OpenAI has released solutions to three math problems, including a result that rejects the Exponential Time Hypothesis (ETH), a durable assumption in computer science regarding what is difficult to compute, TechCrunch reported.[1]

In an open letter, the group says it could “change an important part of theoretical computer science”, New Scientist reported.[2]

Anthropic mathematician Levent Alpöge suggested OpenAI based its solutions on pre-existing work, after which Anthropic released its own refutation of ETH, TechCrunch reported.[1] Magnus Wahlström at Royal Holloway described the phenomenon as mathematics done by press, expressing concern that proprietary models could suddenly overturn months of research, New Scientist reported.[2]