It has been a 12 months and a half since OpenAI launched its most up-to-date “basis” giant language mannequin (LLM), GPT-4, and the rumor mill has been in excessive gear the previous month with hypothesis on when a next-generation mannequin will seem.
Showing at a enterprise convention this week, the lead government for OpenAI’s Japan operations, Tadao Nagasaki, teased a forthcoming advance in LLMs from the corporate, which he known as “GPT Subsequent”.
“The AI mannequin referred to as GPT Subsequent, which is able to finally come out, will evolve almost 100 instances primarily based on previous achievements,” stated Nagasaki, primarily based on an automatic translation of the Japanese-language write-up of the convention presentation by Tachiki Matsuura for the net publication ITmedia.
A blurry picture of Nagasaki’s slide deck displayed on the monitor behind him confirmed a timeline of GPTs, beginning with “GPT-3 Period” in 2021, adopted by “GPT-4 Period” in 2023, “GPT Subsequent” period in 2024, and eventually “Future Fashions” with a date of “202x”.
It’s not clear from Matsuura’s write-up whether or not or not Nagasaki explicitly cited the supply of a “Subsequent” mannequin this 12 months in his remarks.
OpenAI was not in a position to present a remark to ZDNET at press time. If any remark ought to arrive, it is going to be included in an replace to this text.
In a gathering with journalists final November, OpenAI CEO Sam Altman indicated it could be some time earlier than a successor to GPT-4 seems.
“The variety of issues we have gotta work out earlier than we make a mannequin that we’ll name GPT-5 remains to be so much,” stated Altman, who added: “It’s totally exhausting to foretell timelines for this stuff.”
A weblog submit from the corporate in Could of this 12 months stated: “OpenAI has just lately begun coaching its subsequent frontier mannequin, and we anticipate the ensuing methods to carry us to the subsequent stage of capabilities on our path to AGI [artificial general intelligence].”
The inherent uncertainty of these remarks has solely fueled hypothesis, together with what appear to be additional clues.
Reuters’s Anna Tong and Katie Paul reported in July that the corporate is “engaged on a novel method to its synthetic intelligence fashions in a challenge code-named ‘Strawberry’,” citing a number of unnamed sources.
Rumors had been stoked additional final month when Altman posted on the social community X an image of a potted strawberry plant with the caption: “i really like summer time within the backyard”.
i really like summer time within the backyard pic.twitter.com/Ter5Z5nFMc
— Sam Altman (@sama) August 7, 2024
