It’s been a weird week at OpenAI…

First, there have been a series of high-profile departures that have longtime AI watchers skeptical that everything is going well at the company.

Second, CEO Sam Altman has been publishing and engaging with cryptic posts online that reference the company’s top-secret “Strawberry” project that was leaked earlier this year.

Together, these developments could hint at something big brewing at the world’s most famous AI lab.

What’s really going on here?

I got the scoop from Marketing AI Institute founder and CEO Paul Roetzer on Episode 110 of The Artificial Intelligence Show.

The Great OpenAI Exodus?

First, let’s look at the departures. Here’s who left:

1. Greg Brockman, OpenAI president and co-founder, announced he’s taking an “extended leave of absence” until the end of the year.
2. John Schulman, another co-founder and key AI leader, left to join rival company Anthropic.
3. Peter Deng, a product leader who joined OpenAI last year from Meta, has also departed.

The exodus has led some industry observers to question OpenAI’s progress. As AI researcher Benjamin De Kraker put it on X: “If OpenAI is right on the verge of AGI, why do prominent people keep leaving?”

“All any of us can really do at this point is speculate,” says Roetzer. “It’s complicated, and I think we just gotta let it play out. I do have a lot of unanswered questions about the timing of Greg’s leave.”

It’s entirely possible that Brockman’s leave is a well-deserved break a long time in the making.

But it’s hard not to speculate, given what Sam Altman was up to online at the same time as these announcements, that something bigger is going on.

The Mystery Around Project “Strawberry”

On August 7, Sam Altman posted on X a picture of strawberries from his garden, a possible nod to the rumors that have been swirling for months about an internal project named Strawberry at the company. Many speculate that Strawberry refers to a major breakthrough in advanced reasoning that will be present in OpenAI’s next model. 

Just hours later, an anonymous account on X with strawberry emojis as its profile name posts the following, which Sam Altman himself promptly engages with:

“Level 2”, says Roetzer, is referring to the five-level scale that OpenAI came up with to track its progress towards artificial general intelligence (AGI), which was first reported in July 2024.

Level 2 on that scale is the next step beyond the AI models we have today: Models that can reason and engage in human-level problem-solving.

This mysterious account, featuring a profile picture of Joaquin Phoenix from the AI movie “Her” and tons more, then began flooding Twitter with AI-related posts, many referencing Altman’s strawberry tweet and implying some type of major release is coming.

Some AI watchers speculated that the account was being run by an OpenAI insider to tease a major launch. Others believe it could itself be an advanced form of AI running the account.

“My guess is that it is some form of AI,” says Roetzer. “I think there’s some human in the loop here, but I don’t think a human is managing this.”

Whose model is it? Roetzer doesn’t know. But he does suspect that this weird series of events means that OpenAI’s next act is about to unfold.

“I think the work of building the next model is complete, and they’re just finalizing timing and plans for the release of that model.”