Vitaliy Kaurov (@superflow) header
Vitaliy Kaurov (@superflow) profile picture

Curious. Entangled. Director | Chief Editor | Physicist @ Wolfram Research

Joined May 2009
๐Ÿš€ Today 1st time in human history! Took spaceship 49 years what light does in 1 day. The ๐“ต๐“ธ๐“ท๐“ฐ๐“ฎ๐“ผ๐“ฝ distance call: 2 days to chat. ๐–๐จ๐ฅ๐Ÿ๐ซ๐š๐ฆ AstroGraphics builds unusual view: Voyager now can "see" the whole solar system including the Pluto orbit - and it's now tiny in its view. ๐Ÿ”ด ๐–๐จ๐ฅ๐Ÿ๐ซ๐š๐ฆ code & article: community.wolfram.com/t/2861โ€ฆ 22 watts. 26 billion km. NASA still hears Voyager. 17 km/s. One light-year: another 17,700 years. A 1977 machine. Ten watts buy 2.5 more years. ๐–๐จ๐ฅ๐Ÿ๐ซ๐š๐ฆ ๐‹๐š๐ง๐ ๐ฎ๐š๐ ๐ž turns Voyager 1 into a computable object: its trajectory, speed, launch date, distance from Earth. Calculate when its radio signal takes exactly 24 hours to arrive; cross-check with NASA's JPL data. Plot its journey and discover a paradox: Voyager keeps flying away, yet Earth sometimes moves closer! Our orbit changes the distance, and calculations show why the one-light-day boundary is crossed only once. Then flip the perspective: ๐€๐ฌ๐ญ๐ซ๐จ๐†๐ซ๐š๐ฉ๐ก๐ข๐œ๐ฌ places us at Voyager, drawing the planetary orbits against the distant stars. Not an artist's imagination or AI-generated scene โ€” an astronomical view computed from real trajectories. The same notebook even calculates how an ordinary radio transmitter can still be heard across 26 billion kilometers. Voyager entered interstellar space in 2012. Yet storms from our Sun still reach it, sending shock waves through space between stars. Only two scientific instruments remain active, detecting these disturbances. In 40,000 years, Voyager will pass another star โ€” and still miss it by nearly 2 light-years! Its power will die long before then. But Voyager won't stop. On board: Golden Record carrying Earth's music, sounds of nature and greetings in 55 languages. A message from humanity traveling through the galaxy long after its last radio signal.
1
4
208
๐Ÿš€ NEW research that took a century to arrive. 1902 - a newspaper puzzle. 4 weeks later - a solution. ๐Ÿ๐Ÿ๐Ÿ ๐ฒ๐ž๐š๐ซ๐ฌ later - a proof that the solution is optimal. ...or the MINIMAL possible number of pieces needed to cut an equilateral triangle into a square is 4. Henry Dudeney posed the challenge and credited reader C. W. McElroy with a four-piece solution. Demaine, Kamata and Ueharaโ€™s optimality proof appeared as a preprint in 2024 and was published in 2026. Showing four pieces work needs one successful arrangement. Proving three cannot means excluding every possible design. Even three pieces can have wildly complicated outlines! Few pieces, enormous freedom of construction. The result covers straight-edged pieces rearranged without flipping. Note how pure 2D transformation in plane must break the loop. But in 3D we keep the loop whole. I explore history of this problem, references, animation construction and symbolic math in this piece on Wolfram Community: ๐Ÿ”ด Wolfram code & article: community.wolfram.com/t/2827โ€ฆ
1
2
145
There is a ๐ฅ๐ข๐ง๐ค๐š๐ ๐ž signing ๐ฒ๐จ๐ฎ๐ซ ๐ง๐š๐ฆ๐ž โœ๏ธ Rigid sticks link. One crank spins. Your name written. What does the mechanism look like? Compute it! With Wolfram and noncommutative algebra. Even a ๐ฌ๐ญ๐ซ๐š๐ข๐ ๐ก๐ญ ๐ฅ๐ข๐ง๐ž was once a hard mechanical problem. Watt's 1784 steam-engine linkage could only approximate one. In 1864 Peaucellier finally forced hinged rods to draw it exactly. Then an absurd jump: 1876, Alfred Kempe argued that linkwork could draw ANY algebraic plane curve. Even stranger: his proof had a flaw; the universality result survived and was rigorously established much later. But universal does NOT mean practical. Kempe himself warned that perfect generality produced hopelessly complicated machinery. Take an ellipse: unoptimized Kempe construction - 158 links, 235 joints. Modern method by Christoph Koutschan and colleagues - 8 links, 10 joints. Same ellipse. To prove a machine exists is one problem; to compute a ๐ฌ๐š๐ง๐ž ๐ฆ๐š๐œ๐ก๐ข๐ง๐ž is another. They also took "signs your name" almost literally: approximate the giant J in John Hancock's signature on the Declaration of Independence, feed the curve to the algorithm โ†’ 26 links, 37 joints. It draws the J. And collides with itself seven times. See the marvelous machine here: ๐Ÿ”ด Wolfram Community notebook + code: community.wolfram.com/t/2191โ€ฆ
8
166
Grasping OpenAI $1M math ๐ฐ๐ข๐ญ๐ก๐จ๐ฎ๐ญ ๐€๐ˆ ๐Ÿš€ ~10,000 AI agents. ~88 hours. 166-page proposed proof. Now: turn the key idea into clear demo with Wolfram Language. ๐๐š๐ฏ๐ข๐ž๐ซโ€“๐’๐ญ๐จ๐ค๐ž๐ฌ is basically a prediction machine for fluids: give it the state of air or water now, calculate how it moves later. The $1M Millennium question asks something much stranger: if a 3D fluid starts perfectly smooth, must the math stay smooth forever - or can motion concentrate so violently that velocity becomes infinite in finite time? Here is OpenAIโ€™s proposed solution... Start from rest. Apply an external force that stays smooth. Yet the fluid builds an increasingly concentrated vortex until velocity grows without bound. The force stays well behaved; the fluid creates the blow-up internally. Watch the vortex core. It becomes progressively thinner and more elongated as the fluid concentrates into an ever smaller region. At the same time, velocity keeps growing; in OpenAIโ€™s construction it becomes unbounded at a finite time. That infinite-velocity blow-up is the singularity. If AI makes proofs abundant, understanding may become the bottleneck. Formal proof settles the mathematics; executable computation lets us enter the mechanism, modify it and see what is going on. Jayanta Phadikarโ€™s new Wolfram article compresses 166-pages proof into intuition: ๐Ÿ”ด Full Wolfram code + article: community.wolfram.com/t/2801โ€ฆ
92
How many ๐ž๐š๐ซ๐ญ๐ก๐ช๐ฎ๐š๐ค๐ž๐ฌ last month ? Roughly. Think of a number. Expand for answer ๐Ÿ‘‡... Not all tiny shakes are detectible. A more meaningful question is, how many fully detectible earthquakes happened on Earth last month? Those well detectible are roughly above magnitude M5+ and so, per Wolfram one-liner: In[]:= EarthquakeData[All, 5, {2026, 7}] // Length Out[]= 251 Animation is snapshot of 2025 with magnitude M2+ and data visualizing depth and magnitude. Look closer - you can see some points under the Earth surface indicating deeper location. ๐Ÿ”ด Wolfram code & article by artful @ : community.wolfram.com/groupsโ€ฆ Earthquakes are fascinating. Have you ever thought: - Can earthquakes be helpful and how? - Can we detect earthquakes on other planets? - Can an earthquake change Earthโ€™s rotation? - Can one earthquake trigger another across the planet? - Why can a smaller earthquake can be more harmful? - Does Earth ring like a bell? - Can we predict earthquakes? ๐‹๐ž๐ญ ๐ฆ๐ž ๐ค๐ง๐จ๐ฐ ๐ฒ๐จ๐ฎ๐ซ ๐œ๐จ๐จ๐ฅ ๐Ÿ๐š๐œ๐ญ๐ฌ ๐จ๐ซ ๐š๐ง๐ฌ๐ฐ๐ž๐ซ๐ฌ.
1
114
What's common in ๐ฆ๐š๐ซ๐ค๐ž๐ญ๐ฌ & ๐ฉ๐ž๐ง๐๐ฎ๐ฅ๐ฎ๐ฆ? Even a simple equation can show chaos - no randomness needed. Best forecasts will fail eventually due to errors amplified exponentially. Chaos here is not randomness. It is exact dynamics amplifying incomplete knowledge of the present. Randomness from coupled environment might add complexity, but is not required for a minimal math model to be chaotic. This distinction is critical, but is lacking in basic education. Simple model does not mean simple behavior. A simple supply chain can fail due its fundamentally flawed setup. Many simple systems can venture into chaos in a particular parameter range - the danger is naive human assumptions of predictability there just because system looks simple. Double pendulum is deterministic: exactly the same state produces exactly the same motion. Yet no experiment can measure its two angles and two angular velocities with infinite precision. In chaotic regions, the missing decimals matter. A tiny initial error grows roughly exponentially. Neighboring pendulums first move alike, then lose resemblance. Better measurement postpones that split; it cannot remove it. No pendulum acts on another. Each is solved separately. Apparent coordination comes from how the square is built: horizontal and vertical position encode the two starting angles. Neighboring cells therefore begin from neighboring states. The waves come from organized starting conditions, not interaction between pendulums. ๐Ÿ”ด Wolfram code & article: community.wolfram.com/groupsโ€ฆ
1
2
95
Replying to @fermatslibrary
Sphere is easy. What if ๐ ๐ž๐จ๐ฆ๐ž๐ญ๐ซ๐ฒ ๐ข๐ฌ ๐š๐ซ๐›๐ข๐ญ๐ซ๐š๐ซ๐ฒ, ๐ง๐จ ๐œ๐จ๐ง๐ฏ๐ž๐ง๐ข๐ž๐ง๐ญ ๐Ÿ๐จ๐ซ๐ฆ๐ฎ๐ฅ๐š or parameterization? In Wolfram just define the geometry (region): โญ RandomPoint[region, n] โญ and the the uniform sampling is handled automatically. The deeper point: uniformity belongs to the geometry, not the coordinates used to describe it. The same RandomPoint workflow applies to parametric, implicit, mesh and transformed regions: reference.wolfram.com/languaโ€ฆ
1
1
19
1,505
Replying to @fermatslibrary
Not just funny names. ๐‰๐„๐‘๐Š matters in elevators and trains; ๐’๐๐€๐ matters in robot arms and drones, where sudden changes cause vibration or unstable motion. Tiny Wolfram code view of the full chain: Plot[NestList[D[#, t] &, Sin[t] + Cos[6 t/5], 6],{t, 0, 10}]
3
1
20
981
Replying to @fermatslibrary
With 3D ๐ž๐ฆ๐›๐ž๐๐๐ข๐ง๐  ๐๐ข๐ฆ๐ž๐ง๐ฌ๐ข๐จ๐ง animating gives nice 3D space "sense". Done with one-liner: CantorMesh[4, 3] // Tour3DVideo
1
6
1,002
Replying to @fermatslibrary
The Cantor set is a ๐Ÿ๐ซ๐š๐œ๐ญ๐š๐ฅ and can be specified in higher ๐ž๐ฆ๐›๐ž๐๐๐ข๐ง๐  ๐๐ข๐ฆ๐ž๐ง๐ฌ๐ข๐จ๐ง. For example, the Wolfram code below uses CantorMesh to set 3D ๐ž๐ฆ๐›๐ž๐๐๐ข๐ง๐  ๐๐ข๐ฆ๐ž๐ง๐ฌ๐ข๐จ๐ง. The important subtlety is that embedding dimension and fractal dimension are not the same thing: embedding dimension is the space where the object is placed, while fractal dimension describes how much space the object itself fills as we keep zooming in. The ordinary Cantor set lives on a line, but its fractal dimension is only Log[2]/Log[3] โ‰ˆ 0.631, so it is more than isolated dust but less than a full line. In 3D Cantor dust, the embedding dimension is 3, while the fractal dimension is 3 Log[2]/Log[3] โ‰ˆ 1.893, so it lives in 3D space but still does not fill volume. This is why the Cantor set is such a beautiful example: โ€œsizeโ€ can mean length, number of points, embedding dimension, or fractal dimension - and they do not have to agree. Check details here: - mathworld.wolfram.com/Cantorโ€ฆ - reference.wolfram.com/languaโ€ฆ
1
23
1,221
Replying to @fermatslibrary
I think it is absolutely stunning when works from the ๐ฉ๐ž๐ง-๐š๐ง๐-๐ฉ๐š๐ฉ๐ž๐ซ ๐š๐ ๐ž are so impactful they become part of vocabulary of modern programing languages in the ๐š๐ ๐ž ๐จ๐Ÿ ๐ข๐ง๐ญ๐ž๐ฅ๐ฅ๐ข๐ ๐ž๐ง๐ญ ๐ฆ๐š๐œ๐ก๐ข๐ง๐ž๐ฌ many hundreds years later. Especially serendipitous as he helped open the long road toward this age with his calculator. PascalDistribution and PascalBinomial are built functions in Wolfram Language eponymous with Blaise Pascal. reference.wolfram.com/languaโ€ฆ
2
5
370
โ™จ๏ธ ๐ก๐ž๐š๐ญ-๐ฆ๐š๐ฉ boosted molecule diagram: Topologically crowded atoms glow brighter. Lights roll from most to least busy spots. Various chemical metrics can be visualized with color fields (see Wolfram MoleculeValue) ๐Ÿ”ด Wolfram code & article: community.wolfram.com/groupsโ€ฆ Exact metric name used here: Topological Steric Effect Index (TSEI). Each atom gets a TSEI value: a bond-network estimate of local crowding around that atom. The animation is sorted by TSEI and unrolled from highest to lowest values. Each glow is a smooth function centered on an atom. TSEI sets its amplitude: higher value, brighter light, broader spread.
1
8
13
1,557
๐Œ๐š๐ซ๐ญ๐ข๐š๐ง ๐ƒ๐ข๐œ๐ก๐จ๐ญ๐จ๐ฆ๐ฒ got two mechanisms: a giant impact and internal magma flow. Over four billion years ago, a Pluto-sized object likely smashed into the northern hemisphere, blasting away the crust to leave a massive basin. Alternatively, a single giant plume of hot magma inside the planet pushed upward, thickening the southern crust. The most accepted theory combines both: the colossal northern asteroid strike generated a thermal shockwave that forced the planet's internal magma to rise and build the elevated southern highlands. Here is also static high resolution poster. Full screen video recommended for elevation texture. Full animation code: community.wolfram.com/groupsโ€ฆ
2
216
Mars is lopsided - called ๐Œ๐š๐ซ๐ญ๐ข๐š๐ง ๐ƒ๐ข๐œ๐ก๐จ๐ญ๐จ๐ฆ๐ฒ. Oceans in North, continents in south - if ones imagines water on Mars at the same 71% surface area as water on Earth. Mars topography is like Yin-Yang symbol: its highest spot Olympus Mons (solar system tallest mountain) is in lowlands North, and its lowest spot Hellas Planitia (one of the largest craters in solar system) is in highlands South. Amazingly simple programing trick gives minimal Wolfram code below to visualize all this fascinating and unique Mars topography. The trick: sample Mars geo-elevation uniformly from equal-area map projection and then quantile points at 71% - the value you get splits lowlands and highlands. If you flood lowlands to that value it yields 71% global water surface. ๐Ÿ”ด WOLFRAM CODE: elSamp = Flatten @ QuantityMagnitude @ GeoElevationData[ GeoProjection -> "CylindricalEqualArea", GeoZoomLevel -> 1, GeoRange -> "World", GeoModel -> "Mars" ]; tHeight = Rescale[Quantile[elSamp, 0.71], MinMax[elSamp]]; GeoGraphics[ GeoModel -> "Mars", GeoRange -> "World", GeoProjection -> "VanDerGrinten", GeoBackground -> GeoStyling["ReliefMap", ColorFunction -> (If[# < tHeight, StandardBlue, StandardOrange] &) ] ]
4
11
1
41
10,086
Replying to @fermatslibrary
Wolfram one-liner code verification: In[]:= Complement[Alphabet[], Characters["" <> CommonName@ElementData[]]] Out[]= {"j", "q"} This dataset only checks current official names. "Q" is absent today, but it was used historically in temporary IUPAC placeholders (e.g., ununquadium), leaving "J" as the only letter to never appear in any capacity.
3
692
This structure can be made w/ Wolfram 1-liner: p=Tuples[Range[-2,2],4] . I^{0,1,4/3,7/3}; RelationGraph[Abs[#1-#2]==1&,p, VertexCoordinates->ReIm@p] That's static full graph. RandomSample links, add them 1 by 1, and you get this animation.
Following up on the suggestion from Will Sawin, here is an illustration of the new configurations that disprove Erdos' unit distance conjecture (made with the help of ChatGPT 5.5 Thinking).
2
6
3
65
6,372