OpenAI Drama: When Life Imitates Art, But Falls Short

*** Spoiler alert: The first paragraph contains spoilers from the show “Silicon Valley”***

Have you watched or fondly remember the hilarious HBO show Silicon Valley? The series follows a technology startup, Pied Piper, founded by the young and brilliant tech optimist, Richard Hendricks. Over the show’s six seasons, viewers witness Richard attempting to navigate the complex balance between his ideals of technology “making the world a better place” and the harsh realities of capitalist Silicon Valley. Ultimately, he creates a new AI-based internet that’s significantly more efficient and is ready for its big release (deemed as transformative as the iPhone launch). However, just before the release, the core team realizes that their AI engine has the power to breach any firewall or security system, potentially causing chaos in society. Opting for the greater good over personal fortune, they decide to scrap the product and the company, turning the launch into such a colossal disaster that it dissuades competitors from following their path. It’s a bittersweet ending that prompts reflection on what a real-life Richard Hendricks might do in a similar situation.

Last month, a strikingly similar scenario unfolded. OpenAI, the company behind ChatGPT, found itself akin to the show’s Pied Piper. Initially established as a non-profit with a mission to develop AI tools that benefit all of humanity, free from corporate greed, OpenAI faced the challenge of balancing deliberate AI adoption (to ensure safety precautions) with the potential windfall from ChatGPT products. This internal struggle came to its head when the company’s board, concerned about AI’s rapid evolution without adequate guardrails, surprised everyone by dismissing its celebrity CEO, Sam Altman, on Nov 17th. However, within days, Altman was reinstated as CEO, accompanied by the resignation of most board members. While the story is still unfolding and is shrouded in secrecy, let’s explore what likely happened in those four fateful days in November.

Founded in 2015 by Elon Musk, Sam Altman, and ten others, OpenAI started as a non-profit. The board comprised individuals worried about AI’s dangers, such as Musk and Dr. Sutskever, as well as those believing in its benefits, like Altman. The balance shifted towards the latter as Musk departed in 2018, Altman assumed the CEO role, and the company established a for-profit unit, attracting a $1 billion investment from Microsoft. At times, these actions seemed critical to fund OpenAI’s rapidly increasing expense base to sustain GPT models. However, enough board members remained concerned about AI’s dangers, keeping the company’s profile under check. 

All of that changed with the introduction of ChatGPT in Nov ’22, propelling OpenAI and CEO Altman into the spotlight of a revolutionary technology. Microsoft increased its investments to $13 billion to capitalize on future profits through GenAI. As Altman steered the company further into the for-profit arena, including potential investments in AGI (artificial general intelligence that can mimic or exceed human brains), safety concerns among board members intensified. The friction resulted in Altman’s dismissal on Nov 17th, replaced by interim CEO Mira Murati.

Recognizing Altman’s support within the company, including from interim CEO Murati, the board faced increased pressure when Microsoft announced the hiring of Altman and some OpenAI execs to establish a new AI lab. While it was not disclosed, it was clear that Microsoft’s chess move openly threatened OpenAI’s future. Concerns over the company’s future prompted 700+ of its 770 employees to collectively threaten resignation until Altman’s reinstatement and the removal of board members leading the charge. The board had no choice but to capitulate. On Nov 21, three board members (Dr. Sutskever, Ms. Toner, and Ms. McCauley) announced their departure, and Altman returned as CEO. The non-profit’s shift toward capitalism was complete, with new board seats going to an ex-Twitter board member, Microsoft (as an observer), and Larry Summers (former president of Harvard University).

Altman emerged victorious over those concerned about AI’s safety. OpenAI is now poised to immerse itself in AI development, with discussions of the next version of GPT models or models capable of solving unfamiliar mathematical problems. Moreover, OpenAI’s success in GenAI has triggered an arms race, with companies like Google and Meta launching their own GPT-like large language models or LLMs. Not to be left behind, virtually everyone else would invest heavily in developing smarter AI systems, bringing us significantly closer to singularity (or when machines will become smarter than humans). That said, the story is far from over, and it remains to be seen whether there will be a counterbalance to these advancements—perhaps through government regulations or a public movement advocating for safe AI.