Khaberni - OpenAI gives priority to developing the chatbot "ChatGPT" at the expense of long-term research, prompting senior employees to resign, amid the company’s efforts, which is valued at 500 billion dollars, to adapt to fierce competition from companies such as Google and Anthropic.
According to ten current and former employees, the startup, based in San Francisco, has reallocated resources originally reserved for experimental work in favor of developing the large language models that power its main chatbot.
Among those who have left "OpenAI" in recent months due to this strategic shift are the vice president of the research department, Jerry Towrk, model policy researcher Andrea Fallon, and economist Tom Cunningham, according to a report by "Ars Technica" a website specialized in technology news, reviewed by "Al Arabiya Business".
This shift in "OpenAI" represents a significant transition for the group from which "ChatGPT" emerged in 2022 before it sparked a generative artificial intelligence boom.
The company, led by CEO Sam Altman, is transitioning from a research lab to one of the largest companies in Silicon Valley. This means that the company must demonstrate its ability to generate the revenues needed to justify its valuation of 500 billion dollars.
One insider familiar with the company’s research ambitions said: "OpenAI is now treating language models as an engineering problem where they are scaling up computing, algorithms, and data, and making very significant gains from that."
He added: "But if you want to conduct original, innovative research, it’s very challenging. And if you do not find yourself within one of the main teams, it becomes more complicated politically."
Mark Chen, chief research officer at "OpenAI", rejects this description, saying, "Fundamental long-term research remains pivotal to OpenAI and continues to absorb the majority of our computational resources and investments, with hundreds of projects ranging from grassroots to pinnacle, which explore long-term issues beyond any single product."
Like other major tech companies, researchers at "OpenAI" need to apply to executives for "credits" for computing and access to technology to initiate their projects.
Several individuals close to the company said that in recent months, researchers' requests who did not work on large language models were often denied or granted insufficient size to validate their research.
The teams working on video and image generation models like "Sora" and "DALL-E" felt neglected and resource-starved, as their projects were considered less relevant to "ChatGPT", according to people familiar with the matter.
Over the past year, other projects unrelated to the language models have been shut down, according to one person. Others mentioned that teams have been reorganized within the company, as "OpenAI" streamlines its structure around enhancing the popular chatbot "ChatGPT", which is used by 800 million people.
