OpenAI is developing new strategies to address the slowdown in artificial intelligence. This was reported to The Information by sources among the company’s employees. According to them, the new neural network, codenamed Orion, barely surpasses its predecessor GPT-4.
Author:
https://rb.ru/author/bmuzichenko/
Subscribe to RB.RU on Telegram
Although Orion performs better than OpenAI generative models on the market, the improvement in quality was not as significant as when moving from GPT-3 to GPT-4, employees involved in testing the new model said.
According to them, the neural network is worse at programming than its predecessors, but shows improvements in language capabilities. Orion may also be more expensive when used in data centers, The Information’s sources noted.
One of the reasons why the new OpenAI model does not live up to expectations is the lack of high-quality training data sets, most of which were already used for training in previous versions. To solve this problem, OpenAI has formed a team that will train neural networks using synthetic data created by AI, The Information writes.
In late October, The Verge reported that OpenAI planned to launch Orion in December. In response to this, the company’s director, Sam Altman, wrote on his X social network page that “the fakes are out of control,” without directly confirming or denying the information.
Author:
Bogdan Muzychenko
Source: RB
I am a professional journalist and content creator with extensive experience writing for news websites. I currently work as an author at Gadget Onus, where I specialize in covering hot news topics. My written pieces have been published on some of the biggest media outlets around the world, including The Guardian and BBC News.
