The OpenAI Talent Departure Provides Opportunities for Competitors

OpenAI established a reputation for taking daring research risks ahead of others. Following a recent exodus of talent, it must draw in a fresh wave of researchers to sustain this endeavor.

WHEN INVESTORS INVESTED $6.6 billion in OpenAI last week, they appeared mostly unfazed by the recent turmoil, which led to the sudden departure of the company’s chief technology officer, Mira Murati, along with chief research officer Bob McCrew and vice president of research Barret Zoph.

Nonetheless, those three exits were merely the most recent in a continuing departure of vital technical expertise. In recent years, OpenAI has experienced the departure of multiple researchers who were instrumental in creating the algorithms, methods, and systems that established it as a global leader in AI and a well-known entity. Multiple former OpenAI employees mentioned that the ongoing transition toward a more commercial emphasis remains a point of contention.

“Individuals who enjoy conducting research are being compelled to focus on products,” states a former employee at a competing AI firm who has acquaintances at OpenAI. This individual mentions that certain contacts at the company have made inquiries about job opportunities in recent weeks. According to data gathered by Lightcast, a firm that monitors job postings to assess labor trends, OpenAI appears to have adjusted its hiring focus. In 2021, 23 percent of its job listings were for general research positions. In 2024, general research constituted only 4.4 percent of job advertisements.

The exodus of talent may have enduring effects on OpenAI’s trajectory and prospects for success. Professionals and ex-employees claim the firm retains a strong pool of talent, yet competition is ramping up, making it tougher to keep a competitive advantage.

The most recent high-profile exit, announced on Thursday, is Tim Brooks, who led OpenAI’s Sora AI video generation initiative. Brooks announced on X that he would be joining Google DeepMind, one of OpenAI’s primary competitors.

“It might begin to alter circumstances,” states a former OpenAI employee, who currently works in academia, regarding the losses. They requested to stay unnamed due to worries about damaging cooperative ties with the AI sector.

Currently, this individual mentions that numerous students continue to rank OpenAI as their preferred choice among potential employers. It is viewed as being several months ahead of its rivals, and potential employees frequently tolerate the visible drama and conflict to join in. However, candidates are frequently attracted to collaborating with specific researchers or teams, and their considerations may shift as more prominent researchers depart for competing AI firms or establish their own startups.

An examination of several key studies from OpenAI reveals the significant loss of talent. Out of 31 individuals identified as authors of an earlier version of OpenAI’s GPT large language model, less than half are still employed at OpenAI, based on employment information obtained from LinkedIn or other public social media profiles. In 2021, multiple team members involved in creating GPT departed from OpenAI to establish Anthropic, which has become a significant competitor. Approximately a third of the individuals mentioned in the acknowledgements of a technical blog post about ChatGPT have since departed.

A crucial question is if OpenAI possesses a sufficiently strong team of AI researchers to remain at the forefront of important advancements in AI, which is vital for backing its soaring valuation.

According to a report by The New York Times referencing a document provided to potential investors, OpenAI seeks to boost its annual revenues to $100 billion by 2029. The firm anticipates that ChatGPT will generate $2.7 billion in revenue this year, an increase from $700 million in 2023, along with an additional $1 billion from other channels. The competition is also heating up, as Google and others are investing significant resources in rival AI products. The social media powerhouse Meta provides complimentary AI models and tools that frequently match the quality of those offered by profit-driven companies.

Among the largest responsibilities that OpenAI faces are those of Ilya Sutskever, a cofounder and the technical visionary behind a significant portion of the company’s initial projects. Sutskever left in May to establish a new AI company named Safe Superintelligence.

Sutskever was a trailblazer in deep learning and was an early proponent of the idea that increasing the computational power for model development would result in more advanced AI—a forecast that turned out to be strikingly accurate. He also importantly recognized that a neural network structure created by Google researchers called the transformer could be vital for enhancing AI’s language abilities.

Sutskever was part of the board that momentarily removed CEO Sam Altman in November 2023, although he later apologized and advocated for Altman’s reinstatement a few days afterward. The incident appeared to ignite internal disputes regarding Altman’s authority over the company, its business priorities, and willingness to take risks. According to a report in The Wall Street Journal, Murati frequently found herself at the heart of internal disputes, particularly regarding the rush to launch the company’s voice interface.

The exit of Sutskevar was quickly succeeded by Jan Leike’s departure, who co-headed a team focused on overseeing the long-term risks associated with AI. In August, John Schulman, an OpenAI cofounder and esteemed research scientist, additionally declared his departure. Schulman focused on, among other aspects, methods for fine-tuning extensive language models with human input through a process known as reinforcement learning, an approach that was crucial for creating a proficient and talkative chatbot like ChatGPT. Another notable figure to depart this year was Andrej Kaparthy, a cofounder who returned to the company in 2022 after serving as director of AI at Tesla. In February 2024, Kaparthy departed to establish Eureka Labs, a startup focused on education AI.

Another early OpenAI staff member, who requested anonymity as they were not permitted to address the issue by their present employer, states that these exits are noteworthy due to the crucial role the research conducted by these individuals has played for OpenAI. Nonetheless, this individual points out that OpenAI retains several exceptionally skilled young researchers—and can also provide substantial financial incentives to prospective hires due to its rising valuation.

OpenAI’s most recent investment round occurs while the organization strives to transform into a for-profit entity. The initial founders featured Elon Musk, who contributed $45 million in initial financing. Once Musk left the company, Altman transformed it into a capped for-profit organization managed by a board of directors that is not accountable to investors. The coup attempt last November demonstrated that this structure is in direct opposition to a rapidly expanding, well-funded technology firm.

A former OpenAI employee, who also requested anonymity, verified that the product emphasis still led to friction, noting they had been informed by current employees to anticipate further unrest within the company.

Former employees claim that the internal dynamics of the company are intricate. “One individual remarked, ‘The atmosphere is gloomy for some, while others feel differently.’”

When questioned about the situation, OpenAI directed Press to a message Altman sent to the staff following the exits of Murati, McGrew, and Zoph, which he also published on X. In the message, Altman mentions various researchers who will take on more senior positions, including Mark Chen, who will assume the role of senior vice president of research, and Jakub Pachocki, who will be the company’s chief scientist.

“Changes in leadership are a normal aspect of companies, particularly those that expand rapidly and have high demands,” Altman stated. “I certainly won’t act like it’s normal for this to be so sudden, but we aren’t a typical company.”

“According to Simon Johnson, a professor at MIT’s Sloan School of Management, if you believe that a select few possess unique skills and that they have recently left, then that poses an issue.”

Nonetheless, Johnson asserts that OpenAI’s setback could potentially transform into an advantage for the rest of the technology sector.

“I believe we require significantly more competition in AI; we need a greater variety of business models; and we need a larger number of intelligent individuals developing positive solutions,” he states. “I’m quite at ease with the individuals departing.” After OpenAI’s board removed CEO Sam Altman last November, employees threatened to resign in protest, and many used the phrase “OpenAI is nothing without its people” on X. The company appears intent on finding out how interchangeable certain key individuals actually are.