The Aftermath of the OpenAI Board of Directors Drama

Earlier this week, I planned to publish a special edition of The Artificially Intelligent Enterprise based on events around OpenAI and the ousting of Sam Altman, the former and once again CEO. However, every time I finished a draft, the story kept unwinding, but I had to keep updating it. With every draft, there was a new twist or turn. Frankly, it could have been the basis for a Michael Lewis novel and probably will be someday.

Correction to last week’s newsletter

Thanks to one of my readers for pointing out that my link to The Artificially Intelligent Prompt Creator was broken. I apologize for the error. I didn’t want to clog your inboxes with a correction mid-week.

On November 17th, 2023, the board of OpenAI abruptly ousted Sam Altman, the co-founder and CEO of OpenAI (the company that makes ChatGPT). The board cited a lack of consistent candor in Altman's communications with them, hindering their ability to fulfill their responsibilities, as the reason for his dismissal. Since then, Microsoft announced that Altman would be joining Microsoft along with then Chairman of the OpenAI Board Greg Hartmann; in another twist, Altman was reinstated as CEO but not invited to join the board. However, he was not re-added to the board of directors.

I’ve tried to provide a summary as I sifted through tons of data. I also added to the trove of data the Wall Street Journal quoted me:

The Firing of Sam Altman and Microsoft

Sam Alman was fired by the board Friday a week ago, citing a lack of candor as the reason for his firing, but the details point to a lack of candor around specific capabilities of OpenAI, not financial malfeasance or the typical reasons for a CEO’s ousting. In fact, by any reasonable measure, the business success of OpenAI is fantastic.

Here’s what’s crazy: virtually any other board of directors would be thrilled with a company founded eight years ago to be valued at $86 billion. Altman’s been the worldwide face of AI for the last year, testifying in front of Congress during senate hearings on AI and becoming AI’s de facto corporate ambassador. He also had no equity in the company and lacked the dual-share class protections made famous by the founders of Meta and Alphabet. The whole corporate structure of OpenAI is unique.

The root of this blow-up probably lies in the OpenAI structure and a board focused on something other than purely business performance. OpenAI's board last week was very atypical, featuring Open AI Chief Scientist Ilya Sutskever, Quora CEO Adam D’Angelo, ex-CEO of GeoSim Systems Tasha McCauley, and Helen Toner, Director of Strategy at Georgetown’s Center for Security and Emerging Technology. This board stands apart from conventional corporate boards in a significant way: they are not focused on maximizing shareholder value, and notably, they do not possess equity in OpenAI. Their primary objective is to guide the development of artificial general intelligence (AGI) in a universally beneficial direction.

One factor that may have led to the firing was concerns raised in the paper published by Toner, Decoding Intentions, via The Center for Security and Emerging Technology, warning of the potential dangers of super-powerful AI. Insiders have revealed to Reuters reporters that shortly before Altman was removed, OpenAI staff was notified about a project called Q*. Mira Murati, a long-serving executive, informed employees about specific media reports but did not comment on their accuracy.

According to some at OpenAI, Q* (pronounced Q-Star) could be a breakthrough in the startup's search for Artificial General Intelligence (AGI). OpenAI defines AGI as autonomous systems that can surpass humans in most economically valuable tasks.

An anonymous source claims that the new model could solve specific mathematical problems using vast computing resources. Although it currently performs math problems at the level of a grade-school student, it does so with great accuracy. This achievement has made researchers optimistic about the future success of Q*. However, the individual who provided this information was not authorized to speak on behalf of the company.

Five Days of Chaos at Open

The firing of Altman was atypical, but then again, almost everything at OpenAI is. Microsoft, which has invested billions in OpenAI, learned that OpenAI was ousting CEO Sam Altman just a minute before the news was shared with the world, according to a person familiar with the situation. That’s not typical of how companies typically treat their most significant stockholder. Shortly after, Greg Brockman, OpenAI president and board chairman, announced he was quitting the firm. Brockman, a co-founder of OpenAI, shared his message to the firm's staff in a post on X, formerly Twitter:

Then, another plot twist: Microsoft scooped up Altman as their new Head of AI Research lab. Before this drama, there was speculation that Microsoft was pushing for a board seat to provide better oversight of OpenAI, which Microsoft owns just under 50% of the for-profit company. Ironically, it would make perfect sense for Microsoft to ask that Altman rejoin the board as an outsider. But that’s not the end of it….

Altman’s firing would then lead to an OpenAI internal revolt, with 745 of the approximately 770 employees at OpenAI demanding the return of Sam Altman or intending to leave. Amid this turmoil, a rash of other companies looking to poach OpenAI talent started to surface. Salesforce CEO Marc Benioff made an overture from his Twitter account.

But that was not to be…

Atlman’s Return

On Wednesday morning, Altman returned as CEO, announcing a new board. The newly established board of directors at OpenAI now features prominent figures, including Bret Taylor, previously Twitter's board chair before Elon Musk's acquisition, and Larry Summers, ex-U.S. Treasury Secretary. Adam D'Angelo, CEO of Quora, is also on the board and was involved in the decision to replace Sam Altman as OpenAI's leader.

Larry Summers brings a critical perspective on AI to the board. He has openly expressed concerns about AI's potential to disrupt the job market sooner than many experts predict. Besides his role at OpenAI, Summers is also a board member of Block, the payment company founded by Twitter's Jack Dorsey, and Skillsoft Corp, a software firm. These positions provide him extensive connections and insights into the Silicon Valley investment scene.

This board stands apart from conventional corporate boards in a significant way: they are not focused on maximizing shareholder value, and notably, they do not possess significant equity in OpenAI; from what I have read, Altman certainly had no equity, according to previous reports. Their primary objective is to guide the development of artificial general intelligence (AGI) in a universally beneficial direction. Also, no investors are on the board like Microsoft, which has roughly a 49% equity stake in OpenAI.

has A Short History of OpenAI, a good read that outlines the creation and structure of OpenAI and covers the history of OpenAI nicely.

The circumstances surrounding Altman's firing and the subsequent fallout have raised questions and concerns about the future of OpenAI and its impact on the AI industry. The sudden nature of the decision and the resulting turmoil within the company have left many in the industry wondering about the future of OpenAI. I think things have recovered, but here is my take….

My Takeaways on the OpenAI Chaos

• The OpenAI blow-up, in my opinion, was caused by terrible board decisions that lack the business decision-making skills for enterprises to continue to invest in them. However, the new board seems to be better suited to manage the complexities of OpenAI.

• While ethical and explainable AI is important, AI will move forward despite the governance of OpenAI. I struggle with the changes of OpenAI from a foundation to a foundation with a for-profit entity that is so transformational to the industry. This misalignment of the foundation's purpose and the company's will likely come at odds again.

• I think OpenAI will likely be successful, but I don’t think it will be the only player in this space and maybe not even the leader in AI in the industry. If you look at Google, it wasn’t the first search engine to outlast incumbents like Yahoo!, WebCrawler, and many others.

• While I think many companies will continue to and should use OpenAI, keeping an eye on other foundation models that compete with OpenAI makes a lot of sense - LLaMa 2, Anthropic Claude, Mistral, Falcon, and numerous others. Also, consuming OpenAI models through Microsoft might be a reasonable strategy.

Prompt of the Week: Research Paper Summarizer

ChatGPT can help save a lot of time summarizing research papers. For example, in the paper I mentioned, Decoding Intentions, you can use ChatGPT 4 to access the URL directly and the following prompt to create a summary of the paper.

Act as a research paper summarizer. I will provide you with a research paper on a specific topic, and you will create a summary of the main points and findings of the paper.Your summary should be concise and accurately and objectively communicate the paper's key points.You should not include opinions or interpretations in your summary but instead focus on objectively presenting the information from the paper. Your summary should be written in your own words and not include direct quotes from the paper. Please ensure that your summary is clear, concise, and accurately reflects the content of the original paper. The objective is to rephrase the provided text clearly, concisely, and formally without plagiarizing the original authors. The paraphrased text should include all information presented in the original source, use technical jargon and scientific terminology appropriate for scientists who are the intended audience, have inline citations for referenced ideas as (Author, date), consider breaking down complex sentences into simpler ones for better readability while ensuring that the meaning of the original text is not lost, and use examples wherever possible to clarify key points. After the revised text, please include a list of cited references in the appropriate citation format (APA, MLA, etc.), including full details of the sources used in creating the paraphrased text.