Mathverse

BSc maths student. I post the things your professor skips daily.

[Mathverse @Zero__Maths]

Good summarising isn't about making things shorter. It's about knowing what matters and what doesn't.
The best mathematicians write short proofs — not because they skip steps, but because they know which steps really count.
#MathTwitter

[Mathverse @Zero__Maths]

#MathTwitter #MathChat

[Mathverse @Zero__Maths]

You know 1+2+3+...+n = n(n+1)/2.
Now generalize:
What's 1+3+5+...+(2n-1)? (sum of first n odd numbers)

Drop your formula below
#MathTwitter #Mathematics #Mathchat

[Mathverse @Zero__Maths]

@codek_tv Remarkable 🤩 observation

[Mathverse @Zero__Maths]

Generalization is a skill that student should develope but rarely do.
#MathTwitter

[Mathverse @Zero__Maths]

3/8
The theorem's answer: if you can't already win almost always, no cleverer strategy exists.
Not "we haven't found it yet." PROVABLY doesn't exist.
If the game is hard, it's hard. Permanently.
#MathTwitter #Computer #GameTheory

[Mathverse @Zero__Maths]

1/8
There's a game where every answer has exactly 2 correct responses.
If you can't win it almost every time, math proves NO strategy ever will.
This is the 2-to-2 Games Theorem — and it just won the 2026 Held Prize.
Let me break it down.

[Mathverse @Zero__Maths]

2/8
Real-life version: 100 locks.
Each of your keys opens exactly 2 locks. Each of your friend's keys does too.
No communication. You both try to open the same lock.
Can you find a strategy that works almost every time?

[Mathverse @Zero__Maths]

Quick one for you:
What's the last digit of 7^2026?
No calculator. Just pattern recognition.
(Hint: powers of 7 cycle in a pattern — find it, and this takes 10 seconds instead of 10 minutes)
#Maths #MathTwitter

[Mathverse @Zero__Maths]

The Unique Games Conjecture has puzzled computer science for 20+ years.
2026 Held Prize just went to 5 mathematicians for proving the 2-to-2 Games Theorem — the strongest evidence yet that UGC is true.
Massive win for approximation algorithms.
#MathTwitter #Computer

[Mathverse @Zero__Maths]

Graph theory's 50-year-old Cycle Double Cover Conjecture just got proved.
By OpenAI's GPT-5.6 Sol Ultra.
Noga Alon: "surprisingly, the proof is short."
Still under peer review. But that's 2 major open conjectures AI has cracked in months.
Wild times to be doing math.
#MathTwitter

[Mathverse @Zero__Maths]

no matter how slowly a series converges, you can always build another series that converges even slower.
#MathTwitter #Matheatics

[Mathverse @Zero__Maths]

@Math_files If you add a unit in zero.
It's happened.

[Mathverse @Zero__Maths]

When a mathematician stares at a blank page, their motivation isn't to write something brilliant immediately; their motivation is to translate the abstract problem into concrete, workable pieces. The first step is usually an act of setting up the board before playing the game.

[Mathverse @Zero__Maths]

Euler's constant isn't e.
It's this: take the harmonic series, subtract the natural log that's chasing it forever, and watch what's left.
1 + 1/2 + 1/3 + ... + 1/n - ln(n)
As n → ∞, this doesn't blow up. It doesn't hit 0.
It settles at γ ≈ 0.5772...
#Mathtwitter

[Mathverse @Zero__Maths]

1/6
Sum of 1/n² has a closed form. So does 1/n⁴. So does every even power.
But sum of 1/n³?
Nobody has ever found a closed form. Not Euler. Not anyone since.
Welcome to one of math's weirdest asymmetries.

[Mathverse @Zero__Maths]

#MathTwitter #Maths

[Mathverse @Zero__Maths]

OpenAI's chief research officer just said it out loud:
AI could make a Nobel-worthy discovery within 2 years.
Not "assist with." Make.
We already have AI that cracked a 1946 Erdős problem so cleanly that a Fields Medalist said no AI proof has come close before.

[Mathverse @Zero__Maths]

Reports say OpenAI's new model just solved a math problem that stumped humans for 50 years.
In under an hour.
No peer review yet. No public proof. Just a claim.
But if true? We're watching the line between "AI helps mathematicians" and "AI does math" disappear in real time.

[Mathverse @Zero__Maths]

Everyone knows the Fibonacci sequence.
Almost nobody knows what happens when you sum the reciprocals.
1/1 + 1/1 + 1/2 + 1/3 + 1/5 + 1/8 + 1/13 + ...
You'd think it explodes. It doesn't.
It converges to a number even weirder than π: the Reciprocal Fibonacci Constant ≈ 3.359885...