OpenAI’s subsequent flagship mannequin may not symbolize as massive a leap ahead as its predecessors, in keeping with a brand new report in The Info.
Staff who examined the brand new mannequin, code-named Orion, reportedly discovered that regardless that its efficiency exceeds OpenAI’s current fashions, there was much less enchancment than they’d seen within the soar from GPT-3 to GPT-4.
In different phrases, the speed of enchancment appears to be slowing down. In reality, Orion may not be reliably higher than earlier fashions in any respect in some areas, reminiscent of coding.
In response, OpenAI has created a foundations staff to determine how the corporate can proceed to enhance its fashions within the face of a dwindling provide of latest coaching information. These new methods reportedly embrace coaching Orion on artificial information produced by AI fashions, in addition to doing extra enhance fashions through the post-training course of.
OpenAI didn’t instantly reply to a request for remark. In response to earlier studies about plans for its flagship mannequin, the corporate stated, “We don’t have plans to launch a mannequin code-named Orion this yr.”
