OpenAI announced new research results in mathematics, computer science, and physics on Monday, claiming to adopt a more collaborative method for sharing findings with academic communities. The company stated it is now employing expert shepherds to guide how these breakthroughs are communicated. This move follows a period of friction where researchers received hundreds of preprints without adequate context or explanation in early October. While the firm describes this as a course correction, several academics interviewed by The Verge express growing fatigue with the organisation’s handling of its scientific output. They note that the initial flood of unverified papers created significant confusion within the mathematical community. The current approach aims to address these concerns by ensuring external experts help shape the narrative before publication. However, the underlying issue of trust remains unaddressed by the mere presence of advisors. The results themselves show improvements in solving complex problems, but the process of sharing them continues to strain relationships with the broader scientific field. Researchers are now watching closely to see if this new strategy proves sustainable or if it is merely another temporary adjustment to public perception.
- Three specific areas of advancement were highlighted in the latest announcement.
- OpenAI is reportedly using external experts to vet how results are presented.
- Academics remain concerned about the initial lack of transparency regarding preprints.
