange liking anything that isn't ORV | Urbanism ?

[deep mango]

i wrote the blurb after she won the award though

[cold needle]

this is unheard of

[narrow rock]

I nominated my REU advisor for a mentor award

He didn’t win

None of the math mentors won a prize

All the winners were in experimental sciences

And humanities

Every single year

Idk why

[cold needle]

hmm

is the winning criteria hidden or no

[narrow rock]

hidden

They just ask you to write an essay about your experience

[cold needle]

i see

maybe it's easier to be more... colorful or something with those subjects in mentoring idk

lmfao

confusion

[limber perch]

slim out here writing rec letters for profs...

[narrow rock]

No

They’re probably just biased lol

[cold needle]

kekw

[astral marsh]

TRUE

i am experiencing first-hand the anti-math discrimination from the hands of academic professionals on a daily basis

chemistry professor shitting on math (read: simple arithmetic in the literal sense) and physics TA shitting on math rigor while talking about infinitesimal masses and changes in time

[cold needle]

that's annoying

i was accepted into a program that tries to foster undergraduate research

and i told the lady i was a math major and she just

didn't say anything

i heard afterwards from some older people that she doesn't believe in mathematics research

[narrow rock]

lol

[cold needle]

eventually i figured out that i was not going to get anything from this program so i left

[astral marsh]

i mean have you ever seen a mathematics research irl

[cold needle]

it doesnt matter though since i still keep the money

[astral marsh]

thats what i thought

[narrow rock]

the life sciences department over here doesn’t like the way the math department runs their courses

[astral marsh]

🤔

how come

[cold needle]

odd

why should they care

[narrow rock]

They tried to create their own math courses

[astral marsh]

[narrow rock]

So that the life science students don’t have to take courses offered by the math department

[limber thunder]

this can't end well lol

[narrow rock]

Essentially a super watered down calculus sequence

But this takes away jobs from math grad students

[limber thunder]

I mean if they're willing to torture themselves that way

[narrow rock]

So the math department created a “math for life sciences” course sequence

[cold needle]

[sharp mulch]

Oh Berkeley has a math for bio sequence

[cold needle]

🧋

bio math is lit

[sharp mulch]

But it was designed by a math and bio prof

So it's good and stuff

[cold needle]

i just don't know any biology

[vast surge]

Systems of nonlinear ODEs I think is what most bio math is, but maybe that's just what the bio math people at my school do

[sharp mulch]

That is very much not all of bio math

[cold needle]

o

[vast surge]

Then it must just be what those specific professors work on.

[sharp mulch]

What a depressingly narrow perspective

[vast surge]

Precious bodily fluids

I can not see the word "fluid" without thinking of Dr Strangelove

[narrow rock]

Biology with a view toward homotopy theory

[vast surge]

Actually wait now that I think about it I know people who do alg geo in genetics

Which I guess is technically biology

[sharp mulch]

"technically"

[narrow rock]

What did gromov work on

[cold needle]

etalneenya

[devout nacelle]

The bio-math course at my place seems to be the same, ODEs+some dynamical systems in the context of population models, etc.

etalneenya

[vestal river]

[cold needle]

truly

[bronze pelican]

:sullivan:

[sharp mulch]

I would recommend Pachter and Sturmfel's book on alg geo for genetics

[limber perch]

ange seems to be back on promoting algebraic genetics

also what does "doesn't believe in math research" mean

[narrow rock]

this is a really cool paper

[crystal stone]

what does that mean

I don't believe in math research

[cold needle]

she doesn't think it's valuable or something

[crystal stone]

Is she a dean

Is she neglecting some duty

[cold needle]

shes a professor in the dept of business here i think

[crystal stone]

By playing favorites for majors

[forest jackal]

tell her you don't believe in business

[mortal oasis]

tell her she is capitalist swine

[cold needle]

lol gomez

yes

[astral marsh]

what even is business research

gotem

[blazing pawn]

Theft

[ancient flame]

$\Sigma$ male

[fathom swallow]

gmod

[inner finch]

[mortal oasis]

Id assume like financial market stuff

case studies

etc

[astral marsh]

no i didnt actually care

i was just got getting them

[mortal oasis]

Hahhahaha my bad

[cold needle]

chat ded lmao funni

im gonna cry

and log off

sorry

[sharp mulch]

[sharp mulch]

@deep mango reviewing the meeting notes I took last time, it seems like most of the stuff is for you to do?

[deep mango]

ok, fine

[sharp mulch]

It's mostly just the mixed precision stuff and BRGEMM

And of course on going notational difficulties

[deep mango]

yeah

that's what i'm seeing

[blazing pawn]

:bearjak:

[sharp mulch]

Ok I've fixed my notation and pushed the changes

[rancid meadow]

now i think ryc is tterra

this is awful

no pfp changing allowed pls

[sharp mulch]

Lol

[neat lintel]

[cold needle]

lol

[bronze pelican]

https://mathemalchemy.org

Everything in this art piece is mathematically inspired

[scenic narwhal]

✋

[sharp mulch]

Those look like || anal beads ||

[rancid meadow]

can we mute ange for a day for that one

[cold needle]

im fine with that

[rancid meadow]

😄

[onyx pagoda]

I mean he's not wrong 👀

[astral marsh]

[velvet dagger]

Yeah mods get on it

[deep mango]

every time i change my pfp, i make compact mode a little more desirable

🙂

[rancid meadow]

this is attention seeking behaviour

[deep mango]

🙂

[deep mango]

Holy shit bigg boss is doing an OTT season 😌

We win bigg boss army!

I have never seen bigg boss but big brother OTT was so good

So i could be convinced

[summer nest]

Who watches big boss

[deep mango]

https://en.wikipedia.org/wiki/Over-the-top_media_service

so basically, catered to live feeders ✨

idk, is it ever in english?

i'm sure it's always in hindi

time to fire up duolingo

really?

[summer nest]

Wait you guys are watching Hindi big boss?

[deep mango]

dubs

of big brother

wow

that just sounds awful

[summer nest]

😌 we copy a lot of music from you and remake it to sound more cringe

You meaning the English music industry

I've heard a couple of the remakes and they all are horseshit

[deep mango]

they totally mean crazy strategy right

and not like

that this is actually a bad show

[summer nest]

They probably would air salman running over innocent bystanders

[deep mango]

should i make salman my husbando

is salman husbando material

[inner finch]

hindi sub of hindi dub of people speaking hindi

[deep mango]

ok i see

so the format is like BB australia

wait no

it's just BB UK

garbage

[inner finch]

my friend told me she watches big boss religiously and i watched the trailer

and i was like bruh moment

lamha e bruh even

[deep mango]

how is it a real competition if the audience votes for the people to be evicted

that's so boring

[inner finch]

[deep mango]

why are only big brother US, canada, and australia good

[summer nest]

The people do monke things to look more interesting

[inner finch]

on this topic

[summer nest]

Then audience decides who's a more entertaining monke

[deep mango]

well it's more fun when they have to strategize against each other and vote each other out

that's what makes it so good

[inner finch]

i wonder if there are any half decent dramas produced here 🧠

[summer nest]

Even better if it's just classic hunger games 😌

[inner finch]

[vagrant kestrel]

I mean sometimes the audience votes on someone to bring back

hopefully that person never wins though

[deep mango]

yes, i'm fine with that

ok

like

do you know the game your turn to die

[vagrant kestrel]

oh yes I do

[deep mango]

fucking tera would have won bbcan 9

and gary was supposed to win bbcan 1

thank goodness, someone fucked up their vote and he lost by that vote LOL

[vagrant kestrel]

I don't need this bbcan venting

[deep mango]

well it's not venting because both times the right player won

[inner finch]

shut the fuck up ryc

[deep mango]

hey mirza

are you the fucking bishop?

no

[summer nest]

I want to find the BBC series Connections

I heard it's great

[deep mango]

you're probably the left horse pawn

[inner finch]

i am the chess board you piece of shit

[deep mango]

talk to me when you're the bishop, then you can comprehend ty's masterful game

[summer nest]

Yes she walks diagonally

Suits her personality

😌

[deep mango]

wait

if the audience votes for who is evicted on bigg boss

then how does salman know that it's going to be the craziest season ever or whatever

they can only prerecord it when the game is isolated

[summer nest]

He's Channing Tatum of bollywood. Anything with expressions is crazy to him

[deep mango]

yeah i'm not touching bigg boss, it like completely overlaps with BB US

[inner finch]

Imagine even trying to assign reason to what salman khan says

he's beyond us

[deep mango]

he sounds based.

[inner finch]

the only communication we can understand from him is his rapid tearing of shirts

[deep mango]

unlike you dorkos.

[inner finch]

I mean he's a literal murderer

[deep mango]

damn, hella based.

[inner finch]

so yeah very based

[blazing pawn]

is salman khan the guy that ran someone over

[inner finch]

yes

i've heard from my mom apparently he had a hand in the whole sushant thing though im not sure how accurate that is

might just be fake news

[ancient flame]

salman khan is my lord and savior

he teaches me all my calculus

[inner finch]

i thought they were the same person as a kid

[ancient flame]

LOL

[inner finch]

i can never take anyone named salman seriously anymore

just not possible

[ancient flame]

to me it just sounds like salamander

[errant kiln]

salmon

[untold sapphire]

there's also Salman Rushdie

that's who I think of

[sharp mulch]

Ah yes the satanic verses

[inner finch]

yeah i read the first page of that book and the language was weird

was it translated

[sharp mulch]

In to English?

Surely

[bronze pelican]

Has anyone seen any successful educational content on tiktok?

[toxic sapphire]

Yes

[bronze pelican]

examples?

[toxic sapphire]

I know a bunch of history tiktok

I don’t do math

[blazing pawn]

she has a nice youtube channel too

[deep mango]

@sharp mulch if all goes right...

i'm about to push the appendix with mixed precision!

at least for the single processor case

i didn't handle the bad precision tuples yet, but there are only 2 if we just stick to 16/32/64, and those are one array 16, one 32, one 64, or two arrays 16, one 64.

which are super weird cases anyway

but i will still investigate the actual constants for these cases

[bronze pelican]

I wasn't looking for anything in particular. Just curious to see what is there.

Sal Khan mentioned on the 3b1b podcast that "Youtube back in 2006 felt a lot like TikTok does today, although I'm now discovering TikTok is a great place to educate as well."

So I wanted to see some examples

[sharp mulch]

Are you going to do parallel mixed precision

[deep mango]

sure

just not tonight

[sharp mulch]

Ok yeah

No rush I guess

[deep mango]

i can try to work it out tomorrow

maybe on paper

or i can work on these special tuples

we'll see

@neat lintel @leaden torrent in other news, i just finished lift your skinny fists like antennas to heaven

[sharp mulch]

From a practical pov I def think that parallel mixed precision is more important than weird combos of precisions

[deep mango]

still a little weird to me

but it was very fun

[deep mango]

yeah but i also want to have a finished theorem if possible

instead of an unfinished one

[blazing pawn]

No i want everything unfinished and terrible

[deep mango]

hey ultra, do you ever plan on sending 16-bit numbers through a 32-bit number filter, and storing the outputs in a 64-bit number array?

[bronze pelican]

ah I see

[deep mango]

ok

[blazing pawn]

Ultraproduct pro-representability is very cute i think

[deep mango]

moth, do you plan to do that ever?

[blazing pawn]

I plan on eating ice cream

[deep mango]

that's close enough

[blazing pawn]

Does that involve sending 16-bit numbers through a 32-bit number filter, and storing the outputs in a 64-bit number array?

[deep mango]

yes

well

i'm convinced

that's 2 people who want it ange

[vivid halo]

boy am I regretting agreeing to give a talk tomorrow

[deep mango]

i'm a people pleaser. are you gonna have me say no?

[bronze pelican]

F

nG what are you talking on

[vivid halo]

it should be fine but my notes aren't totally done and I'm exhausted

I'm talking about D-modules in characteristic p

[bronze pelican]

oh no

[deep mango]

lol. d-module

[vivid halo]

differential equations in char p are terrible

[bronze pelican]

not D modules

[deep mango]

deez modules

[blazing pawn]

D module... need

Lmao

No idk what that is

[vivid halo]

example: the ring of differential operators on k[z] for k char 0 is 1-dimensional generated by d/dz

[sharp mulch]

I think you might have a 16 bit number times a 64 bit number and get a 32 bit number?

Perhaps

[vivid halo]

on the other hand, the ring of differential operators on k[z] for k perfect of char p is like

[sharp mulch]

Alternatively you can just add another assumption to your theorem

[vivid halo]

infinitely generated

[blazing pawn]

oh

thats gross

[sharp mulch]

So that they do not precisions do not differ too much

[vivid halo]

yea since you can't like

[blazing pawn]

keep adding

yea

[vivid halo]

you can't divide by p!, p^2!,... in your Taylor expansions or whatever

[blazing pawn]

oh

[vivid halo]

so it's infinitely generated by d/dz, "(d^p/dz^p)/p!", "(d^p^2/dz^p^2)/p^2!",...

now the question is, how much of Riemann Hilbert can you still get to work

since flat connections still make sense in char p

[blazing pawn]

wowee i sort of understand that conceptually

[vivid halo]

[blazing pawn]

i cant wait to talk to my prof so i can figure out what im going to lie about to judges panels and pretend is useful

[vivid halo]

hell yea

[blazing pawn]

lying is so fun why does no one acknowledge this

smh

im not an AG memer yet ok

so true

I know what panic! at the disco is

smh

lying is the most fun

is the song

right

is it uh

what finishes it

most fun a girl can have without something something clothes?

taking off her clothes?

panic at the disco had a lot of long song titles

what modern rock bands do the same thing?

car seat headrest does i think

[Joe Gets Kicked Out of School for Using] Drugs With Friends (But Says This Isn't a Problem)

thats pretty long

[deep mango]

well that's what i did for now

in fact, the assumption is really strange and elegant

[sharp mulch]

Oh that's good

[deep mango]

"the precisions form the side lengths of a triangle"

it's literally that

[sharp mulch]

Oh nice

This is elegant, I agree

[deep mango]

it's p_1 <= p_2 + p_3, p_2 <= p_1 + p_3, p_3 <= p_1 + p_2

which is of course what i wrote

but

still, kind of cool

(the p's are proportional to the number of bits each entry takes up, right now i have them in units of 32 bits = words)

turns out with p's being 1/2, 1, and 2 (or 1, 2, 4 to make life easy) we find that the only violators are permutations of 1, 2, 4 and 1, 1, 4.

[sharp mulch]

Right

That's how triangles work

When I do communication counting I will probably actually multiply by 16/32/64 as appropriate

[deep mango]

Ok, I can do that too

I just would change the word "words" to "bits" everywhere

since the whole thing is just variables

and everything is homogeneous

oh, so the funny thing is that the large filter bound is proportional to (sum of precisions)^2 and the small filter bound is proportional to (product of precisions)^(1/2)

so they're totally different

like

that's so strange

[sharp mulch]

It's fine

It's all constant factors

[deep mango]

no, it's great

[sharp mulch]

In reference to words/bits

[deep mango]

it's very satisfying

oh

yeah

[sharp mulch]

Oh I see

Interesting

[deep mango]

anyway, this actually feels like i did something that wasn't just copying what demmel and grace did now

which is gratifying

[sharp mulch]

Yes

[deep mango]

i guess the brgemm was that too

[sharp mulch]

Original work

[deep mango]

but like

that was also

or

this should be in the paper before that

[sharp mulch]

Our paper becomes better by the minute

[static loom]

I don't know but if I had to guess it'd be the volume contained in the intersection of a cylinder with copies of itself rotated to be parallel to the x, y, and z axes

no, apparently not, what I described is called a Steinmetz solid or a tricylinder

[astral marsh]

@vast cipher fun things to try while jittery/very excited

reaction time test failed bc i was getting too nervous by anticipating the cue lmao

bro i want this class to end so i can hit the gym

i wanted to start going to the gym in the student athletic center today

[cold needle]

thats good arch

[astral marsh]

yeah

metal i took my adhd meds and accidentally also drank coffee

[cold needle]

def find friends to go with regularly its very fun

[astral marsh]

small amount of both

so its not thew orst

[cold needle]

ono

oic

[astral marsh]

but im feeling quite stimulated

i dont drink coffee at all so

the coffee is really making me jittery

[cold needle]

yeah might be good to go to the gym to get rid of the excess feeling

maybe

ah

[astral marsh]

yeah im in a class now

shit im shaking both legs

[cold needle]

yeah i don’t drink coffee either idk how i would be if i had caffeine properly

[astral marsh]

lol

I JUST WANNA GO

[vast cipher]

the gym is great all stimmed out

[astral marsh]

let's gooo

that kind of feeling

[vast cipher]

dude you gotta play in a band

i loved that shit all hyped up

[astral marsh]

i see

[cold needle]

math server band when

[astral marsh]

i mean i can do music on my own for sure