Show HN: val – An arbitrary precision calculator language
(github.com/terror)
val (eval) is a simple arbitrary precision calculator language built on top of chumsky and ariadne.
val (eval) is a simple arbitrary precision calculator language built on top of chumsky and ariadne.
Attention Spans for Math and Stories (2019)
(jeremykun.com)
There was a MathOverflow thread about mathematically interesting games for 5–6 year olds. A lot of the discussion revolved around how young age 5 really is, and how we should temper expectations because we don’t really remember what it’s like to be 5.
There was a MathOverflow thread about mathematically interesting games for 5–6 year olds. A lot of the discussion revolved around how young age 5 really is, and how we should temper expectations because we don’t really remember what it’s like to be 5.
Cross-Entropy and KL Divergence
(thegreenplace.net)
Cross-entropy is widely used in modern ML to compute the loss for classification tasks. This post is a brief overview of the math behind it and a related concept called Kullback-Leibler (KL) divergence.
Cross-entropy is widely used in modern ML to compute the loss for classification tasks. This post is a brief overview of the math behind it and a related concept called Kullback-Leibler (KL) divergence.
Color Is a Mathematical Nightmare
(theverge.com)
Understand how to paint by number without your brain exploding.
Understand how to paint by number without your brain exploding.
Gemini 2.5 gets 24.4% on MathArena USAMO beating previous top score of 4.7%
(matharena.ai)
MathArena is a platform for evaluation of LLMs on the latest math competitions and olympiads.
MathArena is a platform for evaluation of LLMs on the latest math competitions and olympiads.
Proof or Bluff? Evaluating LLMs on 2025 USA Math Olympiad
(arxiv.org)
Recent math benchmarks for large language models (LLMs) such as MathArena indicate that state-of-the-art reasoning models achieve impressive performance on mathematical competitions like AIME, with the leading model, o3-mini, achieving scores comparable to top human competitors.
Recent math benchmarks for large language models (LLMs) such as MathArena indicate that state-of-the-art reasoning models achieve impressive performance on mathematical competitions like AIME, with the leading model, o3-mini, achieving scores comparable to top human competitors.
Limits of Smart: Molecules and Chaos
(dynomight.substack.com)
Take me. Now take someone with the combined talents of Von Neumann, Archimedes, Ramanujan, and Mozart. Now take someone smarter again by the same margin and repeat that a few times.
Take me. Now take someone with the combined talents of Von Neumann, Archimedes, Ramanujan, and Mozart. Now take someone smarter again by the same margin and repeat that a few times.
Mathup: Easy MathML authoring tool with a quick to write syntax
(mathup.xyz)
Easy MathML authoring tool with a quick to write syntax.
Easy MathML authoring tool with a quick to write syntax.
Statistical Formulas for Programmers (2013)
(evanmiller.org)
Being able to apply statistics is like having a secret superpower.
Being able to apply statistics is like having a secret superpower.
A new Sudoku layout with 81 uniquely shaped cells
(danielchasehooper.com)
Something productive finally came from my daily Sudoku habit: I invented a new type of puzzle that I call “Cracked Sudoku”. It’s named after cracked dirt:
Something productive finally came from my daily Sudoku habit: I invented a new type of puzzle that I call “Cracked Sudoku”. It’s named after cracked dirt:
Math Academy pulled me out of the Valley of Despair
(bearblog.dev)
When it comes to learning a new skill such as how to drive a car, playing a sport, or an academic discipline, there is a unique relationship between a person’s confidence and their level of competence at different points of the journey.
When it comes to learning a new skill such as how to drive a car, playing a sport, or an academic discipline, there is a unique relationship between a person’s confidence and their level of competence at different points of the journey.
MathB.in Is Shutting Down
(susam.net)
Thirteen years ago, on a quiet Saturday night, I sat down and began developing MathB.in.
Thirteen years ago, on a quiet Saturday night, I sat down and began developing MathB.in.
Making any integer with four 2s
(thegreenplace.net)
There's a cute math puzzle that can be interesting to folks on very different levels:
There's a cute math puzzle that can be interesting to folks on very different levels:
A simple geometry question that fools almost everyone
(theguardian.com)
A triangle and a rectangle walked into a pub
A triangle and a rectangle walked into a pub
Ask HN: Books or games to teach kids math
(ycombinator.com)
Anything that can teach a 3 years old kid math, assuming he knows how to count to 10. But also interested in resources that that would take him beyond that and get him to fall in love with math as he grows up.
Anything that can teach a 3 years old kid math, assuming he knows how to count to 10. But also interested in resources that that would take him beyond that and get him to fall in love with math as he grows up.
Just give the man the fish (i.e. just answer the question)
(plover.com)
Last week I complained about a Math SE pathology in which OP asks a simple question, and instead of an answer gets an attempt at a socratic dialog. I ended by saying:
Last week I complained about a Math SE pathology in which OP asks a simple question, and instead of an answer gets an attempt at a socratic dialog. I ended by saying:
Explorable Flexagons: Learn to create and flex flexagons (2020)
(loki3.com)
Learn to create and flex flexagons
Learn to create and flex flexagons
I compared my daughter against SOTA models on math puzzles
(michalprzadka.com)
I created an AI math reasoning benchmark using puzzles from this year’s GMIL competition — a long-running international mathematical challenge that I participated in myself back in 1998. The results are quite interesting: some of the most advanced AI models performed comparably to my 11-year-old daughter, while others struggled significantly. This experiment gives some amusing insights into current AI capabilities in mathematical reasoning, especially when compared to human performance at the middle school level.
I created an AI math reasoning benchmark using puzzles from this year’s GMIL competition — a long-running international mathematical challenge that I participated in myself back in 1998. The results are quite interesting: some of the most advanced AI models performed comparably to my 11-year-old daughter, while others struggled significantly. This experiment gives some amusing insights into current AI capabilities in mathematical reasoning, especially when compared to human performance at the middle school level.
Orbit Spirograph (2019)
(redblobgames.com)
Inspired by John Carlos Baez’s post about the Pentagram of Venus[1]. This is the position of Venus relative to the Earth. I got distracted by fun spirograph style images.
Inspired by John Carlos Baez’s post about the Pentagram of Venus[1]. This is the position of Venus relative to the Earth. I got distracted by fun spirograph style images.
Why is zero plural? (2024)
(stackexchange.com)
For example, if we choose two 2s, zero 3s, and one 5, we get the divisor
For example, if we choose two 2s, zero 3s, and one 5, we get the divisor
rStar-Math: Small LLMs Can Master Math Reasoning with Self-Evolved Deep Thinking
(arxiv.org)
We present rStar-Math to demonstrate that small language models (SLMs) can rival or even surpass the math reasoning capability of OpenAI o1, without distillation from superior models.
We present rStar-Math to demonstrate that small language models (SLMs) can rival or even surpass the math reasoning capability of OpenAI o1, without distillation from superior models.
Fidget
(mattkeeter.com)
Fidget is a library for representing, compiling, and evaluating large-scale math expressions, i.e. hundreds or thousands of arithmetic clauses.
Fidget is a library for representing, compiling, and evaluating large-scale math expressions, i.e. hundreds or thousands of arithmetic clauses.
Ask HN: Math (Academy) Discord Group
(ycombinator.com)
I first heard about Math Academy last year here on HN. It took me some time to pull the trigger and I am now a happy subscriber looking to connect with other users - email's in my profile.
I first heard about Math Academy last year here on HN. It took me some time to pull the trigger and I am now a happy subscriber looking to connect with other users - email's in my profile.
Benchmarking RSA Key Generation
(filippo.io)
RSA key generation is both conceptually simple, and one of the worst implementation tasks of the field of cryptography engineering. Even benchmarking it is tricky, and involves some math: here’s how we generated a stable but representative “average case” instead of using the ordinary statistical approach.
RSA key generation is both conceptually simple, and one of the worst implementation tasks of the field of cryptography engineering. Even benchmarking it is tricky, and involves some math: here’s how we generated a stable but representative “average case” instead of using the ordinary statistical approach.
U.S. math scores drop on major international test
(chalkbeat.org)
U.S. fourth graders saw their math scores drop steeply between 2019 and 2023 on a key international test even as more than a dozen other countries saw their scores improve.
U.S. fourth graders saw their math scores drop steeply between 2019 and 2023 on a key international test even as more than a dozen other countries saw their scores improve.