In a significant announcement, OpenAI has claimed that its latest reasoning model has provided an original proof disproving a famous geometric conjecture first proposed by mathematician Paul Erdős in 1946. This assertion follows nearly eight months after a controversial claim about the AI's ability to solve multiple Erdős problems sparked backlash within the mathematical community.
OpenAI's earlier attempt to showcase its AI capabilities faced criticism when former VP Kevin Weil claimed that GPT-5 had solved ten unsolved Erdős problems. It was later revealed that the model had only identified solutions that were already known. This misstep drew backlash from prominent figures in the AI field, including Yann LeCun and Demis Hassabis, resulting in the deletion of Weil's post. In contrast, OpenAI's latest claim has support from mathematicians such as Noga Alon, Melanie Wood, and Thomas Bloom, who affirm the validity of the new proof.
According to OpenAI, the proof indicates that for nearly 80 years, mathematicians have been misled into believing that optimal solutions resembled square grids. The AI has allegedly discovered a new family of constructions that surpass these long-held assumptions. OpenAI highlighted that this breakthrough marks the first instance where AI has independently addressed a significant open problem in mathematics—not just by solving equations but by fundamentally reshaping understanding in the field.
This development has broader implications beyond mathematics, suggesting a potential shift in how AI systems can connect various disciplines such as biology, physics, engineering, and medicine. OpenAI argues that its models' ability to maintain complex reasoning chains and link disparate ideas could lead to new insights and discoveries across scientific fields. Thomas Bloom noted, “AI is helping us to more fully explore the cathedral of mathematics we have built over the centuries. What other unseen wonders are waiting in the wings?”
The ramifications of this claim are substantial for both the AI and mathematical communities. While there is excitement about the potential applications of such advancements, skepticism remains grounded in recent history. Industry experts will likely scrutinize OpenAI's proof and the model's reasoning processes to evaluate the validity of its assertions. The ongoing dialogue between AI development and traditional fields like mathematics will be crucial in shaping the future of both disciplines.
Quick answers
What conjecture did OpenAI claim to disprove?
OpenAI claims to have disproved a geometric conjecture posed by Paul Erdős in 1946.
What were the criticisms of OpenAI’s previous claims?
OpenAI faced backlash when it was revealed that its earlier model had only found existing solutions rather than solving new problems.
How does this development impact other scientific fields?
The ability of AI to connect ideas may lead to breakthroughs in biology, physics, engineering, and medicine.
