Two former Google AI specialists, Llion Jones and David Ha, have chosen Tokyo as the base for their new venture, Sakana AI. This startup aims to push the boundaries of generative AI technology. Both Jones and Ha are recognized figures in the AI sector, and with their combined expertise, they are poised to challenge both tech behemoths and emerging startups in this competitive arena.
Jones and Ha: The Visionaries
Llion Jones, a Welsh technologist, recently left his position at Google, paving the way for this exciting collaboration. David Ha, previously the head of Google's AI research division in Japan and now Sakana AI's CEO, joins Jones in this endeavor. In a Medium post, AI News reports the goal is to delve deeper into the possibilities of AI-driven solutions. Jones, who was instrumental in developing the “transformer” software, joins forces with Ha to explore AI's untapped potential.
Central to Sakana AI's foundation is the “transformer” technology. In June 2017, Llion Jones collaborated with seven other experts at Google on a research paper that set the trajectory for generative AI. This paper catalyzed a new wave of AI innovations, leading to the creation of AI applications like ChatGPT and image generators such as Stability AI and Dall-E. This shift also marked a trend where AI professionals began pursuing independent ventures, driven by the quest for generative AI dominance.
Reasons for Leaving Google
Jones, who had been with Google for nearly 12 years and was a co-author of the pivotal Transformers research paper, expressed that the company's size hindered him from pursuing the work he was passionate about. He mentioned facing challenges due to the bureaucracy at Google, which made him feel unproductive. Jones's academic background includes studying AI and holding a master's degree in advanced computer science from the University of Birmingham.
Jones observed that Google is channeling its entire company around generative AI, which he believes can be a limiting framework for innovation. David Ha, who left Google the previous year to lead research at startup Stability AI, shared that while they have discussed working on large language models (LLMs), they haven't finalized their plans. Both Jones and Ha expressed reservations about OpenAI, noting its growth and similarities to bureaucratic structures within Google. Jones felt that OpenAI wasn't particularly innovative, suggesting that they expanded on his research from Google.
Significance of Choosing to Base in Japan
While the decision to base Sakana AI in Japan may seem insignificant, it is actually a noteworthy one. Not least because Japan has been reportedly falling behind in AI development. Unlike the US, China, and the EU, which have powerful chatbots such as ChatGPT and Bing Chat, or Baidu's Ernie, Japan lacks the skills and resources to develop large language models (LLMs), the backbone of generative AI.
One reason for this gap is that Japan does not have enough software engineers to build the infrastructure and applications needed for deep learning, the technique behind LLMs. According to Japan's Ministry of Economy, Trade and Industry (METI), the country will face a shortage of nearly 800,000 software engineers by 2030.
Another reason is that Japan does not have its own AI supercomputers, such as IBM's Vela or Microsoft's Azure-hosted system, which can train LLMs efficiently. Instead, it has to rely on foreign machines or cloud services.
However, Japan does not want to be left out of the AI market and is already developing supercomputers to drive AI research. Being able to attract startups such as Sakana AI will be seen as a minor victory for Japan's growing AI sector.