OpenAI CEO Ousted from Board of Directors
OpenAI CEO Sam Altman has recently been ousted from the board of directors, with CTO Mira Murati stepping in as interim CEO. This sudden change is expected to have a significant impact on the upcoming ‘GPT-5’ release, with the company experiencing shifts in both leadership and financing strategies.
Implications for ‘GPT-5’ Development
Prior to this board shake-up, CEO Altman had announced the commencement of GPT-5 development, outlining ambitions to create a next-generation AI model targeting artificial general intelligence (AGI). OpenAI’s plans included the building of a dataset for learning AGI and securing significant investments from Microsoft to support the necessary computing resources for AGI development. However, Altman’s sudden departure signifies a pivotal shift in OpenAI’s direction and development approach.
Shortly after Altman’s departure, Chairman Greg Brockman and three senior developers also left the company as Altman’s exit resulted in the termination of a $90 billion stock agreement intended to finance new model development. OpenAI’s challenges in raising immediate capital cast uncertainty on the future of ‘GPT-5’ and its associated development timelines. The company faces the dilemma of delaying or scaling back the model’s development without adequate funding.
Industry-Wide Developments
The repercussions of OpenAI’s leadership changes extend beyond the company itself. In light of broader industry developments, NVIDIA is reported to be developing new low-end AI chips for export to China, while YouTube and Meta have implemented policy changes in anticipation of AI-generated content disruptions ahead of the US presidential election.
As the AI landscape continues to evolve, industry experts like Bill Gates foresee significant transformations driven by AI agents, marking the potential for computing to be entirely reshaped within the next five years. With the heightened focus on AI development, advances in language models (LLMs) and AI infrastructure have become prominent across global technology companies.
Global AI Landscape
In South Korea, Naver’s ‘Sejong’ supercomputer has made waves by securing the 22nd position in the world’s supercomputer rankings. On the domestic front, Saltlux and NHN have formed an ‘AI Alliance,’ akin to Microsoft and OpenAI’s partnership, signaling an increasing commitment to enhancing the country’s AI infrastructure and collaborative efforts within the tech industry.
Similarly, startups like Scatter Lab and Upstage are venturing into the emotional chatbot and mathematics LLM development spheres, expanding the horizons of AI applications beyond traditional bounds.
As global AI models continue to proliferate, China has taken center stage with the largest number of LLMs, eclipsing the United States in sheer volume, despite questions surrounding the practical utilization of these models domestically.
(Photo = Shutterstock)
There was shocking news that OpenAI CEO Sam Altman has been ousted from the board of directors. And Open AI’s board appointed CTO Mira Murati as interim CEO.
Whatever the reason, it is clear that this will have an impact on the upcoming ‘GPT-5’. Shortly before this event occurred, CEO Altman announced, “We have begun development of GPT-5,” and its release next year seemed like a foregone conclusion.
In a recent interview, Altman said, “We have started development of GPT-5, a next-generation artificial intelligence (AI) model targeting AGI (artificial general intelligence), and we need more investment from Microsoft for the massive computing resources that needed to build AGI.” Yes.
In fact, OpenAI officially started building a dataset for learning AGI last week, and is known to be preparing a large sum of money to secure key development personnel.
In addition, the company is preparing to sell stock to cover the huge costs of AGI development, and has also requested financial support from its partner Microsoft (MS). They also revealed that they are securing GPUs for model training.
Therefore, it was normal to assume that OpenAI, which had recently completed the development of GPT-4 Turbo, had started the development of GPT-5. The fact that Altman quit at this very moment speaks volumes.
First of all, OpenAI’s board of directors seems to have felt dangerous about Altman rushing to develop GPT-5 by investing large amounts of money. Through official statements and internal meetings, we have repeatedly emphasized that “Open AI’s mission is to create AGI that is beneficial to humanity.”
Chief scientist Ilya Sutskever, who played a central role in Altman’s ouster, leads a team working to manage AGI at OpenAI. However, a month ago, word emerged that the scope of responsibility had been reduced and discord had arisen.
Therefore, it seems certain that with Altman’s departure, OpenAI will focus on providing adequate security measures for future model development. CTO Murathi also said at a recent event that “GPT-5 will have almost no hallucinations.”
AI Open Chairman Greg Brockman (from left), Minister of Small and Medium Businesses and Startups Lee Young, and CEO Sam Altman have a chat at an event held in Seoul last June.
In addition, because of this incident, not only Altman, but also Chairman Greg Brockman and three senior developers left the company. These are people who have played an important role in the development of the new model.
In addition, Altman was in the process of signing an agreement to sell the company’s stock worth $90 billion to finance the development costs of a new model. After he left the company, OpenAI said it was ending this contract. If sufficient costs cannot be raised immediately, model development will have no choice but to delay or reduce it.
As such, it seems inevitable that GPT-5 will change in some way. And as OpenAI’s board of directors said, it seems difficult to find innovative models until there is confidence that ‘safe AI’ can be built.
What will happen to Open AI and GPT-5?
Next is big news at home and abroad.
(Photo = Naver)
■ Naver ‘Sejong’ is ranked 22nd in the world supercomputer rankings
‘Sejong’, which Naver recently revealed, was ranked 22nd in the world’s supercomputer rankings. In addition, there were 12 domestic systems in the top 500. Although its specifications differ from the world’s best, domestic AI infrastructure is also gradually expanding.
■ Saltlux-NHN also forms ‘AI Alliance’
LLM Saltlux and NHN have joined hands. Instead of using NHN Cloud, Saltlux accepted investment and even acquired NHN subsidiaries. It is similar to the ‘MS-Open AI’ partnership, and both companies intend to create synergy through this.
(Photo = Diffusion Lab)
■ Scatter Lab “More fun than knowledge… We’ll open up the emotional chatbot market”
Scatter Lab, famous for its ‘Iruda’ chatbot, has unveiled ‘Pingpong-1′, a large lightweight language model (sLLM). The intention is to create an ’emotional chatbot’ that is closer to F (emotional) than T (rational).
■ Upstage “Challenge to develop the world’s best ‘mathematics GPT’ with Qanda”
Upstage has started developing a dedicated LLM for mathematics. The idea is to take the challenging LLM in Mathematics based on Qanda’s excellent data. Attention will be paid to see if Upstage, which has the highest level of refined technology, will produce good results.
(Photo = Humain)
■ ‘Wearing AI’ created by Apple core members using ChatGPT technology appears
Humain has unveiled ‘AI Pin’, a wearable device equipped with ChatGPT technology. It simply attaches to your clothes and acts as a personal assistant and can even make phone calls. In Korea, a new company called VTouch won a CES Innovation Award for its ‘AI Voice Chat Ring’.
■ NVIDIA is developing three types of low-end AI chips for export to China
There is news that NVIDIA is releasing low-end AI chips again to avoid the US government’s stronger ban on exports to China. As the practice of releasing low-end products to avoid regulations continued, the phrase ‘Catch me if you can’ came out. Meanwhile, NVIDIA announced that it will release the ‘H200’, a GPU that is twice as fast as the ‘H100’, in the second quarter of next year.
(Photo = Shutterstock)
■ YouTube “AI-generated content requires display”
YouTube has ordered the display of AI-generated content ahead of next year’s US presidential election. The intention is to prevent disruption caused by deep forgeries. Meta, which owns social media platforms such as Facebook and Instagram, also announced the same policy.
■ China, the world’s largest number of LLMs with 238… “Any further development is a national waste.”
The number of LLMs in China was claimed to be 238, which is more than that in the United States (about 160). However, only a small number of models are actually used, and some predict that the majority will disappear.
(Photo = Shutterstock)
■ Bill Gates “AI agents will completely change computing within 5 years”
The future predictions made by Microsoft founder Bill Gates attracted a lot of attention. He predicted that ‘AI agents’ will completely change the way computers are used within five years, and will ultimately revolutionize business and society.
AI Times news@aitimes.com
#Open #leaves #Altman…what #happen #GPT5
