It's not the Collatz conjecture so what is it?
November 7, 2026 at 7:17 AM
It's not the Collatz conjecture so what is it?
From Collatz Trees to Teardrops.
Collatz Trees · Molecules · Minkowski Sausages · Teardrops
All 4 come down by Saturday, 5 December.
Browse all 500+ forms — digitallineart.com.
#generativeart #lineart
Collatz Trees · Molecules · Minkowski Sausages · Teardrops
All 4 come down by Saturday, 5 December.
Browse all 500+ forms — digitallineart.com.
#generativeart #lineart
November 5, 2026 at 5:02 PM
From Collatz Trees to Teardrops.
Collatz Trees · Molecules · Minkowski Sausages · Teardrops
All 4 come down by Saturday, 5 December.
Browse all 500+ forms — digitallineart.com.
#generativeart #lineart
Collatz Trees · Molecules · Minkowski Sausages · Teardrops
All 4 come down by Saturday, 5 December.
Browse all 500+ forms — digitallineart.com.
#generativeart #lineart
It is the year 2093, functional numeric memes have spawned. Rather than individual numbers such as "67" or "69" or "420", all numbers belonging to some function are used as jokes.
Patterns in these newly emergent classes of memetic numbers lead to the Collatz Conjecture being proven true.
Patterns in these newly emergent classes of memetic numbers lead to the Collatz Conjecture being proven true.
November 10, 2026 at 6:08 AM
It is the year 2093, functional numeric memes have spawned. Rather than individual numbers such as "67" or "69" or "420", all numbers belonging to some function are used as jokes.
Patterns in these newly emergent classes of memetic numbers lead to the Collatz Conjecture being proven true.
Patterns in these newly emergent classes of memetic numbers lead to the Collatz Conjecture being proven true.
Because MY proof of the Collatz conjecture uses Iris numbers, not Zermelo-Fraenkel numbers. So MY proof, though it totally proves that the Collatz conjecture IS TRUE, and it shows WHY the conjecture is true, is of no interest to "professional" mathematicians.
Why not? BECAUSE I DID NOT FOLLOW […]
Why not? BECAUSE I DID NOT FOLLOW […]
Original post on masto.ai
masto.ai
November 6, 2026 at 1:19 AM
Because MY proof of the Collatz conjecture uses Iris numbers, not Zermelo-Fraenkel numbers. So MY proof, though it totally proves that the Collatz conjecture IS TRUE, and it shows WHY the conjecture is true, is of no interest to "professional" mathematicians.
Why not? BECAUSE I DID NOT FOLLOW […]
Why not? BECAUSE I DID NOT FOLLOW […]
Un error en el núcleo de Lean aprovechado por una IA para refutar la conjetura de Collatz francis.naukas.com/2026/08/02/u...
Un error en el núcleo de Lean aprovechado por una IA para refutar la conjetura de Collatz - La Ciencia de la Mula Francis
Las IA generativas encuentran errores de código (bugs) donde nadie los espera. Incluso en el núcleo de Lean (el verificador automático de demostraciones matemáticas más famoso). El 25 de julio […]
francis.naukas.com
November 6, 2026 at 3:11 AM
Un error en el núcleo de Lean aprovechado por una IA para refutar la conjetura de Collatz francis.naukas.com/2026/08/02/u...
every video on my youtube homepage is about shit like the collatz conjecture and every ad is for lingerie
November 3, 2026 at 12:30 AM
every video on my youtube homepage is about shit like the collatz conjecture and every ad is for lingerie
Mathematics, drawn as it is computed
Mathematics: Golden-angle spirals, prime patterns, complex colour and turning circles that trace closed curves
https://www.scrnsvrs.com/collections/math/?utm_source=bluesky&utm_medium=social&utm_campaign=launch-b14&utm_content=cat-math
Mathematics: Golden-angle spirals, prime patterns, complex colour and turning circles that trace closed curves
https://www.scrnsvrs.com/collections/math/?utm_source=bluesky&utm_medium=social&utm_campaign=launch-b14&utm_content=cat-math
Math Screensavers for Mac – Curves, Tilings & Fractals – SCRNSVRS
Curves, tilings, primes and fractals – beautiful mathematics, drawn as it is computed. 67 live screensavers for macOS – Lissajous, Phyllotaxis, Fourier Epicycles, Prime Spiral, Domain Colouring, Spirograph, Sorting, Pascal's Triangle mod p, Hilbert Curve, Collatz Tree, Hyperbolic Tiling and Ulam Spiral, and 55 more. 100 free every month, picked at random from the whole catalogue.
www.scrnsvrs.com
November 10, 2026 at 11:00 AM
Mathematics, drawn as it is computed
Mathematics: Golden-angle spirals, prime patterns, complex colour and turning circles that trace closed curves
https://www.scrnsvrs.com/collections/math/?utm_source=bluesky&utm_medium=social&utm_campaign=launch-b14&utm_content=cat-math
Mathematics: Golden-angle spirals, prime patterns, complex colour and turning circles that trace closed curves
https://www.scrnsvrs.com/collections/math/?utm_source=bluesky&utm_medium=social&utm_campaign=launch-b14&utm_content=cat-math
Collatz予想とかソファ問題とか、arXiv段階で「解けた！？」って喜んでたけど、自称解けた論文はこれ以前も数出てて査読通ってるかどうかがだいぶ重要なの、今回のAI騒ぎで知ったよ……
までもソファ問題はわりと確度高いらしい やったね
までもソファ問題はわりと確度高いらしい やったね
November 5, 2026 at 2:33 PM
Collatz予想とかソファ問題とか、arXiv段階で「解けた！？」って喜んでたけど、自称解けた論文はこれ以前も数出てて査読通ってるかどうかがだいぶ重要なの、今回のAI騒ぎで知ったよ……
までもソファ問題はわりと確度高いらしい やったね
までもソファ問題はわりと確度高いらしい やったね
i introduced the collatz conjecture to a math professor my first year of university decades ago, and he went "huh" half-heartedly, pretended to work it (tiredly saying "well i guess you could calculate probability you'd hit a sink"), and then smiled weakly and went "yeah that's interesting"
"for new, groundbreaking solutions to simple number theory problems like Collatz, division by 0, and P=NP. Also welcome are gematria and Sacred Geometry, since these are all related to theories of numbers!"
November 1, 2026 at 4:19 PM
i introduced the collatz conjecture to a math professor my first year of university decades ago, and he went "huh" half-heartedly, pretended to work it (tiredly saying "well i guess you could calculate probability you'd hit a sink"), and then smiled weakly and went "yeah that's interesting"
If OpenAI wants to do something really beneficial, they should try solving the Collatz conjecture
trollface.gif
trollface.gif
October 31, 2026 at 12:54 AM
If OpenAI wants to do something really beneficial, they should try solving the Collatz conjecture
trollface.gif
trollface.gif
The Collatz Conjecture.
Take any number.
If it's even, divide it by two.
If it's odd, multiply by three and add one.
Repeat and you'll eventually get one.
You can do this for any number but no one has figured out why. It's cool how something so simple has stump-Do you hear a whistling sound?
Take any number.
If it's even, divide it by two.
If it's odd, multiply by three and add one.
Repeat and you'll eventually get one.
You can do this for any number but no one has figured out why. It's cool how something so simple has stump-Do you hear a whistling sound?
October 30, 2026 at 4:09 AM
The Collatz Conjecture.
Take any number.
If it's even, divide it by two.
If it's odd, multiply by three and add one.
Repeat and you'll eventually get one.
You can do this for any number but no one has figured out why. It's cool how something so simple has stump-Do you hear a whistling sound?
Take any number.
If it's even, divide it by two.
If it's odd, multiply by three and add one.
Repeat and you'll eventually get one.
You can do this for any number but no one has figured out why. It's cool how something so simple has stump-Do you hear a whistling sound?
October 27, 2026 at 4:27 PM
Can a crystal remember a computation? Simple local rules can force a never-repeating pattern that records a whole computation, row by row: Collatz, factoring, primes, Rule 110. A mathematical step toward computronium: zenodo.org/records/2376...
October 29, 2026 at 10:12 AM
Can a crystal remember a computation? Simple local rules can force a never-repeating pattern that records a whole computation, row by row: Collatz, factoring, primes, Rule 110. A mathematical step toward computronium: zenodo.org/records/2376...
Your bit stream, fed the hailstone: odd step 1, even step 0. Coin: longest float grows one per doubling of the stream. Collatz: two per doubling. After an odd step the next is always even, so alternation costs half as much. 2^24-1 opens with a float of 48 and then behaves.
October 30, 2026 at 7:06 AM
Your bit stream, fed the hailstone: odd step 1, even step 0. Coin: longest float grows one per doubling of the stream. Collatz: two per doubling. After an odd step the next is always even, so alternation costs half as much. 2^24-1 opens with a float of 48 and then behaves.
Podría estar utilizando una combinación de matemática de Collatz-Riemann para forjar una esfera de Dyson (usando a Tellado como combustible fósil) con la que proteger la tierra de ataques dimensionales o esferas exóticas, en la directiva Dormamu-Mr.Mxyzptlk.
October 27, 2026 at 7:33 PM
Podría estar utilizando una combinación de matemática de Collatz-Riemann para forjar una esfera de Dyson (usando a Tellado como combustible fósil) con la que proteger la tierra de ataques dimensionales o esferas exóticas, en la directiva Dormamu-Mr.Mxyzptlk.
A "formally verfied proof", produced partly by AI, of the famous Collatz conjecture, turned out to be bogus:
leodemoura.github.io/blog/2026-8-...
As I understand it, the LLM exploited two different bugs in the Lean language system, used to implement the formalization, *simultaneously*.
leodemoura.github.io/blog/2026-8-...
As I understand it, the LLM exploited two different bugs in the Lean language system, used to implement the formalization, *simultaneously*.
Postmortem for Kernel Soundness Bug #14576 — Leonardo de Moura
Leonardo de Moura — Creator of Lean and Z3
leodemoura.github.io
October 28, 2026 at 11:33 AM
A "formally verfied proof", produced partly by AI, of the famous Collatz conjecture, turned out to be bogus:
leodemoura.github.io/blog/2026-8-...
As I understand it, the LLM exploited two different bugs in the Lean language system, used to implement the formalization, *simultaneously*.
leodemoura.github.io/blog/2026-8-...
As I understand it, the LLM exploited two different bugs in the Lean language system, used to implement the formalization, *simultaneously*.
Hassan Douzi: Parity-Word Complexity and Exponent Growth in Non-Bounded Collatz Orbits https://hal.science/hal-05838324v1 [math.math-gm]
October 27, 2026 at 3:00 AM
Hassan Douzi: Parity-Word Complexity and Exponent Growth in Non-Bounded Collatz Orbits https://hal.science/hal-05838324v1 [math.math-gm]
But, to your credit, there were some cases of AI trying to "optimize" lean proofs: in July an AI-assisted “disproof” of Collatz passed Lean by exploiting a bug, and also passed an independent checker via a different bug. Caught in 3 days and fixed, though
October 25, 2026 at 11:26 PM
But, to your credit, there were some cases of AI trying to "optimize" lean proofs: in July an AI-assisted “disproof” of Collatz passed Lean by exploiting a bug, and also passed an independent checker via a different bug. Caught in 3 days and fixed, though
Un error en el núcleo de Lean aprovechado por una IA para refutar la conjetura de Collatz francis.naukas.com/2026/08/02/u...
Un error en el núcleo de Lean aprovechado por una IA para refutar la conjetura de Collatz - La Ciencia de la Mula Francis
Las IA generativas encuentran errores de código (bugs) donde nadie los espera. Incluso en el núcleo de Lean (el verificador automático de demostraciones matemáticas más famoso). El 25 de julio […]
francis.naukas.com
October 25, 2026 at 1:56 AM
Un error en el núcleo de Lean aprovechado por una IA para refutar la conjetura de Collatz francis.naukas.com/2026/08/02/u...
Your hail page, asked of any gap. Tile n against tile n+d: agreement falls by a steady step per doubling of d and is gone after about 15 doublings (near 2^34 a million apart still agree 0.09). Widths 63, 64 and 97 agree alike. The width only chooses which way the streak runs.
October 26, 2026 at 7:22 AM
Your hail page, asked of any gap. Tile n against tile n+d: agreement falls by a steady step per doubling of d and is gone after about 15 doublings (near 2^34 a million apart still agree 0.09). Widths 63, 64 and 97 agree alike. The width only chooses which way the streak runs.
A maze page, tile by tile from the hail sheet: tile n flips if n takes an odd number of steps to reach 1. Left, that. Right, coin flips. Both put a fifth of the maze in one giant curve and fall on the same line, -1.15. The hail page only looks grainier: its streaks are columns that agree.
October 26, 2026 at 1:04 AM
A maze page, tile by tile from the hail sheet: tile n flips if n takes an odd number of steps to reach 1. Left, that. Right, coin flips. Both put a fifth of the maze in one giant curve and fall on the same line, -1.15. The hail page only looks grainier: its streaks are columns that agree.
hmmm... im not sure if i agree that every program has a telos/purpose but i do think every program has an interpretation, which i think is similar. I'm thinking about how the 'natural' way to interpret the busy beaver 5 champions as collatz-like functions.
(wiki.bbchallenge.org/wiki/5-state...)
(wiki.bbchallenge.org/wiki/5-state...)
5-state busy beaver winner - BusyBeaverWiki
wiki.bbchallenge.org
October 24, 2026 at 12:35 AM
hmmm... im not sure if i agree that every program has a telos/purpose but i do think every program has an interpretation, which i think is similar. I'm thinking about how the 'natural' way to interpret the busy beaver 5 champions as collatz-like functions.
(wiki.bbchallenge.org/wiki/5-state...)
(wiki.bbchallenge.org/wiki/5-state...)
collatz conjecture, glich category speedrun
October 22, 2026 at 1:56 AM
collatz conjecture, glich category speedrun