'OpenAI Investment' MS's Suleyman AI CEO
Shows Differences with Altman on AGI Realization Timing
Joined MS, Which Acquired AI Company in March This Year
"Artificial General Intelligence (AGI) at a human-like level could take up to 10 years depending on how the situation unfolds."
Mustafa Suleyman, CEO of Microsoft (MS) AI, said this in an interview with IT media The Verge on the 9th (local time). AGI refers to general-purpose AI capable of performing various intellectual tasks like humans and adapting to new situations. Despite safety concerns, AGI is regarded as an innovative technology because it possesses AI capabilities similar to or beyond human levels, going beyond simple mechanical learning.
CEO Suleyman assessed that the current AI hardware, Nvidia's AI chip 'GB200,' is first or second generation and that AGI development is impossible with this hardware immediately. He said AGI realization would be possible only at some point after the 2nd to 5th generations in the future. Considering that the time required to surpass one generation is about 18 to 24 months, he estimated it would take at least 5 to 7 years, and depending on the situation, AGI might appear only after 10 years.
Suleyman's interview attracted attention as it conflicted with remarks made by Sam Altman, CEO of OpenAI, at an event hosted by The New York Times (NYT) just five days earlier. Altman said OpenAI plans to gradually release more powerful technology within 12 months and predicted, "AGI will be created sooner than expected, and although there are significant safety concerns with AGI development, the problems arising from it will be much less than anticipated."
The public clash between the two AI leaders, Altman and Suleyman, over the timing of AGI realization has drawn significant industry interest, raising questions about whether cracks might form in the AI partnership between the two companies. OpenAI and MS have been AI technology development partners for the past five years, jointly leading the AI market. Backed by MS's massive funding, OpenAI accelerated technology development and, through ChatGPT, established itself as a market leader alongside MS.
Although the two seemed aligned in their vision for AI technology development, their differing views may impact collaboration. When asked in the interview whether tension exists between MS and OpenAI, CEO Suleyman replied, "Every partnership has tension. It is very healthy and natural," adding, "They (OpenAI) operate independently, and partnerships evolve over time."
Who is Suleyman, 'DeepMind Co-founder'?
The market questions how long the collaboration between MS and OpenAI can continue smoothly. Especially, CEO Suleyman is a prominent figure who has been in a de facto competitive relationship with CEO Altman over the past decade of intense AI technology development. The British entrepreneur co-founded DeepMind, acquired by Google in 2014. DeepMind developed AlphaGo, which played a Go match against 9-dan player Lee Sedol in 2016. One media outlet described Suleyman as a figure similar to Altman before Altman emerged.
Altman co-founded OpenAI in 2015 with Elon Musk, CEO of Tesla, and others after Google acquired Suleyman's DeepMind. At that time, Altman and Musk were concerned that Google would monopolize AGI technology by acquiring DeepMind, which had AI capabilities. Thus, the two were inevitably in competition while founding and running AI companies.
CEO Suleyman played a role in integrating Google's products and AI technology at DeepMind after its acquisition, launching services like DeepMind Health and serving as a Google vice president. However, he left Google in 2022 to co-found another AI startup, Inflection AI. In March this year, MS, a key investor, acquired Inflection AI, making Suleyman MS's AI CEO. Many engineers and researchers from Inflection AI also moved to MS. During this period, CEO Altman launched ChatGPT, capturing global attention.
MS's reason for recruiting CEO Suleyman is clear: to reduce OpenAI's influence somewhat and build its own technological capabilities in AI development. MS reportedly suffered a significant shock from the CEO Altman ousting incident at OpenAI last year. Additionally, internal questions arose at MS about whether OpenAI's excessive investment demands for technology development were appropriate. In this context, CEO Suleyman became a key figure for MS to develop AI technology with a long-term vision without relying solely on OpenAI.
MS Checks OpenAI, Suleyman and Altman Clash in Many Areas
"It is true that we are in fierce competition. But we have a deep partnership, and Sam and I are close friends with mutual respect and trust," CEO Suleyman said on CNBC in June. While saying they get along well, he did not deny the fact that tension exists. MS listed OpenAI among its competitors in its annual report this year, alongside Amazon, Apple, Google, and Meta Platforms.
Last November, Sam Altman, CEO of OpenAI (left in the photo), and Satya Nadella, CEO of Microsoft, were speaking on stage at an event. Photo by AP Yonhap News
CEO Suleyman and CEO Altman are reportedly clashing in various areas, reflecting their ongoing competitive relationship. The NYT reported in October concerns about OpenAI's financial pressure, partnership stability, and discord among employees of the two companies, pointing out "signs that the close partnership between MS and OpenAI is unraveling."
According to reports, CEO Suleyman shouted at OpenAI employees who were angry about their failure to promptly disclose new technology developments to MS. Also, there is a protocol agreed upon for collaboration between MS and OpenAI, but MS employees violated it by downloading important OpenAI software, causing problems.
MS and OpenAI, navigating both competition and collaboration, are mobilizing all resources to ultimately develop AGI technology. OpenAI demands necessary resources from MS for technology development, while MS, having already invested $14 billion (about 20 trillion KRW) in OpenAI, receives technology and simultaneously strengthens its own capabilities.
In this process, OpenAI is reportedly considering granting MS access rights to AGI. Founded as a non-profit, OpenAI has so far refrained from providing related technology to investors or allowing commercial use due to concerns about AGI misuse. However, recognizing the enormous funding needed for technology development, OpenAI is pursuing a transition to a for-profit entity and is contemplating changing these principles to allow its largest investor, MS, to use the technology. OpenAI expects a $5 billion deficit this year and anticipates continued losses for several years.
© The Asia Business Daily(www.asiae.co.kr). All rights reserved.
