0 reply
0 recast
0 reaction
OpenAI and Google fight for gold
Experimental AI models from OpenAI and Google have both achieved gold-medal scores in the annual International Math Olympiad (IMO)—the world’s biggest, most challenging math competition for high school students. However, OpenAI didn’t officially enter the competition: They chose to complete the competition in their own time and publish the results before the IMO published the official scores. This sparked controversy and a public backlash from Google, who was invited to participate in the competition and waited for the IMO to publish its results, so they didn’t “steal the spotlight from the students.”
Google and OpenAI both used “informal” AI systems, which translate, understand and answer questions, without human input to compete (Google scored silver last year, using a model that needed human input).
Both models followed the same rules as the students. This included sitting two, 4.5-hour exams, over two days, to solve a total of six questions, with no access to the internet or any other tools.
IMO officially verified that Google’s model answered 5 out of the 6 questions correctly (scoring higher than most IMO students) and OpenAI announced that their model also scored the same.
Putting whether OpenAI did or didn’t formally enter the IMO competition aside, the gold-level results highlight a significant breakthrough for AI reasoning models and their ability to solve advanced math problems. It also shows how close Google and OpenAI are in terms of progress, which is something that will undoubtedly intensify the already fierce competition to hire the industry’s best AI talent.
https://in.mashable.com/tech/97400/openai-claims-gold-medal-performance-at-prestigious-math-competition-drama-ensues 0 reply
0 recast
0 reaction
