« first day (2819 days earlier)      last day (2197 days later) » 
00:00 - 19:0019:00 - 00:00

7:00 PM
But with no calculators ¬¬
 
you approximate it by hand
if there is something nice it has to be an integer
not that many possible cases
 
If I knew that $\sqrt[3]{9+4 \sqrt{5}}$ is of the form $a+b \sqrt{5}$ it would be easy to find $a$ and $b$
 
well then you suppose it is, you try to find it, if you do it"s great if you don't you know there is no nice thing
 
I can't suppose that
what if it's a different problem
what will I suppose?
 
the same kind of thing
 
7:03 PM
I need to be sure before trying it
 
the point is that wether something nice happens is completely decidable
why ?
 
I don't want it to look like magic
 
I guess you could compute the norm of $9+4\sqrt 5$ and check that it's a cube
that's a necessary but not sufficient condition for a cube root to exist
 
nope, that's magic, how did I know that?
 
or you could read a book on algebraic number theory ?
 
7:05 PM
what's the name of the method?
 
I'm not sure there is a name ?
I wrote a post here on this though
 
are you sure there's a method?
 
Does anyone know where to get a cheaper version of this book?
 
6
A: How does one evaluate $\sqrt[3]{x + iy} + \sqrt[3]{x - iy}$?

mercioAs I said in a previous answer, finding out that such a simplification occurs is exactly as hard as finding out that $35+18\sqrt{-3}$ has a cube root in $\Bbb Q(\sqrt{-3})$ (well actually, $\Bbb Z[(1+\sqrt{-3})/2]$ because $35+18\sqrt{-3}$ is an algebraic integer). Suppose $p,q,d$ are integers. ...

 
lol
xD
it was more expensive
 
7:07 PM
xD
 
@mercio Suppose I'm a high school student...
I don't know what $i$ means...
 
$i$ is $\sqrt{-1}$ but it doesn't really matter for the methods
 
and to prove that $\sqrt[3]{9+4 \sqrt{5}}+\sqrt[3]{9-4 \sqrt{5}}=3$?
we go back to the first problem
 
you guess that the cube root is of the form $3/2 + b \sqrt 5$, so you cube that and solve for $b$ and you find a rational solution for $b$ and all is well
 
If I do that I can get $a$ and $b$ and get $\sqrt[3]{9+4 \sqrt{5}}$ directly
but I don't wanna suppose that, why should I suppose that?
 
7:18 PM
you should because it is a result of algebraic number theory that if the sum is equal to $3$ (or any rational number) then there is a cube root of the form $3/2 + b\sqrt 5$ with $b$ rational
I don't think there is any method for this kind of thing that will not seem cheat-y to a high schooler
 
but I don't know if the sum is 3 nor if $\sqrt[3]{9+4 \sqrt{5}}$ is of the form $a+b\sqrt{5}$
 
Try to multiply by conjugate.
 
O.O
 
how exactly is your problem stated ?
 
expand $\sqrt[3]{9+4 \sqrt{5}}$?
 
7:26 PM
o. .O
then you use a meta-theorem that says that a hig hschool exercise always have nice solutions
so you look at the sum and conjecture that it is $3$ then guess the cube root is of the form blablabla
and actually, it is not that easy to find $a$ and $b$ such that $(a+b\sqrt 5)^3 = 9+4\sqrt 5$. However it becomes easy if you know that $a=3/2$
 
@Eric finally done with GMT!
 
ok I'm gonna do it that way thanks
and how can I solve the equation $x^3-3x-18=0$?
 
you use a meta-theorem that says that high school exercises have nice solutions
and you find $x=3$
(by trying every integer until you find a solution)
 
lol ¬¬
but with a method
 
actually you should use the rational root theorem
it limits the possible cases
(and you still have to assume that there is a rational root)
 
7:40 PM
Let $x= \sqrt[3]{9+4 \sqrt{5}}$ and $y=\sqrt[3]{9-4 \sqrt{5}}$. Find $(x+y)^3$ and $(x-y)^3$.
 
basically if there is no rational root you would be totally screwed
and they wouldn't ask that to a high schooler
 
now suppose I'm not in highschool
 
then you are still totally screwed if there is no rational root
and i suppose you could use Cartan's formulas but you will still hate them
 
You will find that $(x+y)^3 = 27$ i.e. $x+y = 3$.
 
but there is one
 
7:43 PM
the thing is that cartan's formula is actually useless
because it doesn't show you when there is a simple solution
@user537566 that supposes you are able to solve $x^3-3x-18 = 0$
how do i solve $x^3 - 3x -18 = 0$ ? with cartan's formula ! wait it tells me to simplify an ugly sum of cube root ! how do i simplify the cube roots ? by solving $x^3-3x-18 = 0$ ! yayyyyy !
 
I wouldn’t say it’s useless, just that the utility is far less than the quadratic formula
 
it was useful to invent the complex numbers
so i forgive it
because complex numbers are beautiful
 
If you want to get explicit roots in radical form, it’s usable to that purpose
But that’s so rarely the actual goal
 
7:48 PM
In my above picture, it seems like image is spanned by $e_2,...,e_7$, so dim of image is 6, hence nullity 2, am i right?
@0celo7
 
When it comes to high school cubics, though, I agree
Rational root or bust
 
@Silent Look at $Te_8$
 
(If one is rational, then the other two can be done with the quadratic formula. If not, good f****** luck)
 
@AlessandroCodenotti oh! forgot. So, nullity is 1 and dimension of image is 7, right?
 
Yeah that looks correct to me
 
7:53 PM
thank u
 
@Semiclassical there is also a whole kind of evil exercises of cubics "magically solvable with trigonometry" : math.stackexchange.com/questions/1980693/…
 
Ew
Devilry, that
 
I think I had another one in mind
but maybe it was a quintic
 
I suppose one way to cook up examples is to use Chebyshev polynomials
 
10
Q: Solve this tough fifth degree equation.

Recep$$x^5+x^4-12x^3-21x^2+x+5=0$$ I think it can be solved by trigonometric ways but how?

I have no idea who would actually ask that and say "btw it's solvable with trigonometry" (whatever that means)
 
7:58 PM
Yeah
 
What would be an easy example of a continuous function $f_n$ on a compact set $[a,b]$ whose integral over $a,b$ goes to zero for $n \to \infty$ but for some $x \in [a,b]$ it is always $f(x)=1$?
I thought about the sawtooth function first, but it's not continuous
 
a triangular spike that gets thinner and thinner
 
oh right it is continuous
 
For a fancier/smoother example, you can probably do something with convolution
But that’s overkill for that problem
 
I also thought about a hat function but yeah thats too complicated
How would I define such a triangular spike on the interval $[a,b]$ with the top of the spike always being $1$?
 
8:06 PM
:)
a solution of $x^3-3x-18=0$ is $x=3$, and the other two?
how can I prove that they're not real numbers?
 
If x=3 is a root, then x-3 is a factor
So factor that out and examine the quadratic that remains
@philmcole take the midpoint to have f=1, and use lines for the sides?
That’s tedious but straightforward to write out
 
@mercio didn't realise that was an old question but that's a cool answer
 
it uses magical powers of gambling that the prof who made it up used a basic sum of cos and not any other more complicated combination
it would have foiled me
 
what if the division of polynomials is not exact?
 
then you double check everything
 
8:13 PM
:D
 
@Semiclassical I don't even get it working for $[0,1]$... I split $[0,1]=[1,1/n]+(1/n,1]$ and define $f(x)=nx$ for $x \in [1,1/n]$ and $f(x) = - ((n-1)/n) x$ for $x \in (1/n,1]$?
The last part is wrong but that's what I have so far
 
@Daminark sick
wait @Daminark are u doing the may-reu
 
@AkivaWeinberger Hm, don't they? Take the plane which cuts the three spheres in equal halves each (exists: nuke the fuck out of that by Ham-Sandwich theorem if you want). Consider the heights of the spheres from that plane, on both sides. That gives three "zeniths" and "nadirs" of the spheres. A plane passes through each of those two pairs of three points.
I like the perspective drawing picture!
@AkivaWeinberger I knew you'd immediately come up with a higher dimensional generalization, for some reason.
:P
 
8:29 PM
I got it working for $[0,1]$, now I need to scale it to $[a,b]$ somehow..
 
@BalarkaSen this lil girl in the new pdp video is baller
 
Wait I haven't seen it
Lemme have a lookie
 
hey @BalarkaSen
 
Hey hey hey
 
how's the exam prep going
 
8:37 PM
It's okay. :P I'd enjoy it more if it wasn't a college admission test that I was preparing for, I guess, is how I'd describe it
@0celo7 oh that. ugh.
 
@BalarkaSen i ship lil tay and lil xan
 
@BalarkaSen Those aren't necessarily tangent to the spheres
 
@AkivaWeinberger I tried to answer your question but need more info to be sure
 
I think your answer was right
@BalarkaSen Consider that^ sort of situation
 
@AkivaWeinberger Oh.
Yikes.
 
8:45 PM
are you doing PDE now?
 
@EricSilva yeah
 
Tiny sphere in the "cone hull" of the big spheres
 
Might very well have him as my mentor, we'll see
 
I mean I'm sure you can apply Zariski bullshit if you want a nuke
but my perspective argument doesn't suffer from that
 
Ok I see now
Thanks! This was helpful
 
8:47 PM
I'm visiting Yale, and I sat in on one of the classes, and it turns out I know the professor's brother
Yair Minsky; I know Yaron Minsky
The latter is a great guy; the former is an OK teacher
(The latter might be an OK teacher and the former might be a great guy but I don't know)
(That was worded confusingly)
 
tfw writing former after latter and then claiming latter after former
 
"Latter" meant the same person both times
Here's a puzzle someone gave me
 
@Akiva one of our best math profs is moving to yale next year
if you have the opp to take a class with wilhelm schlag you should take it
 
Plane, taxicab metric. What's the maximum number of equidistant points? Pairwise equidistant
@EricSilva Sure, will do
 
@Daminark start dressing in his iconic pink button down and beige khakis and see if it takes him long to notice that uve cloned him
 
9:00 PM
I need puzzles to share with math people at Yale
 
here's a fucking puzzle
bah, I can't be bothered to type it
but rest assured it is horrible
 
is it a PDE puzzle
 
it's an inequality involving 3 metrics, some eigenfunctions, and some other stuff
 
do not want
 
well, more like 3(n-k) metrics
it's been tormenting me for a while now
 
9:04 PM
my pde puzzle is figuring out how to get mathematica to do nonlinear hyperbolic pde numerically
without it having a nervous breakdown
 
with hard work
thats the answer
you're welcome
 
what I don't understand is that this is obviously an error in the paper
but these guys shouldn't make such a large error
so what am I missing
 
"This can't be right...but how could they get it so wrong??"
 
well it'd better be right in the end or the whole paper is wrong
but the argument is obviously wrong...or is it
 
gotcha
 
9:07 PM
it's not mentioned in the notes I have, and has an incorrect proof in the paper
very fishy
 
If I multiply $\begin{bmatrix}x\\y\\z\end{bmatrix}$ with $\begin{bmatrix}x&y&z\end{bmatrix}$ I get $\begin{bmatrix}x^2&xy&xz\\ xy&y^2&yz\\ xz&yz&z^2 \end{bmatrix}$. I feel like if I multiply together three copies of $(x,y,z)$, each one rotated to a different axis, I should get a 3x3x3 matrix of all possible triples of products
 
Define "multiply together three copies of (x,y,z)" :/
I guess you may have in mind $v_i v_j v_k$ as a rank 3 tensor?
where $v=[x,y,z]^T$
 
I know, it makes no sense
or maybe it makes sense in tensor-land I dunno
 
I'm just not even sure what you're looking for
 
There's a map from $\Bbb R^3$ to $\Bbb R^{27}$ and I kinda want to calculate with it without stuffs being annoying but that might be too much to ask
 
9:16 PM
What's the map?
something like $v_i\mapsto v_i v_j v_k$?
 
Yeah I guess
 
@0celo7 so I think I just ran into a physics instance of a weak solution
 
if I understand your notation right
 
so that's...fun
("physics instance" is probably a stretch. more like "advisor-indicated")
 
everyone needs weak solutions in their life
 
9:19 PM
lol
 
I am trying to prove the existence of one right now...
 
@0celo7 Chief Keef invented music, apparently
 
I'm not actually certain it's relevant for the full problem, but it definitely matters for the 'easy' problem
 
I haven't watched that video
I wasn't planning to either
 
It's worth it only because of the comments below
Just so smacking lips good
 
9:20 PM
I've got a pair of coupled 1+1 nonlinear PDEs
 
"Chief Keef invented Tupac"
 
@Semiclassical Earlier I was musing that $v\mapsto \frac1{\sqrt2}vv^\top$ is an isometry on $S^2$. I was just wondering if I could generalize it to higher dimensions
 
from the depths of the comment section
 
weak solutions for nonlinear PDEs are hard
 
9:21 PM
The above isometry gives an isometric embedding of $\rm\Bbb RP^2$ in (a five-dimensional subspaces of) $\Bbb R^9$
 
the numeric idea was to make an initial guess for one of the functions, evolve it forward in time, use a condition on t=T to figure out a new guess for the other function, and evolve backwards
then rinse and repeat until convergence
 
ew
 
it worked for the older case
but uh
 
can you prove convergence with some contraction mapping principle?
 
Hell if I know
This is not a very reasoned approach.
I'm pretty sure that, ignoring nonlinear terms, the other example that 'worked' involved the heat equation in one direction
and as a consequence one could hope to avoid too much crap
but I find myself dubious of it being useful here
maybe it'll be fine once one includes the coupling between the PDEs? but that seems sketchy AF
And I have no idea whether to trust what mathematica is giving me numerically
 
9:26 PM
"we then have the following easy observation" ...I'm pretty sure what's written after is not true
I want to lie down
ugh
 
lol
$\partial_t \phi = -(\partial_x \phi)^2/2+g(\rho-\rho_0)$, $\partial_t \rho = \partial_x (\rho \partial_x \phi)$
blehhhh
 
9:43 PM
@Eric also did you get the last problem in complex? I got the bound of 8 instead and Marianna said it was the right idea but I'm not sure how to make it work fully
 
@Daminark accurate
 
Amazing
 
Who even upgrades these instant regret playlists
They get better with every iteration
 
@BalarkaSen on that note: youtube.com/watch?v=y9YfSTJbUyQ
huzzah for troll games
 
Oh I have seen that one
It's on my top 10 list of best playthroughs
 
9:56 PM
lol
i've been watching a few mario maker vids lately and "Cat Mario" has popped up a few times
so I guess that's what inspired it
 
"This game is full of asshole"
My favorite description of the game is prolly "A guy with no dicks made this game, so he could fuck people in it"
 
lol
my fav lately is this series: youtube.com/…
 
Bookmarked. I need more rage game playthroughs in my life
 
Oh god I put so much time into Cat Mario
 
they get better as they go on tbh, as they start trolling each other harder and harder
 
10:01 PM
I think my life count was -281 in the play through that finally won
 
with the last few ones being just pure cruel gold
 
But over the course of my playing the game, throughout my life, I think I've probably died thousands of times
 
I didn't play the classics but I spent like a month's worth of time on "I Wanna Be The Guy: The Movie: The Game"
(Yes, that's the title)
 
So... the action of the group $S^1$ (i.e. the circle group) on $\mathbb{R}^2$ is now officially part of the precalculus curriculum for next Fall. :)
 
Sure, we all have
 
10:04 PM
@XanderHenderson ...so, trigonometry?
 
it comes after trig, but yes
first trig, then a more general discussion of group actions on $\mathbb{R}^2$
 
well, I think that basically boils down to "what happens to sine/cos if you shift by an arbitrary angle"
though it is prettier once you write it via matrix multiplication
 
yeah, but it also puts vectors (the action of $\mathbb{R}^2$ on Euclidean space) and rotations (the action of $S^1$ on Euclidean space) into the same context
and the goal is to avoid complex multiplication, as it is a pre-calc class that leads into a traditional calc sequence, which is in the context of real analysis
of course, the action of $S^1$ on the unit circle is complex multiplication, modulo the notation
 
True
I guess in my eyes the circle action is implicit in the any geometric derivation of the angle addition formulae
 
it is
that is, actually, why we are bothering to introduce the notation and language
we give the angle addition formulæ from some kind of abstractly defined "product" of points on the circle, then go back and make it a bit more rigorous
time permitting, at any rate
quarters go by too fast :\
 
10:15 PM
Hmm, sure
 
Reminds me of the proof that multiplication by a complex number $\alpha$ with $|\alpha|=1$ is a rotation, by showing that it's an isometry and that it fixes only one point
and thus it's a rotation, and since it takes $1$ to $\alpha$, it's a rotation by the angle from $1$ to $\alpha$.
(Isometry 'cause $|\alpha x-\alpha y|={|\alpha|}{|x-y|}=|x-y|$)
 
Could someone be so kind to help me with this problem? I'm designing a game and I have to calculate if the enemy ship has in range of view the player ship in order to shoot him but I am having problems with the maths.

I have two figures: $A$ (the enemy) and $B$ (the ally).
This is my information of the figures (ships):

From A (enemy):
$rotation$ = Rotation of the figure on radiants of figure A.
$Xa$, $Xa$ = Coordinates
$range$ = Range of view of figure A.
$angle$ = Angle of view in radiants of figure A.
That is a visual representation of what I am trying. The green triangle is the enemy an the red cicle is the player
 
Both >?
 
@AkivaWeinberger Yes, I know, a bit strange. But if I use $<$ on the second one it doesn't work in any case
If I put $<$ on the second one instead of work only the upper left corner it works the bottom right one
 
10:45 PM
The most meta one so far
 
I don't... even... wha?
 
it all started with
 
 
1 hour later…
11:47 PM
preparing this talk for tomorrow from my 2 week-old notes
bad idea
what on earth is past me trying to tell present me
 
Yeah, don't do that...
 
11:58 PM
@XanderHenderson you're not the boss of me
 
00:00 - 19:0019:00 - 00:00

« first day (2819 days earlier)      last day (2197 days later) »