OpenAI's unreleased AI model 'Astra' has made 10 major mathematical breakthroughs in just one week, using less than £1,500 worth of computing power. The discoveries include solutions to problems that have stumped mathematicians for over 60 years, such as a geometry puzzle first posed in 1964 by French mathematician Eugène Ehrhart.
Experts stunned by speed of breakthroughs
Experts have expressed shock at the pace of the advancements. Chris Peikert, a professor at the University of Michigan, called the results 'jaw-dropping', while Columbia University's computer science professor Henry Yuen admitted: 'I'm in awe.' Another mathematician wrote on an online forum that several of the discoveries would be 'career-defining for any mathematician'.
Fears of AI eclipsing human researchers
However, not everyone is celebrating. Tasmin Chu, a PhD student at the California Institute of Technology, posted: 'This ruined my morning.' Alex Gerko, a billionaire mathematician and founder of trading firm XTX Markets, predicted the technology would soon achieve 'superhuman performance in all aspects of maths'. He added: 'We are going to see 50 years' worth of maths progress in the next two years, but there won't be nearly enough human mathematicians around to process and understand all of it.'
OpenAI calls for collaboration
OpenAI stated that Astra is far more powerful than any publicly available system, though it plans to release the technology in the future. The company insisted that the breakthrough raises questions that cannot be answered by technology firms alone and called on the mathematical community to help shape the future of AI research. This development comes just months after thousands of mathematicians signed a declaration urging that the future of their subject should remain in the hands of academics rather than Silicon Valley companies.
