We have questions.
OK…
A German Microsoft executive has, for some reason, claimed that OpenAI’s next large language model (LLM) will drop imminently.
“We will introduce GPT-4 next week, there we will have multimodal models that will offer completely different possibilities — for example, videos,” claimed Microsoft Germany CTO Andreas Braun during a digital kickoff event yesterday, per German tech news site Heise Online.
According to the report, Braun didn’t elaborate during the one-hour-long event, other than claiming that GPT-4 will work in “all languages” because the companies want to “make the models comprehensive.”
The report leaves us with plenty of questions. Largely unfounded rumors surrounding OpenAI’s next-generation LLM, which could lay the groundwork for the next version of ChatGPT as well, have been swirling for quite some time now.
While we have, of course, reached out to both Microsoft and OpenAI to corroborate this claim, we nevertheless have to wonder: why would a Microsoft Germany executive be the one to make such an announcement, and indeed, is this even Microsoft’s news to share?
Loose Lips
We’d be remiss to note that this isn’t the first time in recent months that news about Microsoft and OpenAI’s multi-billion-dollar partnership seems to have leaked to the press.
When anonymously-cited reports in January suggested that the deal was in the works, a Microsoft spokesperson was rather short when Futurism contacted the company for on-record confirmation.
“We do not comment on speculation,” the spokesperson said on January 11 — less than two weeks before OpenAI itself confirmed the deal.
Given the somewhat strange nature of the scoop, Heise journalist Silke Hahn took to Twitter to corroborate her story and her sources — and claims that someone from Microsoft contacted her after the story was published to thank her for it.
Altman About It
This isn’t the first news we’ve heard about OpenAI’s forthcoming language model, though what we have heard is also kind of bizarre: in a January interview with StrictlyVC, the firm’s CEO, Sam Altman, quipped that people may be disappointed by GPT-4.
Of its release date, Altman was cagey.
“It’ll come out at some point,” the CEO said, “when we are confident we can do it safely and responsibly.”
Has that point arrived? It’s too soon to tell — and it seems like some German Microsoft employees may have jumped the gun.
More on Microsoft: Microsoft Released an AI That Answers Medical Questions, But It’s Wildly Inaccurate