irving

ZachWeinersmith, 1 month ago to comics

AI
http://smbc-comics.com/comic/ai-11 (bonus panel here)
#smbc #hiveworks #comics #webcomics #ai

ZachWeinersmith, 2 months ago to random

I have an idea for a version of rock paper scissors. here's the idea: it's the normal game, only you start with 6 tokens. Two for rock, two scissor, two paper. when you throw one, you have to give up a token and reassign it to one of your other piles. What do you think would happen?

Inclination is it gets boring because nobody wants to go below 1 token. Maybe there's some way the opponent can force? E.g. if they have 3 on one pile they can kill 2 to destroy one of yours?

gregeganSF, 2 months ago to random

“Kenan Malik is an Obserrver columnist”

Come on, Grauniad, we’re glad you don’t have AIs writing your columns, but nobody is in such desperate need of an automated spellchecker as you are.

https://www.theguardian.com/commentisfree/2024/mar/10/ai-wont-destroy-us-but-tech-giants-use-fear-it-will-to-evade-scrutiny

irving, 2 months ago to random

The Mandelbrot set, with infinity in the center.

j_bertolotti, 9 months ago (edited 9 months ago) to random

Before becoming a father, I didn't realize how crazy powerful reverse psicology is with kids. By far the best way to have my 5y/o to do anything is to tell him to do it in the most absurdly wrong way possible 😆

j_bertolotti, 9 months ago to physics

A little classical mechanics problem you can solve without doing any calculation:
Consider the hyper-simplified problem of a bell-shaped hill, and a point rock that can slide without friction up and down the hill. If you start with the rock at the bottom, and give it exactly the kinetic energy needed to arrive to the top and stop there without sliding on the other side, how long will it take to arrive there?
#Physics #ITeachPhysics

gregeganSF, 9 months ago to random

Great article in @QuantaMagazine.

What’s so striking about this to me is that if I hadn’t been told the history of the problem, I’d have assumed that something so structured would have an easy solution, as a sum of 2^n terms, or a recurrence relation.

https://www.quantamagazine.org/ninth-dedekind-number-found-by-two-independent-groups-20230801/

irving, 9 months ago to random

@henryseg Do you know if there is a sensible definition of the game theoretic value density of Go on the infinite 2D lattice, using an appropriate limiting mean value?

It seems like optimal play would want to be up to some $c\omega$ ordinal for some reasonable c?

zenorogue, 9 months ago to Baduk

A game of Go on a hyperbolic manifold, played in the HyperRogue discord server, ended like this.
#noneuclidean #noneuclideangeometry #rogueviz #bringsurface #baduk

dpiponi, 9 months ago to random

One way to view automatic differentiation is to think of it as adjoining an "infinitesimal" element d to the reals, such that d²=0, to the reals, ie. forming ℝ[d]/(d²). If f is a polynomial then f(x+d)=f(x)+df'(x) giving a nice way to compute derivatives on a computer - especially as it can be extended to rational and even transcendental functions f. It doesn't form a field though. For example you can't always divide by d.

TIL There is a field, named after Levi-Civita, that generalises ℝ[d]/(d²) quite a bit. Every element is a kind of power series in an infinitesimal ε. More precisely, each element is a "formal" sum ∑aᵢεⁱ where the sum is over some subset S of the rationals which is left-finite, ie. for any z, S has only finitely many elements less than z. Addition and multiplication work in the way you might guess.

This means we can form things like ε^(1/2) or even the "infinite" 1/ε. It's not just a field, it's an ordered field so we have, for example, that 1 > ε^(1/2) > ε > ε² > 0.

You can even construct a Dirac delta-like function δ(x) = ε/π(x²+ε²).

It's all very similar to non-standard analysis but this particular construction doesn't require anything non-standard.

https://en.wikipedia.org/wiki/Levi-Civita_field

ZachWeinersmith, 9 months ago to random

Every expression in this painting is priceless

gregeganSF, 9 months ago to random

A sphere is the nicest surface with constant Gaussian curvature κ=1, but these surfaces also have κ=1, except at the cusps.

In cylindrical coords:

z = √(1-χ²) E(arcsin(ρ/χ) | χ²/(χ²-1))

z is measured from a cusp
χ is radius of equator
E is an elliptic integral of the 2nd kind

A sphere deforms, stretching along one axis and shrinking in width into something like a prolate ellipsoid, but with a cusp at the poles.

ZachWeinersmith, 9 months ago to random

http://smbc-comics.com/comic/paradigm

johncarlosbaez, 9 months ago (edited 9 months ago) to random

Phoenix is a sprawling city with a population of 1.6 million. The daily high temperature there has exceeded 43° C (110° F) for 25 consecutive days. For two weeks, the low has exceeded 32° C (90° F).

So, Phoenix is a glimpse into the hellish future we are sleep-walking into.

Like most American cities, Phoenix is addicted to cars. They've tried to make a lot of the highways concrete, which is whiter. But still, there's a lot of black asphalt. Now this asphalt routinely reaches 82° C (180° F), so it can burn the bottom of your feet - even when you're wearing shoes.

Outdoor jobs like construction work are no good when it gets this hot. So Phoenix is starting to try container storage housing, which can be built indoors in air-conditioned environments and then installed on site using a crane.

The mayor there says:

"Well, our priority is to get people into indoor shelter. Thanks to our partnership with the Biden administration, we now have hundreds of millions of dollars that we can put towards indoor, air-conditioned shelter. So that is our top priority."

Air conditioning. Lots of air conditioning! So lots of electricity. Last year 42% of the electric power in Arizona came from natural gas, 29% from nuclear power, 12% from coal, and 10% from solar, 5% from hydroelectric and 1% from wind.

I'm glad there's a lot of nuclear power. All of this comes from just one plant southwest of Phoenix, with three reactors and a capacity of 4 gigawatts. Imagine tripling this, and boosting solar too. This wouldn't save Arizona from global warming, but it would be a step in the right direction.

gregeganSF, 9 months ago to random

TIL about the sverdrup:

“As a consequence, the resulting Gulf Stream is a strong ocean current. It transports water at a rate of 30 million cubic metres per second (30 sverdrups) through the Florida Straits. As it passes south of Newfoundland, this rate increases to 150 sverdrups.”

https://en.wikipedia.org/wiki/Gulf_Stream

“In oceanography, the sverdrup (symbol: Sv) is a non-SI metric unit of volumetric flow rate, with 1 Sv equal to 1 million cubic metres per second (264,172,052 US gal/s).”

https://en.wikipedia.org/wiki/Sverdrup

j_bertolotti, 9 months ago to random

When you see a "list of academics on Mastodon", but Physics is not even one of the possible options 🤷🏻‍♂️
https://t.co/29WkGr0ABw

gregeganSF, 9 months ago to random

This @QuantaMagazine article:

https://www.quantamagazine.org/to-move-fast-quantum-maze-solvers-must-forget-the-past-20230720/

is great, but it left me wondering what it could mean to “solve” a maze if the solution isn’t a description of a complete path from entry to exit.

The linked paper https://arxiv.org/abs/quant-ph/0209131 has the answer. Vertices are given random labels, and a maze oracle, queried with one vertex’s label, tells you the labels of the vertices you can reach from it. To “solve” the maze means identifying the label of the exit.

So a quantum algorithm that won’t remember any specific sequence of vertices through the maze can still tell you the name of the exit vertex.

ZachWeinersmith, 9 months ago to random

Just figured out how to rebrand smbc for the modern moment

gregeganSF, 9 months ago to random

[On the IMAX print of Oppenheimer. Good grief. At least he didn’t demand doubling the frame rate ...]

"It costs around $80,000 and weighs 260 kilograms; without the upgrade, the print would have fallen off the edges of the film platter.

"Previously, Nolan's Interstellar was the longest you could run at 170 minutes … Nolan didn't want to compromise his vision of Oppenheimer and edit it down, so he said to IMAX 'you need to extend the equipment'.

https://www.abc.net.au/news/2023-07-20/imax-oppenheimer-christopher-nolan-imax/102620264

ZachWeinersmith, 9 months ago to random

Found a solution to AI alignment