It wasn’t just a math problem.
It was a ghost story.
In the 1930s, a German mathematician named Werner Burau turned geometry into algebra. He took braids—those tangled vertical strands we associate with hair or friendship bracelets—and flattened them into grids of numbers.
Matrices.
Spreadsheets of topology.
The catch?
Nobody knew if the translation worked perfectly.
Did the matrix always match the braid?
Or did the math lie?
If the matrix lost information, it was “unfaithful.” If it kept the truth intact, it was “faithful.”
For ninety years, nobody could pin it down.
Then, in 2026, Joan Birman solved it.
She was ninety-nine years old.
How the 4-Strand Braid Mystery Was Finally Solved
To understand why this took a century, you have to look at the strands.
One strand?
Trivial. It goes straight down. Faithful.
Two strands?
They cross over each other in a repeating pattern. Easy to track. Faithful.
Three strands?
For decades, mathematicians proved this was also faithful. The matrix mirrored the knot exactly.
Five or more strands?
Here, the math broke down.
Geometric proofs showed the matrices were completely unfaithful.
Information evaporated during the translation.
But what about four?
The number four sat in the middle. The outlier.
For years, the consensus shifted.
Many researchers believed four strands would join the five-plus club. They thought it was unfaithful. They spent decades hunting for a counter-example, trying to prove the matrix lied.
They were looking for a lie that didn’t exist.
Birman, along with Tara Brendle (University of Glasgow) and Vasudha Bharathram (Princeton University), flipped the script.
Instead of proving the matrix was broken, they proved it held true.
Why Mathematicians Struggled With Burau Representations For Decades
Birman didn’t start with matrices.
She started with knitting.
“It wasn’t until I arrived at graduate school that I discovered the beautiful mathematical side to my pastime.” — Joan Birman
In the 1970s, she wrote Braids, Links, and Mapping Class Groups.
The book revived Burau’s cold case.
Birman showed that for four strands, the problem reduced to searching for relationships in specific three-by-three matrices.
It sounded promising.
It turned out to be a red herring.
Other mathematicians, like Emmanuel Breuillard and Oleksandr Kosyak, tried for over twenty years.
Even artificial intelligence failed them.
They spent six months feeding prompts into AI chatbots using Birman’s matrix approach.
The bots couldn’t crack it.
On a Wednesday morning in 2026, the human team announced the solution.
Which Mathematical Approach Proves Braid Faithfulness?
The team stopped looking for flaws in the matrix.
Bharathram had a hunch.
Maybe the assumption of unfaithfulness was the error.
Maybe the four-strand braid was faithful.
They adapted a geometric method pioneered by John Moody.
Moody’s original approach used this geometry to prove five-plus strands were unfaithful.
Birman’s team inverted the logic.
Imagine points on a piece of paper.
These points represent strands.
Draw loops around them.
The loops capture how the strands interact.
The inside of the loop?
That’s a disk.
Brendle explained the mechanics simply:
The disks calculate the Burau matrices.
With three strands, the loops are limited.
The matrix has no choice but to be faithful.
With five or more strands, the number of possible disks explodes.
Different loops create similar matrix effects.
Information is lost.
The matrix becomes unfaithful.
But four?
Four is the phase transition.
Like water freezing into ice.
The possibilities grow, but not enough to break the link.
The team found “nasty scenarios” within the four-strand geometry.
They handled them.
They proved the matrix stays faithful.
Did AI Fail to Solve The Burau Conjecture?
Yes.
And no.
The failure wasn’t the tool’s fault.
It was the approach.
The researchers were asking the AI to solve a problem using the wrong framework.
They were digging for unfaithfulness.
They were looking for cracks in a wall that stood solid.
Yang-Hui He from the London Institute for Mathematical Sciences called the result “crazy.”
Not because the math is complex.
But because it took seventy years of effort from so many brilliant minds.
“It’s one of the most interesting stories in recent—not just that there’s this proof,” He says, “but that so many people have worked on it.”
Why This Proof Matters For Group And Knot Theory
Braids aren’t just hairdos.
They appear in protein folding.
In string theory.
In quantum computing.
If the map of the braid to the matrix is flawed, the physics might be too.
By confirming the four-strand Burau representation is faithful, the team closed a gap in the foundational map of topology.
Birman was surprised by the attention.
“I’m very surprised that it’s gotten this much attention,” she says.
“People were looking for a counter-example.”
The trio didn’t throw a party.
Birman is ninety-nine.
Brendle and Bharathram are scattered across continents.
The work is done.
“It’s intrinsically interesting,” Bharathram says.
The braid is tied.
The matrix matches.
But the next tangle?
Someone else has to figure that out.
