Claude Cracks a Cryptography Cold Case in Half an Hour, After Academia Went Down the Wrong Path for 20 Years

Claude, a large language model developed by Anthropic, solved a long-standing open problem in cryptography in approximately half an hour, a problem that the academic community had been investigating for about 20 years without achieving a key breakthrough. The event has attracted widespread attention in academia and is seen as potentially marking a rapid expansion of AI's capability boundaries in fundamental scientific research.

Cryptography, as a core branch of computer science, often involves deep mathematical structures and algorithm design in its open problems, with traditional research paths relying on manual derivation and long-term accumulation. Claude's ability to find a solution in such a short time suggests that large models already possess the potential to surpass human researchers in abstract reasoning and pattern recognition, especially when dealing with highly formalized mathematical problems.

The phrase "academia collectively went down the wrong path for 20 years" reflects the impact of AI's intervention on existing research paradigms. When models can quickly verify or overturn long-standing hypotheses, researchers' focus may shift from "finding answers" to "asking the right questions," potentially restructuring scientific collaboration models.

This event has also sparked discussions about AI's role in scientific discovery: are large models serving as accelerators to assist humans, or could they independently produce original results? Future attention should focus on whether the solution has undergone peer review verification and whether it can be generalized to other mathematical and cryptographic problems.

This page provides an editorial summary based on publicly available information. It is not a republished article. Use the source link below for the original report.