OpenAI cracks million-dollar math problem and sparks scientific scandal
OpenAI embroiled in plagiarism scandal (photo: Google DeepMind)
Astrologers have prepared a Tarot forecast for September 11 for each zodiac sign. The cards indicate where you should be more cautious and where you can confidently take a chance and move forward.
RBC-Ukraine reveals what you should pay special attention to this Friday.
An AI breakthrough in 88 hours
To solve a fundamental mathematical problem related to the physics of water and other fluids, OpenAI used a new, previously unannounced model that is more powerful than GPT-6 Astra.
During the process, around 10,000 AI agents were divided into groups, exchanged nearly 3 million messages, and used more than 130 billion tokens.
The main part of the computation took approximately 88 hours, while further verification of the proof using the Lean 4 language took another 17 hours.
Interestingly, OpenAI said it was declining the cash prize offered by the Clay Mathematics Institute.
Accusations of using other researchers’ data
The scandal erupted after a statement by New York University (NYU) mathematics professor Tristan Buckmaster.
According to him, OpenAI began its work only after learning about rumors of a breakthrough in a related topic (forced Euler), where he had been working with Levant Alpogé from Anthropic.
The researchers used OpenAI Codex and Anthropic Claude tools. Buckmaster suspects that OpenAI may have used their private training logs.
In response, OpenAI said that its developers and agents did not have direct access to user data, but acknowledged that it “cannot completely rule out the use of de-identified data” from its products.
Reaction from the scientific community
Prominent mathematician and Fields Medalist Terence Tao expressed deep concern.
According to him, when rumors about scientific research prompt technology giants to spend millions of dollars on computing power in an attempt to get ahead of researchers, it undermines the tradition of open science.
As a result, researchers may stop sharing hypotheses, which could harm the long-term development of mathematics.
