After OpenAI Exit, Ilya Sutskever Makes AI Safety Play with Safe Superintelligence
In a significant shift within the artificial intelligence (AI) landscape, Ilya Sutskever, the co-founder and former chief scientist at OpenAI, emerges as a central figure in the formation of a groundbreaking AI initiative, Safe Superintelligence (SSI). Following a notable departure from OpenAI—an organization he played a foundational role in—Sutskever’s venture into SSI marks a pivotal moment, spotlighting the continual evolution of AI towards the paramount goal of safety.
Safe Superintelligence is not just another AI company. With a distinct ambition, SSI is set on the monumental task of developing superintelligent machines, those which not only surpass human intellect but do so with an unwavering commitment to safety. “I am starting a new company,” Sutskever revealed in a social media post, indicating a single-minded pursuit towards achieving safe superintelligence. It’s a pursuit that underscores the critical nature of the endeavor—a straight shot towards creating an existence where machines significantly more intelligent than humans harmonize with safety and security.
At its core, artificial superintelligence represents the frontier of software-based AI systems, possessing cognitive abilities and a depth of thinking far beyond human scope. This venture into the unknown is spearheaded by Sutskever, along with co-founders Daniel Gross, a former partner at Y Combinator, and Daniel Levy, an ex-OpenAI engineer. Together, they champion the belief that constructing a safe superintelligence is the most crucial technical challenge of our era.
The foundation for SSI has been laid in Palo Alto and Tel Aviv, branching out in search of the world’s most skilled engineers and researchers. This ambitious recruitment drive focuses on gathering a uniquely talented team, all while maintaining an aura of mystery surrounding the company’s investors or potential financial backers.
SSI’s mission is deeply rooted in Sutskever’s longstanding advocacy for AI safety. Alongside Jan Leike, co-lead of OpenAI’s Superalignment team, Sutskever previously emphasized the imminent emergence of AI systems surpassing human intelligence and the urgent need for research dedicated to managing and mitigating the risks associated with potentially malevolent AI. This foresight into AI safety remains a cornerstone of SSI’s philosophy, aiming to develop technologies free from the distractions of conventional management or commercial pressures, as stated on the company’s website.
In conversations with the media, Sutskever further delineated the essence of SSI’s mission, highlighting its singular product focus on achieving safe superintelligence. This dedicated approach aims to circumvent the pitfalls of rapid, competitive advancement in favor of a meticulously insulated research and development process.
The backdrop of Sutskever’s new venture is his departure from OpenAI in May, preceding his involvement in the CEO transition saga involving Sam Altman. Sutskever’s exit, triggered by concerns over AI safety and the pace of advancement at OpenAI, eventually led to Altman’s temporary removal and subsequent reinstatement as CEO—a series of events underscoring the complex interplay of innovation, governance, and ethical considerations within the AI community.
SSI’s inception not only signals a new chapter for Sutskever but also serves as a testament to the evolving discourse on artificial intelligence. By focusing singularly on the safe advancement of superintelligent AI, Sutskever and his team at Safe Superintelligence aim to navigate the intricate balance between groundbreaking innovation and the ethical complexities that accompany it, marking a bold stride towards the future of AI.
