Independent newsroom

Sunday, October 18, 2026

The File

Every claim traces to a primary source. Uncertainty is labeled. Corrections are public.

Research · Sunday, October 18, 2026

Anthropic says Claude traced old encryption seeds

The File desk · Oct 18, 2026, 11:30 AM UTC

Status

Confirmed as Anthropic's research post of Oct. 16, 2026. The findings are the company's. The post says this is not a break in the curves.
  • Confirmed
    Anthropic posted that Claude found the origins of the NIST P-192, P-224, P-256, P-384, and P-521 seeds in under half an hour.
  • Confirmed
    The company says the seeds were output from a C library random routine, not a hash, and that this is not a cryptographic break.
  • Confirmed
    Anthropic says the $36,864 bounty will go to the Internet Security Research Group.
Anthropic said Friday that its model Claude found where five famous encryption-curve seeds came from. The company says the work took less than half an hour.1
The seeds are for curves known as P-192, P-224, P-256, P-384, and P-521. Anthropic says they were not the result of a hash. They were the output of an old random routine in a C library, started from a clock time.2
The company says this is not a break, a new attack, or a weakness in the curves. It says one seed, called B-163, is still unexplained, and so is the way several base points were made.3
Why it matters (Estimate). People have wondered for years whether those seeds hid a back door. Anthropic's account, if it holds, says they came from a clock and a library routine.
The bounty for finding the seeds is going to the Internet Security Research Group.

Notes

  1. 1.Anthropic, “Claude reveals ECDSA curve origins,” Oct. 16, 2026. “In under half an hour, Claude discovered the origins of the NIST P-192, P-224, P-256, P-384, and P-521 seeds.” https://www.anthropic.com/research/ecdsa-curve-origins
  2. 2.Same post. “None of them were the results of any hash algorithm at all, but rather the output of the C library’s lrand48.” The post describes that routine as seeded from a Unix timestamp. The curves are the ones listed in the federal Digital Signature Standard. This desk cites them through Anthropic's post and did not open a separate NIST page.
  3. 3.Same post. “This is not a cryptographic break, new attack, or weakness in the ECDSA curves.” “There is also one seed we were unable to recover. The fifth binary seed, B-163, fits nothing Claude has tried and remains unexplained.” The post also says it does not know how the base points for P-224, P-256, P-384, and P-521 were generated. “The $36,864 for finding the seeds will be donated to the Internet Security Research Group.”

What is still unknown or disputed

Primary sources

Every claim in this story is drawn from the documents below. If a fetch failed, that is recorded on the card.

  1. Source 1

    Claude reveals ECDSA curve origins

    Anthropic · October 16, 2026

    In under half an hour, Claude discovered the origins of the NIST P-192, P-224, P-256, P-384, and P-521 seeds. This is not a cryptographic break, new attack, or weakness in the ECDSA curves. The $36,864 for finding the seeds will be donated to the Internet Security Research Group.

    https://www.anthropic.com/research/ecdsa-curve-origins