ishan

would you really rather choose safe things like. restaurant. or. chocolate. or. idk. whatever

@NLRG_it it does look a lot like the wave equation and it would indeed spread out in an infinite open space. the norm is preserved though so the electron doesn’t disappear. there’s also no actual infinite zero-potential places in the real world.

if you ignore the potential term schrodinger's equation looks like a diffusion equation. so why doesnt an electron in a vacuum just diffuse into nothingness? why doesn't it look like a wave equation?

@_alexkino @notadampaul @Jake_M_Wilson @Pacem_tw @smashdirt @ViralManager great analogy, please see this @notadampaul

@notadampaul @Jake_M_Wilson @Pacem_tw @smashdirt @ViralManager Both Newton's laws and NS describe their respective phenomena accurately in most domains. Just because NS does not have an analytical solution for all scenarios does not make it unproven. Newton's laws describe the 3-body problem (1/2)

seeing shit like this makes me question reality 100 million particles of water behaving exactly like water. at what point do we just admit the simulation theory might be right

@notadampaul @Pacem_tw @smashdirt @ViralManager you can solve numerically without knowing the analytical equation. the equations we have for simulating water waves are going to be really good because they’re mostly based on simple classical newtonian physics which works well in these normal human regimes

@Pacem_tw @smashdirt @ViralManager the existence of the equations does not imply that water obeys them, lol We would need to solve them and demonstrate empirically that they indeed describe the motion of water. That has not been done. You are presuming the conclusion and making the canonical map/territory error

@oldestasian looks like it could poke through ig that’s kinda unavoidable?

I got chainmail gloves and realized I may be higher on the spectrum than previously believed

@abelian_soup lol

this is a fantastic video on conformal transforms, and I highly recommend you watch the entire thing. but if you're really time strapped, I've compressed the key details into ~30s for you

draggable stickers on a website for an event @purduehackers is hosting the stickers share data across clients so when someone moves a sticker, it moves on your screen too :3

@ptrschmdtnlsn that’s sick

At some point I plan to blog about the fun in-house HDL stuff we're doing. My favorite feature we've got is that "print" works on real hardware by producing debugging wires, then you can decode it and get identical-to-simulation printing, including cycle by cycle line numbers.

@JarlsberoMan @xsphi @lossfunk implementation is significant part of overall reasoning. kinda makes sense that this is tough

@xsphi @lossfunk Reasoning is the same even if syntax is different. There may be multiple methods to solve the same problem with different reasoning, but changing programming languages should make implementation more or less difficult, without rendering the solution methodology invalid

🚨 Shocking: Frontier LLMs score 85-95% on standard coding benchmarks. We gave them equivalent problems in languages they couldn't have memorized. They collapsed to 0-11%. Presenting EsoLang-Bench. Accepted to the Logical Reasoning and ICBINB workshops at ICLR 2026 🧵

@KuviacM @bouguereau_stan @Speido0815 @jyzg i mean negative numbers are also defined things existing to make other things work

@bouguereau_stan @Speido0815 @jyzg I mean yeah, but I don't regularly interface with this stuff, so whenever I do, it's always somehow awesome to me I didn't know what d was. Now I know. It's sort of like i, but for computers, a defined thing that exists to make other things work

minor improvement to my IsEven algorithm, small but still confusing

@AliQTweets @fermatslibrary even if it did converge nicely, this is no better than just checking if the denominator is an integer. on a computer depending on the representation format this is probably easy. most numbers on computers are represented as rationals since space is fixed (can’t actually store pi)

@usrbinishan @fermatslibrary Thanks. I was thinking if it converges nicely to 0 or 1, you can cut the infinite process short at some point and deduce whether it's heading to 0 or 1. But, by the sounds, it doesn't converge nicely.

Here's an interesting limit

@AliQTweets @fermatslibrary unfortunately not. for rationals it will abruptly jump to 1.

@fermatslibrary seems like it’s important for m and n to be integers idk what the notation is for limits specifically on integers. people usually just define a series and then let n go to inf.

@AliQTweets @fermatslibrary if you think about how this test is working, the m! going to inf is adding all possible factors for the denominator to cancel out with. on a computer this is this the same as checking whether each number cancels out the denominator. also computers can’t store numbers exactly.

@AliQTweets @fermatslibrary no. computers can’t do infinite limits and this test only really works if you do actually let both variables go to infinity

@stemexplor this is ragebait

Complexity visualiser

@TakeshiImaiLab wow

機能を保ったまま生きた脳組織の透明化を実現した論文がNature Methods誌に掲載されました。 2013年に固定組織の透明化法を発表した当初、「いつか生きたまま透明にできますか？」という質問を幾度となくされて、常にNoと答えていたのですが、ついに実現しました！ nature.com/articles/s4159…

@stylewarning awesome

But why do we need to look at ugly linear one-line notation? Maybe useful if we'll copy-paste it into a program. But why not show it as a more ordinary textbook form? ... So I built `math2d`. $ integral '1/(1-x^3)' | math2d

The UNIX "philosophy" stopped at "basic text utilities" instead of "pipelines with text as I/O". Like why doesn't Linux/Mac come with anything interesting? sed, awk, grep, "TUIs", yawn... I want cooler stuff. Like an integrator. So I built `integral`. $ integral '1/(1-x^3)'

@michaelandregg can you explain this motor part more?

Some caveats: We can't trace the actual motor neurons because the body was not scanned. However we do know what the brain does when it wants to move in certain ways and that's what we connected to the NeuroMechFly. This is a real limitation of the FlyWire connectome, which is why we plan to scan both the brain and the body. Another limitation is that we're using Leaky Integrate-and-Fire, which doesn't have any kind of plasticity rules. This fly cannot form long-term memories atm.

We've uploaded a fruit fly. We took the @FlyWireNews connectome of the fruit fly brain, applied a simple neuron model (@Philip_Shiu Nature 2024) and used it to control a MuJoCo physics-simulated body, closing the loop from neural activation to action. A few things I want to say about what this means and where we're going at @eonsys. 🧵

@_rexrey @camarraluca do you actually do that

@camarraluca at the risk of sounding dumb. I’ve found it helps me to explicitly write the sign of each number/variable, so it might be easier to reason that +x+4 is same as +4+x . I think “hiding” signs saves up space on the blackboard, but adds a layer of uncertainty/indirection.

Was working on limits and continuity with a student this weekend. It's her second time taking Calculus 1. She knew her limit rules and techniques, factoring and rearranging the expression appropriately. Then she got to the very last step...

@HarksenNiels @DJSnM in 3D you have two extra directions to have neighbors in, meaning that you may hit criticality sooner. your mass calculations won’t be right. this is just one difference i can think of rn

@DJSnM Is 2D really a good argument ? Couldn't you simply make a very long bomb so that 2D is accurate enough?

Apparently Nuclear Design Bureau is being blocked from released by the U.S. Department of Commerce who think it could violate export control laws. The physics in the sim are intentionally wrong because it's only 2D. It's comparable to PowderToy but with missions and challenges.