OpenAI Founders' Journey - A Transformer Organization, Transformed

Welcome to our "Founders' Journey" series by CoffeeSpace, where we explore the remarkable stories and cofounder journeys behind the world’s most successful startups.

In this edition, we explore the journey of OpenAI, better known as the company that introduced ChatGPT to the masses. OpenAI is a leading organization specializing in artificial intelligence (AI) research and development, which has revolutionized the field and reshaped our understanding of technology's potential. From foundational research to major breakthroughs, we’ll examine how OpenAI has navigated the intricacies of innovation and ethics in the fast-changing world of AI, and what lessons we can learn from their path in the rapidly evolving AI landscape.
‍

The Need for OpenAI: A Conversation between Musk and Page

In an interview with Fox News, Elon Musk shared a conversation between himself and Google’s founding CEO, Larry Page, about AI safety. The two tech entrepreneurs had differing visions for AI development—Page focused on pursuing digital superintelligence as soon as possible, while Musk was more concerned with the safety of humanity and the need for actions to minimize the risks of AI. Public statements have also indicated that Google is researching and developing Artificial General Intelligence (AGI), a type of AI that aims to mimic the cognitive abilities of the human brain.

Musk, a vocal advocate for ethical AI development, described AGI as a potential existential threat to humanity and civilization, emphasizing the need for global cooperation and caution. He recounted being called a "speciesist" by Page during their conversation. Musk’s concerns led him to co-found OpenAI in December 2015 along with Sam Altman, Greg Brockman, Ilya Sutskever, Wojciech Zaremba, and John Schulman. The AI research non-profit organization received backing from investors like Reid Hoffman, Jessica Livingston, Peter Thiel, Amazon Web Services (AWS), Infosys, and YC Research, with a $1 billion pledge.

The name “Open” reflects OpenAI's initial commitment to openness and transparency, with a promise to freely share research and patents with the broader AI community and ensure that its work would be open-source. Initially established as a non-profit organization, OpenAI aimed to prevent AI from being monopolized by a few powerful corporations or governments. This mission was in direct contrast to tech giants like Google and Meta, which were increasingly investing in proprietary AI research.

OpenAI’s Early Development 

In its early stages, OpenAI embarked on research projects spanning various domains within AI, including natural language processing, reinforcement learning, and robotics. Their multidisciplinary approach addressed technical challenges and contributed significantly to the scientific community. One of OpenAI's foundational works was the development of Generative Pre-trained Transformers (GPT), which laid the groundwork for future innovations in language understanding and generation.

A notable project involved AI systems capable of playing complex games like Dota 2, requiring substantial computational resources and showcasing OpenAI's ambition to push the boundaries of AI capabilities. These projects helped establish OpenAI as a thought leader in the AI community.

The journey began with GPT-1, introduced in June 2018. Based on the transformer architecture, GPT-1 employed unsupervised learning techniques to predict the next word in a sentence. With 117 million parameters, GPT-1 demonstrated the potential of large language models (LLMs) to understand and generate human-like text, setting the foundation for future iterations.

OpenAI LP, Microsoft & Musk’s Departure: The Shift to Commercialization

A pivotal moment for OpenAI came in 2018, when Musk resigned from the OpenAI board due to a reported conflict of interest with Tesla’s AI developments for autonomous vehicles. Another reason for his departure, as later reported, was disagreements over OpenAI’s direction and leadership. Musk wanted OpenAI to take a more aggressive stance on preventing AI dangers, while other cofounders, particularly Sam Altman, leaned toward rapidly advancing AI capabilities. Musk later criticized OpenAI’s direction, stating it had drifted from its original mission.

In March 2019, OpenAI transitioned from a non-profit to a "capped-profit" entity, OpenAI LP, which allowed investors to make returns up to 100x their investment, with profits beyond that directed toward funding OpenAI’s mission. This controversial shift raised skepticism about the organization's future.

In July 2019, OpenAI announced a $1 billion investment from Microsoft to support the development of AGI. The partnership also included a collaboration in which OpenAI would exclusively use Microsoft Azure as its cloud platform. This funding enabled OpenAI to scale its compute power significantly and develop larger language models.

Impressive Growth, Massive Loss: The Groundbreaking Launch of ChatGPT

In February 2019, OpenAI unveiled GPT-2, expanding the model’s capabilities to 1.5 billion parameters. Initially, the release was cautious due to concerns about potential misuse, and the full model was withheld. However, after extensive testing and feedback, GPT-2 was fully released in November 2019, showcasing its ability to generate high-quality text across various contexts.

DALL·E, a groundbreaking text-to-image model, was introduced in January 2021, demonstrating AI’s potential in creative fields. But it was the launch of ChatGPT on November 30, 2022, that truly captured public attention. Within five days, ChatGPT reached 1 million users, breaking records for the fastest-growing consumer application in history. By January 2023, it had surpassed 100 million active users.

Despite impressive user growth, OpenAI faced enormous operational expenses, leading to a $540 million loss in 2022. With rising hardware costs and the need for top talent in a competitive market, analysts projected further losses of up to $44 billion between 2023 and 2028. With their user number continuing to grow in 2024, at its peak in April 2024,  the site was receiving nearly 2 billion visits alone according to a tabulation by Exploding Topics. 

Altman as CEO, To Be or Not To Be? 

On November 17, 2023, the OpenAI board of directors composed of  Helen Toner, Ilya Sutskever, Adam D'Angelo and Tasha McCauley removed Sam Altman as CEO, citing a unanimous agreement over the lack of confidence in his leadership. Cofounder and president of OpenAI Greg Brockman was also removed from his position as chairman of the board but retained his role, while Mira Murati, the CTO of that time took over as interim CEO. The president walked out of the company after Altman got forced out of the company shortly thereafter. This outburst also led to significant senior OpenAI researchers resigning and leaving the company.

The next day came reported pressure amid the board from major investors and stakeholders such as Microsoft and Thrive Capital, objecting to the decision and urging the return of Altman. In the midst of negotiations, the board initially contacted Dario Amodei, a former OpenAI executive and CEO of rival company Anthropic, about replacing Altman, and proposed a merger of the two companies, but both offers were declined. 

Microsoft CEO Satya Nadella announced on the following day that Altman and Brockman would join Microsoft’s team to lead an advanced AI research team, but the company would still stay committed to OpenAI. Before sealing the deal, in an open letter among the employees, about 738 of OpenAI's 770 employees expressed their intentions of resigning and joining Microsoft if the board did not rehire Altman, forcing the board to reconsider their decision and hence the negotiations continued between Altman and the directors.

On November 21, 2023, the final decision came: Altman and Brockman were to return to the company in their prior roles along with a reconstructed board with new members. Microsoft then took a non-voting, observer position on OpenAI's board after Altman’s return. 

Musk vs Altman: Lawsuit Drama 

The drama continues following the firing and hiring of Altman, with the legal drama unfolding between Elon Musk and Sam Altman, both co-founders of OpenAI. The saga began when Musk filed a lawsuit in February 2024, alleging that OpenAI and Altman had strayed from their founding mission to develop artificial intelligence for the benefit of humanity. Musk claimed that OpenAI, following its partnership with Microsoft, had become overly profit-driven, which he argued transformed the organization into a "closed-source de facto subsidiary" focused on maximizing profits rather than public good. He contended that this shift violated a supposed founding agreement that prioritized transparency and safety in AI development.

However, the lawsuit took an unexpected turn when Musk abruptly withdrew it on June 11, 2024, just one day before a scheduled hearing regarding OpenAI's request to dismiss the case. This withdrawal came after Musk had previously expressed intentions to revive the lawsuit, indicating ongoing frustration with OpenAI's trajectory. His legal team had filed objections in court just weeks prior, suggesting that the lawsuit was still very much alive. Musk's decision to drop the case left many puzzled, especially given his vocal criticisms of OpenAI's direction and his establishment of his own AI venture to compete against OpenAI, xAI.

In the aftermath of the lawsuit's withdrawal, Musk's relationship with Altman remained strained. The two had co-founded OpenAI in 2015 but parted ways in 2018 due to disagreements over control and direction. Their professional relationship deteriorated further as Musk publicly criticized Altman and OpenAI for not adequately addressing potential risks associated with AI technologies. In August 2024, Musk revived his legal complaint against Altman and others, reiterating claims that they had manipulated him into co-founding OpenAI under false pretenses regarding its nonprofit structure. This latest ongoing suit accused Altman and Greg Brockman of prioritizing commercial interests over their original mission.

The OpenAI Founding Team: Where Are They Now?

Sam Altman

Former Y-Combinator president, Altman, who is the CEO of OpenAI since 2019 after the initial Microsoft investment, upon his tumultuous journey with the board of investors ousting and reinstating his position as OpenAI’s CEO, continues to serve in the company.

Greg Brockman

Brockman, former CTO of Stripe, is now the cofounder and president of OpenAI. First announcing his sabbatical leave from the company in August 2024, he is still seen tweeting and sharing updates about the company’s new launches and developments. 

Wojciech Zaremba

Zaremba started his journey with the company in 2016 after ending his roles as a research scientist in Facebook AI Research and Google Brain prior to joining. He is currently still with the company.

Elon Musk

The tech billionaire who provided most of OpenAI’s early funding left the company in 2018 due to a reported conflict of interest between OpenAI and Tesla, his own automotive company’s AI development. There are also reports alleging Musk’s departure may be due to the clash with OpenAI’s CEO Altman over the direction of research the company is heading. Musk then launched his own competitor, xAI, in 2023 with claims to overhaul OpenAI's development and lead in the field.

Ilya Sutskever

OpenAI Chief Scientist and cofounder Sutskever left the company after 8 years in June 2024 to start his own venture, Safe Superintelligence Inc, with its aim to develop AI systems that surpass human intelligence safely. He played a key role in the brief ouster of Sam Altman as the CEO, and reportedly his leaving is due to the clash with Altman on the company’s direction to develop AI. 

Vicki Cheung

Previously the founding engineer for Duolingo and soon later OpenAI in 2015, Cheung left the company in 2017 for an engineering manager position at the ride-hailing startup Lyft, in which she expressed in an interview with Hashicorp that she wanted to “spend some time working on diversity and inclusion issues and tech” and that led her to try out her new role at Lyft. Shortly thereafter, she started her own venture at Gantry, where she is the cofounder and CTO building ML systems and models.

Andrej Karpathy

Slovak-Canadian research scientist and founding member Karpathy helped develop OpenAI from 2015 to 2017. However, in 2017, he left the company to lead an autopilot team as a Senior AI Director in Tesla. He soon returned to OpenAI in 2023 in which he helped improve GPT-4 on ChatGPT. Karpathy left in 2024 upon the completion of the project, and in a statement addressing his departure published on X, he mentioned that there was no bad blood and expressed his interest to pursue his personal projects, which he then founded Eureka Labs, a new AI+Education company.

John Schulman

Schulman was a part of OpenAI for more than 9 years and he announced his exit from the company on X recently this August as he also sets to join OpenAI's rival Anthropic, stating that he  hopes to “gain new perspectives and do research alongside people deeply engaged with the topics I'm most interested in” at his new position.

Durk Kingma

Angel investor and machine learning researcher focusing on generative models, Kingma left OpenAI in 2018 to join Google DeepMind as a research scientist where he also led several projects on generative models. He too joined the Anthropic pipeline in October 2024 as a ML researcher, claiming that Anthropic’s approach and mission of developing powerful AI systems resonated significantly with his own beliefs.

Trevor Blackwell

Blackwell was a partner at Y-Combinator and helped start OpenAI before leaving the company shortly after in 2017 with little to no explanations.

Pamela Vagata

Vagata is one of the lesser known cofounders of OpenAI, with no mentions of her role in the company on her LinkedIn profile. In 2016, she allegedly left the company to join Stripe with no mentions of the reason for her departure. Since 2021, she is currently leading Pebblebed, an early stage venture firm she cofounded which specializes in early-stage AI tech investments. 

The Future of OpenAI 

So, this brings us to the question: what does the future hold for OpenAI? Both promising and precarious, OpenAI anticipates that its revenue could soar to $11.6 billion in 2025, with a hopeful estimation of $3.7 billion in 2024, which is a remarkable increase from the $200 million revenue in 2022. This optimistic forecast is underpinned by the growing demand for generative AI technologies across various sectors.

Figures and numbers aside, the future of AI development remains uncertain, with more apparent and booming growth of interest in AGI from other tech giants like Google and Meta. While these companies are making significant investments and strides in research, ethical considerations and operational challenges complicate the playing field. As OpenAI and its competitors navigate this complex terrain, it will be essential for all stakeholders to prioritize responsible development practices and engage in open dialogue about the implications of advanced AI technologies on society. The journey toward AGI promises to be both transformational and fraught with challenges, requiring careful stewardship to ensure that its benefits are realized for mankind, and not a dystopia controlled in the hands of the dictated few. 

Conclusion: The Takeaways for Founders

Prioritize Core Values

OpenAI's mission centers around developing safe and beneficial AI technologies. When leadership changes occur, it is crucial to reaffirm commitment to these core values. The erosion of trust among employees regarding the organization's dedication to safety illustrates how vital it is for leaders to consistently demonstrate alignment with foundational principles.

Agility in Leadership

In the startup world, there is rapid evolution and the changes in its leadership demonstrate the necessity for leaders to be agile. The ability to adapt to new challenges and market conditions is crucial to founders when facing new challenges and pressure from external parties. Embracing change and encouraging pivot strategies effectively are also important to grow a startup and reach for greater heights.

Embracing change

As all early stage startups could relate to, the initial stages are volatile with many uncertainties and unpredictability, hence it is crucial to be strapped in and prepared for change. Keeping an open mind and taking every pivot as an opportunity to grow is a healthy growth mindset that all founders should adopt. 

‍If you’re inspired by this story and want to start exploring your own ideas and find someone to get off the ground with, join us at CoffeeSpace.