Claude's Fable 5 just solved an 87-year-old math problem, and it matters for bitcoin

Claude Fabel 5 disproved the Jacobian conjecture over the weekend, landing days after the capabilities of China's Kimi AI became the biggest force moving bitcoin markets.

Claude's Fable 5 just solved an 87-year-old math problem, and it matters for bitcoin

Tech

Claude's Fable 5 just solved an 87-year-old math problem, and it matters for bitcoin

Claude Fabel 5 disproved the Jacobian conjecture over the weekend, landing days after the capabilities of China's Kimi AI became the biggest force moving bitcoin markets.

著者

Shaurya Malwa

|

編集者

Aoyon Ashraf

2026年7月21日 午後4:19

AIによる翻訳

2

min read

Make

preferred on

共有

この記事を共有

リンクをコピー

X icon

X (Twitter)

LinkedIn

Facebook

メール

Make

preferred on

Claude Fable 5 just solved an 87-year-old math problem, and it matters for bitcoin. (Getty Images)

Summary

Show

An Anthropic AI model, Claude Fable 5, helped disprove the 87-year-old Jacobian conjecture, a famous open problem in mathematics dating to 1939.

The model produced a concrete counterexample that mathematicians could verify by hand within a day, showing that a certain kind of mathematical “machine” can pass standard checks yet still not be reversible.

The breakthrough underscores how rapidly advancing AI capabilities are drawing speculative capital and attention away from crypto and toward computing power, chips and model builders.

An artificial intelligence model has helped disprove a mathematical conjecture that had gone unsolved for 87 years, and the answer was confirmed within a day because anyone can now check it by hand.

Levent Alpöge, a number theorist who works at Anthropic and previously held a fellowship at Harvard, posted the result on X on Sunday night, crediting the company's Claude Fable 5 model.

The problem, called the Jacobian conjecture, had been open since 1939 and sits on mathematician Stephen Smale's influential list of unsolved problems for the century.

hello there the jacobian conjecture is false thanx to my close friend akhil for asking about it and my other close friend fable for working during the world cup final

((1+xy)^3 z + y^2 (1+xy) (4+3xy), y + 3 x (1+xy)^2 z + 3 x y^2 (4+3xy), 2 x - 3 x^2 y - x^3 z): \C^3\to \C^3,…

— levent (@__alpoge__)

July 20, 2026

That an AI produced original mathematics of this difficulty is the kind of capability

leap that now moves markets

, crypto among them. Bitcoin has spent months trading on a narrative about artificial intelligence, moving with chipmakers and memory stocks rather than any catalyst of its own.

It

fell hard last Friday

when Chinese lab Moonshot AI released a model that rattled the semiconductor industry, and it recovered this week as those stocks bounced.

Part of that link is direct. Bitcoin's biggest holders, its miners, have

rebuilt themselves

into AI data-center operators, so their fortunes now rise and fall with demand for computing power.

But the larger pull is simpler: AI is where the speculative money and investors’ attention are now going. The capital that once chased crypto is now chasing compute, chips and model builders, and every leap in what these systems can do widens that appeal.

Each result like Fable’s finding of the Jacobian conjecture strengthens the case for pouring capital into AI, and poses a difficult conundrum for crypto investors: Why hold a token that trades as a sidecar to the AI cycle when someone can own the vehicle itself?

AI’s capability curve is steep, and the steeper it gets, the more of the market's risk appetite it draws away from everything else, crypto included.

What the problem actually was

Think of a machine that takes two numbers and gives back two new numbers, using only adding and multiplying. The question, first asked in 1939, was whether the machine can always be run backward: given only its answer, can the original two numbers be recovered every time?

Mathematicians had a quick way to check whether a machine looked reversible. The Jacobian conjecture said that if a machine passed that check, it should always be reversible.

Why the 87-year-old Jacobian conjecture is false, in one picture. (Shaurya Malwa/CoinDesk)

For 87 years, nobody could prove it was true, and nobody could find a machine that broke the rule.

The AI found one. It built a machine that passes the check but still can't be reversed, because three different inputs all give the exact same answer. If three inputs give one answer, it can't work backward from that answer. One counterexample is all it took, and the rule is forever broken.

Latest Crypto News

1

Galaxy sets up $5 million fund to help shield Bitcoin against quantum computing threats

59 分前

2

Russia’s parliament passes crypto market law with $3,800 annual cap for retail investors

1 時間前

3

Augustus raises $180 million to build a clearing bank for the AI and stablecoin era

2 時間前

4

MoneyGram's CEO says blockchain works best when customers don't know it's there

3 時間前

5

Jack Mallers steps down as XXI Capital CEO as Tether's plans to merge three bitcoin firms falls

4 時間前

6

Bitcoin rally has broad-based support as institutions, whales, options traders pile in

5 時間前

7

Clarity odds jump to 43% on Polymarket after unverified reports Trump agreed to ethics deal

5 時間前

8

Crypto markets rally on Clarity progress report, Asian chip-stock rebound

6 時間前

9

XRP jumps 4% as traders watch 'triangle breakout' toward $1.35

9 時間前

10

Live markets: Bitcoin rises to five-week high near $67,000 on boosted hope for Clarity Act

10 時間前

Latest Research

TRON Network - Q2 2026

TRON Network - Q2 2026

In Q2; TRON's stablecoin dominance rose to 28.7%, USDT supply on TRON hit $89B ATH, $89M in protocol fees (2nd to Hyperliquid), TRX +3%, and deepening institutional & agentic reach.

By

CoinDesk Research

4 hours ago

Commissioned by

Tron

In Q2; TRON's stablecoin dominance rose to 28.7%, USDT supply on TRON hit $89B ATH, $89M in protocol fees (2nd to Hyperliquid), TRX +3%, and deepening institutional & agentic reach.

Why it matters

:

In Q2; TRON's stablecoin dominance rose to 28.7%, USDT supply on TRON hit $89B ATH, $89M in protocol fees (2nd to Hyperliquid), TRX +3%, and deepening institutional & agentic reach.

View Full Report

More From

Tech

Galaxy sets up $5 million fund to help shield Bitcoin against quantum computing threats

MoneyGram's CEO says blockchain works best when customers don't know it's there

Live updates: Bitcoin rises to one-month high of $65,700 before pulling back with stocks

Crypto

CD20

$1,801.64

CD20 up 1.50 percent

1.50%

BTC

$66,425.35

BTC up 1.35 percent

1.35%

ETH

$1,924.79

ETH up 0.98 percent

0.98%

XRP

$1.15

XRP up 3.12 percent

3.12%

SOL

$77.93

SOL up 0.34 percent

0.34%