Urgent.News

What's breaking now, across hundreds of outlets.

AI

"It can take years": Why the AI Riemann proof is far from settled

AI claimed another major scalp in maths on Friday, as Anthropic mathematician Levent Alpöge published what he says is a proof of the Riemann hypothesis, a Millennium Prize Problem, in 17 million lines of code that no human has read in full. Physicist Tony Padilla tells Tech 24 that verification could take years, and that Alpöge should share how he got there.

"It can take years": Why the AI Riemann proof is far from settled

A solution to the Riemann hypothesis, achieved by a human-AI team, may seem like a monumental breakthrough affecting our lives. However, the implications are murkier than they may initially appear. The proof pertains to the generalized Riemann hypothesis, a broader application of Riemann's insights to a larger class of mathematical functions related to prime numbers.

This advancement would not only have ripple effects throughout mathematics but also reveal that many interlocking assumptions built on its presumed veracity are indeed true.

Although cryptography relies heavily on prime numbers, a direct impact on security cannot be inferred from the proof. While it would enhance mathematicians' understanding of prime distribution, it wouldn't provide a shortcut for factoring large numbers essential to encryption algorithms like RSA. Therefore, your online transactions and digital secrets remain secure under current encryption techniques.

The most consequential impacts of this proof wouldn't center around prime numbers or pure mathematics but would signal AI's growing potential for groundbreaking scientific discoveries. The Riemann hypothesis, notorious for resisting solutions despite 150 years of relentless effort from top mathematicians, was solved using Claude, an AI developed by Anthropic researchers. This achievement demonstrates AI's ability to transcend mere theorem explanation and create genuinely new knowledge.

This breakthrough raises intriguing questions about AI's capacity to revolutionize various scientific disciplines, including theoretical physics, engineering, and computer science. If AI can prove such a complex theorem, it implies that machines could autonomously execute entire lines of theoretical investigation. The implications are profound, hinting at a future where AI might outstrip human capabilities for fundamental intellectual discoveries.

However, the extent to which humans could comprehend or even utilize these AI-generated advancements remains uncertain.

Rewritten by urgent.news; may contain errors — check the original.

Sources: nature.com · alpo.ge · scientificamerican.com · techcrunch.com · businessinsider.com · khaosodenglish.com · thehindu.com · france24.com

This story

This is one outlet's version. Read the fullest account.

Every outlet that reported this, and when →

Read the original at france24.com →

Report a problem with this story

More in AI

Design Notes for the Claude Agent SDK

Design Notes for the Claude Agent SDK I'm building an agent with the Claude Agent SDK for the first time, so I'm collecting what I learn and the design decisions I need to think through.

  • Verify success independently of SDK result; examine deliverables and acceptance criteria
  • Set execution limits (turns, budget, time, retries); enforce permissions at runtime

COP31 Concludes Without Consensus; Peng Chi-ming: Taiwan Will Accelerate Promotion of International Carbon Markets and AI Carbon Reduction Resilience Cooperation

COP31 concluded in Turkey today (22nd). Amid divergent positions among various countries regarding the issue of fossil fuel transition, the "Antalya Agreement" failed to reach a consensus and was not passed; however, the trend of global climate governance moving toward concrete implementation remains unchanged.

More from Sunday 22 November →