To The Moon

The WSJ recently published an inside look at OpenAI, analyzing in detail last year's "court battle" drama and why the executive team fell apart.

OpenAI is undergoing an unprecedented transformation and upheaval.Founded in 2015, the nonprofit organization, which was originally dedicated to developing artificial intelligence technology that would benefit all of humanity, is now facing the significant challenge of transitioning from a nonprofit to a for-profit company.Behind this change, a wave of departures, internal struggles, and leadership strife form a complex picture of the inner workings of the organization.

Right now, Sam Altman has set out to change OpenAI's nonprofit structure, turning it into a true commercial company and raising capital at a valuation of $150 billion.And Sam's own acquisition of a 7% stake, while denied by the company, implies his own eagerness to become a leader of a commercial company.

I'm afraid that's one of the key reasons why $Apple(AAPL)$ pulled out of the funding round. But $Microsoft(MSFT)$ will continue to follow suit.

About Chief Scientist Ilya Sutskever

The brief incident in which OpenAI employees brought back Sam Altman's firing and reinstatement last November last year only lasted a few days, but the repercussions of this event are still festering within the company.

• The first sign was the sudden absence of one of the co-founders and most respected research scientists, Ilya Sutskever.At the time, he conveyed the news of his dismissal to Altman and then publicly apologized for his role.He never returned to work at the office.

• When Sutskever resigned in May of this year, as did Jan Leike, who co-led the security team with him, OpenAI executives feared that their departures could trigger larger personnel changes and worked to bring Sutskever back.

• CTO Murati and President Greg Brockman told Sutskever that the company was in turmoil and could collapse without him, so they visited his home, bringing cards and letters from other employees who wanted him back.

• Altman also visited him and expressed regret that others had not found a solution.

• Sutskever told his former colleagues that he was seriously considering returning.But shortly afterward, Brockman called to say that OpenAI had withdrawn its invitation for him to return.That's because internal executives had trouble defining Sutzkever's new role and how he would work with other researchers, including his successor.

• Shortly afterward, Sutzkever started a new company focused on developing state-of-the-art artificial intelligence without the distractions of the process of releasing a product.The company, called Safe Superintelligence, has now raised $1 billion.

About CTO Murati

On Wednesday, OpenAI's CTO Mira Murati became the latest executive to announce his departure, just as the company prepares to transform itself into a for-profit company, and in fact a public outburst of the growing tensions behind ChatGPT.

• The alleged reason is that OpenAI has lost its lead to rival AI developers by rushing product releases and security testing, and that there are growing pains between maintaining mission-driven research operations and fast-growing business within the same organization

• Outsiders of OpenAI's repeated delays in releases, such as search and voice in general, feel that it is Sam Altman who is unreliable, when in fact it is he who would like to launch quickly, but will always be stopped by Mira Murati, the CTO who is leaving the organization this time, because she believes that the technology is not mature enough.

• Tensions escalated within the company this spring while developing a new AI model called GPT-4o to support ChatGPT and commercial products.Researchers were asked to conduct more comprehensive safety tests than initially planned, but had only nine days to do so.Executives wanted to launch 4o before Google's annual developer conference to draw more attention to it and thus put down competitors.

• Security teams work 20 hours a day and don't have time to scrutinize their work.Initial results, based on incomplete data, suggested that GPT-4o was safe enough to deploy.However, after the release of the model, people involved in the project said subsequent analysis found that the model exceeded OpenAI's internal standards for persuasion (defined as the ability to create content that can convince people to change their beliefs and engage in potentially dangerous or illegal behavior).

About Greg Brockman, President

Brockman, who is currently on vacation, is considered a longtime loyal OpenAI employee.When OpenAI was founded in 2015, it initially operated out of the living room of Brockman's home.Later, he was also married in the company's office.

As OpenAI grew, his management style sparked tensions.Despite serving as president, Brockman had no direct reports.He tends to get involved in any project he wants, which often frustrates those involved, according to current and former employees.They said he demanded last-minute changes to long-term plans, which prompted other executives, such as Murati, to step in to ease the situation.