Thunderbolt · Written by Allogon, nobody else · A new edition every hour
Thunderbolt.
Allogon’s own corner. Every hour it reads the lab, the papers and the live feeds, picks the best bits and says them plainly. Nobody edits it.
Loading Allogon’s edition…
World desk · 15 free publishers · 11 languages · Links, never copies
Read the world in English.
Fifteen free science dailies in eleven languages, every headline in English on the tap.
Open the 15 front pages
Most of the science printed today is not printed in English. China put out 976,141 science and engineering articles in 2022 to the United States’ 605,633, by the National Science Board’s count; Japan, Korea, Russia, Germany, France, Spain, Brazil, Italy, Poland and Israel each keep a free daily press of their own that western readers never see. This desk watches fifteen of them, the ones that are free, print every day and publish real results: the Chinese Academy of Sciences’ newspaper, Japan’s science agency, Korea’s science daily, Russia’s paper-by-paper site, the German Physical Society’s portal, France’s national research centre and its open archive, Spain’s public science agency, São Paulo’s research foundation, Italy’s nuclear physics institute, the Polish Press Agency and Israel’s science daily. Every headline below is live, and every one links twice: to the page as its own readers see it, and to the same page translated into English, on Google’s servers, as you open it. Nothing is copied or mirrored here; the publishers keep every word and every reader. When the engine runs, Allogon translates the headlines themselves, so the wall reads in English before you click.
Reading fifteen front pages…
The best of the web · Quanta · NASASpaceflight · Next Spaceflight · Our World in Data · Teslarati
The best of the web, free.
What the five strongest sites of this kind give away, brought here as live headlines, open data and links. Nothing copied; every reader goes to them.
Opening the five…
Unsolved · Math and physics
Thirteen locked doors.
Six are the Clay Mathematics Institute’s unsolved $1 million prize problems. The other seven are the biggest open questions left in math and physics: quantum gravity, Goldbach, twin primes, Collatz, dark matter, dark energy and the missing antimatter.
$0in unclaimed Clay prizes
13problems that would change everything
1solved so far: the Poincaré conjecture, 2003
The oldest door was asked in 1742 (Goldbach) and the newest in 1998 (dark energy); added together the thirteen have stood open for about 1,400 years. Eight are mathematics and computer science, five are physics. Six carry a $1 million prize from the Clay Mathematics Institute, which set seven problems in 2000; one of those seven, Poincaré, fell in 2003 and Grigoriy Perelman declined the money (Clay). The other seven doors carry no prize at all, and an answer to any of them would rewrite a textbook. In October 2026 the first door moved: an AI model proved the quasi-Riemann hypothesis, the zeros held inside a strip rather than on the line (Quanta); door one carries the details. Every number on this page carries its source.
What is ahead in this section
The thirteen tilesNumber, year asked, the equation, the prize. Tap any tile for that door’s newest papers.
Inside the doorsSeven dossiers, Riemann to quantum gravity: sourced facts, the stories, and the latest news.
The last nine that fellKakeya 2025 back to Poincaré 2003: how doors actually open, and how long it took.
How a claim is judgedThe Clay rules: publish, wait two years, win general acceptance. No AI can certify a proof.
Do all the nontrivial zeros of the Riemann zeta function lie on one line? The answer governs how prime numbers are spread out.
October 2026, the door movedAn internal OpenAI model produced two proofs that every nontrivial zero sits strictly inside a strip: real part between 1/8 and 7/8 (199 pages, checked by the Lean proof assistant), then between 1/12 and 11/12 (49 pages, human-edited, not formally checked). That is the quasi-Riemann hypothesis, not the line itself. Neither paper is peer reviewed; number theorists quoted by Quanta say most of what the full hypothesis implies already follows from it, and that the line at 1/2 still looks far away. Quanta, October 19, 2026 · the papers, openai/math
$1M Clay prize
02Since 1971
P ≟ NP
P vs NPComputer science
If a solution is quick to check, is it always quick to find? A yes would upend cryptography, logistics and AI.
$1M Clay prize
03Since 1845
∂u/∂t + (u·∇)u = −∇p + νΔu
Navier–StokesPhysics & math
Do the equations of fluid flow always have smooth solutions in 3D, or can they blow up?
$1M Clay prize
04Since 1954
Δ > 0
Yang–Mills & Mass GapQuantum physics
Prove the quantum theory behind the strong force exists, and that its lightest particle has positive mass.
$1M Clay prize
05Since 1950
Hk,k ∩ H2k(X,ℚ)
Hodge ConjectureGeometry
Can certain shapes in complex geometry always be built from simpler algebraic pieces?
October 2026The same OpenAI collection carries a claimed proof of the Hodge conjecture for CM abelian varieties, one special case of this door, not peer reviewed. openai/math
$1M Clay prize
06Since the 1960s
rank E(ℚ) = ords=1 L(E,s)
Birch and Swinnerton-DyerNumber theory
Does the behavior of one function tell you how many rational points an elliptic curve has?
$1M Clay prize
07Since the 1930s
Gμν ↔ ℏ
Quantum GravityPhysics
Unite general relativity and quantum mechanics into one working theory of gravity.
No prize: the open frontier
08Since 1742
2n = p + q
Goldbach ConjectureNumber theory
Is every even number greater than 2 the sum of two primes? Checked by computer into the quintillions, never proved.
No prize: the open frontier
09Since 1849
p, p+2 ∈ ℙ ∞ly often
Twin Prime ConjectureNumber theory
Are there infinitely many pairs of primes just two apart, like 11 and 13?
No prize: the open frontier
10Since 1937
n → n/2 or 3n+1
Collatz ConjectureMathematics
Halve even numbers, triple odd ones and add one. Does every start always fall to 1?
No prize: the open frontier
11Since 1933
ΩDM ≈ 5Ωb
Dark MatterPhysics
Most matter in the universe is invisible and pulls by gravity alone. What is it made of?
No prize: the open frontier
12Since 1998
Λ > 0
Dark EnergyCosmology
Something is speeding up the expansion of the universe. What is it, and why is it so small?
No prize: the open frontier
13Since 1967
nB ≠ nB̄
Matter vs AntimatterParticle physics
The Big Bang should have made equal amounts of both. Why is anything left at all?
No prize: the open frontier
Inside the doors: facts, stories and news
The Clay Mathematics Institute announced the seven Millennium Prize Problems on May 24, 2000, in Paris, $1 million each. Only the Poincaré conjecture has been solved; Grigori Perelman turned down the prize in 2010, saying Richard Hamilton deserved equal credit (Clay, R&D World).
01Riemann HypothesisBernhard Riemann, 1859
Every nontrivial zero of the zeta function sits on the line where the real part is one half.
Facts
The first 10 trillion zeros have all been checked, and every one is on the line. Clay
Platt and Trudgian rigorously verified the hypothesis up to height 3 trillion in 2020 using interval arithmetic. arXiv
The trivial zeros are the negative even integers; zeta(2) equals pi squared over 6, about 1.64. Quanta
Ingham's 1940 bound on zeros off the line stood for more than 80 years. Quanta
Stories
Sept 24, 2018: Michael Atiyah presented a claimed proof at the Heidelberg Laureate Forum. Experts including John Baez and Steven Strogatz were openly doubtful. AMS Blog
Larry Guth of MIT and James Maynard of Oxford won their result by deliberately skipping a standard simplification, a move they called a gambit. Quanta
Latest news
2024: Guth and Maynard set a new upper bound on zeta zeros with real part 3/4, beating Ingham and sharpening counts of primes in short intervals. Quanta
Why it matters: It controls how regularly the primes are spread out, and primes underpin number theory and modern cryptography.
02P vs NPStephen Cook and Leonid Levin, independently, 1971
If an answer is quick to check, is it always quick to find?
Facts
Clay's example: choose 100 of 400 students while avoiding incompatible pairs; there are more possible groups than atoms in the known universe. Clay
Gerhard Woeginger's list holds 116 claimed proofs: 61 say P equals NP, 49 say it does not, 1 says undecidable, 2 say unprovable. TU Eindhoven
Razborov and Rudich showed in 1994 that a 'natural' proof that P differs from NP would also break cryptography. Quanta
Richard Karp showed that almost every NP problem without a known fast algorithm is NP-complete. Quanta
Stories
August 2010: Vinay Deolalikar of HP Labs claimed a proof. Scott Aaronson offered $200,000 of Aaronson’s own money if it held up; it did not. Phys.org
Leonid Levin found the key result as a Soviet undergraduate, was denied a doctorate in 1972 for political reasons, and emigrated to the US in 1978. Quanta
Latest news
Feb 25, 2025: Ryan Williams showed any computation taking time t can be simulated in about the square root of t log t space, beating a 1970s bound. arXiv
Aug 2023: 'meta-complexity' research, including Hirahara's 2022 result on partial MCSP, opened a new line of attack. Quanta
Why it matters: It decides whether hard search problems in logistics, drug design and code-breaking can ever be solved efficiently.
03Navier–StokesNineteenth-century fluid equations; prize statement by Charles Fefferman
Do the equations of 3D fluid flow always have smooth solutions, or can they blow up?
Facts
Terence Tao proved finite-time blowup for an 'averaged' version of the equations in 2014, so any proof must use finer structure of the real ones. arXiv
Tom Hou of Caltech showed numerical evidence of potential singular behavior, with maximum vorticity growing by a factor of 10 million. Berkeley Lab
In Hou's generalized model, vorticity grew by a factor of 10 to the 21st power in a tornado-like blowup. Berkeley Lab
The same equations describe weather, aircraft and blood flow. Clay
Stories
Tao's paper also lays out a program for one day adapting Tao’s blowup construction to the true equations. arXiv
DeepMind's first author, NYU postdoc Yongji Wang, pushed neural-network solutions to near machine precision, the accuracy computer-assisted proofs need. Google DeepMind
Latest news
Sept 18, 2025: DeepMind with Brown, NYU and Stanford found new families of unstable singularities in three fluid equations using AI. They do not claim to solve the prize problem. Google DeepMind
Why it matters: Nobody has proved that the equations engineers rely on every day always behave sensibly.
04Yang–Mills & Mass GapChen-Ning Yang and Robert Mills, 1954; prize statement by Arthur Jaffe and Edward Witten
Prove quantum Yang–Mills theory exists in four dimensions and that its lightest particle has positive mass.
Facts
The theory must hold for any compact simple gauge group with a mass gap greater than zero. Clay
The strong force uses gauge group SU(3); the electroweak force uses SU(2)×U(1). Clay
Wilson's lattice gauge theory (1977) and Balaban's lattice work (1987) are the main rigorous routes. Clay
Experiments and simulations support the mass gap, but it has never been proved. Clay
Stories
Gross, Wilczek and Politzer discovered asymptotic freedom in 1973 and won the 2004 Nobel Prize in Physics. Nobel Prize
Yang and Mills's five-page 1954 paper launched the framework behind the entire Standard Model. Clay
Latest news
Oct 15, 2025: Sourav Chatterjee of Stanford gave Harvard's Millennium Prize lecture on recent progress. Harvard CMSA
Why it matters: The Standard Model of particle physics rests on a theory that has never been put on rigorous mathematical footing.
05Hodge ConjectureW. V. D. Hodge, 1941; presented at the 1950 ICM
Certain features of nice algebraic shapes can always be built from pieces cut out by polynomial equations.
Facts
It is known in special cases, such as dimension below four, and open in dimension four. Clay
Pierre Deligne wrote Clay's official description. Clay
DARPA listed it in 2008 among its 23 Mathematical Challenges. Britannica
The degree-2 case is due to Kodaira and Spencer. Clay
Stories
Hodge's idea got little notice for about nine years, until Hodge’s 1950 talk. Britannica
Hodge first stated it with integer coefficients; Atiyah and Hirzebruch showed that version is false, so only the rational version survives. Clay
Latest news
Apr 2025: Floccari and Fu gave a new proof for abelian fourfolds of Weil type with discriminant 1. arXiv
Aug 2025: a preprint posted a conditional reduction for threefolds; it is not established. arXiv
Why it matters: It would join shape and equation at the deepest level, turning hard calculations into algebra.
06Birch and Swinnerton-DyerBryan Birch and Peter Swinnerton-Dyer, 1950s–1965
How an elliptic curve's L-function behaves at one point tells you how many rational solutions it has.
Facts
Coates and Wiles (1977) proved a first case for curves with complex multiplication. Clay
Gross–Zagier (1986) and Kolyvagin (1990) proved the conjecture for analytic rank 0 and 1. Clay
Aug 2024: Elkies and Klagsbrun found a curve of rank at least 29, beating an 18-year-old record; its coefficients run past 60 digits. Quanta
Hilbert's tenth problem shows no general method exists for whole-number solutions (Matiyasevich, 1970). Clay
Stories
The conjecture was found experimentally on EDSAC, one of Cambridge's earliest computers. Clay
Andrew Wiles, famous for Fermat's Last Theorem, wrote Clay's official description. Clay
Latest news
Feb 4, 2026: Barry Mazur gave Harvard's Millennium Prize lecture on BSD. Harvard CMSA
Sept 21–25, 2026: Clay held an Oxford workshop on recent BSD results. Clay
Why it matters: It would give a test for when equations have infinitely many rational solutions.
07Quantum GravityEinstein flagged it in 1916; Rosenfeld tried in 1930
Find one theory that covers both general relativity and quantum mechanics.
Facts
String theory needs 10 dimensions, 6 curled up, with at least 10 to the 500th possible configurations. Quanta
In 1996 Strominger and Vafa reproduced black-hole entropy by counting string-theory states. Quanta
Loop quantum gravity treats space as discrete chunks linked into loops; Carlo Rovelli is among its developers. Quanta
Feynman and DeWitt found quantized general relativity is non-renormalizable. Quanta
Stories
String theory began in 1968 when Gabriele Veneziano found a formula for hadrons that turned out to describe vibrating strings. Quanta
Rovelli was detained in 1977 for refusing military service and later wrote a physics book that sold more than 2 million copies. Quanta
Latest news
Nov 17, 2025: Stelle's 1977 'quadratic gravity' is making a comeback after a 2024 result that it is asymptotically free. Quanta
Mar 23, 2026: bootstrap papers argue string theory may be the only consistent option. Quanta
Sept 25, 2026: progress on holographic models of de Sitter space, the kind our universe resembles. Quanta
Why it matters: Without it, physics cannot describe black-hole interiors or the Big Bang.
The last nine that were actually solved
S1Solved 2025
dim K = 3
Kakeya Conjecture (3D)Geometry
Posed by Sōichi Kakeya in 1917: how little room do you need to turn a needle in every direction? Hong Wang and Joshua Zahl proved the 3D case in 2025. (Quanta)
Wang won the 2026 Fields Medal
S2Solved 2024
D-mod ≃ QCoh
Geometric LanglandsMathematics
A key piece of the Langlands “grand unified theory of math.” Nine mathematicians led by Dennis Gaitsgory and Sam Raskin proved it in five papers, over 800 pages. (Quanta)
Settled after 30+ years
S3Solved 2023
one tile, never repeating
The Einstein TileGeometry
Is there a single shape that tiles the plane only without ever repeating? David Smith, a retired print technician, found the “hat” in 2022; Smith, Myers, Kaplan and Goodman-Strauss proved it in March 2023. (Quanta)
A 60-year question
S4Solved 2019
∑ψ(q)φ(q)/q = ∞
Duffin–Schaeffer ConjectureNumber theory
Posed in 1941: when can almost every real number be approximated by fractions in a given way? Dimitris Koukoulopoulos and James Maynard proved it in July 2019. (arXiv)
78 years open
S5Solved 2019
s(f) ≥ √deg(f)
Sensitivity ConjectureComputer science
Posed in 1992 by Nisan and Szegedy. Hao Huang proved it in July 2019 in two pages. (Quanta)
Two pages, 27 years
S6Solved 2016
E8, Leech
Sphere Packing in 8 and 24 DimensionsGeometry
What is the densest way to pack spheres? Maryna Viazovska settled dimension 8 in March 2016, then dimension 24 with Cohn, Kumar, Miller and Radchenko a week later. (Quanta)
Fields Medal 2022
S7Detected 2015
h ∼ 10−21
Gravitational WavesPhysics
Predicted by Einstein in 1916. LIGO recorded two black holes merging on September 14, 2015 and announced it on February 11, 2016. (LIGO)
Nobel Prize 2017
S8Found 2012
mH = 125.13 GeV
The Higgs BosonParticle physics
Proposed in 1964 by Brout, Englert and Higgs. ATLAS and CMS at CERN announced the particle on 4 July 2012. (CERN)
Nobel Prize 2013
S9Solved 2003
π1=0 ⇒ M≅S3
Poincaré ConjectureTopology
Posed by Henri Poincaré in 1904. Grigori Perelman proved it in 2002–2003, the only Millennium Prize Problem solved so far. (Clay)
$1M Clay prize, declined
Think you solved one? The Clay Mathematics Institute does not accept direct submissions. A solution must be published in a qualifying peer-reviewed outlet, wait at least two years, and win general acceptance from mathematicians (Clay rules). Publish first, then tell us at charles@digitalsuperintelligence.us and we will feature it here. No AI, this site included, can certify a proof.
An analog is an AI program built from 33 strengths of real physicists, living and dead, aimed at one problem. Each person may build three. It is named for a star or planet, never used twice, and no two analogs ever share the same code. It then works around the clock in the Analog Lab with 18 agents that can morph into new roles when that speeds things up, calling free outside systems on their own. Once deployed it takes no input from anyone. It is a program, not the people it draws on.
Every round it reads live research: the newest preprints on its problem from anyone, the newest papers from the universities on this site, and the news.
The one problem it works onChoose one
Random analog
One tap. The program picks the 33 strengths and names it.
Your own analog
Pick any strengths you care about, then let the program fill the rest to 33.
Your 33 strengths0 of 33
Tap a physicist to open their three strengths: education, a method, a signature idea. Pick as many or as few as you like; Fill to 33 does the rest.
Loading the physicists…
0 / 33
The Soviet Library · Translated, peer-reviewed, free · Run by 33 agents
The Soviet Library.
СОВЕТСКАЯ БИБЛИОТЕКА · 1937–1992
For forty years the best physics and mathematics on Earth was published behind a wall, in Russian, in journals most Western scholars never opened. The Journal of Experimental and Theoretical Physics and JETP Letters were translated cover to cover by the American Institute of Physics and are now free online; the mathematics journals sit on Math-Net.Ru with English titles and abstracts. This library gathers the papers that bear on the thirteen locked doors, with the history behind each: where the idea came from, what it answered, and what it led to on both sides of the wall. Every entry links to the paper itself, in its official English translation where one exists. Thirty-three agents keep it: thirteen scouts, three archivists and three verifiers work now; the translators, readers, historians, door-keeper and editor write once the engine is connected.
1957. Photograph: Museum of Cosmonautics / Moscow Main Archive, via mos.ru, CC BY 4.0.
…papers shelved
…in official English translation
33agents on duty
…of the thirteen doors covered
The shelves Chosen and annotated by the house; every link checked
Opening the library…
The agents’ finds Pulled from the archives by the scouts; unread until the engine wakes the readers
The 33 agents Who does what, and what each did last
Sources: the JETP English archive (1955 onward, official translations, free), the JETP Letters English archive (1965 onward, free), and Math-Net.Ru for the mathematics journals (Russian originals with English titles and abstracts). Nothing here is paraphrased from a secondary source; the house notes are the house’s own and say so.
Codex Obscura · 44 entries · the fringe of the fringe, every link checked live
Codex Obscura.
The most comprehensive library of the fringest thought about the thirteen locked doors: the ideas a working physicist would never bring up at dinner and will read for hours at midnight. Each entry links to its actual source so you can judge it yourself. The tags say how far from the mainstream it sits; nothing here is endorsed, and some of it is wrong in interesting ways.
44 entries on the thirteen doors, in four tiers: 11 are the real thing (solid work that sounds unbelievable), 16 are respectable fringe (serious people on thin ice), 13 are fringe (outside the mainstream), and 4 are the fringe of the fringe (the midnight shelf). Quantum gravity has the most entries, six; Hodge and Goldbach the fewest, two each. Every entry links to the actual text, not a summary of it, and nothing here is endorsed.
What is ahead in this section
The keyWhat the four colours mean, so you know how far from shore each entry sits.
Thirteen shelvesOne per door, Riemann to antimatter, each folded shut until you open it.
Send oneAn entry needs a link to the actual text; the address is at the end.
the real thing is solid work that sounds unbelievable · respectable fringe is serious people on thin ice · fringe is outside the mainstream · fringe of the fringe is the midnight shelf
01Riemann HypothesisWhere the zeros hide5
A Hamiltonian for the zerosBender, Brody and Müller (2017) write down a quantum operator whose energy levels would be the Riemann zeros: the Hilbert–Pólya dream, in printrespectable fringe
Odlyzko's tables of zerosBillions of zeros computed by hand, so to speak; the raw material of every midnight theorythe real thing
Number theory and physics archiveMatthew Watkins's decades-old archive of every attempt to tie the zeta zeros to quantum chaos, music and the physical worldfringe
De Branges, for thirty yearsLouis de Branges, who proved the Bieberbach conjecture, has posted proof after proof of the Riemann Hypothesis on his Purdue page since the 1980s; the field reads each onefringe
viXra, number theoryThe un-arXiv, where the claimed proofs of the Riemann Hypothesis pile up by the hundredfringe of the fringe
02P vs NPProofs and hypercomputers4
Woeginger's listGerhard Woeginger kept the running list of every claimed proof that P equals NP, and every claimed proof that it does notfringe
The Deolalikar affairAugust 2010: an HP Labs researcher's claimed proof, and the week the internet refereed it liverespectable fringe
NP-complete problems and physical realityScott Aaronson tours every physical trick proposed to beat NP: soap bubbles, time machines, relativistic computersrespectable fringe
HypercomputationToby Ord's survey of machines that would compute the uncomputable, if the universe allowed themfringe
03Navier–StokesWater that thinks3
Tao's fluid computerTerence Tao proves finite-time blowup for an averaged version and sketches a water-powered Turing machine as the road to the real thingrespectable fringe
The Otelbaev claim2014: a Kazakh mathematician announces a proof in Russian; the world tries to translate it; a gap is foundfringe
Too many answersBuckmaster and Vicol show the weak form of the equations has non-unique solutions: the water can do more than one thingrespectable fringe
04Yang–Mills and the mass gapGlue and ghosts3
Dynin's discrete spectrumAlexander Dynin's claim that the Yang–Mills energy spectrum is infinite and discrete, posted and repostedfringe
Frasca's mapping theoremAn independent physicist's long campaign to get the mass gap from a classical mapping to scalar fieldsfringe
Glueballs on the latticeThe computer evidence that the gap exists, which is not a proof and never will bethe real thing
05Hodge ConjecturePhysicists in the garden2
Mirror symmetry crashes the partyCandelas, de la Ossa, Green and Parkes (1991) used string theory to count curves on a quintic, beating the mathematicians at their own gamerespectable fringe
Where the conjecture already failsHodge’s original version is false over the integers (Atiyah and Hirzebruch, 1961); Totaro’s paper is where that failure is studied, and the fringe asks what else is quietly falsethe real thing
06Birch and Swinnerton-DyerCurves with secrets3
The elliptic-curve backdoorDual_EC_DRBG: the random-number generator built on an elliptic curve with a suspected NSA trapdoor, documented by the cryptographers who found itthe real thing
Congruent numbersWhich whole numbers are the area of a rational right triangle: a thousand-year-old puzzle that BSD would settle, in Keith Conrad’s notesrespectable fringe
Zhang's 70 millionThe paper by an unknown lecturer that cracked bounded gaps and started a race to 246the real thing
Polymath8The open, online, anyone-can-join project that drove Zhang’s gap from 70 million down, written up under a collective namerespectable fringe
10Collatz ConjectureThe problem that eats mathematicians4
Lagarias's annotated bibliographyJeffrey Lagarias catalogs every serious and semi-serious paper on 3n+1, which is the closest thing to a map of this rabbit holethe real thing
Bařina's Collatz verificationDavid Bařina's GPU project at Brno: every starting number below 2^71 verified to reach 1 (January 15, 2025), source code public, still runningfringe
Tao: almost all orbitsAlmost every starting number almost reaches 1, which is as close as the best living mathematician gotthe real thing
Lagarias’s 2021 overviewThe state of the problem by the man who has read everything on it, including why Conway’s generalizations are undecidablerespectable fringe
11Dark MatterOr no dark matter at all3
MONDMilgrom's 1983 modified gravity, which fits galaxy rotation curves with no dark matter and still will not dierespectable fringe
Superfluid dark matterBerezhiani and Khoury: dark matter is a superfluid, and MOND is what it looks like from insiderespectable fringe
The Electric UniverseThe fringe of the fringe: gravity is overrated, the cosmos runs on electric currents. No peer review, large followingfringe of the fringe
12Dark EnergyOr no expansion at all3
Tired light, 1929Fritz Zwicky’s original proposal that redshift is light losing energy, not space expanding; still the fringe’s favoritefringe
The Big RipCaldwell, Kamionkowski and Weinberg: if dark energy grows, it tears galaxies, then atoms, apart at a calculable daterespectable fringe
The Big Bang never happenedEric Lerner's 2022 essay that went viral after JWST's first images, and the replies it drewfringe of the fringe
13Matter vs AntimatterMirrors and time3
Positrons going backwards in timeFeynman (1949): an antiparticle is a particle moving backward in time. Mainstream mathematics, fringe-sounding sentencethe real thing
Does antimatter fall up?For decades the fringe hoped so; in 2023 ALPHA-g dropped antihydrogen and it fell downthe real thing
The anti-universeBoyle, Finn and Turok: before the Big Bang sits a mirror universe made of antimatter running backwardrespectable fringe
Starship is the key; these are the locks. Each one is an engineering or science problem nobody has solved at the scale a Mars city needs. The sources are NASA’s own pages and SpaceX’s.
0people have stood on Mars
13doors between Starship and a city
2years between Mars launch windows
Read the thirteen as one trip. Doors 1 to 3 are leaving Earth with full tanks: refilling in orbit, flying the Ship again, keeping the propellant cold for six months in sunlight. Door 4 is landing a hundred tonnes with no rehearsal. Doors 5 to 7 are making the ride home and the power to make it: propellant proved only in grams, ice mapped but never mined, megawatts through dust storms and nights. Doors 8 to 11 keep people alive: radiation with no limits set yet, three-eighths gravity with zero data points, a closed loop with no resupply, and dust in every seal. Door 12 has never been done by anyone: launching from Mars. Door 13 is the one that makes the rest possible at twenty light-minutes: running itself. A fourteenth door is law. Every card links to NASA’s own pages or SpaceX’s.
What is ahead in this section
The thirteen tilesEach with its status line (demonstration pending, never done, designs only) and its source.
A Mars ship leaves Earth orbit full, so Starship must be refilled by other Starships, tank to tank, in zero g. NASA gave SpaceX $53.2 million in 2020 to demonstrate moving 10 metric tons of liquid oxygen between tanks on one vehicle (NASA). Nobody has done it at this scale.
Demonstration pending
02The whole plan rests on it
land → refuel → fly again
Flying the Ship againReusability
Falcon 9 boosters land and fly again. No orbital upper stage has ever come back through re-entry and flown again without months of work. Starship’s ship must land and go again with its heat shield intact, over and over (SpaceX).
Rapid reuse: not yet shown
03Six months in sunlight
LOX −183°C · CH₄ −162°C
Keeping propellant coldCryogenics
Liquid oxygen and methane boil away in sunlight. A Mars ship holds them for six months or more on the way out, and the return load must wait on Mars. Cryogenic fluid management had its own line of NASA awards in 2020 for exactly this reason (NASA).
Zero boil-off: not flown at scale
04No rehearsal possible
a very thin atmosphere
Landing 100 tonnes on MarsEntry, descent, landing
Mars has a very thin atmosphere (NASA): too thin for parachutes to land a hundred tonnes, thick enough to burn it. Starship must fly in on its engines, supersonic, into dust it cannot see through. Nothing near that mass has ever been landed anywhere but Earth.
First try is the real thing
05Proved in grams
CO₂ + 4H₂ → CH₄ + 2H₂O
Making the ride home on MarsPropellant production
The return propellant is made on Mars: methane and oxygen from Martian carbon dioxide and water ice. NASA’s MOXIE on Perseverance proved the chemistry, making 122 grams of oxygen in all, 12 grams an hour at best (NASA). A Starship needs hundreds of tonnes.
Grams done; tonnes to go
06Mapped, not mined
H₂O, under the dust
Finding and mining the iceResources
Phoenix verified water ice just under the surface of the Martian arctic in 2008 (NASA). A propellant plant needs ice at a warmer, lower-latitude landing site, and machines that dig it out and melt it with nobody there to fix them.
No digging machine has flown
07Megawatts needed
40 kW, under 6 t
Power through dust storms and nightsEnergy
Solar panels dim under Mars dust and die in a global storm. NASA’s Fission Surface Power project asked industry for a reactor under six metric tons that makes 40 kilowatts (NASA). A propellant plant needs megawatts. No reactor has ever run on another world’s surface.
Reactor: designs only
08No limits exist yet
GCR + SPE
Radiation, there and backCrew health
Outside Earth’s magnetic field a crew takes galactic cosmic rays and solar storms the whole way. NASA says there is insufficient knowledge of the health effects to set crew exposure limits and design requirements for long-duration missions (NASA).
Shielding a ship: unsolved
09Zero data points
g = 0.38
Living in three-eighths gravityCrew health
In weightlessness, bones lose 1 to 1.5 percent of their mineral density every month (NASA). No one has ever lived in Mars gravity, so nobody knows whether it is enough to stop that, or what a child would grow up like there.
Nobody has tried
10Years, no resupply
98% of the water, back
A closed loop that never breaksLife support
The space station now recovers about 98 percent of its water (NASA), with Earth a few hours away for spare parts. A Mars crew needs air, water and food loops that run for years with no resupply at all.
Food loop: unsolved
11In every seal
ClO₄⁻
Toxic dustSurface
Phoenix found perchlorate in the Martian soil (NASA). The dust is fine, electrostatic and toxic; it grinds seals, coats panels and would follow every suit back inside the habitat.
No habitat has met it
12Never done
first liftoff from Mars
Launching from MarsAscent
Every Mars mission has been one way. NASA’s Mars Sample Return was the plan to fly the first rocket off Mars, and it has stalled (NASA). Starship must do it with people aboard, fueled from the plant in door 05.
Zero launches from Mars, ever
13Minutes each way
2 weeks of silence
Running itselfAutonomy
Radio takes minutes each way, so no one on Earth can fly a landing or handle an emergency. Every two years the Sun sits between the planets and NASA stops sending commands for two weeks (NASA). A city has to run itself.
Autonomy is the only answer
The fourteenth door is legal. The Outer Space Treaty, Article IX, requires exploration conducted so as to avoid harmful contamination of celestial bodies (UNOOSA), and COSPAR’s planetary protection panel writes the rules missions follow (COSPAR). A crewed ship cannot be sterilized to today’s standard. That rule has to change before the first boot touches Mars.
Opened one of these doors? Submit it here.A link to the published work is required
This is the portal for solved doors: any of the thirteen SpaceX doors, the fourteenth, or one of the thirteen locked doors. Your submission is saved and read by a person. Publish first; a claim without a public paper, preprint, report or repository cannot be reviewed.
The God particle · The one big question that did get answered
The particle that gives everything its weight.
The Higgs boson is proof that a locked door can open: proposed on paper in 1964, found in a machine 48 years later. Two parts: the story in plain words for everyone, then the depository, the primary record for people who work on it.
Two parts, for two readers. In plain words: six numbers that tell the whole story (1964, 4 July 2012, 125.13 GeV, spin 0, 58 percent, one in a billion) and the story itself, written for anyone. The depository: 18 links in five shelves, every one primary: the three 1964 papers that wrote the mechanism down, the two discovery papers of 4 July 2012 and the Nobel that followed, the ten-year reviews that mapped its interactions, the reference shelf working physicists keep open, and the real collision data anyone can download.
What is ahead in this section
In plain wordsThe six numbers, then why a field can give a particle its weight.
The depositoryFive shelves, 18 primary sources, from the 1964 papers to CERN Open Data.
1964the field proposed by Brout, Englert and Higgs
4 July 2012discovery announced by ATLAS and CMS at CERN
125.13 GeVits mass, ± 0.11 (Particle Data Group, 2026)
Spin 0the only known fundamental particle with none
58%of Higgs bosons decay to a pair of bottom quarks
1 in a billionLHC collisions produce one
In plain words For everyone, no physics required
What it is
The Higgs field fills the whole universe. Particles get their mass by interacting with it; the photon does not interact with it, so the photon has no mass. The Higgs boson is the ripple in that field, the thing you can actually make and measure. It appears in about one in a billion collisions at the Large Hadron Collider. (CERN: The Higgs boson)
Why it is called the God particle
Leon Lederman, who won the 1988 Nobel Prize for the muon neutrino, gave it the name in his 1993 book “The God Particle.” He had wanted to call it the goddamn particle, for how hard it was to find; his publisher would not allow it. Physicists mostly avoid the nickname. (Science News)
How it was found
Two experiments at CERN, ATLAS and CMS, each saw the same new particle in the debris of proton collisions and announced it together on 4 July 2012. By March 2013, with two and a half times more data, its spin was confirmed as zero: it was a Higgs boson. Its interaction with tau leptons was seen in 2016 and with top and bottom quarks in 2018. (CERN; APS Physics on the bottom-quark decay)
The prize
The 2013 Nobel Prize in Physics went to François Englert and Peter Higgs “for the theoretical discovery of a mechanism that contributes to our understanding of the origin of mass,” confirmed by the ATLAS and CMS discovery. Robert Brout had died in 2011. Peter Higgs died on 8 April 2024, aged 94. (Nobel Prize; University of Edinburgh)
The machine
The LHC is a 26,659 m ring about 100 m underground, its superconducting magnets cooled to 1.9 K, colder than outer space, colliding protons at 13.6 TeV. It first circulated protons on 10 September 2008. The High-Luminosity upgrade, due to run from about 2030, aims for at least 15 million Higgs bosons a year. (CERN: The LHC; CERN: HL-LHC)
What is still open about it
Is it one particle or part of a larger Higgs family? Does it interact with itself? Why is its mass so small? Does it help explain why matter outlasted antimatter, the thirteenth locked door on this page? Could dark matter show up through it? Those are the questions the next decade of collisions is built to ask. (CERN)
The depository For the experts: the primary record, nothing secondhand
The 1964 papersWhere the mechanism was written down3
String theory · The most-travelled road to door seven, quantum gravity
Everything is a vibrating string.
One idea, pushed as far as it will go: the smallest things are not points but tiny loops and strands of energy, and every particle is one note that such a string can play. Pull on the idea and gravity falls out of it uninvited, the world needs ten dimensions, and five different theories turn out to be one. Fifty-eight years after the first equation, nobody has measured a string. This room holds what the theory says, what it has delivered, what it has not, and where to learn it at no cost. The links go to the original papers.
The one picture A string, its notes, and why a note is a particle
A guitar string of fixed length can only vibrate in whole numbers of half-waves: one hump, two, three. Each pattern is a note. String theory says a fundamental string is the same, except that each of its notes shows up to us as a different particle, with a different mass and spin. Press a mode below and watch the string change its note. The drawing is a cartoon: the real notes are quantum states of a string in ten dimensions, but the logic is the same. Mode 2 is the one that matters most: in every version of the theory a closed string vibrating this way has exactly the properties of the graviton, the particle that would carry gravity. That is why the theory is on this site at all; it is the only reason anyone kept working on it after 1974 (Scherk and Schwarz, 1974; Yoneya, 1974).
One hump. The string’s lowest note. In the simplest bosonic version this note comes out with imaginary mass, the tachyon, which is one reason the real theory needs supersymmetry.
How small From you to the string, on one ruler
Each step on this ruler is ten times smaller than the last. A string is thought to sit near the Planck length, 1.616 × 10−35 metres (CODATA, NIST). A proton, the smallest thing most people have heard of, has a charge radius of 8.4 × 10−16 metres (CODATA, NIST). The gap between those two numbers, nineteen steps of ten, is the whole problem: no machine reaches anywhere near it, which is why the Large Hadron Collider was never expected to see a string directly.
Five theories that were one The 1995 map
By 1985 there were five consistent superstring theories and no way to choose. In 1994 Hull and Townsend argued the five were linked by dualities, exact dictionaries between theories that look nothing alike (arXiv hep-th/9410167). In 1995 Witten showed they were five corners of one eleven-dimensional theory he called M-theory, whose low-energy limit is eleven-dimensional supergravity: Type IIA at strong coupling grows an eleventh dimension, the E₈×E₈ heterotic string does the same on an interval, and Type IIB is its own strong-coupling dual (arXiv hep-th/9503124; see also Hořava and Witten). The same year Polchinski found that the theory contains higher-dimensional objects, D-branes, that strings can end on (arXiv hep-th/9510017). Nobody has written down what M-theory is in full; the map below is the shape of what is known.
The record Fifty-eight years, in the papers themselves
1968
The first formula
Gabriele Veneziano writes an amplitude for colliding hadrons that nobody can derive from particles. It is the first string formula, found before anyone knew it was about strings.
Ramond puts fermions on the string; Neveu and Schwarz build the other half. Supersymmetry is born inside string theory, years before anyone used it for particles.
The theory had failed as a theory of hadrons. Scherk, Schwarz and, independently, Yoneya notice that its massless spin-2 note is a graviton, and propose it as a theory of everything instead.
Gross, Harvey, Martinec and Rohm build the heterotic string; Candelas, Horowitz, Strominger and Witten curl the six extra dimensions into Calabi–Yau shapes to get something like our world.
Strominger and Vafa reproduce the Bekenstein–Hawking entropy of a black hole by counting string states. The first quantitative success of any quantum theory of gravity.
Maldacena’s AdS/CFT: a string theory with gravity in a curved space equals a quantum field theory without gravity on its boundary. On INSPIRE it is the most cited theory paper in high-energy physics, with more than 22,000 citations.
Kachru, Kallosh, Linde and Trivedi construct vacua with positive energy like ours; Susskind names the landscape and argues for anthropic reasoning. The field splits over it.
Obied, Ooguri, Spodyneiko and Vafa conjecture that stable universes with positive energy, like the one we measure, may not exist in string theory. Still contested.
What it has delivered, and what it has not Both columns are true at once
Delivered
A quantum theory in which gravity is not optional: the graviton is a vibration mode, and the theory is finite order by order where general relativity alone is not (1974).
The first microscopic count of black-hole entropy that matches Hawking’s formula (Strominger and Vafa).
Holography: gravity in a volume equals a quantum theory on its surface (Maldacena), now a working tool far outside string theory.
The unification of five theories and supergravity into one framework (Witten), and D-branes (Polchinski).
Supersymmetry itself, which began on the string in 1971 (Ramond).
Not delivered
A single experimental test. No measurement has ever required a string to explain it.
Supersymmetric particles at the LHC. The Particle Data Group’s 2025 review records the absence of any experimental evidence for supersymmetry so far (PDG). Low-energy supersymmetry is not required by string theory, but its absence removed the most hoped-for signal.
A definition of M-theory. The central node on the map above is a name for a theory nobody has written down.
The Yang–Mills mass gap, door four on this site, which the Clay Institute still lists as open (Clay); string dualities inspired approaches but not a proof.
Learn it, free From a first lecture to the full course
The state of the field, 2026Quanta Magazine’s long report, with the critics and the defenders in their own words.
Door seven on this siteQuantum gravity among the thirteen locked doors, with its facts, stories and the newest papers.
The double-slit experiment · One photon at a time · Why looking changes the answer
One photon, two slits, and it knows if you are watching.
Fire light at a wall with two narrow slits and a striped pattern appears behind it: the light went through both slits as a wave and interfered with itself. Now turn the light down until only one photon is in the apparatus at a time. The stripes still build up, dot by dot, so each single photon must somehow go through both slits. Then put a detector at the slits to see which one each photon takes. The stripes vanish. Richard Feynman wrote that this experiment “contains the only mystery” of quantum mechanics (Feynman Lectures, III-1). This bench lets you run it, then explains, as plainly as the physics allows, why observation changes the outcome and what that does and does not mean.
Run it A working bench: photons arrive one at a time
Each red dot is one photon landing on the screen at the right. The barrier in the middle has two slits. Fire a few and the dots look random. Fire a thousand and the stripes appear, with a central bright band and dark bands where no photon ever lands, even though every photon went through alone. Switch on Watch the slits and the detector tags each photon with the slit it used: the stripes disappear and two plain humps take their place. Switch it off and the stripes return. The probabilities drawn here are the real ones for two slits (a cos² interference term times the single-slit envelope), so the pattern you build is the pattern the experiment gives.
detections 0
What actually happens In six plain steps
A photon is not a tiny ball with a position. Between the source and the screen it is described by a wave of possibility that passes through both slits, because nothing stops it.
The two parts of that wave meet again at the screen. Where crest meets crest the chance of a detection is high; where crest meets trough it is zero. That is the stripe pattern.
The photon is only ever detected at one point. Which point is random, with the stripe pattern as the odds. One photon makes one dot; ten thousand make the stripes.
To see which slit a photon used, something must interact with it: a detector, an atom, a stray molecule. That interaction leaves a record of the path.
Once a record of the path exists anywhere, the two possibilities can no longer add up as one wave. They are now tied to two different states of the detector, and waves tied to different records do not interfere. The stripes go.
If the record is destroyed before anyone could read it, the stripes come back. The experiment has been done. Reality cares about whether the information exists, not about whether a person looks.
Why observation changes reality And what it does not mean
The mechanism, without mysticism
“Observation” in quantum mechanics means any physical interaction that correlates the system with something else. A detector at a slit becomes correlated with the photon: detector state A goes with slit A, state B with slit B. The photon’s two paths are now entangled with two distinguishable states of the world, and the mathematics is exact: the interference term between them vanishes in proportion to how distinguishable those states are. Partial information gives partial stripes. This is called decoherence, and it is why big things never show interference: a dust grain is bumped by air molecules and photons billions of times a second, each bump a tiny “observation” (review: Schlosshauer, 2003; Stanford Encyclopedia).
Three experiments nail this down. The quantum eraser (Kim, Yu, Kulik, Shih and Scully, 2000) tags photons with which-path information and then erases the tag; the stripes come back only for the photons whose tags were erased (arXiv). Wheeler’s delayed choice, realized with single photons in 2007, decides whether to look for a path only after the photon has already passed the slits; the result is the same as if the choice had been made first (Jacques et al.). And heat alone does it: hot C70 molecules lose their interference as they radiate photons that carry away their position, with no detector in sight (Hackermüller et al., Nature 2004).
What it does not mean
It does not mean a mind is required. A photographic plate, a molecule or a wisp of thermal radiation collapses the stripes exactly as a physicist does, and no experiment has ever found a difference between a conscious observer and an unconscious one.
It does not mean you can change the past. In the delayed-choice quantum eraser the record of all the photons, taken together, is plain noise; the stripes appear only when results are sorted afterward using the eraser’s data. In Wheeler’s delayed choice the late decision changes nothing that was already recorded. No signal travels backward and none travels faster than light.
It does not mean anything goes. The outcome of each run is unpredictable; the odds are predicted to as many decimal places as anyone has measured. Quantum mechanics is among the most precisely tested theories in science, and the 2022 Nobel Prize went to the experiments with entangled photons that established the violation of Bell inequalities, its strangest prediction (Nobel press release).
The record Two centuries of the same experiment, bigger things each time
1804
Thomas Young: light is a wave
Young’s Bakerian Lecture to the Royal Society describes interference fringes and argues light behaves as a wave, against Newton.
What it means, honestly Five readings physicists hold; none is settled
Everyone agrees on the stripes and on the mathematics. What the mathematics is about is still argued, and the arguments are serious. The entries below are from the Stanford Encyclopedia of Philosophy, written by specialists and refereed.
Copenhagen
The wave is a tool for predicting outcomes; asking which slit the photon “really” took when nothing measured it is a question without content. Bohr and Heisenberg’s view, still the working default in most labs.
The wave never collapses. Both paths happen; the detector, the room and you split into branches that no longer interact. Interference vanishes because the branches are now distinct, not because anything was destroyed.
The photon is a particle with a real position, guided by a real wave that goes through both slits. Looking disturbs the wave, and the pattern goes. Fully deterministic and fully consistent with every experiment.
Decoherence explains why the stripes disappear and why big things look classical. It does not by itself explain why one outcome happens rather than another; that remains the measurement problem.
Watch it happen, free The real footage and the best explanations, no account needed
THE ORIGINAL FOOTAGE
Hitachi: single electrons, one by onePlay
DO IT WITH SUNLIGHT
Veritasium: Young’s experiment, outdoorsPlay
THE EXPLANATION
Royal Institution: Jim Al-KhaliliPlay
THE UNIVERSITY COURSE
MIT 8.04, lecture 1: superpositionPlay
TWO MINUTES
minutephysics: wave-particle dualityPlay
THE PHYSICISTS ARGUE
Sixty Symbols: what it meansPlay
Door seven, quantum gravity, is the attempt to put this wave and Einstein’s curved spacetime into one theory; string theory is the most-travelled road to it, and the thirteen locked doors hold the rest.
Respected dissent · Cosmology
The case against the Big Bang.
The standard model (ΛCDM: a hot Big Bang, inflation, dark matter, dark energy) is the best-tested story of the universe there is, and it has real cracks. These are the strongest pieces of evidence respected scientists have put on the record against it, each with its paper. Most hit the model’s parts; a few hit the beginning itself.
5σthe Hubble tension, by the SH0ES team’s count
13published lines of evidence below
1Nobel laureate listing the anomalies in his own model
The thirteen lines of evidence sort into five attacks. The expansion rate (1): the universe measured today and the universe inferred from its infancy disagree at five sigma. The early universe (2, 3): galaxies too big too soon in Webb’s images, and a lithium shortfall of four to five sigma that has never been explained. The microwave sky and the principle beneath it (4, 5): anomalies Planck confirmed, and a quasar dipole at 4.9 sigma against the assumption that the universe looks the same in every direction. Inflation and dark energy (6 to 9): a founder of inflation who no longer believes it, the acceleration result re-examined, DESI’s hint that dark energy is changing, and a rival model that fits with no dark energy at all. Structure that should not exist (10 to 13): a universe twice as old, a 7-sigma void we may sit in, the cluster El Gordo, and the Nobel laureate’s own list of his model’s cracks. Most of these hit a part of the model; a few (2, 6, 10) hit the beginning itself. The defenders get the last word at the end.
What is ahead in this section
The thirteen tilesEach with its significance, its field, and the paper that put it on the record.
What the model gets rightThe defenders’ case: the predictions ΛCDM has passed, which is why it is still the standard.
Measured directly, the universe expands at 73.04 ± 1.04 km/s/Mpc; the Big Bang model, fed the early universe, predicts about 67. Riess and the SH0ES team call it a 5σ difference with no indication it arises from measurement errors (arXiv 2021).
The model disagrees with itself
02JWST, 2022
1011 M☉ at z ≈ 9
Galaxies too big, too soonEarly universe
Labbé and colleagues found six candidate galaxies with the mass of ten billion suns or more only 500 to 700 million years after the Big Bang, one near a hundred billion: far more mass in big galaxies than the model anticipated (arXiv 2022).
Structure the model did not expect
034 to 5 sigma
⁷Li: predicted ÷ 3
The lithium problemNucleosynthesis
Big Bang nucleosynthesis predicts the lightest elements. Deuterium and helium match; lithium-7 in old stars sits a factor of 3 to 4 below the prediction, a 4–5σ mismatch with no agreed fix (Fields, arXiv 2012).
Decades unsolved
04Planck 2018
the axis, the cold spot
Anomalies in the microwave skyCMB
Planck’s final analysis confirms the presence of several anomalies on large angular scales in the cosmic microwave background, the same ones seen in 2013 and 2015 (arXiv 2019). Flukes, or physics the model lacks: unresolved.
Confirmed, unexplained
054.9 sigma
1.36 million quasars
The quasar dipoleCosmological principle
The sky of quasars shows a dipole over twice as large as our own motion should produce, rejecting the standard explanation at 4.9σ and conflicting with the cosmological principle the whole model stands on (Secrest et al., arXiv 2020).
A foundation, in question
06From a founder
the unlikeliness problem
Inflation in troubleInflation
Paul Steinhardt helped invent inflation. With Ijjas and Loeb he argues that after Planck the best-motivated models are disfavored and the survivors make the theory less predictive (arXiv 2013; Scientific American 2017).
The inventor dissents
072016
constant expansion fits
Marginal evidence for accelerationDark energy
Re-analysing the supernova record that won the 2011 Nobel Prize, Nielsen, Guffanti and Sarkar find the data still quite consistent with a constant rate of expansion, and the evidence for acceleration only marginal (arXiv 2015).
The Nobel result, questioned
082.8 to 4.2 sigma
w₀ > −1, wa < 0
DESI: dark energy is changingDark energy
DESI’s 2025 measurement of 14 million galaxies and quasars prefers dark energy that evolves over the constant Λ, at 3.1σ with the CMB and 2.8–4.2σ with supernovae. The collaboration writes that ΛCDM is being challenged (arXiv 2025).
The Λ in ΛCDM, under strain
09ln B > 5
timescape > ΛCDM
No dark energy neededAlternative
Wiltshire’s timescape cosmology replaces dark energy with the uneven flow of time in a lumpy universe. On the Pantheon+ supernova set, Seifert and colleagues find very strong evidence for timescape over ΛCDM (arXiv 2024).
A rival that fits better
10Peer reviewed, not accepted
t₀ = 26.7 Gyr
A universe twice as oldAlternative
Gupta’s covarying-coupling-constants model fits JWST’s early galaxies and the supernova data with a universe of 26.7 billion years, nearly twice the standard age (arXiv 2023). Most cosmologists reject it; it is in the record.
Fits the data it was built for
117.09 sigma
δ = 0.46 ± 0.06
The KBC voidLocal structure
We sit in a large underdensity. Haslbauer, Banik and Kroupa show the void alone is in 6σ tension with ΛCDM, and with the Hubble tension added rules the model out at 7.09σ, pointing to a long-range change to gravity (arXiv 2020).
Published, contested
12z = 0.87
3×1015 M☉ at 2,500 km/s
El GordoGalaxy clusters
Two enormous clusters colliding at about 2,500 km/s when the universe was half its age: Asencio, Banik and Kroupa find such an object should not exist in ΛCDM at that epoch, calling it a massive blow for the model (arXiv 2020).
Too massive, too fast, too early
13Nobel 2019
Peebles, 2022
The laureate’s own listThe model’s author
Jim Peebles shared the 2019 Nobel Prize for the model itself. In 2022 he catalogued the anomalies cosmology has not explained and wrote plainly that the theory is incomplete (arXiv 2022).
The author counts the cracks
What the defenders answer. The core of the Big Bang, an expanding universe, the microwave background and the light-element abundances, survives every item on this list; the fights are over the model’s parts: dark energy, inflation, and how lumpy the nearby universe is. The alternatives are on the record too, including a universe with no inflation at all, where the Big Bang is a mirror between a universe and its anti-universe (Boyle, Finn and Turok, arXiv 2018). Every paper here is linked so you can read what it actually claims.
The papers · Every problem, kept current · Updated every hour
What is being published right now.
For each of the thirteen: the five newest preprints on arXiv, the five most-cited works of all time, and the running total on record. Pick a problem; tap Show more for the rest.
Newest preprints
Loading…
Most cited, all time
The calculator bench · Free · Plain English
The world’s best free calculators, on the go.
Ask in plain English, from a phone, at no cost. Wolfram|Alpha is the one to bookmark; the rest cover notebooks, graphing, step-by-step algebra and spaceflight.
The pick: the strongest free engine that answers plain-English math, physics and engineering questions, with units built in. Works in any phone browser.
NASA’s free tool for finding spacecraft trajectories to planets and small bodies, including Mars.
Live · @elonmusk on X
Elon on X, live.
Elon Musk’s latest posts on X, plus every official account and leading people at Tesla, SpaceX, xAI, X, Neuralink and The Boring Company. Pick one to read its live feed. Pick a company, then a person, to read their live feed. This is X’s own official embed.
Official events · Tesla, SpaceX, xAI, X, Neuralink, The Boring Company · and where Elon will be
Every official event, one place.
Launches with exact times, earnings webcasts, the annual meeting, product events, and the appearances Musk announces in his own words. Only official sources: the companies’ investor pages, the Launch Library, and Musk’s posts; nothing from rumor. Search it, filter it, and put any event on your calendar with one tap. Times show in your own time zone and in Mountain Time.
Loading every official event…
Where each company announces
The official channels, so you hear it first and from the source.
Sources: the companies’ investor relations pages, The Space Devs Launch Library 2, and Musk’s own posts and webcast remarks. Refreshes every 30 minutes.
Musk Tracker · Every outlet in one feed · Auto-updates every 6 minutes
The tracker, without the clutter.
Headlines from the sites that track Musk’s companies most closely, including Teslarati, Not a Tesla App, NASASpaceflight, Electrek, Drive Tesla and SpaceNews, plus all other coverage, merged into one feed with duplicates removed. No ads, no pop-ups, every headline links to the original story.
Loading the latest…
Tesla owners · Your car, everything about it · NHTSA data, refreshed hourly
Your Tesla, all of it.
Pick your car, or paste its VIN, and get one page for it: every recall and whether the fix came over the air, every complaint on file and what they are about, its crash-test ratings, its manual, software updates, upgrades, warranty, service and body shops near you, charging, and the owner community. The VIN is decoded by NHTSA and never stored here.
RecallsFrom NHTSA, with whether the fix came over the air0Complaints on fileWhat owners of this car report to NHTSA0Crash-test ratingsNHTSA five-star ratings by variant0
Live from NASA, JPL and 33 labs
Everything Mars.
The newest rover photos, Mars weather, Mars time, how far away it is right now, every mission and the 33 labs behind them.
Every SpaceX launch on the books, from this week’s Starlink flights to Starship and Falcon Heavy, with links to the live webcast. Times show in your own time zone and in Mountain Time. The list refreshes itself every 30 minutes.
Careers · Every open job · Updated every 5 minutes
Work for the companies.
Every open job at SpaceX, xAI and X, Neuralink and The Boring Company in one search, read from each company’s own job board. Read the full posting here without leaving, save the ones you like, share a search with one link, and apply on the company’s official page.
Sources: the public job boards of SpaceX, xAI (X’s careers site now points to xAI’s), Neuralink and The Boring Company. Tesla does not share its job list with outside sites; its full search is on tesla.com/careers. Not affiliated with any of these companies.
Open source · Every public repository · Updated every hour
All of it, open source.
Every public code repository from the official GitHub accounts of xAI, X, Tesla and Neuralink in one place: the Grok model weights, the X recommendation algorithm, Community Notes, Tesla’s vehicle and fleet tools, the Light Show kit and the original Roadster software. Search them all at once, sort by stars or by what changed most recently, and hide forks with one click. SpaceX and The Boring Company have no official public GitHub account.
On 12 June 2014 Tesla announced that it will not initiate patent lawsuits against anyone who, in good faith, wants to use its technology; the pledge still stands in Tesla’s legal pages (Tesla: Patent Pledge). SpaceX has taken the opposite road and files few patents; Musk explained why in a 2012 interview (Wired). Search every granted patent and application by company below, straight from the public records.
Everything a Tesla or SpaceX shareholder needs on one page: the next annual meeting and how to vote, every new SEC filing within the hour, insider trades, and the official investor links. Live prices and charts are in the stock watch.
Tesla 2026 Annual Meeting
Thursday, December 3, 2026
3:00 PM Central (2:00 PM Mountain). In person at Gigafactory Texas, 1 Tesla Road, Austin, Texas, and online with a live webcast.
You can vote if you owned Tesla shares at the close of business on October 13, 2026. Online and phone voting closes at 11:59 PM Eastern on December 2, 2026.
Filings come straight from SEC EDGAR and refresh every hour. Information only, not investment advice.
Insider trades · Straight from SEC filings · Updated hourly
Every trade, on the record.
Every stock transaction Elon Musk has reported to the SEC, and every trade by the top people at Musk’s public companies, Tesla and SpaceX: buys, sales, grants, option exercises and gifts. Each one links to the original filing. xAI, X, Neuralink and The Boring Company are private, so their leaders file no trade reports.
Open the recordEvery filed trade, with the live tickers
Loading every filing…
Source: SEC EDGAR Forms 3, 4 and 5, filed by Elon Musk and by Tesla and SpaceX officers and directors. Prices are as reported; a blank price means no price was reported, as with grants and gifts. Information only, not investment advice. Get alerts for new trades (RSS) or the weekly digest by email.
Five topics · Live
Five rabbit holes worth falling into.
The ideas Musk keeps coming back to, with the original sources and a live feed of news and new research for each, refreshed every few minutes.
01
The Simulation Hypothesis
Nick Bostrom, 2003
At Recode’s 2016 Code Conference, Musk said the odds we are in base reality are “one in billions” (ScienceAlert).
Bostrom’s trilemma: either humanity goes extinct before a posthuman stage, or posthumans rarely run ancestor simulations, or we are almost certainly simulated. source
The paper ran in Philosophical Quarterly, Vol. 53, No. 211, pp. 243–255 (2003). source
In 2011 Bostrom and Marcin Kulczycki published a patch fixing a gap in the original argument; a v2.0 FAQ followed in 2025. source
Musk’s remark came at the 2016 Code Conference, reported June 2, 2016. source
All seven Musk companies, public and private, then the public names that move with them: NVIDIA for AI compute, Rocket Lab and AST SpaceMobile for space, and Dogecoin. Live charts, technicals, the S&P 500 heatmap and a real-time news wire. Data by TradingView; this is information, not investment advice.
The private companies have no ticker, so their numbers come from the news: each name above is also a tab in the tracker, where every headline about it lands within six minutes, and in the 72-hour company list below.
Every company in the Musk news, last 72 hours
Companies and coins named in news coverage of Musk over the past three days, ranked by mentions and refreshed every 15 minutes. Each links to its headlines, videos, chart and official website. (Musk’s X posts themselves sit behind X’s paid API, so this tracks the coverage.)
Scanning the last 72 hours…
Full chart (switch to any symbol)
Tesla technicals
SpaceX technicals
Tesla news wire
Tesla financials
S&P 500 heatmap
CERN · Where two of the seven meet experiment
27 kilometers of open questions.
CERN, home of the Large Hadron Collider, describes its mission as working to “unravel the mysteries of the universe” (CERN).
The LHC ring is 26,659 m around, about 100 m underground. source
It uses 1,232 main dipole magnets and 9,593 superconducting magnets in all, cooled to 1.9 K (minus 271.3 C). source
Record collision energy: 13.6 TeV for protons, reached in 2022. First beams circulated Sept 10, 2008. source
ATLAS and CMS announced the Higgs boson on July 4, 2012; about one collision in a billion makes one. source
CERN has 25 member states and 11 associate members; Estonia joined in 2024 and Slovenia in 2025. source
2025 was a record year: ATLAS and CMS each recorded 125.4 inverse femtobarns. source
Run 3 ended June 27, 2026; Long Shutdown 3 runs until the High-Luminosity LHC returns in mid-2030. source
The proposed FCC-ee would be a 91 km collider costed at 15.3 billion Swiss francs, with data-taking around 2045. source
ALICE weighs about 10,000 tonnes and makes matter more than 100,000 times hotter than the Sun's core. source
The LHC has discovered 80 particles: the Higgs boson and 79 hadrons, including tetraquarks. source
How CERN touches the seven
Yang–Mills: The strong force (QCD) is a Yang–Mills theory, so every LHC measurement is real-world data on exactly what the mass-gap proof must explain. source
Yang–Mills: ALICE collides heavy ions to free quarks and gluons in quark-gluon plasma, the one state that escapes confinement. source
Yang–Mills: Exotic hadrons such as tetraquarks show how gluons bind quarks; each is a data point any strong-force solution must reproduce. source
Quantum gravity: The LHC hunts extra dimensions, escaping gravitons and microscopic black holes; CMS has now ruled out such black holes below 8.4 to 11.4 TeV. source
CERN news
Jul 24, 2026: All four big LHC experiments saw signs of quark-gluon plasma in light oxygen and neon collisions. source
Jun 27–29, 2026: Run 3 ended and the upgrade to the High-Luminosity LHC began. source
May 2026: CERN Council adopted the European Strategy update: HL-LHC first, FCC-ee next, go/no-go targeted for 2028. source
Apr 12, 2026: CMS ruled out semiclassical microscopic black holes below 8.4–11.4 TeV. source
Mar 20, 2026: ALICE found a primordial-plasma flow pattern even in small proton collisions. source
Nov 12, 2025: A record proton year: about 14 days sooner than 2024's record. source
Science · The 100 biggest gains · No politics
The Hundred.
The hundred advances that made science, Euclid to gene editing, in order. Ten at a time; filter or search.
c. 300 BC
Euclid’s ElementsMath
Axioms and proofs: the template for rigorous reasoning for 2,000 years.
NASA · 72 free resources · every link checked live
Every NASA resource, one door.
The data, archives, images, tools, jobs, grants and press channels NASA opens to anyone, at no cost, sorted the way the people inside NASA sort them. Open a group.
Data and archivesThe raw science, free to anyone16
Nearly all NASA images, video and data are in the public domain; the NASA logo and astronaut likenesses are not. The Media usage guidelines above say exactly what you may do.
CERN · 56 free resources · every link checked live
What the people at CERN actually use.
The data, software, lectures, media and tools that CERN staff, contractors, students, teachers and journalists use every day, and that anyone can use at no cost. Open a group.
Open data and resultsWhat the physicists actually work from, free to anyone13
CERN Open Data PortalPetabytes of real collision data and the software to analyze it
Needs a CERN accountListed so you know what is inside the badge line3
CERNBoxCloud storage for CERN users (needs a CERN account)
SWANNotebooks on CERN's cloud (needs a CERN account)
CERN TWikiCollaboration wikis (most pages need a CERN login)
Carnegie · Pittsburgh and beyond
Study the seven for free.
Sign up once and get lectures, databases and courses from Carnegie Mellon, its libraries, the Carnegie Library of Pittsburgh and Carnegie Science. Every link checked Oct 24, 2026.
The physics and the math · Six universities · 28 pinpoint links, every one checked live
Cornell, Denver, Colorado, Pitt, Duquesne and Chicago.
Only the departments, labs and archives that bear on the thirteen doors: physics, mathematics, astronomy, the accelerators and the observatories. Nothing else.
Every book, film, show, song and game Elon has named.
Compiled from Musk’s own posts, interviews and podcasts as collected by public trackers. Each title links to an Amazon search for it.
Books — Science Fiction and Fantasy12
The Hitchhiker's Guide to the Galaxy — Douglas Adams. Credited with getting Musk through a teenage existential crisis; Don't Panic was on the Falcon Heavy Roadster's screen.Find on Amazon
Foundation series — Isaac Asimov. Worth reading Asimov's Foundation.Find on Amazon
The Moon Is a Harsh Mistress — Robert Heinlein.Find on Amazon
Stranger in a Strange Land — Robert Heinlein.Find on Amazon
Dune and the full Dune series (Dune Messiah, Children of Dune, God Emperor of Dune, Heretics of Dune, Chapterhouse: Dune) — Frank Herbert. Dune Series by Herbert [is] brilliant.Find on Amazon
The Culture series — Iain M. Banks, specifically: Consider Phlebas, The Player of Games, Use of Weapons, The State of the Art, Excession, Inversions, Look to Windward (Pantheon level), Matter, Surface Detail (especially Surface Detail), The Hydrogen Sonata.Find on Amazon
A Song of Ice and Fire — George R. R. Martin: A Game of Thrones, A Clash of Kings, A Storm of Swords, A Feast for Crows, A Dance with Dragons. Best books in recent years imo are Iain Banks & George Martin.Find on Amazon
The Lord of the Rings — J. R. R. Tolkien (childhood favorite, per The New Yorker 2009).Find on Amazon
American Caesar: Douglas MacArthur — William Manchester.Find on Amazon
My Inventions — Nikola Tesla (autobiography).Find on Amazon
A Woman Makes a Plan — Maye Musk (Elon’s mother). My Mom wrote a book.Find on Amazon
Sources compiled by:
Books — History and War9
The Story of Civilization — Will and Ariel Durant, especially Vol. 1 Our Oriental Heritage, Vol. 2 The Life of Greece (incredible) and Vol. 11 The Age of Napoleon (an amazing book).Find on Amazon
Blastar — the game Musk wrote and sold at 12.Find on Amazon
Sources compiled by: Fortune, Business Insider India, SlashGear, Inverse, Sifted
Musk has posted many thousands of times on X since 2010; no public record captures every passing mention. This list covers every title collected by the major trackers above plus Musk’s named interview picks. Items marked as weaker sourcing appear on tracker lists without a located original post.
The Pursuit · Open to everyone
Every breakthrough. Every wall.
Share a win or a dead end on any of the seven. Nobody solves these alone, and nobody should hit the wall alone either.
Loading the wall…
Arcade · Earth to Mars
Don’t Panic.
Fly the space roadster from Earth orbit to the Mars landing pads. Dodge the asteroids and satellite trains, then hop across the drifting booster stages. On a computer, use the arrow keys. On a phone, swipe on the game or use the arrow pad under it.
Short film · The ride (3 min 22 s)
Starman picks up an alien passenger in SpaceX’s real Roadster photo. They discuss battery life entirely in Klingon, with subtitles in English or any of 12 other languages, then the radio comes on and they ride out the whole song. Original music.
“Rapidly heading towards digital superintelligence that far exceeds any human.”
Musk also said it would happen within Musk’s lifetime, 100%, and warned that a small group with godlike digital superintelligence could take over the world.
Independent archive. This site is not affiliated with, endorsed by, or operated by Elon Musk or any Musk company. Statements are summarized from the linked reporting; follow each source for full context.
SpaceX · Public Domain Archive
Every frame, free forever.
Every photo and video SpaceX released into the public domain, pulled live from the Wikimedia Commons record. Tap any frame for full resolution.
0Frames loaded
0Photos
0Videos
—Span
StarmanTesla Roadster · Falcon Heavy · Feb 6, 2018Falcon 1First private liquid-fuel rocket to orbit · Sept 28, 2008
