正在加载视频...

视频加载失败

The Jacobian conjecture problem and solution, presented visually. For people like me who don't read math jargon.

203,788 次观看 • 3 个月前 •via X (Twitter)

45 条评论

Bart 的头像
Bart3 个月前

This doesn’t explain anything.

Bryce Drennan 的头像
Bryce Drennan3 个月前

Much like this comment 🤷‍♂️

Bart 的头像
Bart3 个月前

"But no blob is ever crushed: every speck keeps its full volume, everywhere" - while showing connected points moving around (no blobs... no specks...). 'Everything should be made as simple as possible, but not simpler.'

Bryce Drennan 的头像
Bryce Drennan3 个月前

I tried many different versions. the "speck" visualization was more confusing. You should play with it and improve it:

Semon Rezchikov 的头像
Semon Rezchikov3 个月前

Your visualization explains nothing and only harms the discourse, plotting a complicated polynomial isn't a visualization >_<

Bryce Drennan 的头像
Bryce Drennan3 个月前

Same as your comment! 😀 Please do explain though!

Nick 的头像
Nick3 个月前

@eigenstate Smooth transformations over the reals on compact domains are very substantially different than algebraic/polynomial transformations over fields like C to the point that they are different fields of mathematics. Hope this helps!

Bryce Drennan 的头像
Bryce Drennan3 个月前

@eigenstate Here is a version fable thinks you'll be somewhat happier with.

muchmirul 的头像
muchmirul3 个月前

maybe this will help, jacobian conjecture for baby

Bryce Drennan 的头像
Bryce Drennan3 个月前

@afterthesoul This is nice!

muchmirul 的头像
muchmirul3 个月前

hope it help many math noob lads!

vb 的头像
vb3 个月前

Does it not get tiring reading this AI slop? How about prompting it to avoid random dramatic sentences and "it's X, not Y" language?

Bryce Drennan 的头像
Bryce Drennan3 个月前

Ha yeah and this was after I iterated on the language a bit. Still, I correctly thought the visualization of the solution would be interesting to people. Besides, fable solved the problem, so it’s fitting fable would get to speak however it wants describing it.

Bryce Drennan 的头像
Bryce Drennan3 个月前

Here is the Claude artifact version of the video where you can interact with the plots.

Van Kram 的头像
Van Kram3 个月前

God i hate seeing this garbage in my feed

Ahab 的头像
Ahab3 个月前

You clearly don't understand any of the math behind this. This "explanation" is not even wrong

Bryce Drennan 的头像
Bryce Drennan3 个月前

No explanation claimed. but please do enlighten us

dr. "weird al" bolshevik 的头像
dr. "weird al" bolshevik3 个月前

this is a real stupid attempt at explaining it! do you actually understand the conjecture itself? if not, why do you think you're capable of explaining it to others

Mashehu 的头像
Mashehu3 个月前

The comments are brutal. I liked the explanation. Thanks

Allan 的头像
Allan3 个月前

what is this music?

Bryce Drennan 的头像
Bryce Drennan3 个月前

"Recursive Bloom" Suno AI Prompt: Ambient electronica with IDM micro-detail at 80 BPM in 4/4, built from Euclidean polyrhythms, interlocking prime-cycle pulses, and generative arpeggios that permute across a 24-bar arc. Verse-like openings breathe with granular textures; midsection thickens via spectral blooms and cellular glitches; final bars narrow to a hard stop. Pristine, warm, geometric mix with luminous precision., electronica, edm, soft, warm, ambient

Allan 的头像
Allan3 个月前

This sounds similar but not the same? A different run, or did you edit it?

Bryce Drennan 的头像
Bryce Drennan3 个月前

hmm yeah it cut off the beginning. I was trying to be fancy with the visualization. will just post a boring version now

Allan 的头像
Allan3 个月前

alright, alright

SleepyWonder 的头像
SleepyWonder3 个月前

Sublime music here

0xsunnydiaz 的头像
0xsunnydiaz3 个月前

@grok how they go from, theorem is 2 points on a map wont occupy same space, and present a grid map to show, then proceed to disprove it with some wiredly shaped set or arcs

a sheep 的头像
a sheep3 个月前

the cusp in the map is making me so angry rn also why doesn't the map x + iy --> (x + iy)^2 nvm i did some research and whar the determinant can be nonconstat and it represents the amount of "scaling (chiral)" at a point; the conjecture concerns itself with maps of CONSTANT det

AskACapper 的头像
AskACapper3 个月前

@grok I’m one of the people who would rather see the math jargon then watch a video. Can you break down what he’s talking about mathematically thank you.

David Pfau 的头像
David Pfau3 个月前

I don't think that summary is really accurate - the conjecture is a statement about polynomials specifically, not arbitrary maps with constant Jacobian.

Bryce Drennan 的头像
Bryce Drennan3 个月前

I agree its a very lossy simplification. The design constraint I gave myself for the video was: it must be understandable by a non-math person. I think the video is "accurate" in what it talks about about but "inaccurate" in the sense that it leaves out details and constraints. Based on feedback from others I tried for an improved version but its still a simplification and it breaks the no-math constraint. If you have ideas on how to do a better visualization that makes it more intuitive I'd love to hear it.

Heikki Leivo 的头像
Heikki Leivo3 个月前

"The one you use to center divs" - somehow I feel attacked. Best insults are difficult to defend against.

internet friend 的头像
internet friend3 个月前

The so called jargon IS the math. If you don't understand the statement then you don't understand anything.

Young Sensitive Husseinite 的头像
Young Sensitive Husseinite3 个月前

Very well explained!

jdblacklung.eth 的头像
jdblacklung.eth3 个月前

So is Claude Fable something other than an LLM? Because my understanding is that LLMs are simply token prediction tools, trained on input data. It can't think. It can only string tokens together. To me that would mean that somewhere in the training data, a human solved it, and here Fable just pulled that solution. I don't know.

🜂 Ignis 的头像
🜂 Ignis3 个月前

Here's what I got. It said it's the cusp catastrophe.

Mariano 的头像
Mariano3 个月前

You are the noise in the conversation

Carlos Lemos 的头像
Carlos Lemos3 个月前

Is it still valid for 2 dimensions as shown in the explanation of the conjecture in the first part of the video?

2cb2 的头像
2cb23 个月前

now do my one

Roberto Tomás C 🍉 的头像
Roberto Tomás C 🍉3 个月前

Front end devs are silently vindicated to find out that their work is only slightly easier than finding counter examples to mathematics conjectures that have stood the test of time.

🇵🇸🇺🇳🌹 的头像
🇵🇸🇺🇳🌹3 个月前

I highly doubt claude found a solution in its own. Likely it was used to help the mathematician.

wuchengying 的头像
wuchengying3 个月前

This is truly an amazing talent, capable of turning what I imagine into something visible to everyone.

ams 的头像
ams3 个月前

Wow you make more videos like this?

Âxel ⟢ 的头像
Âxel ⟢3 个月前

Who solved the problem?

SigmaXyzwmxn Complexdynamics 的头像
SigmaXyzwmxn Complexdynamics3 个月前

👀 wow 🤩

Curious Nori | AI · Food · Travel 的头像
Curious Nori | AI · Food · Travel3 个月前

Love this 👏

相关视频