iCreatAI News
Skip to content
Trending storyUpdating

HN questions whether AI's RH proof or assembled training data

1 reports1 sources1 hr agoUpdated

Understand this story

AI overview

Hacker News users posted doubts that Claude's proof of RH may not have been completed independently, but rather assembled a complete proof from scattered proof attempts by others on platforms such as GitHub and Google Docs; these fragments may have already entered the training data. The poster said that assembling existing proof fragments and independently writing a proof are two different things. The user also mentioned that AI had a similar performance on the earlier Ergos problem, and Terence Tao once commented that it merely connected existing dots and had no truly new ideas.

AI overview based on source reports · Updated 25 min ago

Coverage timeline

Follow the coverage to explore different perspectives.

Nov 23
  1. Hacker News AI
    Ask HN: Could AI's RH proof have used human work found in its training data?

    An Ask HN post asks whether an AI's Riemann Hypothesis proof could have assembled fragments of human proof attempts stored on GitHub or Google Docs that ended up in training data, rather than authoring the proof itself. The author notes the distinction matters because assembling existing parts seems different from original authorship, and cites an earlier Ergos problem where Terence Tao said the AI connected dots but didn't think of anything truly novel.

Activity for this story

Not enough continuous observations to show a trend yet.